US20110045610A1

US20110045610A1 - Uv treatment for carbon-containing low-k dielectric repair in semiconductor processing

Info

Publication number: US20110045610A1
Application number: US12/940,324
Authority: US
Inventors: Bart van Schravendijk; William Crew
Original assignee: Individual
Current assignee: Individual
Priority date: 2006-10-30
Filing date: 2010-11-05
Publication date: 2011-02-24
Also published as: US7851232B2; US20100261349A1

Abstract

A method for the ultraviolet (UV) treatment of carbon-containing low-k dielectric enables process-induced damage repair. The method is particularly applicable in the context of damascene processing. A method provides for forming a semiconductor device by depositing a carbon-containing low-k dielectric layer on a substrate and forming a trench in the low-k dielectric layer, the trench having sidewalls ending at a bottom. The trench is then exposed to UV radiation and, optionally a gas phase source of —CH₃groups, to repair damage to the carbon-containing low-k material of the trench sidewalls and bottom caused by the trench formation process (generally etching, ashing, and wet or dry cleaning). A similar treatment, with or without the gas phase source of —CH₃groups, may be applied to repair damage caused in a subsequent planarization operation.

Description

CROSS REFERENCE TO RELATED APPLICATION

This application is a divisional of prior application Ser. No. 11/590,661 filed on Oct. 30, 2006, entitled UV TREATMENT FOR CARBON-CONTAINING LOW-K DIELECTRIC REPAIR IN SEMICONDUCTOR PROCESSING, now pending, the disclosure of which is incorporated by reference herein in its entirety.

BACKGROUND OF THE INVENTION

The invention relates to low dielectric constant layers in semiconductor processing for use in various applications. More specifically, the invention relates to the UV treatment for repair of process-induced damage of low dielectric constant dielectric materials in, for example, damascene processing.
Ultrafine feature sizes and high performance requirements have necessitated the integration of low dielectric constant (low-k) insulating materials, that are mechanically weaker than previous generation materials, into semiconductor devices. The inherently weak nature of the low-k dielectric material can pose significant challenges for downstream electronic-packaging processes and material compatibility.
Low-k materials are, by definition, those semiconductor-grade insulating materials that have a dielectric constant (“k”) lower than that of SiO₂, i.e., 3.9. Various types of low-k materials can have dielectric constants ranging from about 3.8-3.6 (e.g., fluorosilicate glass (FSG)), to less than about 3.2 (e.g., (carbon doped oxide (CDO)), to as low as 2.2 (e.g., spin-on glass (SOG)) or even lower, and encompass low-k dielectrics referred to as “ultra low-k” (ULK) and “extreme ultra low-k” (ELK). In many CDO carbon-containing low-k implementations, such as are described herein, suitable carbon-containing low-k materials have a dielectric constant of about 2.7 or lower. To further reduce the size of devices on integrated circuits, it has become necessary to use conductive materials having low resistivity and insulators having low dielectric constants to reduce the capacitive coupling between adjacent metal lines. Low-k materials are being integrated into the devices to improve device performance and allow for device scaling.
Low-k materials are less dense than standard insulating materials such as SiO₂. This introduces a host of process integration and material compatibility difficulties. The balance between maintaining the film's integrity and integrating it properly and performing the necessary stripping, cleaning, and conditioning gets increasingly precarious. Patterning processes (etching, stripping, deposition, and cleaning) can also have a drastic impact on the integrity of carbon-containing low-k materials, in particular SiOC-based low-k materials.
The properties that give carbon-containing low-k dielectric materials their desirable low dielectric constants are the very same properties that are leading to significant integration challenges. Carbon-containing low-k materials achieve lower dielectric constants through the incorporation of non-polar covalent bonds (e.g., from the addition of carbon) and the introduction of porosity to decrease film density. Introducing porosity or the incorporation of terminal bonds, such as Si—CH₃, breaks the continuity of the rigid Si—O—Si lattice of traditional oxides, yielding a lower dielectric constant film that is both mechanically and chemically weaker. Because of the mechanical weakness, carbon-containing low-k films are susceptible to kinetic plasma damage that can undesirably densify the film and thus increase the film's effective k value.
Furthermore, chemical plasmas can modify carbon-containing low-k films where bonds such as Si—CH₃are readily broken. The susceptibility of carbon-containing low-k materials to plasma modification poses a serious integration challenge since plasma processes are routinely used to etch, clean, and deposit films in the manufacturing of a semiconductor device. Moreover, in damascene processing, prior to metal barrier deposition, process induced carbon-containing low-k dielectric damage can be incurred from etch, dry resist strip, wet cleaning and dry cleaning Carbon-containing low-k materials are also susceptible to the intercalation of plasma species, residues, solvents, moisture, and precursor molecules that can either adsorb into, outgas from, or chemically modify the film.
Damage to the carbon-containing low-k dielectric material on the sidewalls of the via trenches or in the inter-layer dielectric (ILD) regions during copper (Cu) damascene processing can compromise the integrity of dielectric spacing, leading to increased leakage, higher capacitance, and reduced performance and reliability. The damaged low-k layers can absorb moisture which remains trapped in the dielectric. This can oxidize the barrier material leading to Cu diffusion. Damage of the low-k dielectric material is linked to the loss of methyl groups (CH₃) in the film during processing. Thus, dielectric repair to prevent the unwanted absorption of moisture is important.
Carbon depletion occurs when, for example, a Si—CH₃bond is broken leaving a silicon dangling bond. Reaction with absorbed water results in the formation of silanol (Si—OH) groups. Slurry from chemical mechanical polishing (CMP) can also chemically remove carbon groups on the surface of the low-k dielectric film. Again, a reaction with the water present in the CMP slurry results in the formation of highly polarizable silanol (Si—OH) groups, which leads to an increase in k value for the damaged portion of the film, increasing the effective k value of the dielectric significantly.
Because of this, semiconductor manufacturers have developed methods to eliminate carbon depletion or replenish (repair) the depleted carbon. One method is the use of chemicals called “Toughening Agents” (TA) to repair carbon depletion damage. Another method is to use sacrificial capping layers to protect the low-k films from carbon depletion. However, both methods narrow the required process window and lead to higher manufacturing costs. Furthermore, an increase k effective may still result. Thus, improved methods for low-k dielectric repair in semiconductor processing are needed.

