Patent Application
20240018678
This invention relates generally to methods, compositions, and additives for electrolytic copper metallization in the field of microelectronics manufacture.
Electrolytic copper metallization is employed in the field of microelectronics manufacture to provide electrical interconnection in a wide variety of applications, including, for example, manufacturing semiconductor integrated circuit (IC) devices. The demand for semiconductor IC devices, such as computer chips with high circuit speed and high packing density, requires the downward scaling of feature sizes in ultra-large-scale integration (ULSI) and very-large-scale integration (VLSI) structures. The trend to smaller device sizes and increased circuit density requires decreasing the dimensions of interconnect features. An interconnect feature is a feature such as a via or trench formed in a dielectric substrate which is then filled with metal to yield an electrically conductive interconnect. Further decreases in interconnect size present challenges in metal filling.
Copper has been used to replace aluminum to form connection lines and interconnects in semiconductor substrates. Copper has a lower resistivity than aluminum and the thickness of a copper line for the same resistance can be thinner than that of a corresponding aluminum line.
However, the use of copper has introduced a number of requirements into the IC manufacturing process. First, copper has a tendency to diffuse into the semiconductor's junctions, thereby disturbing their electrical characteristics. To combat this occurrence, a barrier layer, such as titanium nitride, tantalum, tantalum nitride, or other layers as are known in the art, is applied to the dielectric prior to the deposition of the copper layer. It is also generally necessary that the copper be deposited on the barrier layer cost-effectively while ensuring the requisite coverage thickness for carrying signals between the IC's devices. However, as the architecture of ICs continues to shrink, this requirement has been increasingly difficult to satisfy.
One conventional semiconductor manufacturing process is the copper damascene system, which begins by etching the circuit architecture into the substrate's dielectric material. The architecture comprises a combination of the aforementioned trenches and vias. Next, a barrier layer is laid over the dielectric to prevent diffusion of the subsequently applied copper layer into the substrate's junctions, followed by physical or chemical vapor deposition of a copper seed layer, to provide electrical conductivity for a sequential electrochemical process. Copper to fill into the vias and trenches on substrates can be deposited by plating (such as electroless and electrolytic), sputtering, plasma vapor deposition (PVD), and chemical vapor deposition (CVD).
Electrochemical deposition is a preferred method for applying copper since it is more economical than other deposition methods and can flawlessly fill into the interconnect features (often called “bottom up” growth). After the copper layer has been deposited, excess copper is removed from the facial plane of the dielectric by chemical mechanical polishing, leaving copper in only the etched interconnect features of the dielectric. Subsequent layers are similarly produced before assembly into the final semiconductor package.
Copper plating methods must meet the stringent requirements of the semiconductor industry. For example, copper deposits must be uniform and capable of flawlessly filling the small interconnect features of the device, such as those with openings of 100 nm or smaller.
Electrolytic copper systems have been developed which rely on so-called “superfilling” or “bottom-up growth” to deposit copper into high aspect ratio features. Superfilling involves filling a feature from the bottom up, rather than at an equal rate on all its surfaces, to avoid seams and pinching off that can result in voiding. Systems comprising a suppressor and an accelerator as additives have been developed for superfilling. As the result of momentum of bottom-up growth, the copper deposit is thicker on the areas of interconnect features than on the field area that does not have features. These overgrowth regions are commonly called overplating, mounding, bumps, or humps. Smaller features generate higher overplating humps due to faster superfill speed. Overplating poses challenges for later chemical and mechanical polishing processes that planarize the copper surface. Various additives, including levelers, can be used to reduce the overgrowth.
As chip architecture gets smaller, with interconnects having openings on the order of 100 nm and smaller through which copper must grow to fill the interconnects, there is a need for enhanced bottom-up speed. That is, the copper must fill “faster” in the sense that the rate of growth on the feature bottom must be substantially greater than the rate of growth on the rest of areas, and even more so than in conventional superfilling of larger interconnects.
In addition to superfilling and overplating issues, micro-defects may form when electrodepositing copper for filling interconnect features. One such defect is the formation of internal voids inside the features. As copper is deposited on the feature side walls and top entry of the feature, deposition on the side walls and entrance to the feature can pinch off and thereby close access to the depths of the features, especially with features that are small (e.g., <100 nm) and/or that have a high aspect ratio (depth:width) if the bottom-up growth rate is not fast enough. Smaller feature size or higher aspect ratio generally requires faster bottom-up speed to avoid pinching off. Moreover, smaller size or higher aspect ratio features tend to have thinner seed coverage on the sidewall and bottom of a via/trench where voids can also be produced due to insufficient copper growth in these areas. An internal void can interfere with electrical connectivity through the feature.
Microvoids are another type of defect which can form during or after electrolytic copper deposition due to uneven copper growth or grain recrystallization that happens after copper plating.
Additionally, some local areas of a semiconductor substrate, such as areas where there is a copper seed layer deposited by physical vapor deposition, may not grow copper during the electrolytic deposition, resulting in pits or missing metal defects. These copper voids are considered to be “killer defects,” as they reduce the yield of semiconductor manufacturing products. Multiple mechanisms contribute to the formation of these copper voids, including the semiconductor substrate itself. However, copper electroplating chemistry has influence on the occurrence and population of these defects.
Other defects include, for example, surface protrusions, which are isolated deposition peaks occurring at localized high current density sites, localized impurity sites, or otherwise. Copper plating chemistry has an influence on the occurrence of such protrusion defects. Although not considered as defects, copper surface roughness is also important for semiconductor wafer manufacturing. Generally, a bright copper surface is desired as it can reduce the swirl patterns formed during wafer entry in the plating solution. Roughness of copper deposits makes it more difficult to detect defects by inspection, as defects may be concealed by peaks and valleys of rough surface topography. Moreover, smooth growth of copper has become more important for flawlessly filling of fine interconnect structures as the roughness can cause pinch off of feature and thereby close access to the depths of the feature. It is generally recognized that copper plating chemistry, including suppressor, accelerator, and leveler, has great influence on the roughness of copper deposits.
