Copper separation during wafer-level encapsulation of integrated circuits

Description

FIELD OF THE INVENTION

[0001] The present invention relates generally to the electrodeposition of copper in the manufacture of electronic circuits and plating bath formulations in forming features in wafer level packaging (WLP) applications.BACKGROUND OF THE INVENTION

[0002] In order to take advantage of the progressively finer and denser architecture of integrated circuits, it is necessary to also provide corresponding ultra-miniaturization of semiconductor packaging. Among the structural requirements for this purpose are increases in the density of input / output transmission leads in an integrated circuit chip.

[0003] In flip chip packaging, the leads comprise bumps or pillars on a face of the chip, and more particularly on the side of the chip that faces a substrate, such as a printed circuit board (PCB), to which the circuitry of the chip is connected.

[0004] Input and output pads for flip chip circuitry are often provided with solder bumps through which the pads are electrically connected to circuity external to the chip, such as the circuits of a PCB or another integrated circuit chip. Solder bumps are provided from relatively low melting point base metals and base metal alloys comprising metals such as lead, tin, and bismuth. Alloys of base metals with other electrically conductive metals, such as Sn / Ag alloys are also used.

[0005] In the manufacture of the packaged chip, the bumps are provided as globular molten beads on the so-called under bump metal of the pad, and allowed to solidify in place to form the electrical connector through which current is exchanged between the chip and the external circuit. Unless subjected to lateral or vertical constraint during solidification, solder bumps generally assume a spherical form. As a result, the cross-sectional area for current flow at the interface with the under bump metal or pad may depend on the wettability of the under bump structure by the solder bump composition. Absent external constraints on the extent of lateral growth, the height of the bump cannot exceed its lateral dimension, and is diminished relative to the height as wettability of the under bump metal by the molten solder increases. Thus, dimensions of an unconstrained solder bump are determined mainly by the surface tension of the molten solder, the interfacial tension between the solder and the under bump metal, and the extent to which the volume of the solder drop can be controlled in operation of the solder delivery mechanism used in the process.

[0006] In an array of solder bumps formed on the face of an integrated circuit chip, these factors may limit the fineness of the pitch, i.e., the distances between the centers of immediately neighboring bumps in the array.

[0007] In order to achieve a finer pitch, attempts have been made to substitute copper bumps or pillars for the solder by electrodeposition onto the under bump metal. However, it can be difficult to control the electrodeposition process to provide a copper pillar having the desired configuration. While the shape of the main body of the pillar can be determined by forming it within the confines of a cavity having sidewalls formed from a dielectric material, the configuration of the distal end of the pillar may still be unsatisfactory, e.g., excessively domed, excessively dished, or irregular.

[0008] By comparison with the provision of solder bumps, manufacturing of copper pillars can suffer a further disadvantage in productivity, and in the effect of productivity on manufacturing cost. While a drop of molten solder can be delivered almost instantaneously once a delivery head is brought into registry with the under bump metal, the rate of electrodeposition of a copper pillar is limited by the maximum current density that can be achieved in the electrodeposition circuit. In commercial practice, the current density is limited by various configuration problems, including the problems of doming, dishing, and irregular configuration at the distal end of a copper pillar, which are aggravated if the current density rises above a limiting value, for example, about 40 A / dm 2< , depending on the application, corresponding to a vertical growth rate of no greater than about 7µm / min.

[0009] Although copper bumps and pillars have substantial advantages over tin / lead solder bumps, a small bead of solder is still used in the manufacturing process to bond the end of the bump or pillar to external circuity such as the circuit traces of a PCB. However, to assure proper bonding of copper to the solder, and to prevent formation of Kirkendall voids (i.e., voids formed at the boundary interface in various kinds of alloys to bonding) at the copper / solder interface that may result from migration of copper into the solder phase, it has been necessary to provide a nickel cap on the distal end of the bump or pillar as a barrier between the copper phase and the solder phase, thus adding to the expense and complication of the manufacturing process.

[0010] WO 2017 / 139087 A1 comprises part of the state of the art under the provisions of Article 54(3) EPC and discloses an aqueous electrolytic composition and a process for electrodeposition of copper on a dielectric or semiconductor base structure using the aqueous electrolytic composition. The process includes (i) contacting a metalizing substrate comprising a seminal conductive layer on the base structure with an aqueous electrolytic deposition composition; and (ii) supplying electrical current to the electrolytic deposition composition to deposit copper on the substrate. The aqueous electrolytic composition comprises: (a) copper ions; (b) an acid; (c) a suppressor; and (d) a quaternized poly(epihalohydrin).

[0011] US 4,336,114 discloses a composition and method for electrodepositing copper deposits from an aqueous acidic copper plating electrolyte containing a brightening and leveling amount of a brightening and leveling system comprising (a) a bath soluble substituted phthalocyanine radical, (b) a bath soluble adduct of a tertiary alkyl amine with polyepichlorohydrin, (c) a bath soluble organic divalent sulfur compound, and (d) a bath soluble reaction product of polyethyleneimine and an alkylating agent which will alkylate the nitrogen on the polyethyleneimine to produce a quaternary nitrogen.

[0012] US 4,555,315 discloses an improved electrolyte composition and process for electrodepositing copper deposits on a substrate. The electrolyte contains an additive system comprising controlled relative concentrations of: (a) a bath soluble polyether compound; (b) a bath soluble organic divalent sulfur compound; (c) a bath soluble adduct of a tertiary alkyl amine with epichlorohydrin; and (d) a bath soluble reaction product of polyethyleneimine and an alkylating agent.SUMMARY OF THE INVENTION

[0013] The present invention provides an electrodeposition composition according to claim 1. Optional or preferred features are defined in the dependent claims.

[0014] Described herein are improved compositions and processes for electrodeposition of wafer level packaging (WLP) features. Particularly effective for this purpose is a process in which a feature such as a bump, pillar, or megabump is deposited from a plating solution that comprises a source of copper ions, an acid, a leveler, a suppressor, and an accelerator.

[0015] The process described herein is effective for forming WLP features on semiconductor substrates for interconnection of an electronic circuit of a semiconductor device with a circuit external to the device. This includes, for example, features having diameters as small as about 1 µm or as large as about 240 µm, and heights as small as about 2 µm and as large as about 300 µm; with aspect ratios between about 1:1 and about 2:1, and up to about 10:1 or even up to about 20:1. In one implementation of the process, current is supplied to an electrolytic solution in contact with a cathode comprising a WLP structure such as an under bump structure on a semiconductor assembly, wherein the electrolytic solution comprises a source of copper ions, an acid, a leveler, and a suppressor, thereby electrodepositing a copper bump or pillar on the under bump structure.DETAILED DESCRIPTION

[0016] As described herein, a copper bump or pillar having a suitable distal configuration is deposited at a relatively high rate of vertical growth. By "suitable distal configuration" what is meant is that the copper bump or pillar is not unduly domed, unduly dished, or irregular in shape. The rate of growth of bumps and pillars having suitable distal configurations compares favorably with the rate that is achieved using electrodeposition baths that do not involve the composition and process described herein.