SUMMARY OF THE INVENTION

The present invention provides a method for the ultraviolet (UV) treatment of carbon-containing low-k dielectric, for example, but not limited to carbon-doped oxide (CDO), for the repair of process-induced damage. Inventive methods provide for treatment of a damaged low-k dielectric on a semiconductor substrate with UV radiation to repair processing induced damage. The method is particularly applicable in the context of damascene processing.
In one aspect, the invention relates to a method of forming a semiconductor device by depositing a carbon-containing low-k dielectric (e.g., CDO) layer on a substrate and forming a via and trench in the low-k dielectric layer, the trench having sidewalls ending at a bottom. The trench is then exposed to UV radiation to repair damage (e.g., dangling bonds or highly strained bonds, e.g., Si—O—Si or Si—CH₂—Si, caused by removal of organic (generally —CH₃) groups) to the low-k material of the trench sidewalls and bottom caused by the trench formation process (generally etching, ashing, and wet or dry cleaning)). The repaired damascene trench can then be filled with a conductive material, particularly a diffusion barrier layer followed by copper. The top surface of the semiconductor device can then be planarized, generally by chemical mechanical polishing (CMP). Post-planarization UV repair of planaraization-induced dielectric damage may also be conducted.
The UV exposure may be conducted in vacuum. While the invention is not limited to this theory of operation, it is believed that the UV exposure of the damaged dielectric surface according to this aspect of the invention cross-links the surface Si groups to fill gaps from the departed methyl (—CH₃) groups.
In other embodiments, the UV exposure may be conducted in a partial pressure of a reactive gas that participates in dielectric repair. An appropriate gas will include a gas phase source of methyl (—CH₃) groups during the UV exposure. Exposure time should be limited in order to prevent oxidation of or the stripping of methyl groups from the dielectric. In general, the dose time should be for no more than 10 seconds and/or result in a penetration of no more than four to five monolayers of the dielectric. A preferred dose time is about one to two seconds. Suitable gas phase reactants include, preferably, organo-silanes, -silazanes, and -siloxanes, for example, dichlorodimethylsilane (DCDMS), chlorotrimethylsilane (CTMS), hexamethyldisilazane (HMDS), hexamethyldisiloxane (HMDSO), tetravinyltetramethylcyclotetrasiloxane (TVTMCTS)). Other suitable gas phase reactants include acetaldehyde; alkanes, for example methane and ethane; alkenes, for example ethylene; and alkynes, for example acetylene, may also be used. —H and —O groups may also participate in suitable repair reactions. In that case, —H and —O may be provided in one or more gas phase reactants or may evolve from the film. The gas phase may also include inert carriers such as He, Ar, Ne, N₂, etc.
While the invention is not limited to this theory of operation, it is believed that damage sites, including dangling Si bonds and/or highly strained bonds (e.g., Si—O—Si or Si—CH₂—Si) in the carbon-containing dielectric film are satisfied with a methyl group from methyl-containing molecules of the gas phase source of methyl (—CH₃) groups in a reaction induced by the activation provided by UV radiation, thereby accomplishing low-k dielectric repair without substantial alteration of dielectric properties. In some instances, active methyl (—CH₃) groups may be generated by dissociation of methyl-containing molecules of the gas phase source of —CH₃groups by the UV radiation. Alternatively, methyl groups in methyl-containing molecules in the gas phase source of —CH₃groups can react with damage sites in the film. The reaction of the activated methyl groups with the damaged area of the film when the activation energy is lowered by the UV radiation reduces the energy of film, rendering it more stable.
Anything other than a carbon group reacting with a damage site on the surface of the dielectric will produce a higher k than the original carbon-containing low-k film. However, in instances where a minimal k value of the dielectric is not required, this rise in effective k resulting from non-carbon-containing repair (e.g., UV exposure alone) may be acceptable.
Another aspect of the invention relates to post-planarization (e.g., CMP) carbon-containing low-k dielectric damage repair arising in damascene processing. According to this aspect, following planarization by CMP, Si—OH (silanol) bonds (formed by removal of organic (generally —CH₃) groups by CMP slurry and processing and reaction of water in the slurry with resulting dangling Si-bonds), or other film damage (e.g., highly strained bonds) are repaired by exposing the surface to UV radiation. As described above, it is believed that the UV exposure cleaves the silanol bonds (Si—OH) and cross-links the film Si groups to fill gaps from the departed methyl (—CH₃) groups. The result is a densified surface layer. This UV exposure may optionally also be conducted in the presence of a gas phase source of methyl (—CH₃) groups such that the surface low-k dielectric repair is achieved without substantial alteration of dielectric properties, as described above.
The invention may also be more generally applicable in other semiconductor processing contexts. For example, a method of forming a semiconductor device, may involve depositing a carbon-containing low-k dielectric layer on a substrate, conducting a semiconductor processing operation that damages the low-k dielectric layer, and exposing the low-k dielectric layer to UV radiation such that processing-induced low-k dielectric damage to the dielectric is repaired.
These and other features and advantages of the present invention will be described in more detail below with reference to the associated drawings.