A superior suppressor for use in superfilling of submicron features of a semiconductor substrate by electrodeposition of copper is described in Paneccasio U.S. Pat. No. 7,303,992. The suppressor comprising a combination of propylene oxide (PO) repeat units and ethylene oxide (EO) repeat units present in a PO:EO ratio between about 1:9 and about 9:1 and bonded to a nitrogen-containing species, wherein the molecular weight of the suppressor compound is between about 1000 and about 30,000.
Published applications of BASF, US 2012/0018310, US 2012/0024711 and US 2012/0027948, describe amines having at least three active amine sites substituted with random copolymers of PO and EO.
Described herein are electrodeposition compositions comprising a source of copper ions, in an amount sufficient to electrolytically deposit copper onto the substrate and into the electrical interconnect features, and a suppressor comprising a polyether bonded to a nitrogen of an oligo(alkylene imine) having at least three amine sites. The polyether comprises a block copolymer substituent comprising propylene oxide (PO) repeat units and ethylene oxide (EO) repeat units wherein the ratio of propylene oxide (PO) repeat units to ethylene oxide (PO) repeat units is adjusted to balance the strongly polarizing effect of the PO repeat units with the more hydrophilic and solubilizing effect of the EC) repeat units. The number average molecular weight of the suppressor compound is between about 6,000 and about 20,000, more typically between 6,000 and 12,000, preferably between about 6,500 and about 10,000. For some applications, however, the molecular weight can be substantially lower, e.g., in the range of 1,000 to 3,000, or more particularly 1,500 to 2,000, especially in applications where the ratio of propylene oxide (PO) repeat units to ethylene oxide (EO) repeat units is relatively high.
For various embodiments, preferred alkoxylated oligo(alkylene imine) suppressors correspond to the structure:
wherein x is an integer between 0 and 4, y is an integer between 0 and 4, x+y is an integer between 2 and 6, R1 comprises an alkylene group, at least one of R1, R2, R3, R4, R5 and R6 comprises a polyether substituent comprising a block copolymer of propylene oxide and ethylene oxide. The ratio of ethylene oxide (EO) repeat units to propylene oxide (PO) repeat units in at least one polyether substituent is between 2:8 and 7:3, e.g., between 0.25:1 and 1.4:1, and each of the remainder of R1, R2, R3, R4, R5 and R6 is independently selected from the group consisting of hydrogen, lower alkyl, aminoalkyl, hydroxyalkyl, and a polyether substituent comprising propylene oxide (PO) repeat units, ethylene oxide (EO) repeat units, or a combination of PO and EO repeat units, and the number average molecular weight of the suppressor compound is between about 6,000 and about 12,000. Preferably, the composition comprises a leveler.
In particularly preferred embodiments of the electrolytic deposition composition, the value of x+y in the suppressor structure:
is between 3 and 6 inclusive, the ratio of propylene oxide (PO) to ethylene oxide (EO) repeat units in the block copolymer polyether substituents is between 3:7 and 7:3, and the number average molecular weight is between about 6,000 and about 20,000.
In further preferred embodiments, the value of x+y is between 2 and 6, the electrolytic composition comprises an oligomeric or polymeric dipyridyl leveler, the ratio of PO to EO is between 2:8 and 7:3, the molecular weight is between about 6,000 and about 20,000, and the PO/EO copolymer may be block, random, or other repeating pattern.
In an alternative preferred embodiment, the molecular weight is much lower, e.g., between about 1,000 and about 3,000 and the ratio of PO repeat units to EO repeat units is high, e.g., between about 1:1 and about 9:1.
Also disclosed herein are processes for electrolytically plating copper onto a substrate employing any of the foregoing compositions. More particularly disclosed are processes for electroplating a copper deposit onto a semiconductor integrated circuit device substrate with electrical interconnect features including submicron-sized features having bottoms, sidewalls, and top openings. The process comprises: immersing the semiconductor integrated circuit device substrate including submicron-sized features having bottoms, sidewalls, and top openings wherein the submicron-sized features include high aspect ratio features having aspect ratios of at least about 3:1 into an electrolytic plating bath having the composition described above, and supplying electrical current to the electrolytic composition to deposit copper onto the substrate and superfill the submicron-sized features by rapid bottom-up deposition at a rate of growth in the vertical direction which is greater than a rate of growth in the horizontal direction.
Preferably, the electrodeposition composition comprises an accelerator and also preferably comprises a leveler. Upon supply of electrical current to the electrolytic composition, copper deposits onto the substrate and superfills the submicron-sized features by rapid bottom-up deposition at a vertical copper deposition growth rate from the bottoms of the features to the top openings of the features which is greater than 50% faster than a comparable vertical copper deposition growth rate of comparable process which is equivalent in all respects except that it employs a commercially available suppressor.
Other objects and features will be in part apparent and in part pointed out hereinafter.
The present invention relates generally to compositions that are suitable for plating semiconductor integrated circuit substrates having challenging fill characteristics, including interconnect features that are poorly seeded or not substantially seeded, interconnect features having a complex geometry, and large diameter interconnect features as well as small diameter features (less than about 0.5 μm or even substantially smaller, e.g., less than 200 nm, less than 100 nm, less than 50 nm, less than 25 nm, less than 20 nm, less than 15 nm, or even less than 10 nm), and features with high aspect ratios (at least about 3:1) or low aspect ratios (less than about 3:1) where copper must fill all the features completely and substantially defect-free. The process is particularly useful for filling features that have an entry dimension of 5 to 20 nm and an aspect ratio greater than 3:1, e.g., between about 4:1 and about 10:1), and features with high aspect ratios (at least about 3:1) or low aspect ratios (less than about 3:1) where copper must fill all the features completely and substantially defect-free.
The compositions for copper superfilling of semiconductor integrated circuit substrates having challenging fill characteristics of the present invention generally comprise a suppressor compound and a source of copper ions. These compositions may also optionally, but preferably, comprise one or more levelers, one or more accelerators, and/or chloride. The above-listed additives find application in high copper metal/low acid electrolytic plating baths, in low copper metal/high acid electrolytic plating baths, and in mid acid/high copper metal electrolytic plating baths. The compositions described herein can also comprise other additives which are known in the art including halides, grain refiners, quaternary amines, and polysulfide compounds, by way of example and not limitation. Compositions comprising the suppressor, leveler, and accelerator described herein can be used to fill small diameter/high aspect ratio features.