[0017] The electrolytic plating bath generally comprises: a) a source of copper ions; b) an acid component; c) a suppressor; d) a leveler, wherein the leveler comprises a quaternized poly(epihalohydrin), and e) an accelerator. In accordance with the present invention, a) to e) are as defined in claim 1.

[0018] The source of copper ions in the electrolytic plating bath may be any of a variety of watersoluble copper salts, including, for example, copper sulfate or a copper salt of an alkanesulfonic acid such as methane sulfonic acid or ethane sulfonic acid.

[0019] The acid component is preferably sulfuric acid or an alkane sulfonic acid, such as methane sulfonic acid or ethane sulfonic acid, most preferably methane sulfonic acid. Preferably, the conjugate base of the acid is the same as the counteranion of the copper salt, but mixtures can also be effective, including, for example, copper sulfate and methane sulfonic acid, or copper methane sulfonate and sulfuric acid.

[0020] The concentration of copper and acid may vary over wide limits, and may be for example, from about 20 to about 100 g / L copper and from about 40 to about 300 g / L acid. In many embodiments, the copper ion concentration can be greater than about 30 g / L, 40 g / L, and even up to on the order of 60 or 80, such as from about 25 to about 100 g / L copper (50 g / L copper corresponds to 200 g / L CuSO 4 ·5H 2 O Cu sulfate pentahydrate). The acid concentration in these systems is on the order of about 50 to about 300 g / L, preferably about 80 to about 220 g / L in some embodiments.

[0021] Chloride ions may also be used in the bath at a level up to 100 mg / L, preferably within the range of about 10 to 90 mg / L, more preferably in the range of 30 to 80 mg / L. The addition of chloride ions at these concentration ranges enhances the function of other bath additives. These other bath additives include accelerators, suppressors, and levelers. Table 1 describes various make-up compositions with varying copper and acid concentrations: Table 1. Sample bath make-up compositionsCu 2+< (g / L)Acid (g / L)Cl (mg / L)132.0180 H 2 SO 4 50240.0140 H 2 SO 4 50350.0100 H 2 SO 4 50455.0150 H 2 SO 4 50-80562.5100 H 2 SO 4 506*80.0100 MSA50740.0100 MSA50*Cu 2+< in 5 is from CuMSA; Cu 2+< for others is from CuSO 4 .

[0022] The electrodeposition composition contains a leveler compound in a suitable concentration, such as between about 1.0 and about 100 mg / L, more preferably, between about 5.0 and about 50 mg / L.

[0023] The levelers are quaternized poly(epihalohydrin) polymers, as defined more specifically in claim 1.

[0024] The electrodeposition composition may use as levelers, in the electrodeposition of copper bumps, pillars, and other WLP features, quaternized poly(epihalohydrin) polymers as described in WO 2017 / 139087 A1 (claiming priority from US provisional application Serial No. 62 / 294,643) which comprises part of the state of the art under the provisions of Article 54(3) EPC. In the referenced application and herein, such polymers are referred to by the acronym QPECH.

[0025] The QPECH polymers that are used in the electrodeposition composition of this invention comprise n repeating units corresponding to structure 1N and p repeating units corresponding to structure 1P: wherein Q has a quaternized amine structure selected from the group consisting of: (i) NR a< R b< R c< wherein one of R a< , R b< and R c< is selected from the group consisting of unsubstituted alkyl having at least three carbon atoms, hydroxyalkyl, dihydroxyalkyl, phenyl, hydroxyphenyl, dihydroxyphenyl, benzyl, hydroxyphenylmethyl and dihydroxyphenylmethyl, and each of the others of R a< , R b< and R c< is independently selected from lower alkyl having between 1 and 3 carbon atoms; (ii) an N-substituted and optionally further substituted heteroalicyclic amine wherein the N-substituent is selected from the group consisting of substituted or unsubstituted alkyl, alicyclic, aralkyl, aryl, and heterocyclic; and (iii) a substituted or unsubstituted nitrogen-containing heteroaryl compound; n is an integer between 3 and 35, p is an integer greater than 0 and up to 25, X is a halo substituent; and X -< is a monovalent anion, ordinarily a halide ion. When there are polar substituents on the quaternized amine structure or the quaternized amine structure contains aryl or aralkyl groups, the ratio of n / (n+p) is between 0.40 and 0.60, and otherwise the ratio of n / (n+p) is between 0.70 and 1.0. Where any of R a< , R b< and R c< is substituted, the substituent preferably does not comprise an amino group.

[0026] Examples of suitable QPECH polymers include, but are not limited to: where n and p are as defined above.

[0027] The structure and properties of the quaternized poly(epichlorohydrin) are compatible with other components of the electrolytic composition, especially with respect to solubility in the overall composition. Compatibility is a function of various structural parameters, including the extent of quaternization, the presence or absence of repeating units other than 1N and 1P, and molecular weight. By selection of, for example, the molecular weight of the core PECH, the amine for the quaternization reaction, the extent of quaternization, the optional use of co-monomers in polymerizing epihalohydrin, the identity of any co-monomers, and fraction of co-monomers, the composition can be tailored to assure compatibility while imparting the properties that may be optimal for a particular application. Preferably Q in structures I, 1P, and 1N corresponds to structure 2A, 2B, or 2C: or or wherein: (i) structure 2B is an N-substituted heterocyclic moiety; (ii) structure 2C is a heterocyclic moiety; (iii) each of R a< , R b< , R c< and R d< is independently selected from the group consisting of substituted or unsubstituted alkyl, substituted or unsubstituted alkenyl, substituted or unsubstituted alkynyl, substituted or unsubstituted aralkyl, substituted or unsubstituted alicyclic, substituted or unsubstituted aryl, and substituted or unsubstituted heterocyclic; and (iv) each of R e< , R f< , R g< , R h< and R j< is independently selected from the group consisting of hydrogen, substituted or unsubstituted alkyl, substituted or unsubstituted alkenyl, substituted or unsubstituted alkynyl, substituted or unsubstituted aralkyl, substituted or unsubstituted alicyclic, substituted or unsubstituted aryl, and substituted or unsubstituted heterocyclic. Where any of R a< to R j< is substituted, the substituent preferably does not comprise an amino group.