BRIEF DESCRIPTION OF THE DRAWINGS

The accompanying drawings, which are incorporated into and constitute a part of this specification, illustrate one or more embodiments of the present invention and, together with the detailed description, serve to explain the principles and implementations of the invention.

In the drawings:

FIG. 1 is a process flow chart depicting a method in accordance with an embodiment of the invention.

FIGS. 2A-2D are cross sectional diagrams illustrating the formation of a semiconductor device by a dual damascene process in accordance with an embodiment of the invention.

FIG. 3 is a schematic diagram of an example chamber suitable for implementing the present invention.

DETAILED DESCRIPTION

Embodiments of the present invention are described herein in the context of a UV treatment for carbon-containing low-k dielectric repair in damascene processing. Those of ordinary skill in the art will realize that the following detailed description of the present invention is illustrative only and is not intended to be in any way limiting. Other embodiments of the present invention will readily suggest themselves to such skilled persons having the benefit of this disclosure. Reference will now be made in detail to implementations of the present invention as illustrated in the accompanying drawings. The same reference indicators will be used throughout the drawings and the following detailed description to refer to the same or like parts.
The term “semiconductor device” as used herein refers to any device formed on a semiconductor substrate or any device possessing a semiconductor material. In many cases, a semiconductor device participates in electronic logic or memory, or in energy conversion. The term “semiconductor device” subsumes partially fabricated devices (such as partially fabricated integrated circuits) as well as completed devices available for sale or installed in particular apparatus. In short, a semiconductor device may exist at any state of manufacture that employs a method of this invention or possesses a structure of this invention.
As noted above, the present invention provides a method for the ultraviolet (UV) treatment of carbon-containing low-k dielectric for the repair of process-induced damage. Applicable carbon containing dielectrics typically have SiO-based backbones doped with carbon, in particular CDO (for example, those formed from octamethyl cyclotetrasiloxane (OMCTS), tetramethylcyclotetrasiloxane (TMCTS), dimethyldimethoxysilane (DMDMOS), and diethoxymethylsilane (DEMS) and other known CDO precursors), but may also include hybrid polymers incorporating both C, Si and O in the backbone. Inventive methods provide for treatment of a damaged carbon-containing low-k dielectric on a semiconductor substrate with UV radiation to repair processing induced damage. The method is particularly applicable in the context of damascene processing.
In one aspect, the invention relates to a method of forming a semiconductor device by depositing a carbon-containing low-k dielectric layer on a substrate and forming a via and trench in the low-k dielectric layer, the trench having sidewalls ending at a bottom. The trench is then exposed to UV radiation to repair process induced low-k dielectric damage (e.g., dangling bonds or highly strained bonds, e.g., Si—O—Si or Si—CH₂—Si, caused by removal of organic (generally —CH₃) groups) to the low-k material of the trench sidewalls and bottom caused by etch, dry resist strip, wet cleaning and dry cleaning, for example, involved in the trench formation process. The repaired damascene trench can then be filled with a conductive material, particularly a metal diffusion barrier followed by copper. The top surface of the semiconductor device can then be planarized, generally by chemical mechanical polishing (CMP). Post-planarization UV repair of planaraization-induced dielectric damage may also be conducted.
FIG. 1 is a process flow chart depicting operations in a method in accordance with an embodiment of this invention. A generalized version of a dual damascene technique will be described below with reference to FIGS. 2A through 2D, which depict a partially formed semiconductor device during various stages of this process. However, the invention is not intended to be limited to the use of a dual damascene technique as the invention may be used with other techniques.
Referring now to FIG. 1, a carbon containing low-k dielectric layer is deposited on a substrate at 100. Patterns of conductive features are formed in the dielectric layer, generally by etching, at 102. Etching generally results in damage to the pattern edges, generally trench sidewalls and bottoms, as described above. Other process operations, such as dry resist strip, wet cleaning and dry cleaning, can also cause or contribute to low-k dielectric damage. The conductive features are typically, though not necessarily, metal lines and vias. In one example, they are the interconnects of a metallization layer that is formed from copper. As is known to those of skill in the art, various techniques may be employed to form such layers.
