US20070012576A1

US20070012576A1 - Plating method

Info

Publication number: US20070012576A1
Application number: US11/180,468
Authority: US
Inventors: Robert A. Binstead; Chunyi Wu; Robert D. Mikkola
Original assignee: Rohm and Haas Electronic Materials LLC
Current assignee: Rohm and Haas Electronic Materials LLC
Priority date: 2005-07-13
Filing date: 2005-07-13
Publication date: 2007-01-18

Abstract

Methods of depositing layers of copper that selectively incorporate certain impurities are provided. Such copper layers reduce stress-induced void formation in wide copper lines under vias.

Description

The present invention relates generally to the field of metal plating. In particular, the present invention relates to the electrodeposition of copper.
Copper is used in the manufacture of many electronic devices. For example, in the manufacture of integrated circuits copper damascene processes (including dual damascene) involve the formation of inlaid copper wiring patterns with the simultaneous formation of via connections between metal layers. In such processes, the copper is deposited electrolytically using direct current.
The purity of the electrolytically deposited copper becomes more important as the size of the electronic devices shrink. High levels of impurities in small copper deposits will increase the resistivity of the copper. Accordingly, the trend in the industry is toward copper electroplating baths that provide purer copper deposits in order to reduce the resistivity of the deposits.
Stress-induced voiding occurs in copper deposits in dual damascene structures where voids are formed under vias that connect to wide metal lines. Such voiding leads to failures in the device. One theory attributes the formation of such voiding to vacancies that develop in the copper deposits when the copper is not properly annealed. See, for example, E. T. Ogawa et al., Stress-Induced Voiding Under Vias Connected to Wide Cu Metal Lines, IEEE International Reliability Physics Symposium Proceedings (2002), 40^th, pp 312-321, which discusses the formation of voids under vias due to stress. Regardless of how such voiding occurs, the use of higher purity copper in the wide metal lines exacerbates the formation of such voiding. There is a need in the industry for high purity copper deposits that do not form stress-induced voids.
It has been surprisingly found that impurities can be selectively incorporated into copper metal lines during electroplating of the copper. Such selective incorporation of impurities in wide metal lines reduces the formation of stress-induced voids under vias connected to such metal lines.
In one embodiment, the present invention provides a method of depositing copper including the steps of: a) contacting an electronic device substrate having apertures with a copper electroplating bath including a source of copper ions, an electrolyte, and a disulfide-containing accelerator; b) depositing a layer of copper in the apertures using a duty cycle including 1) applying a first current density for a first period to electrochemically reduce the disulfide-containing accelerator to a thiol compound at a copper surface; and 2) applying a second current density for a second period; and c) repeating step b) until a desired copper deposit is obtained; wherein the second current density is less than the first current density. The present method is useful for incorporating impurities at a desired level within the copper deposit. In particular, the present invention is useful in the manufacture of integrated circuits, and more specifically in the deposition of wide metal lines in the manufacture of integrated circuits.
FIG. 1 is a secondary ion mass spectrogram showing impurity levels as a function of copper film depth for a prior art process.
FIG. 2 is a secondary ion mass spectrogram showing impurity levels as a function of copper film depth for a prior art process.
FIG. 3 is a secondary ion mass spectrogram showing impurity levels as a function of copper film depth for the process of the invention.
As used throughout the specification, the following abbreviations shall have the following meanings: nm=nanometers; g/L=grams per liter; mA/cm²=milliamperes per square centimeter; μm=micron=micrometer; ppm=parts per million, mL/L=milliliter/liter;° C.=degrees Centigrade; sec.=seconds; msec.=milliseconds; g=grams; DC =direct current; Hz=Hertz; and Å=Angstroms.
As used throughout the specification, “feature” refers to the geometries on a substrate. “Apertures” refer to recessed features, such as vias and trenches. As used throughout this specification, the term “plating” refers to copper electroplating, unless the context clearly indicates otherwise. “Deposition” and “plating” are used interchangeably throughout this specification. “Defects” refer to surface defects of a copper layer, such as protrusions and pits, as well as defects within a copper layer, such as voids. “Wide metal lines” refers to metal lines having a width of >1 μm. The terms “layer” and “film” are used interchangeably and refer to a metal deposit, particularly a copper deposit, unless the context clearly indicates otherwise.