[0028] In preferred embodiments of the electrodeposition composition used in the electrodeposition process described herein, Q in structure I is a quaternized amine and corresponds to that which may be obtained by reacting a pendent methylene halide group of poly(epihalohydrin) with a tertiary amine selected from the group consisting of N-methylmorpholine, 1-methylimidazole, 1-benzylimidazole, 2-imidazoline, 3-imidazoline, and NR a< R b< R c< wherein one of R a< , R b< and R c< is selected from the group consisting of unsubstituted alkyl having at least three carbon atoms, hydroxyalkyl, dihydroxyalkyl, benzyl, hydroxyphenylmethyl and dihydroxyphenylmethyl, and each of the others of R a< , R b< and R c< is independently selected from lower alkyl having between 1 and 3 carbon atoms.

[0029] Preferably, Q is a quaternized amine and corresponds to the structure which may be obtained from reacting a pendent methylene halide group with a tertiary amine selected from the group consisting of 3-hydroxypropyldimethylamine, n-butyldimethylamine, di(3-hydroxypropyl)methylamine, 2,3-dihydroxypropyldimethylamine, 3-hydroxypropyldiethylamine, 2-hydroxypropyldimethylamine, 4-hydroxybutyldimethylamine, 2-hydroxyethyldimethylamine, n-propyldimethylamine, 2-hydroxyethoxyethyldimethylamine, di(2-hydroxyethyl)methylamine, benzyldimethylamine, and 4-hydroxybenzyldimethyleamine, 4-methylpyridine, 3-ethylpyridine, 4-propylpyridine, 4-tertbutylpyridine, 4-cyanopyridine, 4-isopropylpyridine, 4-methoxypyridine, 3,4-lutidine, 3-methoxypyridine, and 4-pyridinemethanol.

[0030] Particularly useful tertiary amines subject to quaternization comprise N,N-dimethylalkanol amines such as 3-hydroxypropyl dimethylamine and 2-dimethylamino-1-ethanol. Other preferred tertiary amines include n-butyl dimethylamine, and N,N-dimethylbenzylamine, 4-ethylpyridine, and 1-methylimidazole, 1-benzylimidazole, N-methylmorpholine, and particularly 2-[2-(dimethylamino)ethoxy]ethanol.

[0031] In addition to quaternized epihalohydrin and unquaternized epihalohydrin, the backbone of the quaternized poly(epihalohydrin) polymer (" QPEHH" or "QPECH" where the epihalohydrin is epichlorohydrin) may optionally include repeating units that are residues of alkylene oxides such as ethylene oxide, propylene oxide, butylene oxide, 3,4-epoxy-1-butanol, 2,3-epoxy-1-propanol, or glycidol. Polymers of this structure may be prepared by co-polymerizing one or more alkylene oxides and ephihalohydrin, followed by quaternization of epihalohydrin repeating units. The distribution of epihalohydrin units and alkylene oxide units in the backbone may be random, block, or other pattern. Where plural alkylene oxides are used in the polymerization, the backbone comprises repeating units that are residues of more than one alkylene oxide.

[0032] For convenience and familiarity of nomenclature, the description of the electrodeposition composition and method hereinafter will refer to "QPECH" levelers. However, it will be understood that, unless otherwise stated, the description encompasses polymers derived from other epihalohydrins, including principally epibromohydrin. The use of epichlorohydrin is highly preferred.

[0033] Where the number of alkylene oxide units in the QPECH backbone is q, the ratio of q / n+p+q is preferably not greater than 0.05 in those embodiments where the properties provided by presence of alkylene oxide repeat units is not needed; and typically between about 0.05 and about 0.50, more typically between about 0.10 and about 0.40 where their presence is warranted. In certain preferred embodiments, q is essentially zero and the QPECH consists essentially of repeating units that are residues of epihalohydrin and repeating units that are residues of quaternized epihalohydrin. In such embodiments, the poly(epihalohydrin) may be depicted by the structure: where Q is as described above.

[0034] Regardless of the presence or absence of alkylene oxide repeat units and regardless of the value of q, the ratio of n / (n+p) can have a bearing on the relative efficacy of the leveler. In embodiments of the invention where there are polar substituents on the quaternized tertiary amine, e.g., 3-hydroxypropyl dimethylamine, the ratio of n / (n+p) is between 0.40 and 0.60. This limitation also applies where the quaternized tertiary amine contains aryl or aralkyl groups. In other embodiments wherein the tertiary amine is more hydrophobic, for example, n-butyl dimethylamine, the value of n / (n+p) is between 0.70 and 1.0, preferably at least 0.75, still more preferably at least 0.80, e.g., between 0.80 and 0.95, and most preferably about 0.90.

[0035] Where the poly(epichlorohydrin) contains alkylene oxide units, and there are polar substituent on the quaternary nitrogen and q / n+p+q is at least 0.05, n / n+p+q is preferably at least 0.20, or between 0.40 and 0.60. Where the substituents on the quaternary nitrogen are hydrophobic e.g., where they are all hydrocarbyl groups, n / n+p+q is preferably between 0.60 and 0.95, more preferably between 0.70 and 0.9, most preferably between 0.75 and 0.90.

[0036] Regardless of whether the repeating units include alkylene oxide residues, the chain of repeating units of the polymer may optionally be bonded to a residual oxygen of an alcohol. For example, plural QPECH chains can be bonded to residual oxygens of a monomeric polyol as described in U.S. Pat. Pub. No. 2006 / 0062753 A1 to Naraghi. The preparation of such compositions comprises the polymerization of epihalohydrin and the condensation of the polymer with the hydroxyl groups of glycerin, as catalyzed, for example, by boron trifluoride. Once the PECH / polyol adduct is formed, epihalohydrin repeating units can be quaternized with an appropriate tertiary amine.

[0037] Average molecular weight of the QPECH may vary significantly depending, for example, on the molecular weight of the amine to be quaternized, the fraction of repeating units that comprise quaternary ammonium groups, and the extent to which the polymer backbone comprises repeating units derived from a co-monomer such as an alkylene oxide. These structural features are combined to enhance the polarization imparted by the QPECH while preserving compatibility of the compound with the electrolytic solution, thus achieving a favorable relationship between efficacy and solubility of the QPECH in the solution. Generally, solubility varies inversely with molecular weight while the extent of quaternization enhances solubility at a given molecular weight. The PECH with which the tertiary amine is reacted may have a weight average molecular weight that ranges, for example, between about 300 and about 4,000 but is preferably in the range of about 1,700 to about 2,200. After reaction of the PECH or ECH / alkylene oxide copolymer with the tertiary amine, the weight average molecular weight of the quaternized polymer may typically range from about 500 to about 10,000, or even higher depending on the overall structure of the QPECH and the overall composition of the electrodeposition bath, more preferably between about 800 and about 6,000, or between about 1,200 and about 3,000, or still more preferably between about 2,500 and about 4,000.