The etched trenches are then exposed to ultraviolet (UV) radiation at 104. The UV exposure may be conducted in vacuum. While the invention is not limited to any particular theory of operation, it is believed that the UV exposure of the damaged dielectric surface according to this aspect of the invention cross-links the surface Si groups to fill gaps from the departed methyl (—CH₃) groups.
In other embodiments, the UV exposure may be conducted in a partial pressure of a reactive gas that participates in dielectric repair. An appropriate gas will have a gas phase source of methyl (—CH₃) groups during the UV exposure. Exposure time should be limited in order to prevent oxidation of or the stripping of methyl groups from the dielectric. In general, the dose time should be for no more than 10 seconds and/or result in a penetration of no more than four to five monolayers of the dielectric. A preferred dose time is about one to two seconds. Suitable gas phase reactants include, preferably, organo-silanes, -silazanes, and -siloxanes, for example, dichlorodimethylsilane (DCDMS), chlorotrimethylsilane (CTMS), hexamethyldisilazane (HMDS), hexamethyldisiloxane (HMDSO), tetravinyltetramethylcyclotetrasiloxane (TVTMCTS)). Other suitable gas phase reactants include acetaldehyde; alkanes, for example methane and ethane; alkenes, for example ethylene; and alkynes, for example acetylene, may also be used. —H and —O groups may also participate in suitable repair reactions. In that case, —H and —O may be provided in one or more gas phase reactants or may evolve from the film. The gas phase may also include inert carriers such as He, Ar, Ne, N₂, etc.
While the invention is not limited to this theory of operation, it is believed that damage sites, including dangling Si bonds and/or highly strained bonds (e.g., Si—O—Si or Si—CH₂—Si) in the carbon-containing low-k dielectric film are satisfied with a methyl group from the gas phase source of —CH₃groups in a reaction induced by the activation provided by UV radiation, thereby accomplishing low-k dielectric repair without substantial alteration of dielectric properties. In some instances, active methyl (—CH₃) groups may be generated by dissociation of methyl-containing molecules in the gas phase source of —CH₃groups by the UV radiation. Alternatively, methyl groups in methyl-containing molecules in the gas phase source of —CH₃groups can react with damage sites in the film. The reaction of activated methyl with the damage surface site can occur when UV radiation excites electrons into anti-bonding states, lowering the thermal activation energy of the reaction. This renders the film more stable.
Anything other than a carbon group reacting with a damage site on the surface of the dielectric will produce a higher k than the original low-k film. However, in instances where a minimal k value of the dielectric is not required, this rise in effective k resulting from non-carbon-containing repair (e.g., UV exposure alone) may be acceptable.
The UV light may be irradiated on the sidewalls and bottom of the trench to repair the damage from the etching of the trenches. The photon energy supplied by the UV treatment effectively lowers the activation barrier for reaction, and depleted methyl sites within the films are filled by a reaction with the active methyl groups derived from the gas phase source of methyl (—CH₃) groups. Satisfying the Si dangling bonds in the damaged film repairs the damage done to the low-k dielectric during the trench formation process (e.g., etching, ashing, and wet or dry cleaning) without substantially altering the dielectric properties. In addition, there may be silanol groups (—OH) in the damaged areas that are formed from Si dangling bonds that are exposed to moisture, either in subsequent processing or from moisture present in the fab ambient atmosphere. These silanol bonds are cleaved during the UV treatment, and the —OH groups leave the film, effectively lowering the dielectric constant and “repairing” the film. Interconnect reliability is thereby improved.
The dielectric film may be exposed to a UV source that produces UV radiation. The choices of UV wavelength(s), total exposure time and intensity, etc., depend on a number of factors, including the thickness of the dielectric film and the composition of the dielectric film. Suitable UV treatment parameters are in the power intensity range of about 1 mw-20 W/cm², preferably about 500 mW-5 W/cm²; at a wavelength of about 150-500 nm, preferably about 200-400 nm; for up to about 20 minutes, for example less than 10 minutes, or less than 5 minutes, and preferably less than about 2 minutes; at a wafer temperature of between room temperature up to about 450° C., preferably about 200-400° C. A typical UV exposure in accordance with this aspect of the invention has a power density of about 1-3 W/cm²at a wafer temperature of about 350° C. in either inert (e.g., He, Ar, forming gas, or N₂) or oxidizing environments (e.g., in an anneal environment that comprises one or more of oxygen, ozone, peroxide or nitrous oxide). In oxidizing environments, oxygen can promote bond breaking thereby facilitating hydrogen removal for film repair. The UV source can be a single wavelength excimer lamp or broad spectrum source with arc or microwave excitations. The process pressure can range from about 1 mTorr to 760 Torr, preferably from about 1 Torr to 200 Torr. The UV exposure can also be achieved through direct exposure to a plasma of He, Ar, N₂, NH₃, N₂O, O₂, or a mixture of them.