The term “alkyl” includes linear, branched and cyclic alkyl. “Accelerator” refers to an organic additive that increases the plating rate of a metal during electroplating. “Suppressors” (also known as “carriers”) refer to organic additives that suppress the plating rate of a metal during electroplating. “Leveler” refers to an organic additive that is capable of providing a substantially planar metal layer. The terms “leveler” and “leveling agent” are used interchangeably throughout this specification. The term “halide” refers to fluoride, chloride, bromide and iodide. As used herein, “duty cycle” means the relationship between the time period of high current density and the time period of low current density. A 75% duty cycle means that for a given time the ratio of time periods of high to low current density is 3:1 (or that the high current density is applied for 75% of the time and the low current density is applied for 25% of the time).
The indefinite articles “a” and “an” are intended to include both the singular and the plural. All percentages and ratios are by weight unless otherwise indicated. All ranges are inclusive and combinable in any order except where it is clear that such numerical ranges are constrained to add up to 100%.
A wide variety of electronic device substrates may be plated with copper according to the present invention. Suitable substrates include, without limitation: printed circuit board substrates, integrated circuit substrates such as wafers used in the manufacture of integrated circuits, electronic packages such as lead frames and electronic interconnects such as wafer bumps; and optoelectronic device substrates such as hermetic sealing layers.
A wide variety of copper electroplating baths may be used with the present invention. Copper electroplating baths typically contain a source of copper ions, an electrolyte, a source of chloride ions, and a disulfide-containing accelerator. More typically, organic additives such as a suppressor are added to the copper electroplating baths. The copper electroplating baths may optionally contain a leveler.
Typical sources of copper ions are any copper compounds that are soluble in the electroplating bath. Suitable copper compounds include, but are not limited to, copper salts such as copper sulfate, copper persulfate, copper halide, copper chlorate, copper perchlorate, copper alkanesulfonate such as copper methanesulfonate, copper alkanol sulfonate, copper arylsulfonate, copper fluoroborate, cupric nitrate, copper acetate, and copper citrate. Copper sulfate is preferred. Mixtures of copper compounds may be used. Such sources of copper ions are generally commercially available.
The source of copper ions may be used in the present electroplating baths in a relatively wide concentration range. Typically, the copper ion source is present in an amount sufficient to provide an amount of copper ion of 10 to 80 g/L in the plating bath. More typically, the amount of copper source provides 15 to 65 g/L of copper ions in the plating bath. The copper plating bath may also contain amounts of other alloying elements, such as, but not limited to, tin, zinc, indium, antimony, and the like. Such alloying elements are added to the electroplating baths in the form of any suitable bath-solution salt. Thus, the copper electroplating baths useful in the present invention may deposit copper or copper alloy.
The electrolyte may be alkaline or acidic and is typically acidic. Any acid which is compatible with the copper compound may be used in the present invention. Suitable acids include, but are not limited to: sulfuric acid, acetic acid, fluoroboric acid, nitric acid, sulfamic acid, phosphoric acid, hydrogen halide acids such as hydrochloric acid, alkanesulfonic acids and arylsulfonic acids such as methanesulfonic acid, toluenesulfonic acid, phenolsulfonic acid and benzenesulfonic acid, and halogenated acids such as trifluoromethylsulfonic acid and haloacetic acid. Typically the acid is sulfuric acid, alkanesulfonic acid or arylsulfonic acid. Mixtures of acids may be used. In general, the acid is present in an amount to impart conductivity to a bath containing the acidic electrolyte composition. Typically, the pH of the acidic electrolyte of the present invention has a value of less than 7, and more typically less than 2. Exemplary alkaline electroplating baths use pyrophosphate as the electrolyte, although other electrolytes may be employed. It will be appreciated by those skilled in the art that the pH of the electrolyte may be adjusted by any known methods, if necessary.
The total amount of acid electrolyte used in the present electroplating baths may be from 0 to 100 g/L, and typically from 0 to 60 g/L, although higher amounts of acid may be used for certain applications, such as up to 225 g/L or even 300 g/L. It will be appreciated by those skilled in the art that by using copper sulfate, a copper alkanesulfonate or a copper arylsulfonate as the copper ion source, an acidic electrolyte can be obtained without any added acid.