[0038] Advantageously, the synthesis of the QPECH can be integrated into the formulation of the electrodeposition composition. Conventionally, a QPECH is formed by condensation of pendent methylene halide groups of poly(epihalohydrin) with a tertiary amine in an aqueous medium. Because the temperatures required for the quaternizing reaction are relatively high, in excess of about 95°C, for example in the range of about 95° to about 100°C, quaternization in an aqueous medium requires the reaction to be conducted in a pressure vessel and the reaction mass cooled before it is removed from the reactor, thereby limiting productivity and / or requiring a high pressure heat exchanger to cool any of the reaction mass that may be removed immediately after the reaction.

[0039] In the integrated process for formulating the electrodeposition bath, the quaternization reaction is conducted in a medium comprising a relatively high boiling polar organic solvent, preferably having a boiling point of at least about 120°C, for example, between about 120° and about 220°C, more preferably between about 160° and about 200°C. The use of a high boiling solvent allows the reaction to be conducted at high temperature but ambient pressure, the reaction mass to be removed from the reactor immediately on completion of the reaction cycle, and the reaction mass to be cooled by flow of heat transfer fluid through coils in an atmospheric reaction vessel, or in withdrawal from the reactor through an external heat exchanger which need be pressurized only to the extent incident to fluid flow.

[0040] A solution of poly(epihalohydrin) in the polar solvent is initially prepared having a PECH concentration that may typically fall in the range between about 10 and about 35 percent by weight. A tertiary amine is added in a proportion governed by the extent of quaternization that is to be achieved. The quaternization reaction proceeds at a temperature that is normally in the range of about 160° to about 190°C.

[0041] Exemplary polar solvents for the quaternization include, for example, ethylene glycol, propylene glycol, glycerin, ethylene glycol monomethyl ether, ethylene glycol monoethyl ether, ethylene glycol monopropyl ether, and dipropylene glycol monomethyl ether. In a batch reaction system, the reaction cycle is generally between about 0.5 and about 3 hours.

[0042] Where polyol / QPECH adducts having the structure of those described in U.S. Pat. Pub. No. 2006 / 0062753 A1 to Naraghi are prepared, epihalohydrin may be polymerized in the presence of a polyol and, typically, a polymerization catalyst such as BF3. The resulting adduct of polyol and PECH is then reacted with a tertiary amine in a polar solvent medium. Where epihalohydrin is polymerized in the presence of a sufficient excess of polyol, unreacted polyol may optionally function thereafter as the solvent for the subsequent quaternization reaction.

[0043] Once the reaction is complete and the quaternization reaction mass has been cooled, the reaction mass can be directly combined in an aqueous electrolytic plating bath with the copper salt, suppressor, chloride ion. Optionally, depending on the size and geometry of the base structure, an accelerator may also be included. It is not necessary to recover the QPECH from the quaternization reaction medium. In these embodiments, the plating composition contains the polar organic solvent, typically in a concentration between about 5 and about 50 mg / l. The ratio of leveler to polar organic solvent in the electrodeposition solution is typically between about 0.2:1 to about 1.5:1, more typically about 1:1.

[0044] In other embodiments of the process, a polar solvent is not necessarily present. For example, if the QPECH is acquired in neat form from an independent source, there may be no need or occasion for the presence of a polar solvent in the electrolytic plating bath.

[0045] The electrodeposition composition also contains a suppressor. The suppressor component may be an alkoxylated amine, wherein the alkoxylated amine is selected from the group consisting of a tetraalkoxylated alkylene diamine, an alkoxylated diethylene triamine, a tetraalkoxylated triethylene tetraamine, and combinations of two or more of the foregoing. Alternatively, the suppressor component may be a polyether compound comprising a combination of propylene oxide (PO) repeat units and ethylene oxide (EO) repeat units present in a PO:EO ratio between 1:9 and 9:1 and covalently bonded to a nitrogen-containing cationic species selected from the group consisting of primary, secondary, tertiary, and quaternary amines.

[0046] Particularly preferred suppressors for the electrodeposition compositions used in the process described herein comprise polyether groups covalently bonded to a cationic species which comprises a nitrogen atom, the cationic species comprising a nitrogen atom selected from the group consisting of 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.

[0047] The suppressors comprising polyether groups covalently bonded to a cationic species comprise at least one amine functional group, preferably between two amine functional groups and four amine functional groups. Accordingly, the cationic species can be an amine, a diamine, a triamine or a tetraamine. The alkyl group of the alkylamine can be a substituted or unsubstituted alkyl, preferably a short chain hydrocarbon having between 1 and 8 carbons, which may be branched or straight chained. Exemplary alkylamines include, for example, methylamine, ethylamine, propylamine, n-butylamine, isobutylamine, t-butylamine, ethylenediamine, diethylenetriamine, 1,3-diaminopropane, 1,4-diaminobutane, 2-butene-1,4-diamine, and others, 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 cycloalkyl group of the cycloalkyl amine typically comprises a 5- or 6-carbon ring, although bicylic, tricylic, and higher multi-cyclic alkyl amines are usable in the present invention. Exemplary cycloalkyl amines include, for example, substituted or unsubstituted cyclopentylamines, cyclohexylamines, cyclopentyldiamines, cylcoalkyltriamines, and higher cycloalkyl amines. Alkylether amines preferably comprise an ether moiety defined by short chain hydrocarbons typically having between 1 and 8 carbons, such as diethylene glycol diamine and triethylene glycol diamine.

[0048] The suppressor comprises a cationic species in which the polyether chain is bonded directly to the nitrogen of an amine or other nitrogen-containing species. The epoxide monomers are selected from a combination of ethylene oxide monomer and propylene oxide monomer. The polyether comprises a chain of repeat units formed by the polymerization of both ethylene oxide monomer and propylene oxide monomer. Accordingly, the ratio of ethylene oxide (EO) repeat units and propylene oxide (PO) repeat units in the polyether is between 1:9 and 9:1. In especially preferred embodiments, the ratio is between 2:3 and 2:1, more preferably about 3:2. In certain preferred embodiments, the polyether comprises between 1 and 30 EO repeat units and between 30 and 1 PO repeat units, such as between 7 and 15 EO repeat units and between 15 and 7 PO repeat units. In one embodiment, the polyether comprises, for example, about 11 EO repeat units and about 13 PO repeat units. In other preferred embodiments, the polyether comprises about 7 or 8 EO repeat units and about 9 PO repeat units. Accordingly, the molecular weight of the polyether can be between 100 g / mol and 20,000 g / mol, more preferably between 3500 and 15,000 g / mol, more preferably between 6000 and 8000 g / mol.