Exposure time should be limited in order to prevent oxidation of or the stripping of methyl groups from the dielectric, or dielectric shrinking (which causes stress and strain in the film. In general, the dose time should be for no more than 10 seconds and/or result in a penetration of no more than four to five monolayers of the dielectric. A preferred dose time is about one to two seconds.
The apparatus employed to implement the invention can have one or more UV sources. In some embodiments described herein, the apparatus will have a single UV source. Suitable apparatus are described in more detail below.
The trenches may then be filled at 106 with a conductive material such as Cu. When filling the features, excess material will have to be removed from the top (exposed surface) of the dielectric layer. The excess material may be removed by a planarization process to form an exposed pattern of conductive features in the dielectric at 108. As discussed above, one widely-used planarization process is chemical mechanical polishing (CMP).
Particularly if the conductive layer includes copper, it is beneficial to remove oxides from the conductive layer after the planarization process at 110. As known by those of skill in the art, this step is typically accomplished by exposing the copper surfaces to a hydrogen or ammonia plasma. As discussed above, slurry from CMP can chemically remove carbon groups. The hydrogen/plasma used to remove the oxides can also strip carbon or hydroxyl groups from the film, leaving silicon dangling bonds on the surface of carbon-containing low-k dielectric film.
To repair the dangling bonds, the dielectric surface can be exposed to a second UV radiation treatment at 112. The treatment may be conducted under the same conditions as described above with reference to the post-trench etch treatment. The presence of the gas phase source of —CH₃groups is optional. While the invention is not limited to this theory of operation, it is believed that the UV exposure of the surface removes the dangling bonds by cross-linking the surface Si groups to fill gaps from the departed methyl (—CH₃) groups and produces a densified surface layer. In some instances, the alteration of the surface dielectric properties of the carbon-containing low-k dielectric arising from the UV-induced surface cross-linking is acceptable.
In other instances, where the carbon-containing low-k surface properties are more critical, this surface UV exposure may optionally also be conducted in the presence of a gas phase source of —CH₃groups such that the surface low-k dielectric repair is achieved without substantial alteration of dielectric properties, as described above.
Following the damascene processing with repair in accordance with the present invention, a diffusion barrier film, such as a copper diffusion barrier film, may be deposited on the planarized surface of the partially-formed semiconductor device. This layer may serve other purposes aside from that of a diffusion barrier. For example, the diffusion barrier film may also act as an etch stop layer.
Referring now to FIGS. 2A-2D, a typical dual damascene process incorporating the carbon-containing low-k dielectric repair processes of the present invention are illustrated. First and second layers of dielectric are deposited in succession, possibly separated by deposition of an etch stop layer, such as a silicon nitride layer. These layers are depicted in FIG. 2A as first dielectric layer 203, second dielectric layer 205, and etch stop layer 207. As is well known in the art, according to alternative damascene processing techniques a single thicker dielectric layer can be used instead of discrete first and second layers.
After deposition of the second dielectric layer 205, a via mask 211 is formed having openings where vias will be subsequently etched. Next, the vias are partially etched down through the level of etch stop 207. Then via mask 211 is stripped off and replaced with a line mask 213 as depicted in FIG. 2B. A second etch operation is performed to remove sufficient amounts of dielectric to define line paths 215 in second dielectric layer 205. The etch operation also extends via holes 217 through first dielectric layer 203, down to contact the underlying substrate 209.