Preferred suppressors for use in the copper plating compositions of the present invention comprise polyether groups covalently bonded to a cationic species. The cationic polyether suppressor preferably comprises a nitrogen atom. Exemplary cationic species comprising a nitrogen atom include primary, secondary, tertiary, and quaternary amines. By “cationic” what is meant is that the polyether suppressor either contains or can contain a positive charge in solution. Primary, secondary, and tertiary amines are weakly basic and become protonated and positively charged when added to a solution comprising an acid. Quaternary amines comprise four nitrogen-substituents, and a quaternized nitrogen possesses a positive charge regardless of the solution pH. The primary, secondary, tertiary, and quaternary amines can be substituted or unsubstituted alkyl amines, substituted or unsubstituted cycloalkyl amines, substituted or unsubstituted aromatic amines, substituted or unsubstituted heteroaryl amines, substituted or unsubstituted alkylether amines, and substituted or unsubstituted aromatic alkyl amines.
The suppressors used in the novel electrodeposition composition and process generally comprise alkoxylated oligo(alkylene imines), such as, for example, alkoxylated diethylene triamine or alkoxylated triethylene tetramine wherein poly(oxyalkylene) polyether groups are covalently bonded to a cationic nitrogen, and are preferably prepared by reaction of alkylene oxides with the oligo(alklene imine) substrate whose residue constitutes the core amine structure of the suppressor. The substrate amine preferably comprises at least three amine functional groups, more preferably at least four amine functional groups, and typically up to seven amine functional groups. More particularly, the substrate amine that is alkoxylated to form the suppressor can correspond to the structure:
wherein x is an integer between 0 and 4, y is an integer between 0 and 4, x+y is an integer between 2 and 6, R1 comprises an alkylene group, at least one of R1, R2, R3, R4, R5 and R6 comprises a polyether substituent comprising a block copolymer of propylene oxide and ethylene oxide. The ratio of ethylene oxide (EO) repeat units to propylene oxide (PO) repeat units in at least one polyether substituent is between 2:8 and 7:3, e.g., between 0.25:1 and 1.4:1, and each of the remainder of R1, R2, R3, R4, R5 and R6 is independently selected from the group consisting of hydrogen, lower alkyl, aminoalkyl, hydroxyalkyl, and a polyether substituent comprising propylene oxide (PO) repeat units, ethylene oxide (EO) repeat units, or a combination of PO and EO repeat units, and the number average molecular weight of the suppressor compound is between about 6,000 and about 12,000.
In one embodiment, the suppressor can correspond to the structure:
wherein x is an integer between 0 and 4, y is an integer between 0 and 4 and x+y is an integer between 2 and 6, R11 is alkylene, preferably at least two, more preferably at least three, of R12, R13, R14, R15 and R16 are hydrogen, and the remainder of R12, R13, R14, R15 and R16 are independently either hydrogen, alkyl, aminoalkyl or hydroxyalkyl. Preferably, each of R12, R13, R14, R15 and R16 in the oligo(alkylene imine) substrate is hydrogen. Alkoxylation of the amine substrate with propylene oxide (PO) and ethylene oxide (EO) provides the polyether substituents of the suppressor used in the novel electrodeposition process described herein. Particularly preferred amine substrates include diethylene triamine, triethylene tetramine, and tetraethylene pentamine, i.e., where x+y has a value of 2, 3 or 4. For many applications, triethylene tetramine is especially preferred.
In the polyether chain produced by copolymerization of both ethylene oxide monomer and propylene oxide monomers, it is generally preferred that the molar ratio of PO to EO is between 2:8 and 7:3, more preferably between about 0.25:1 and about 1.4:1. A relatively low PO/EO ratio enhances the solubility of the suppressor, increases the cloud point, and, in the course of the electrodeposition process, provides enhanced uniformity of the copper deposit. In particular, the presence of EO units helps minimize formation of defects such as swirl patterns in the copper deposit formed during wafer entry into the plating solution. A relatively high EO content also serves to mitigate the polarization strength which is greatly increased by the use of an alkoxylated diethylene triamine or triethylene tetramine as compared to a conventional alkoxylated ethylene diamine as the core amine. Without being bound to a particular theory, it is thought that the enhanced affinity for the negative polarity of the cathode that is provided by the polyamine core structure allows for a higher relative EO content in the polyether substituent, thus enhancing solubility in the aqueous medium as well. However, it is preferred that the EO content not be too high, because it may unduly compromise the desired enhancement in polarization strength provided by use of the higher amine core structure of the suppressor. In applications where it is desirable to modulate the polarizing effect of the suppressor, a PO/EO ratio in the range of 0.25:1 to 1.1:1, e.g., in the range of 1:3 to 1.0:1 or in the range of 3:7 to 6:5 may be preferred. Where somewhat stronger polarization is preferred, a suitable PO/EO ratio in the range of 4:6 to 6:4 or even 1.0:1 to 1.4:1.
In most embodiments, the number average molecular weight of the suppressor is preferably between about 6,000 and about 20,000, more preferably between about 6,000 and about 12,000, most preferably between about 6,500 and about 10,000, especially in those embodiments wherein the value of x+y is 2, i.e., where the core amine is diethylene triamine. However, in certain preferred embodiments, the molecular weight is between about 1,000 and about 3,000, more preferably between about 1,500 and about 2,000, and the PO/EO ratio is high, e.g., between about 1:1 and about 9:1, more preferably between about 6:4 and about 8:1, still more preferably between about 6:4 and about 5:1. An exemplary species of this embodiment has a molecular weight of about 1,700 and a PO/EO molar ratio of about 8:2. Another exemplary species corresponds to the structure:
which has a molecular weight of 1,700 and a PO/EO ratio of 3.1:1.