[0049] The polyether preferably comprises 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 discernable 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. The co-polymer can be arranged in a block configuration. 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. In one embodiment of the block configuration, each block of repeat units comprises between 1 and 30 repeat units, more preferably between 7 and 15 repeat units. In a preferred embodiment involving a PO block-EO block-PO block tri-block configuration, the first PO-block bonded to the cationic species comprises between 7 and 15 PO repeat units, the second EO-block bonded to the PO-block comprises between 7 and 15 repeat units, and the third PO-block bonded to the second EO-block comprises between 1 and 5 repeat units.

[0050] 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.

[0051] The suppressor compounds comprising polyether groups covalently bonded to a cationic species 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 copper plating compositions useful in the electrodeposition process.

[0052] 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 active repeat unit in the suppressors that are useful in the electrodeposition process. 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.

[0053] Where the polyether substituent is a block copolymer of PO and EO, it is particularly preferred that a PO block is bonded to a nitrogen of an amine or other nitrogen-containing species. The combination of the positively charged nitrogen atom and the hydrophobic PO block enhances the affinity of the suppressor for the cathodic copper substrate. This is believed to enhance the suppressive effect of the alkoxylated amine suppressor. An especially effective suppressor comprises a hexaalkoxylated triethylene tetraamine as follows: where

[0054] Another similar suppressor contains: .

[0055] Another suppressor of interest is the following triamine: where

[0056] Preferably, the electrodeposition composition contains between 100 and 1,000 mg / L, more preferably between 200 and 600 mg / L suppressor.

[0057] The electrodeposition composition also contains an accelerator. The accelerator may include an organic sulfur compound. Preferred organic sulfur compounds include water soluble organic divalent sulfur compounds. In one embodiment, the organic sulfur compound has the following general structure (11): wherein: X is O or S, preferably S; n is 1 to 6; M is hydrogen, alkali metal, or ammonium as needed to satisfy the valence; R 1 is an alkylene or cyclic alkylene group of 1 to 8 carbon atoms, an aromatic hydrocarbon or an aliphatic aromatic hydrocarbon of 6 to 12 carbon atoms; and R 2 is selected from the group of MO 3 SR 1 wherein M and R 1 are as defined above, a thiocarbamate represented by the formula: a xanthate represented by the formula: and an aminoimine represented by the formula: wherein R 3 , R 4 , R 5 , R 6 , and R 7 are independently hydrogen, an alkyl group of 1 to 4 carbon atoms, a heterocyclic group, or an aromatic group. In one preferred embodiment, X is Sulfur, and n is two.

[0058] A preferred organic sulfur compound of Structure (1) has the following general structure (12): wherein M is a counter ion possessing charge sufficient to balance the negative charges on the oxygen atoms. M may be, for example, protons, alkali metal ions such as sodium and potassium, or another charge balancing cation such as ammonium or a quaternary amine.

[0059] One example of the organic sulfur compound of structure (2) is the sodium salt of 3,3' dithiobis(1-propanesulfonate), which has the following structure (13):

[0060] An especially preferred example of the organic sulfur compound of structure (2) is 3,3'-dithiobis(1-propanesulfonic acid), which has the following structure (14):

[0061] Another preferred organic sulfur compound of Structure (1) has the following general structure (5):

[0062] One example of the organic sulfur compound of structure (l6) is 3-(dimethylcarbamothioylthio)propane-1-sulfonic acid:

[0063] Another suitable sulfur compound comprises the following structure:

[0064] The organic sulfur compound may be added in a concentration between 1 mg / L and 100 mg / L (ppm), such as between 25 mg / L and 70 mg / L.

[0065] In other embodiments, the accelerator may be an oxidized version of one of the foregoing accelerator compounds.

[0066] The process described herein is useful for building copper bumps and pillars in flip chip packaging and for other wafer-level packaging features such as through silicon vias and redistribution layers (RDLs) and processes directed to the manufacture of integrated circuits. In wafer level packaging, an array of copper bumps or pillars is provided over a semiconductor substrate for interconnection of an electric circuit of a semiconductor device with a circuit external to the device, for example, to a printed circuit board (PCB) or another integrated chip circuit. Current is supplied to the electrolytic solution while the solution is in contact with a cathode comprising an under bump structure on a semiconductor assembly. The semiconductor assembly comprises a base structure bearing the under bump structure, and the latter comprises a seminal conductive layer that may comprise either under bump metal, which is preferably copper or a copper alloy, or an under bump pad that comprises another conductive material such as, for example, a conductive polymer. An under bump metal structure may comprise, for example, a copper seed layer as provided by physical vapor deposition.

[0067] In the electrodeposition of pillars, and optionally also in the deposition of bumps, the under bump structure is positioned within or extends into a concavity in the surface of the base structure. The configuration of said bump or pillar is defined by the complementary configuration of the concavity.

[0068] In one implementation of the process, the concavity comprises a floor comprising the under bump pad or under bump metal and a sidewall comprising a dielectric material. In another implementation, the base structure comprises a dielectric layer comprising a photoresist, mask, or stress buffer material and the concavity comprises an opening in a surface of the dielectric layer. In this instance, the dielectric layer may be removed after electrodeposition of said bump or pillar.

[0069] In addition, the sidewall of the concavity can be provided with a dielectric liner prior to electrodeposition of the bump or pillar. In other words, the cavity in which copper is to be deposited may first be provided with a dielectric liner such as silicon dioxide or silicon nitride. The dielectric liner can be formed, for example, by chemical vapor deposition or plasma vapor deposition. Alternatively, organic dielectrics can be used to mitigate a coefficient of thermal expansion mismatch. A photoresist wall of the cavity may have sufficient dielectric properties to obviate the need for a further dielectric layer. However, the nature of the vapor deposition process may cause a further dielectric layer to form on the photoresist wall as well. A seminal conductive layer is then provided by either chemical vapor deposition of a seed layer.