It should be noted that the foregoing description is just an example of one dual damascene process with which the present invention may be implemented. One type of via first dual damascene process has been described. In other embodiments, a via first process may involve complete etching of the vias prior to etching of the line trenches. Or, a trench first process, in which the etching of the line trenches precedes the via etching, may be used. These various damascene processing techniques, and other variations thereon, are well know in the art and represent alternative implementation contexts for embodiments of the present invention. The invention is also applicable is single damascene processing, more conventional metal deposition and etch, and essentially any semiconductor processing context where carbon-containing low-k dielectrics are used.
Further in this regard, the term “trench” in the context of damascene processing is commonly understood to describe a feature formed in dielectric and subsequently filled to form a conductive line in a dielectric layer. In a more general semiconductor processing context, the term is also understood to describe a feature formed in dielectric and subsequently filled to form an element of a semiconductor device (e.g., via, line, STI, etc.), and may include a damascene trench, via or combined damascene structure. Unless it is otherwise clear from the context, when used herein, the term should be understood to have its broader meaning
After etching the via hole and trenches, the photoresist is removed in another plasma process, followed by a wet or dry clean and then damage on the low-k dielectric surface is repaired, as discussed above.
Thereafter a thin layer of conductive barrier layer material 219 is formed on the exposed surfaces (including sidewalls) of dielectric layers 203 and 205. Conductive barrier layer material 219 may be formed, for example, of tantalum or tantalum nitride. A CVD or PVD operation is typically employed to deposit conductive barrier layer material 219. Prior to the deposition of the barrier material, a plasma process is typically used to clean the bottoms of the vias to remove oxidation and contaminants from the exposed copper surface on the underlying layer. As is known to those skilled in the art, this barrier “preclean” plasma process can be simply an inert plasma or a reactive plasma of a gas such as hydrogen. The preclean plasma process can also damage the low k dielectric film. An in-situ UV treatment to repair the damaged low k film may employed prior to the PVD operation—for Ta or TaN.
On top of the barrier layer, a conductive metal (typically copper) is deposited in the via holes and line paths 217 and 215. Conventionally this deposition is performed in two steps: an initial deposition of a conductive seed layer followed by bulk deposition of copper by electroplating. The seed layer may be deposited by physical vapor deposition, chemical vapor deposition, electroless plating, etc. Note that the bulk deposition of copper not only fills line paths 215 but, to ensure complete filling, also covers all the exposed regions on top of second dielectric layer 205.
Thus, it becomes necessary to planarize the structure and remove the excess copper from the device. Planarization removes material down to the level of the top of dielectric layer 205. This results in an exposed pattern of conductive lines 221 in dielectric layer 205 and vias in dielectric layer 203. (See the cross-sectional view of FIG. 2C and the simplified top view of FIG. 2D.)
Planarization may be accomplished by various techniques. Typically, the process involves some amount of CMP. It may also involve a combination of electropolishing, to remove most of the excess bulk copper, followed by CMP to remove the remaining copper down to the level of the top surface of dielectric layer 205. As discussed above, slurry from CMP can chemically remove carbon groups, and the hydrogen used to remove the oxides from the conductive layer after the planarization process can also remove carbon groups leaving silicon dangling bonds on the surface of low-k dielectric films.
Another aspect of the invention relates to post-planarization (e.g., CMP) damage repair arising in damascene processing. According to this aspect, following planarization by CMP, Si—OH (silanol) bonds (formed by removal of organic (generally —CH₃) groups by CMP slurry and processing and reaction of water in the slurry with resulting dangling Si— bonds), or other carbon-containing low-k dielectric film damage (e.g., highly strained bonds) are repaired by exposing the surface to UV radiation. As described above, it is believed that the UV exposure cross-links the film Si groups to fill gaps from the departed methyl (—CH₃) groups. The result is a densified surface layer. This UV exposure may optionally also be conducted in the presence of a gas phase source of -methyl (—CH₃) groups such that the surface low-k dielectric repair is achieved without substantial alteration of dielectric properties, as described above.