The polyether may comprise EO repeat units and PO repeat units in random, alternating, or block configurations. In a random configuration, the EO repeat units and PO repeat units have no discernible linear pattern along the polyether chain. In an alternating configuration, the EO repeat units and PO repeat units alternate according to some defined pattern, such as repeating units of EO-PO, PO-EO, and other alternating patterns. In the block configuration, the linear portion of the polyether chain comprises a block of EO repeat units bonded to a block of PO repeat units. The polyether chain may comprise a diblock. That is, the chain may comprise a first block of EO repeat units bonded to a second block of PO repeat units. Alternatively, the chain may comprise a first block of PO repeat units bonded to a second block of EO repeat units. In more complicated block configurations, the polyether chain may comprise a triblock (EO block-PO block-EO block or PO block-EO block-PO block), tetrablock, pentablock, or higher block arrangements. It has been discovered that a PO block-EO block-PO triblock configuration is effective to reduce polyether suppressor foaming in electrolytic solution. Where the polyether chain is in block copolymer configuration, each block of repeat units comprises between about 1 and about 30 repeat units, more preferably between about 7 and about 15 repeat units, especially in those embodiments in which the molecular weight of the suppressor is between 6,000 and 12,000. Where the suppressor molecular weight is in the 12,000 to 20,000 range and either PO or EO predominates (e.g., if the ratio is 2:8 or 7:3), a preferred length of a block comprising the predominant alkylene oxide may be somewhat greater, e.g., 15 to 25 repeat units. In a preferred embodiment involving a PO block-EO block-PO block tri-block configuration, the first PO-block bonded to the nitrogen (terminal interior block) comprises a single PO unit, or alternatively up to about 7 or as many as 15 PO repeat units, the second EO-block bonded to the PO-block comprises between about 7 and about 15 repeat units, and the third (terminal exterior) PO-block bonded to the second EO-block comprises between about 5 and about 20 repeat units. Where the molecular weight is above about 12,000, the exterior PO block and the intermediary EO block may be somewhat longer, e.g., each comprising 15 to 25 repeat units.
As noted above, the polyether may comprise only two blocks, e.g., a terminal interior PO block and a terminal exterior EO block, or may comprise a tri-block or a series of four or more blocks. To enhance solubility, provide a relatively high cloud point, and modulate suppression, the polyether substituents of the suppressor may, e.g., comprises a terminal exterior block comprising at least 5, more preferably at least 10, ethylene oxide (EO) repeat units bonded to a relatively more interior block comprising at least 5, more preferably at least 10, propylene oxide (PO) repeat units. Alternatively, to minimize foaming and ensure strong suppression, the polyether may comprise a terminal exterior block comprising at least 5, more preferably at least 10, propylene oxide (PO) repeat units bonded to a relatively more interior block comprising at least 5, preferably at least 10 ethylene oxide (EO) repeat units.
Optionally, the PO/EO polyethers are capped by a substituted or unsubstituted alkyl group, aryl group, aralkyl, or heteroaryl group. A preferred capping moiety for its ease of manufacture and low cost is a methyl group.
The suppressor compounds comprising polyether groups covalently bonded to a nitrogen comprise a positive charge in acidic solution and repeat units, EO and PO. It is thought that the separate functionalities of the positive charge, the EO repeat units, and the PO repeat units contribute different chemical and physical properties which affect, and thereby enhance, the function of the polyether as a suppressor in the copper plating compositions of the present invention. Without being bound to a particular theory, it is thought that the positive charge of the cationic species enhances the attraction of the suppressor compound to copper deposited into interconnect features, which, during an electrolytic plating operation, functions as the cathode. It is believed that the PO repeat unit is the more active repeat unit in the suppressors of the present invention. That is, the PO repeat unit has suppressor functionality and affects the quality of the copper deposit. Without being bound to a particular theory, it is thought that the PO repeat units, being relatively hydrophobic form a polarizing film over a copper seed layer and electrolytically deposited copper.
A copper seed layer may be deposited over the barrier layer in interconnect features by CVD, PVD, or other methods known in the art. The copper seed layer acts as the cathode for further reduction of copper that superfills the interconnects during the electrolytic plating operation. Copper seed layers can be thin (i.e., less than about 3 nm, such as between 1 and 3 nm). However, the copper thickness on the bottom or sidewall of features is typically much thinner than those on the feature top and unpatterned areas due to the non-uniform deposition rates of PVD processes.
In some extreme circumstances, the copper coverage on the bottom or sidewall can be so thin that the seed layer is discontinuous. Thus, in some instances the substrate comprises surface portions which have a copper seed layer thereon which is less than about 700 angstroms thick, and in some instances the seed layer is discontinuous.
In another case, the seed layer coverage on the top of features is thicker than on other feature areas, which is often called “seed overhang.” Generally, the uniformity of seed layer coverage degrades significantly with shrinking feature size and increasing aspect ratio. However, the inventors of the present invention have found that the present invention performs well, and better than the prior art, even with thin or overhanging seed layers.
The suppressor compound with somewhat hydrophobic PO repeat units covalently bonded to a nitrogen-containing cationic species is able to form a suppressive film over the copper seed layer. In the case of thin copper seed coverage, this polarizing organic film can cause the current to be more evenly distributed over the entire interconnect feature, i.e., the bottom and sidewalls of the via or trench. Even current distribution is believed to promote faster bottom up growth relative to sidewall growth, and may also reduce or eliminate bottom and sidewall voiding.
The strongly suppressive suppressor described herein is also desirable to suppress copper growth at the seed overhang areas on the top of the interconnect features, reducing the formation of internal voids from early pinching off.
The inventors of the present invention have discovered that the suppressor compound comprising a polyether group covalently bonded to a cationic species of the present invention is effective at suppressing copper deposition over thin or thick copper seed layers, and at enhancing nucleation density. In contrast, a polyether comprising of only PO repeat units, being relatively hydrophobic, lacks the solubility necessary to act as an adequate suppressor and lead to unacceptably high defectivity. That is, while PO is a superior suppressor, a polymer consisting of only of PO repeat units may not be soluble enough to go into the copper plating solution so that it can adsorb onto the copper seed layer in a high enough concentration to form a polarizing film. Accordingly, the polyether group preferably comprises EO repeat units to enhance its hydrophilicity and thus its solubility.