[0070] In a process for forming bumps and pillars, the conductive under bump structure may be deposited only at the bottom, i.e., the floor, of the cavity, or in some embodiments, such as those illustrated and described in U.S. Pat. No. 8,546,254 B2 to Lu et al., the conductive under bump structure may extend from the bottom of the concavity for some distance upwardly along the sidewall. Preferably, at least the upper sidewall of the concavity remains non-conductive. The bottom of the concavity can be flat, or may comprise a recess filled with polyimide that promotes better bonding. This embodiment of the process differs from filling TSVs, for example, in which the seminal conductive layer is formed over the entire surface of the cavity, including bottom and sidewalls, and metallization is carried out to deposit copper on both bottom and sidewalls.

[0071] In carrying out the process described herein, current is supplied to an electrolytic circuit comprising a direct current power source, the aqueous electrodeposition composition, an under bump pad, under bump metal, or array of under bump pads or metal in electrical communication with the negative terminal of the power source and in contact with the electrodeposition composition, and an anode in electrical communication with the positive terminal of the power source and in contact with the electrodeposition composition.

[0072] In wafer level packaging, under bump structures are arrayed on a face of a semiconductor wafer, the under bump structure is electrically connected to the negative terminal of the power source, the semiconductor wafer and anode are immersed in the electrodeposition bath, and the power applied. Using the electrodeposition composition described herein within wafer (WIW) uniformity is maintained at a standard deviation not greater than about 10%, for example, while within die (WID) uniformity for dies cut from the wafer is maintained at a standard deviation of, for example, not greater than about 10%. Average feature (WIF) doming is typically about 10%, for example, for baths containing a single leveler. However, greater deviation may be tolerated in situations where productivity gains can be achieved or the device has greater tolerance of the deviation can be remedied downstream by, for example, a mechanical copper removal process. Doming and dishing of bumps and pillars can be minimized, and relatively flat head bumps and pillars can be prepared, using electrodeposition baths containing combinations of levelers as described herein.

[0073] The process can be used to provide the under bump metal pads for flip chip manufacturing in which case the metalizing substrate is generally limited to the faces of the bonding pads. Alternatively, with reference to the under bump metal as the floor, the process can be used to form a copper bump or pillar by bottom-up filling of the cavity formed at its floor by the under bump pad or under bump metal and on its sides by the sidewall of an opening in a stress buffer layer and / or photoresist that allows access to the pad or under bump metal. In the latter application, the aperture size of the cavity is roughly comparable to that of a blind through silicon via, and the parameters of the process for building the bump or pillar are similar to those used for filling blind TSVs. However, the concavity wall provided by openings in photoresist or stress-reducing material is ordinarily not seeded and is therefore non-conductive. Only a semiconductor or dielectric under bump structure at the floor of the cavity is provided with a seminal conductive layer, typically comprising a conductive polymer such as a polyimide. In such implementations, the process is not as dependent on the balance of accelerator and suppressor as it is in the case of bottom filling submicron vias or TSVs.

[0074] During the electrodeposition of a bump or pillar within a concavity in the surface of the base structure, lateral growth thereof is constrained by the sidewall(s) of the concavity, and the configuration of the bump or pillar is defined by the complementary configuration of the concavity.

[0075] In other implementations, a bump may be grown over the under bump metal or pad without lateral constraint, or may be caused to grow above the upper rim of a concavity or other lateral constraint, in which case a bump is formed that typically assumes a generally spherical configuration. However, in these implementations, the configuration of the bump can be influenced by the orientation, configuration and dimension of the anode in the electrolytic circuit.

[0076] As described, for example, in U.S. Pat. No. 8,546,254 B2 to Lu et al., an anode immersed in an electrodeposition bath can be brought into registry with an under bump structure that is also immersed in the bath, or each of an array of anodes can be brought into registry with a complementary array of under bump structures within the bath, and current applied to deposit a bump or pillar on the under bump structure. If growth of the bump is not constrained by the sidewall of a concavity, or if application of current is continued to a point that the growing bump extends outside the concavity or other lateral constraint, growth of the distal end of the bump assumes a spherical or hemispherical shape as illustrated in Fig. 2B of the '254 patent. As further described therein, the anode may be pulled away from the substrate along the axis of the growing bump, and the vertical rate of withdrawal of the anode from substrate can affect the shape of distal end of the bump. Generally, the faster the pulling rate, the higher the tangential angle θ (theta) between a horizontal plane and the growing bump at any given distance between the location of the plane and the under bump metal or pad. As further described in the '254 patent, the pulling rate is not necessarily constant but, if desired, can be varied with deposition time or extent of vertical growth. Alternatively, the under bump structure can be pulled away from the anode instead of the anode being pulled away from substrate. In addition to the pulling rate of the anode, the voltage difference between the anode and the cathode (initially the under bump structure and thereafter the growing bump) can also affect the shape of the bump.

[0077] The inventors of the present invention have found that, where a solder bump is added at the distal end of a copper bump or pillar that has been formed by the process described herein, the solder bump adheres seamlessly to the copper with a minimum of Kirkendall voids. Thus a solder bump comprised of a low melting alloy such as, for example, Sn / Ag or Sn / Pb, can be directly applied to the copper pillar or bump without need for a cap on the copper containing an intermediate layer of nickel or Ni alloy. Also Kirkendall voids are substantially avoided at the juncture between the copper bump or pillar and an under bump metal.

[0078] It has also been shown that the use of the compositions described herein provides a high level of within die (WID) and within wafer (WIW) uniformity in the deposition of arrays of copper bumps or pillars on a wafer that has been provided with an array of under bump structures as also described herein.

[0079] Using the levelers described herein, high current densities can be established and maintained throughout the electrodeposition process. Thus, the rate at which a bump or pillar may be caused to grow in the vertical direction is at least about 0.25 µm / min, more typically at least about 2.5 or about 3 µm / min, and even more typically at least about 3.3 µm / min. Achievable growth rates range up to about 10 µm / min or higher, equating to a current density of at least about 1 A / dm 2< , at least about 12 A / dm 2< , or at least about 20 A / dm 2< , ranging up to about 30 A / dm 2< or higher.

[0080] While the foregoing discussion is directed primarily to embodiments involving bumps and pillars, the compositions and methods have also been proven to be effective in forming other WLP copper features including megabumps, through silicon vias, and redistribution layers. The compositions and processes also apply to heterogeneous WLPs and semiconductor substrates other than Si-based substrates, such as, for example, GaAs-based substrates.

[0081] The invention is further illustrated by the following non-limiting working examples.Example 1 (not covered by the claimed invention)

[0082] Mega Pillar Aspect Ratio ~1:1, plate speed 3-7 µm / min, and feature size 240 µm(w) x 210 µm(h).