Apparatus

The present invention can be implemented in many different types of apparatus. In preferred embodiments, the apparatus will include one or more chambers (sometimes referred to as process vessels) that house one or more wafers and are suitable for wafer processing. At least one chamber will include a UV source. A single chamber may be employed for all operations of the invention or separate chambers may be used. Each chamber may house one or more wafers (substrates) for processing. The one or more chambers maintain the wafer in a defined position or positions (with or without motion within that position, e.g., rotation, vibration, or other agitation) during procedures of the invention. For certain operations in which the wafer temperature is to be controlled, the apparatus may include a heating or cooling platen.
FIG. 3 is a schematic diagram of an example chamber 301 in accordance with the invention. Chamber 301 is capable of holding a vacuum and/or containing gases at pressures above atmospheric pressure. For simplicity, only one station of one chamber 301 is shown. It is noted that in preferred embodiments, chamber 301 is one chamber in a multi-chambered apparatus (entire apparatus not shown), although chamber 301 could alternatively be part of a stand-alone single chambered apparatus. In either case, the chamber(s) may have one or more than one station. Suitable single chamber, multi-station apparatus, for example, include the Novellus Sequel and Vector systems modified for UV treatments.
A substrate holder 303 secures a wafer 305 in a position such that light from a UV light source array 307 can irradiate wafer 305. Substrate holder 303 can have a heater (not shown) that can heat the substrate to defined temperatures, or could be cooled using a chiller and can be controlled by a temperature controller (not shown). Chamber 301 is configured with a gas inlet 315, which is connected to a gas source (not shown), and with a vacuum outlet 313, which is connected to a vacuum pump (not shown). The amount of gas introduced into the chamber 301 can be controlled by valves and mass flow controller (not shown) and pressure is measured by pressure gauge (not shown).
In this example, the UV light source array 307 is mounted outside the chamber 301. In alternate embodiments, the UV light source array may be housed inside the chamber 301. UV light source array 307 includes an array of individual UV sources such as mercury vapor or xenon lamps. Note that the invention is not limited to mercury vapor or xenon lamps as UV light sources and other suitable light sources include deuterium lamps, excimer lamps or lasers (e.g., excimer lasers and tunable variations of various lasers). Various optical elements, such as reflectors, may be required to direct the UV light toward portions of the substrate. Methods for directing the light at different portions of the substrate at different times may be required as well. A scanning mechanism may be used for this purpose. A window 311 made of quartz, sapphire or other suitable material is positioned between UV light source array 307 and wafer 305 to provide isolation. Filters can also be used to remove unwanted spectral components from particular sources to “tune” the sources.
The UV light source array 307 may be comprised of one or more types of UV sources, for example an array of three types of UV sources, each type providing UV radiation with a different wavelength distribution. The UV sources are electrically connected to each other (309 a, 309 b and 309 c) and controlled by control system 310, which controls when each of the various UV sources is illuminated. Control system 310 is typically, but not limited to, a computer processing system such as a PC or workstation. Of course, any number and types of individual light sources in any suitable configuration can be used.
Note that the light source array and control configuration of FIG. 3 is only an example of a suitable configuration. In general, it is preferable that the lamps are arranged to provide uniform UV radiation to the wafer. For example, other suitable lamp arrangements can include circular lamps concentrically arranged or lamps of smaller length arranged at 90 degree and 180 degree angles with respect to each other may be used. The light source(s) can be fixed or movable so as to provide light in appropriate locations on the wafer. Alternatively, an optical system, including for example a series of movable lenses, filters, and/or mirrors, can be controlled to direct light from different sources to the substrate at different times.
The UV light intensity can be directly controlled by the type of light source and by the power applied to the light source or array of light sources. Factors influencing the intensity of applied power include, for example, the number or light sources (e.g., in an array of light sources) and the light source types (e.g., lamp type or laser type). Other methods of controlling the UV light intensity on the wafer sample include using filters that can block portions of light from reaching the wafer sample. As with the direction of light, the intensity of light at the wafer can be modulated using various optical components such as mirrors, lenses, diffusers and filters. The spectral distribution of individual sources can be controlled by the choice of sources (e.g., mercury vapor lamp vs. xenon lamp vs. deuterium lamp vs. excimer laser, etc.) as well as the use of filters that tailor the spectral distribution. In addition, the spectral distributions of some lamps can be tuned by doping the gas mixture in the lamp with particular dopants such as iron, gallium, etc.
It should be understood that the apparatus depicted in FIG. 3 is only an example of a suitable apparatus and that other designs for other methods involved in previous and/or subsequent processes may be used. Examples of apparatus that may be suitable for implementing the present invention are also described in commonly assigned co-pending application Ser. Nos. 11/115,576 filed Apr. 26, 2005, 10/800,377 filed Mar. 11, 2004 and 10/972,084 filed Oct. 22, 2004, incorporated by reference herein.
While the invention has been described primarily in the context of damascene processing, it may also be applicable in other semiconductor processing contexts.
Although the foregoing invention has been described in some detail for purposes of clarity of understanding, it will be apparent that certain changes and modifications may be practiced within the scope of the appended claims. It should be noted that there are many alternative ways of implementing both the process and compositions of the present invention. Accordingly, the present embodiments are to be considered as illustrative and not restrictive, and the invention is not to be limited to the details given herein.