In those embodiments where the cationic species comprises a nitrogen atom, each nitrogen atom can be covalently bonded to one, two, or three PO/EO polyethers. Preferably, the nitrogen atom is covalently bonded to two PO/EO polyethers. In embodiments where the cationic species is a primary, secondary, or tertiary amine, the nitrogen atom can be alkylated to quaternize the nitrogen atom and render it positively charged. Preferably, the alkyl group is a short chain hydrocarbon radical having between 1 and 8 carbons, such as methyl, ethyl, n-propyl, isopropyl, and the like. Preferably the alkyl group is a methyl group. Accordingly, the nitrogen atom can form a quaternary amine having a positive charge where the suppressor comprises, for example, two PO/EO polyethers covalently bonded to a methylated alkylamine.
The preferred class of suppressors comprised by the electrolytic deposition composition used in the novel process correspond to the structure:
wherein x is an integer between 0 and 4, y is an integer between 0 and 4, x+y is an integer between 0 and 6, R1 comprises an alkylene group, and at least one of R2, R3, R4, R5 and R6 comprises a polyether substituent comprising a block copolymer of propylene oxide and ethylene oxide, and each of the remainder of R2, R3, R4, R5 and R6 is selected from the group consisting of hydrogen, lower alkyl, aminoalkyl, hydroxyalkyl and a polyether substituent comprising propylene oxide (PO) repeat units, ethylene oxide (EC)) repeat units, or a combination of PO and EO repeat units. The ratio of propylene oxide repeat units (PO) to ethylene oxide (EO) repeat units in the at least one polyether substituent, preferably in all polyether substituents and in the molecule as a whole, is generally between 2:8 and 7:3, more preferably between 4:6 and 6:4. In certain preferred embodiments, the PO/EO ratio is relatively low, i.e., between 3:7 and 6:5, or between 0.25:1 and 1.4:1, or between 0.25:1 and 1.1:1. Especially in species in which x+y of Structure I has a value of 3, optimally enhanced polarization can be realized at a PO/EO ratio between 1.1:1 and 1.4:1. In certain low molecular weight embodiments as described above, the PO/EO ratio is preferably between 1:1 and 9:1, more preferably between 6:4 and 8:1, most preferably between 6:4 and 5:1. One particularly preferred species has a molecular weight of 1,700 and a PO/EO ratio of approximately 8:2. In another particularly preferred species, x and y are both 0, the PO/EO ratio is between about 2:1 and about 1:1.3 and the species has a molecular weight of between about 4,500-6,000.
Preferably at least two of R2, R3, R4, R5 and R6 comprise a polyether comprising a block copolymer of propylene oxide (PO) and ethylene (EO), more preferably at least three, and most preferably each of R2, R3, R4, R5 and R6, comprises a PO/EO block copolymer. In each case, the PO/EO ratio is preferably between about 2:8 and about 7:3, more preferably between about 4:6 and about 6:4, or any of the other more finely tuned ratios described above.
The suppressor compounds described above can be present in an overall bath concentration between about 10 mg/L to about 1000 mg/L, preferably between about 50 mg/L to about 500 mg/L, more preferably about 75 to about 300 mg/L. Adding the weakly cationic polyether suppressors to Cu plating compositions within these concentration ranges is sufficient to fill complex features in an integrated circuit device, with the added benefits of reducing early pinching off, bottom voiding, or sidewall voiding.
Particularly preferred suppressors correspond to the following structure:
Relatively void-free deposits have also been provided using suppressors corresponding to the structures:
Among the suppressors of Structure VII to XIII, particularly preferred suppressors correspond to structure 11 to 17, respectively:
As noted above, alkoxylated amine groups of the suppressor by optionally be quaternized. To prepare the quaternized amine embodiments, the amines sites are first alkoxylated to provide polyether substituents having the PO/EO relationships and patterns as described above, after which the alkoxylated amine is react with a quaternizing agent such as an aliphatic or aromatic halide or sulfate to quaternize one or more of the amine sites. Generally amine sites separated only by a short alkylene group, e.g., where R1 is ethylene or propylene, are not both quaternized in the reaction. Thus, e.g., where x+y=3 or x+y=4, the quaternized suppressors may typically correspond to the structure:
wherein at least one of R2, R3, R4, R5 and R6 comprises a polyether substituent comprising a block copolymer of propylene oxide and ethylene oxide wherein the ratio of propylene oxide (PO) repeat units to ethylene oxide (PO) repeat units is between 0.25:1 and 1.4:1, each of the remainder of R2, R3, R4, R5 and R6 is selected from the group consisting of hydrogen, lower alkyl, aminoalkyl, hydroxyalkyl and a polyether substituent comprising propylene oxide (PO) repeat units, ethylene oxide (EO) repeat units, or a combination of PO and EO repeat units, each of R7 and R8 is selected from the group consisting of alkyl, aryl, aralkyl, alkenyl, and a proton, at least one of R7 and R8 is other than a proton, R9 is selected from the group consisting of alkyl, aryl, aralkyl, and alkenyl, and the number average molecular weight of the suppressor compound is between about 6,000 and about 12,000.
It has been discovered that electrolytic plating baths containing alkoxylated oligo(alkylene imine) suppressors in combination with accelerators as described elsewhere herein are capable of bottom-up superfilling of vias and trenches having an entry dimension less than 500 nm, or less than 200 nm, or less than 100 nm, or less than 50 nm. It has been found that use of the novel plating composition comprising the suppressor of Structure II is effective for superfilling of features having an entry dimension of less than 25 nm, or less than 20 nm, or less than 15 nm, or even less than 10 nm.
The compositions described herein also optionally, but preferably, include a leveler which can enhance leveling effect without substantially interfering with superfilling of copper into high aspect ratio features. Examples of suitable levelers can be found, for example in U.S. Pat. Pub. No. 2005/0045488 to Paneccasio et al., the subject matter of which is herein incorporated by reference in its entirety. It has been found that this type of leveler does not substantially interfere with superfilling, so the copper bath can be formulated with a combination of accelerator and suppressor additives which provides a rate of growth in the vertical direction which is substantially greater than the rate of growth in the horizontal direction, and even more so than in conventional superfilling of larger interconnects. The leveler is incorporated, for example, in a concentration between about 0.01 mg/L and about 25 mg/L, more preferably about 0.1 to about 15 mg/L, more preferably about 1.0 to about 6.0 mg/L.