[0083] One exemplary preferred electrodeposition composition contains between about 28 and about 35 g / L copper ions in the form of copper sulfate or copper methane sulfonate, between about 150 and about 200 g / L sulfuric acid or methane sulfonic acid, between about 40 and about 50 mg / L SPS, between about 200 and 400 mg / L hexaalkoxylated triethylene tetramine and between 12 and about 15 mg / L leveler of structure IV, VI, VII, VIII, VIA, VIIIA, XXX or combinations of such structures. The weight ratio of hexaalkoxylated triethylene tetramine to leveler dipyridyl polymer is preferably between about 10:1 and about 40:1, more preferably between about 14:1 and about 20:1.Example 2 (not covered by the claimed invention)

[0084] Mega Pillar Aspect Ratio ~4:1, plate speed 3-7 µm / min, and feature size 60 µm(w) x 220 µm(h).

[0085] Another exemplary preferred electrodeposition composition contains between about 55 and about 65 g / L copper ions in the form of copper sulfate or copper methane sulfonate, between about 80 and about 120 g / L sulfuric acid or methane sulfonic acid, between about 40 and about 60 mg / L SPS accelerator, between about 200 and 400 mg / L hexaalkoxylated triethylene tetramine suppressor and between 12 and about 15 mg / L of leveler of structure IV, VI, VII, VIII, VIA, VIIIA, XXX or combinations of such structures. The weight ratio of hexaalkoxylated triethylene tetramine to dipyridyl polymer is preferably between about 10:1 and about 40:1, more preferably between about 14:1 and about 20:1.Example 3 (not covered by the claimed invention)

[0086] Mega Pillar Aspect Ratio ~1.5:1 plate speed 3-7 µm / min, and feature size 150 µm(w) x 220µm(h).

[0087] Yet another exemplary preferred electrodeposition composition contains between about 55 and about 65 g / L copper ions in the form of copper sulfate or copper methane sulfonate, between about 80 and about 120 g / L sulfuric acid or methane sulfonic acid, between about 10 and about 30 mg / L SPS, between about 500 and 700 mg / L hexaalkoxylated triethylene tetramine suppressor and between 3 and about 8 mg / L of leveler of structure IV, VI, VII, VIII, VIA, VIIIA, XXX or combinations of such structures. The weight ratio of hexaalkoxylated triethylene tetramine to dipyridyl polymer is preferably between about 50:1 and about 250:1, more preferably between about 110:1 and about 170:1.Example 4 (not covered by the claimed invention)

[0088] Redistribution layer with Aspect Ratio ~1:1, plate speed 0.5-2 µm / min, and feature size 2 µm(w)x 2 µm(h).

[0089] An additional exemplary preferred electrodeposition composition contains between about 25 and about 35 g / L copper ions in the form of copper sulfate or copper methane sulfonate, between about 170 and about 220 g / L sulfuric acid or methane sulfonic acid, between about 45 and about 55 mg / L SPS, between about 200 and 400 mg / L hexaalkoxylated triethylene tetramine and between 3 and about 6 mg / L of leveler of structure IV, VI, VII, VIII, VIA, VIIIA, XXX or combinations of such structures. The weight ratio of hexaalkoxylated triethylene tetramine to dipyridyl polymer is preferably between about 40:1 and about 90:1, more preferably between about 50:1 and about 70:1.Example 5 (not covered by the claimed invention)

[0090] TSV+RDL (2in1) Aspect Ratio ~1:1, plate speed 0.1-1 µm / min, and feature size 2 µm(w) x 2 µm(h) to 100 µm (h).

[0091] Yet another exemplary preferred electrodeposition composition contains between about 40 and about 60 g / L copper ions in the form of copper sulfate or copper methane sulfonate, between about 80 and about 120 g / L sulfuric acid or methane sulfonic acid, between about 1 and about 10mg / L SPS, between about 10 and 100 mg / L ethoxylated and propoxylated butanol and between 1 and about 10 mg / L of leveler of structure IV, VI, VII, VIII, VIA, VIIIA, XXX or combinations of such structures. The weight ratio of hexaalkoxylated triethylene tetramine to dipyridyl polymer is preferably between about 1:1 and about 100:1, more preferably between about 2:1 and about 50:1.Example 6 (not covered by the claimed invention)

[0092] Microbump / Pillar Aspect Ratio ~1-2:1, plate speed 1-2.5 µm / min, and feature size: 20-80 µm(w)x 20-100 µm(h).

[0093] Yet another exemplary preferred electrodeposition composition contains between about 25 and about 55 g / L copper ions in the form of copper sulfate or copper methane sulfonate, between about 80 and about 200 g / L sulfuric acid or methane sulfonic acid, between about 75 and about 125 mg / L SPS, between about 200 and 600 mg / L hexaalkoxylated triethylene tetramine and between 8 and about 20 mg / L of leveler of structure IV, VI, VII, VIII, VIA, VIIIA, XXX or combinations of such structures. The weight ratio of hexaalkoxylated triethylene tetramine to dipyridyl polymer is preferably between about 10:1 and about 75:1, more preferably between about 30:1 and about 40:1.Example 7

[0094] In order to able to better predict how certain combinations of levelers, suppressors and accelerators would work in plating copper features, a series of experiments were run in which different combinations of levelers, suppressors and accelerators were used. It is not yet understood while certain combinations provided a good result and other combinations provided a result that was not as successful. Table 1 lists various levelers, suppressors and accelerators:CompoundStructureAccelerator 1 Accelerator 2 Leveler 1 Leveler 2 Suppressor 1 Suppressor 2

[0095] Table 2 provides the results of the various combinations of accelerators, levelers and suppressors of Table 1 (Runs 1 to 3 are embodiments not covered by the claimed invention): Table 2.RunAcc. mg / LLeveler mg / LSupp. mg / LAve. H µmWID µmWIF µmComments1Acc. 2 40-50Lev. 1 10-20Supp. 1 800-1200196258.1Need flatter, better fill2Acc. 1 40-50Lev. 1 10-20Supp. 1 800-1200194321.2Better TIR, need better BHR3Acc. 1 40-50Lev. 1 10-20Supp. 2 800-120020531-10Dished is preferred to doming, need better BHR4Acc. 2 40-50Lev. 2 10-20Supp. 2 800-1200217231.2Best system, robustAve. H = average height of plated bump WIF = illustrates how much the bump domes. Positive value is doming, negative is dishing, 0 is flat

[0096] The results showed that while Run 1 initial produced the best result, switching from Accelerator 2 to Accelerator 1 (Run 2) greatly reduced the doming as the different accelerators perfomed differently with the same leveler (and same suppressor). In Run 3, when the suppressor was changed to Suppressor 2, the dishing was large and the WID was too high. By switching the accelerator back to Accelerator 2 and using Leveler 2 and Suppressor 2, it was possible to get both flat bumps WIF and good WID.