Claims

1. A method of forming a semiconductor device in damascene processing, comprising:

receiving in a processing chamber a semiconductor device substrate comprising a planarized surface having conductive features in a carbon-containing low-k dielectric layer;

exposing the planarized surface to UV radiation;

whereby planarization-induced low-k dielectric damage on the surface is repaired without substantially altering the dielectric properties.

2. The method of claim 1, wherein the conductive features are comprised of metal.

3. The method of claim 2, wherein oxide is removed from the metal surface.

4. The method of claim 1, further comprising exposing the planarized surface to a gas phase source of —CH₃groups.

5. The method of claim 4, wherein the gas phase source of —CH₃groups comprises one or more selected from the group consisting of organo-silanes, -silazanes, and -siloxanes; acetaldehyde; alkanes; alkenes; and alkynes.

6. The method of claim 5, wherein the gas phase source of —CH₃groups comprises one or more selected from the group consisting of dichlorodimethylsilane (DCDMS), chlorotrimethylsilane (CTMS), hexamethyldisilazane (HMDS), hexamethyldisiloxane (HMDSO), tetravinyltetramethylcyclotetrasiloxane (TVTMCTS), acetaldehyde, methane, ethane, ethylene, acetylene, and combinations thereof.

7. The method of claim 6, wherein the gas phase source of —CH₃groups comprises one or more selected from the group consisting of dichlorodimethylsilane (DCDMS), chlorotrimethylsilane (CTMS), hexamethyldisilazane (HMDS), hexamethyldisiloxane (HMDSO), tetravinyltetramethylcyclotetrasiloxane (TVTMCTS), and combinations thereof.

8. The method of claim 2, wherein the metal is copper.

9. The method of claim 1, wherein the UV radiation has a power density of about 500 mW-5 W/cm²and a wavelength from about 150-500 nm, and the exposure is conducted at a temperature of about room temperature up to 450° C. for less than 20 minutes.

10. The method of claim 1, wherein the UV radiation has a power density of about 1-3 W/cm²and a wavelength from about 200-400 nm, and the exposure is conducted at a temperature of about 200-400° C. for less than 5 minutes.

11. The method of claim 1, wherein the carbon-containing low-k dielectric comprises CDO.

12. A method of forming a semiconductor device, comprising:

receiving in a processing chamber a semiconductor device substrate comprising a carbon-containing low-k dielectric layer, the semiconductor device substrate having been damaged by a semiconductor processing operation;

exposing the low-k dielectric layer to UV radiation such that processing-induced low-k dielectric damage to the dielectric is repaired without substantially altering the dielectric properties.

13. The method of claim 12, further comprising exposing the low-k dielectric layer to a gas phase source of —CH₃groups.

14. The method of claim 13, wherein the gas phase source of —CH₃groups comprises one or more selected from the group consisting of organo-silanes, -silazanes, and -siloxanes; acetaldehyde; alkanes; alkenes; and alkynes.

15. The method of claim 14, wherein the gas phase source of —CH₃groups comprises one or more selected from the group consisting of dichlorodimethylsilane (DCDMS), chlorotrimethylsilane (CTMS), hexamethyldisilazane (HMDS), hexamethyldisiloxane (HMDSO), tetravinyltetramethylcyclotetrasiloxane (TVTMCTS), acetaldehyde, methane, ethane, ethylene, acetylene, and combinations thereof.

16. The method of claim 15, wherein the gas phase source of —CH₃groups comprises one or more selected from the group consisting of dichlorodimethylsilane (DCDMS), chlorotrimethylsilane (CTMS), hexamethyldisilazane (HMDS), hexamethyldisiloxane (HMDSO), tetravinyltetramethylcyclotetrasiloxane (TVTMCTS), and combinations thereof.

17. The method of claim 12, wherein the UV radiation has a power density of about 500 mW-5 W/cm²and a wavelength from about 150-500 nm, and the exposure is conducted at a temperature of about room temperature up to 450° C. for less than 20 minutes.

18. The method of claim 12, wherein the UV radiation has a power density of about 1-3 W/cm²and a wavelength from about 200-400 nm, and the exposure is conducted at a temperature of about 200-400° C. for less than 5 minutes.

19. The method of claim 12, wherein the carbon-containing low-k dielectric comprises CDO.

20. The method of claim 12, where the device substrate further comprises metal features.