A particularly preferred leveler comprises a dipyridyl polymer. In various embodiments, the leveler corresponds to the structure
In further and alternative embodiments, the leveler comprises the reaction product of:
With regard to accelerators, the accelerators may be bath soluble organic divalent sulfur compounds as described, for example in U.S. Pat. No. 6,776,893 to Too et al., the subject matter of which is herein incorporated by reference in its entirety.
In one preferred embodiment, the accelerator corresponds to the formula (10)
R1—(S)nRXO3M (10),
Where the leveler is a dipyridyl polymer, the polyether substituents can have a PO/EO sequence other than block, e.g., random, but the block copolymer configuration remains preferred.
In another preferred embodiment, the accelerator is 1-propanesulfonic acid, 3,3′-dithiobis, disodium salt according to the following formula (11):
The accelerator is incorporated typically in a concentration between about 0.5 and about 1000 mg/L, more typically between about 2 and about 100 mg/L, such as between about 50 and 90 mg/L.
The inventors of the present invention have found that the electrolytic compositions described herein and containing the described suppressors permit the use of a greater concentration of accelerator, and, in many applications, it must be used in combination with a higher concentration of accelerator than in conventional processes. This allows for enhanced rates of superfilling, as described, for example, in Example 7 below.
Optionally, additional leveling compounds can be incorporated into the bath, including, for example a reaction product of benzyl chloride and hydroxyethyl polyethylenimine as disclosed in U.S. Pat. Pub. No. 2003/0168343 to Commander et al., the subject matter of which is herein incorporated by reference in its entirety.
The accelerator, suppressor and leveler compositions described herein can be used in various combinations and at various concentrations to obtain the desired result of low defects, including the prevalence of very few to no voids. Thus, it is desirable that the combination of accelerator(s), suppressor(s) and leveler(s) in the bath produce a deposit that has substantially no voids, meaning that the deposit contains less than 75 voids per 5 μm2, more preferably less than 50 voids per μm2 and even more preferably less than 10 voids μm2.
The concentration and type of accelerator, suppressor and leveler can be optimized to produce the desired result. For example, good results have been obtained using a plating bath containing 75 to 225 mg/L of any of the suppressors described above, 0.1 to 50 mg/L of a dipyridyl leveler and 2-100 mg/L of an SPS accelerator. Good results have also been obtained using a plating bath containing 75 to 225 mg/L of a suppressor described above and having a number average molecular weight between about 1,000 and about 20,000, 0.1 to 25 m/g, 1.0 to 75 mg/L of a dipyridyl leveler, and 50-100 mg/L of an accelerator comprising an organic divalent sulfur compound.
The components of the copper electrolytic plating bath may vary widely depending on the substrate to be plated and the type of copper deposit desired. The electrolytic baths described herein include acid baths and alkaline baths. A variety of copper electrolytic plating baths are described in the book entitled Modern Electroplating, edited by F. A. Lowenheim, John Reily & Sons, Inc., pages 183-203 (1974). Exemplary copper electrolytic plating baths include copper fluoroborate, copper pyrophosphate, copper cyanide, copper phosphonate, and other copper metal complexes such as methane sulfonic acid. The most typical copper electrolytic plating bath comprises copper sulfate in an acid solution.
The concentration of copper and acid may vary widely, for example, from about 2 to about 70 g/L copper and from about 2 to about 225 g/L acid. In this regard the suppressors of Structure I are suitable for use in all acid/copper concentration ranges, such as high acid/low copper systems, in low acid/high copper systems, and mid acid/high copper systems.
In high acid/low copper systems, the copper ion concentration can be on the order of 4 g/L to on the order of 30 g/L; and the acid concentration may be sulfuric acid in an amount of greater than about 100 g/L up to about 225 g/L. In one high acid/low copper system, the copper ion concentration is about 17 g/L and the H2SO4 concentration is about 180 g/L.
In low acid/high copper systems, the copper ion concentration can be on the order of greater than about 30 g/L, greater than about 40 g/L, and even up to on the order of about 60 g/L copper (it is noted that 50 g/L copper corresponds to 200 g/L CuSO4·.5H2O copper sulfate pentahydrate). The acid concentration in these systems is less than about 50 g/L, less than about 40 g/L, and may even be less than about 30 g/L H2SO4, down to about 2 g/L. In one exemplary low acid/high copper system, the copper concentration is about 40 g/L and the H2SO4 concentration is about 10 g/L.
In mid acid/high copper systems, the copper ion concentration can be on the order of 30 g/L to on the order of 60 g/L, and the acid concentration may be sulfuric acid in an amount of greater than about 50 g/L up to about 100 g/L. In one mid acid/high copper system, the copper ion concentration is about 50 g/L and the H2SO4 concentration is about 80 g/L.
Chloride ions may also be used in the bath at a level up to 200 mg/L, preferably up to 100 mg/L, more preferably about 10 to 90 mg/L. Chloride ions are added in these concentration ranges to enhance the function of other bath additives, including accelerators, suppressors, and levelers. One preferred low copper/low acid electrodeposition bath contains about 5 g/L copper ion, about 10 g/L sulfuric acid and about 50 ppm chloride ion.
A large variety of additives may typically be used in the bath to provide desired surface finishes for the copper plated metal. Usually more than one additive is used with each additive forming a desired function. At least two additives are generally used to initiate bottom-up filling of interconnect features as well as for improved metal plated physical (such as brightness), structural, and electrical properties (such as electrical conductivity and reliability). Particular additives (usually organic additives) are used for grain refinement, suppression of dendritic growth, and improved covering and throwing power. Various additives used in electrolytic plating are discussed in a number of references including Modern Electroplating, cited above. A particularly desirable additive system uses a mixture of aromatic or aliphatic quaternary amines, polysulfide compounds, and polyethers. Other additives include ingredients such as selenium, tellurium, and sulfur compounds.