[0097] Thus, it can be seen that it is not easy to predict the combination of levelers, suppressors and accelerators that are needed to provide plated bumps having good average height (greater than about 190 nm, preferably greater than about 200 nm, more preferably greater than about 210 nm or even 215 nm), values of WID within an acceptable range (less than about 30, more preferably less than about 25), and an acceptable WIF (-5 ≤ TIR ≤ 5, more preferably -1.0 ≤ TIR ≤ 1.0).

[0098] In view of the above, it will be seen that the several objects of the invention are achieved and other advantageous results attained.

[0099] As various changes could be made in the above compositions and processes without departing from the scope of the invention as defined by the appended claims, 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. When introducing elements of the present invention or the preferred embodiments(s) thereof, the articles "a", "an", "the" and "said" are intended to mean that there are one or more of the elements. 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 + / -0.1% 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.

Claims

1. An electrodeposition composition comprising: a) a source of copper ions; b) an acid; c) a suppressor, wherein the suppressor comprises a compound selected from: i) an alkoxylated amine, wherein the alkoxylated amine is selected from of a tetraalkoxylated alkylene diamine, an alkoxylated diethylene triamine, a tetraalkoxylated triethylene tetraamine, and combinations of two or more of the foregoing; and ii) a polyether compound comprising a combination of propylene oxide, PO, repeat units and ethylene oxide, EO, repeat units present in a PO:EO ratio between 1:9 and 9:1 and covalently bonded to a nitrogen-containing cationic species selected from primary, secondary, tertiary, and quaternary amines; d) a leveler, wherein the leveler comprises a quaternized poly(epihalohydrin) compound, wherein the quaternized poly(epihalohydrin) compound comprises n repeating units corresponding to structure 1N and p repeating units corresponding to structure 1P: wherein Q has a quaternized amine structure selected from: i) NRaRbRc wherein one of Ra, Rb and Rc is selected from unsubstituted alkyl having at least three carbon atoms, hydroxyalkyl, dihydroxyalkyl, phenyl, hydroxyphenyl, dihydroxyphenyl, benzyl, hydroxyphenylmethyl and dihydroxyphenylmethyl, and each of the others of Ra, Rb and Rc is independently selected from lower alkyl having between 1 and 3 carbon atoms; ii) an N-substituted and optionally further substituted heteroalicyclic amine wherein the N-substituent is selected from substituted or unsubstituted alkyl, alicyclic, aralkyl, aryl, and heterocyclic; and iii) a substituted or unsubstituted nitrogen-containing heteroaryl compound; wherein n is an integer between 3 and 35; p is an integer greater than 0 and up to 25; X is a halo substituent; and X- is a monovalent anion, optionally a halide ion; when there are polar substituents on the quaternized amine structure or the quaternized amine structure contains aryl or aralkyl groups, the ratio of n / (n+p) is between 0.40 and 0.60, and otherwise the ratio of n / (n+p) is between 0.70 and 1.0; and e) an accelerator.

2. The composition as set forth in claim 1, wherein Q is selected from: or or wherein: i) structure 2B is an N-substituted heterocyclic moiety; ii) structure 2C is a heterocyclic moiety; iii) each of Ra, Rb, Rc and Rd is independently selected from substituted or unsubstituted alkyl, substituted or unsubstituted alkenyl, substituted or unsubstituted alkynyl, substituted or unsubstituted aralkyl, substituted or unsubstituted alicyclic, substituted or unsubstituted aryl, and substituted or unsubstituted heterocyclic; and iv) each of Re, Rf, Rg, Rh and Rj is independently selected from hydrogen, substituted or unsubstituted alkyl, substituted or unsubstituted alkenyl, substituted or unsubstituted alkynyl, substituted or unsubstituted aralkyl, substituted or unsubstituted alicyclic, and substituted or unsubstituted aryl, and substituted or unsubstituted heterocyclic.

3. The composition as set forth in claim 2, wherein the quaternized poly(epihalohydrin) comprises additional repeating units comprising residues of at least one alkylene oxide, or wherein the quaternized poly(epihalohydrin) comprises repeating units that are residues of ethylene oxide, or wherein the quaternized poly(epihalohydrin) comprises repeating units that are residues of propylene oxide.

4. The composition as set forth in claim 3, wherein the total of alkylene oxide repeating units in said quaternized poly(epihalohydrin) is q, and the ratio of q / n+p+q is not greater than 0.05, or is between 0.05 and 0.50.

5. The composition as set forth in claim 1 wherein repeating units of said poly(epihalohydrin) consist of residues of epihalohydrin and residues of quaternized epihalohydrin:

6. The composition as set forth in claim 1, wherein the quaternized poly(epihalohydrin) repeating units and non-quaternized epihalohydrin repeating units in said quaternized poly(epihalohydrin) are arranged in a block, alternating or random configuration.

7. The composition as set forth in claim 1, wherein Q has the quaternized amine structure NRaRbRc.

8. The composition as set forth in claim 1, wherein Q has the quaternized amine structure selected from 3-hydroxypropyldimethylamine, n-butyldimethylamine, di(3-hydroxypropyl)methylamine, 2,3-dihydroxypropyldimethylamine, 3-hydroxypropyldiethylamine, 2-hydroxypropyldimethylamine, 4-hydroxybutyldimethylamine, 2-hydroxyethyldimethylamine, n-propyldimethylamine, 2-hydroxyethoxyethyldimethylamine, di(2-hydroxyethyl)methylamine, benzyldimethylamine, and 4-hydroxybenzyldimethyleamine, 2-methylpyridine, 3-methylpyridine,4-methylpyridine, 2-ethylpyridine,3-ethylpyridine, 4-ethylpyridine, 2-propylpyridine, 3-propylpyridine,4-propylpyridine, 4-tertbutylpyridine, 4-cyanopyridine, 4-isopropylpyridine, 4-methoxypyridine, 3,4-lutidine, 3-methoxypyridine, 3-pyridinemethanol and 4-pyridinemethanol.

9. The composition as set forth in claim 1, wherein the quaternized poly(epihalohydrin) comprises:

10. The composition as set forth in claim 1, wherein the accelerator comprises an organic sulfur compound, optionally wherein the organic sulfur compound has the following general structure (5): wherein M is selected from alkali metal ions and charge balancing cations.