Plating equipment for plating semiconductor substrates are well known in the art as described, for example, in U.S. Pat. No. 6,024,856 to Haydu et al., the subject matter of which is herein incorporated by reference in its entirety. Plating equipment typically comprises an electrolytic plating tank which holds the copper electrolytic solution and which is made of a suitable material such as plastic or other material inert to the electrolytic plating solution. The tank may be cylindrical, especially for wafer plating. A cathode is horizontally disposed at the upper part of tank and the cathode may be any type of substrate such as a silicon wafer having openings such as trenches and vias. The wafer substrate is typically coated first with a barrier layer, which may be titanium nitride, tantalum, tantalum nitride, or ruthenium to inhibit copper diffusion and then with a seminal conductive layer, typically a seed layer of copper or other metal substrate for initiation of copper superfilling. A copper seed layer may be applied, for example, by chemical vapor deposition (CVD), physical vapor deposition (PVD), or the like. The anode may also be circular (i.e., for wafer plating) and be horizontally disposed at the lower part of tank forming a space between the anode and cathode. In one embodiment, the anode is a soluble anode including, for example, copper metal.
The electrolytic compositions comprising the bath additives described herein are also useful in combination with membrane technology. Thus, in one preferred embodiment, the anode may be isolated from the organic bath additives by a membrane in order to minimize the oxidation of the organic bath additives on the anode surface.
The cathode substrate and anode are electrically connected by wiring and, respectively, to a rectifier (power supply). The cathode substrate for direct or pulse current has a net negative charge so that copper ions in the solution are reduced at the cathode substrate forming plated copper metal on the cathode surface. An oxidation reaction takes place at the anode. The cathode and anode may be disposed horizontally or vertically in the tank.
Current is supplied from a power source having its negative terminal in electrical communication with the seminal conductive layer, typically a copper seed layer, in the submicron features. The positive terminal of the power source is in electrical communication with an anode in contact with said electrolytic solution.
During operation of the electrolytic plating system of the invention, copper metal is plated on the surface of a cathode substrate when the rectifier is energized. A pulse current, direct current, reverse periodic current, or other suitable current configuration may be employed. The temperature of the electrolytic solution may be maintained using a heater/cooler whereby electrolytic solution is removed from the holding tank and flows through the heater/cooler and then is recycled to the holding tank.
In the case of thin copper seed coverage, less current will be delivered to the lower portions of interconnect features, which may lead to bottom or sidewall voids and slow bottom-up growth. For features which have seed overhang, the electrolytic copper growth may have early pinching off on the feature tops before the bottom-up growth can reach the surface. Conventional suppressors may not distribute enough current to the bottom of the interconnect feature to promote bottom-up superfilling rapid enough to prevent the pinching off of interconnect features by copper electrolytic deposition leading to the formation of internal voids, especially for features seeded with a thin copper seed layer. Also, conventional suppressors may not have strong enough suppression to suppress copper growth on seed overhang areas to prevent early pinching off.
Without being bound to a particular theory, it is thought that the suppressor compounds of the present invention function to inhibit the formation of internal voids and enhance the bottom-up superfilling deposition rate by up to twice the rate over a typical electrolytic plating solution not comprising the suppressor compounds of the present invention by forming a polarizing film over the copper seed layer. Also, the suppressor compounds described herein possess stronger suppression (more polarizing) than most conventional suppressors, which allows the current to be distributed more evenly over the copper seed layer deposited on the bottom and sidewalls of the interconnect feature leading to the reduction or elimination of bottom and sidewall voids. An even current distribution enhances copper growth at the bottom of the feature relative to deposition at other regions to such an extent that bottom-up superfilling occurs so rapidly that deposition at the side and top of the feature will not cause a pinching off of the deposit and the formation of internal voids. The suppressor compounds of the present invention are effective at rapid bottom-up superfilling over thin or overhanged copper seed layers. For example, the suppressor compounds have been found effective to superfill an interconnect feature seeded with a thin copper seed layer on the bottom and side walls of an interconnect feature having a thickness between about 1 Angstrom and about 100 Angstroms.
An advantage of adding the suppressor compounds of the present invention to electrolytic copper plating solutions in the manner described herein is the reduction in the occurrence of internal voids as compared to deposits formed from a bath not containing these compounds. Internal voids form from copper depositing on the feature side walls and top entry of the feature, which causes pinching off and thereby closes access to the depths of the feature. This defect is observed especially with features which are small (e.g., less than about 100 nm) and/or which have a high aspect ratio (depth:width), for example, greater than about 4:1. Those voids left in the feature can interfere with electrical connectivity of copper interconnects. The suppressor compounds described herein appear to reduce the incidence of internal voids by the above-described rapid superfilling mechanism and strong suppression.
Optionally, the plating system of the invention may be controlled as described in U.S. Pat. No. 6,024,856 to Haydu et al., the subject matter of which is herein incorporated by reference in its entirety, by removing a portion of the electrolytic solution from the system when a predetermined operating parameter (condition) is met. Thereafter, new electrolytic solution is added to the system either simultaneously or after the removal in substantially the same amount. The new electrolytic solution is preferably a single liquid containing all the materials needed to maintain the electrolytic plating bath and system. The addition/removal system maintains a steady-state constant plating system having enhanced plating effects such as constant plating properties. With this system and method the plating bath reaches a steady state where bath components are substantially a steady-state value.
Electrolysis conditions, including, for example, electric current concentration, applied voltage, electric current density, and electrolytic solution temperature, are essentially the same as those in conventional electrolytic copper plating methods. For example, the bath temperature may be maintained at about room temperature such as about 20-27° C., but may be at elevated temperatures up to about 40° C. or higher. The electrical current density is typically up to about 100 mA/cm2, typically about 2 mA/cm2 to about 60 mA/cm2. It is preferred to use an anode to cathode ratio of about 1:1, but this may also vary widely from about 1:4 to 4:1. The process also uses mixing in the electrolytic plating tank which may be supplied by agitation or preferably by the circulating flow of recycle electrolytic solution through the tank. The flow through the electrolytic plating tank provides a typical residence time of electrolytic solution in the tank of less than about 1 minute, more typically less than 30 seconds, e.g., 10-20 seconds.
In a particular preferred electrodeposition schedule for filling features having an entry dimension less than 50 nm, the current density is maintained at at least 3.5 mA/cm2, more preferably at least 5 mA/cm2. The current can be ramped up during the electrodeposition cycle. For example, electrodeposition can be initiated at a current density of at least 5 mA/cm2 for 3 to 8 seconds, subsequently maintained at at least 7 mA/cm2 for a second period of 10 to 30 seconds, and thereafter maintained at at least 15 mA/cm2 for a period of at least 50 seconds. In one alternative, the current density is maintained at from 3 to 10 mA/cm2 for a first period and from 8 to 20 mA/cm2 for a second period in which the electrodeposition is concluded.
The following examples further illustrate the practice of the present invention.
An electrodeposition makeup solution was prepared comprising copper sulfate (5 g/L Cu++ ions), sulfuric acid (10 g/L), and chloride ion (50 mg/L). A dipyridyl polymer leveler corresponding to the structure:
was added to the makeup solution in a concentration of 1.4 mg/L.
From the makeup solution comprising the dipyridyl polymer leveler, a series of experimental electrodeposition baths were prepared by adding an SPS accelerator at concentrations of 61 mg/L or 82 mg/L, and either of four separate suppressors at concentrations of either 100 mg/L or 200 mg/L.
Suppressor 13 had a number average molecular weight of 7,100 and comprised triethylene tetramine substituted at each amine site with a block copolymer of propylene oxide (PO) and ethylene oxide (EO) repeating units in a PO:EO molar ratio of 0.57:
Suppressor 17 also had a molecular weight of 7,100 and comprised triethylene tetramine substituted at each amine site with a block copolymer of propylene oxide (PO) and ethylene oxide (EO) in a PO:EO molar ratio of 0.93.
Suppressor 14 had a number average molecular weight of 9,970 and comprised triethylene tetramine substituted at each amine site with a block copolymer of propylene oxide (PO) and ethylene oxide (EO) in a PO:EO molar ratio of 0.57.
Suppressor 18 had a number average molecular weight of 14,944 and comprised triethylene tetramine substituted at each amine site with a block copolymer of propylene oxide (PO) and ethylene oxide (EU) in a PO:EO molar ratio of 0.51:1
Suppressor 16 had a number average molecular weight of 7,100 and comprised triethylene tetramine substituted at each amine site with a block copolymer of propylene oxide (PO) and ethylene oxide (EO) in a PO:EO molar ratio of 0.32.
Using these plating solutions, twenty five separate electrodeposition runs were conducted in which submicron cavities in commercially available silicon wafers were superfilled. Each of the wafers comprised an isolated via chain region and a more dense via chain region, and the cavities had entry dimensions of 8 to 14 nm and an aspect ratio of 10:1 to 15:1. Ten of the runs were conducted at high initial current density (2.8 mA/cm2 for 4.5 seconds, followed by 10 mA/cm2 for 22 seconds, and 20 mA/cm2 for 78 seconds) while the remaining fifteen runs were conducted at low initial current density (1.6 mA/cm2 for 6 seconds, followed by 10 mA/cm2 for 20 seconds, and 20 mA/cm2 for 78 seconds).
Observations were made of the prevalence of center voids, sidewall voids, and missing metal in the deposits formed in each of the runs. Prevalence of voids was determined from top view inspections of both the isolated via chain region (ISO) and the more dense via chain region of the wafer. Compositions of the electrodeposition baths, current densities, and observation of voids are set forth in Table 1. Void counts for the dense region are graphically illustrated in
The data reported in Table 1 and illustrated in
From the makeup solution described in Example 1, a series of electrodeposition baths were prepared by adding suppressor 13 in a concentration of 200 mg/L, an SPS accelerator in a concentration of 82 mg/L, and varying proportions of either of five separate levelers. Structures of the levelers and the conditions under which levelers III, IV and V were synthesized are set out below:
n = 9-10
n = 7-8
Eleven separate electrodeposition runs were made superfilling submicron cavities in commercially available silicon wafers, each wafer comprising an isolated via chain region (ISO) and a more dense via chain region. The cavities had an entry dimension of 8 to 14 nm and an aspect ratio of 10:1 to 15:1.
Observations were made of the prevalence of center voids, sidewall voids, and missing metal in the copper deposits formed from each combination of accelerator concentration, suppressor concentration, leveler selection, and leveler concentration. Prevalence of voids was again determined by top view inspections of both the isolated via chain regions (ISO) and the more dense via chain regions of the wafers. Compositions of the electrodeposition baths, current densities, and observation of voids are set forth in Table 2. Void counts for the dense region are graphically illustrated in
V (19.6 mg/L)
V (25.2 mg/L)
V (30.8 mg/L)
When introducing elements of the present invention or the preferred embodiment(s) thereof, the articles “a”, “an”, “the” and “said” are intended to mean that there are one or more of the elements. For example, that the foregoing description and following claims refer to “an” interconnect means that there are one or more such interconnects. The terms “comprising”, “including” and “having” are intended to be inclusive and mean that there may be additional elements other than the listed elements.
As used herein, the term “about” refers to a measurable value such as a parameter, an amount, a temporal duration, and the like and is meant to include variations of +/−15% or less, preferably variations of +/−10% or less, more preferably variations of +/−5% or less, even more preferably variations of +/−1% or less, and still more preferably variations of +/− or less of and from the particularly recited value, in so far as such variations are appropriate to perform in the invention described herein. Furthermore, it is also to be understood that the value to which the modifier “about” refers is itself specifically disclosed herein.
As used herein, the term “immediately,” including “immediately before” or “immediately after” refers to a time period that is within one day, more preferably within several hours, more preferably within one hour, and still more preferably within several minutes.
As various changes could be made in the above without departing from the scope of the invention, it is intended that all matter contained in the above description and shown in the accompanying drawings shall be interpreted as illustrative and not in a limiting sense. The scope of invention is defined by the appended claims and modifications to the embodiments above may be made that do not depart from the scope of the invention.
The present application claims the benefit of U.S. Provisional Application Ser. No. 62/398,294, filed on Sep. 22, 2016, the subject matter of which is herein incorporated by reference in its entirety.
| Number | Date | Country | |
|---|---|---|---|
| 62398294 | Sep 2016 | US |
| Number | Date | Country | |
|---|---|---|---|
| Parent | 16334168 | Mar 2019 | US |
| Child | 18372236 | US |