A wide variety of disulfide-containing accelerators may be employed in the present copper electroplating baths. Such accelerators may be used alone or as a mixture of two or more. In general, the disulfide-containing accelerators have a molecular weight of 5000 or less and more typically 1000 or less. Disulfide-containing accelerators that also have sulfonic acid groups are generally preferred, particularly compounds that include a group of the formula R′—S—S—R—SO₃X, where R is an optionally substituted alkyl (which include cycloalkyl), optionally substituted heteroalkyl, optionally substituted aryl group, or optionally substituted heteroalicyclic; X is hydrogen or a counter ion such as sodium or potassium; and R′is hydrogen or an organic residue, such as a group of the formula —R—SO₃X or a substituent of a larger compound. Typically alkyl groups will have from 1 to 16 carbons, more typically 1 to 8 or 12 carbons. Heteroalkyl groups will have one or more hetero (N, O or S) atoms in the chain, and typically have from 1 to 16 carbons, more typically 1 to 8 or 12 carbons. Carbocyclic aryl groups are typical aryl groups, such as phenyl and naphthyl. Heteroaromatic groups also will be suitable aryl groups, and typically contain 1 to 3 of one or more of N, O and S atoms and 1 to 3 separate or fused rings and include, e.g., coumarinyl, quinolinyl, pyridyl, pyrazinyl, pyrimidyl, furyl, pyrrolyl, thienyl, thiazolyl, oxazolyl, oxidizolyl, triazole, imidazolyl, indolyl, benzofuranyl, and benzothiazol. Heteroalicyclic groups typically will have 1 to 3 of one or more of N, O and S atoms and from 1 to 3 separate or fused rings and include, e.g., tetrahydrofuranyl, thienyl, tetrahydropyranyl, piperdinyl, morpholino, and pyrrolindinyl. Substituents of substituted alkyl, heteroalkyl, aryl or heteroalicyclic groups include, e.g., C_1-8alkoxy; C_1-8alkyl, halogen such as F, C1 and Br; cyano; and nitro.
More specifically, useful disulfide-containing accelerators include those of the following formulae; XO₃S—R—S—S—R—SO₃X and XO₃S—Ar—S—S—Ar—SO₃X, wherein R in the above formulae is an optionally substituted alkyl group, and typically is an alkyl group having from 1 to 6 carbon atoms, more typically is an alkyl group having from 1 to 4 carbon atoms; Ar is an optionally substituted aryl group such as optionally substituted phenyl or naphthyl; and X is hydrogen or a suitable counter ion such as sodium or potassium. Exemplary disulfide-containing accelerators include, without limitation, bis-sulfopropyl disulfide and bis-sodium-sulfopropyl disulfide.
Optionally, an additional accelerator that does not contain a disulfide group may be used in combination with the present disulfide-containing accelerator. Typical additional accelerators are sulfur-containing and contain one or more sulfur atoms and may be, without limitation, thiols, mercaptans, sulfides, disulfides and organic sulfonic acids. In one embodiment, such additional accelerator compound has the formula XO₃S—R—SH, wherein R is an optionally substituted alkyl group, and typically is an alkyl group having from 1 to 6 carbon atoms, more typically is an alkyl group having from 1 to 4 carbon atoms and X is hydrogen or a suitable counter ion such as sodium or potassium.
Exemplary additional accelerators include, without limitation, N,N-dimethyl-dithiocarbamic acid-(3-sulfopropyl)ester; 3-mercapto-propylsulfonic acid-(3-sulfopropyl)ester; 3-mercapto-propylsulfonic acid (sodium salt); carbonic acid-dithio-o-ethylester-s-ester with 3-mercapto-1 -propane sulfonic acid (potassium salt); 3-(benzthiazolyl-s-thio)propyl sulfonic acid (sodium salt); pyridinium propyl sulfobetaine; 1-sodium-3-mercaptopropane-1-sulfonate; sulfoalkyl sulfide compounds disclosed in U.S. Pat. No. 3,778,357; the peroxide oxidation product of a dialkyl amino-thiox-methyl-thioalkanesulfonic acid; and combinations of the above. Additional suitable accelerators are also described in U.S. Pat. Nos. 3,770,598; 4,374,709; 4,376,685; 4,555,315; and 4,673,469.
The amount of the disulfide-containing accelerators present in a freshly prepared copper electroplating bath is typically from 0.1 to 1000 ppm. More typically, the disulfide-containing accelerator compounds are present in an amount of from 0.5 to 300 ppm, still more typically from 1 to 100 ppm, and even more typically from 2 to 50 ppm. Any additional accelerators present in the copper electroplating bath is used the amounts described for the disulfide-containing accelerators.
In general, the copper electroplating baths also contain water. The water may be present in a wide range of amounts. Any type of water may be used, such as distilled, deionized or tap.
It will be appreciated by those skilled in the art that one or more other components may be added to the copper electroplating baths of the present invention, such as, e.g., suppressors, levelers, halide ions, and other alloying materials.
Any suppressor may optionally be used in the present electroplating baths. Suppressors, as used throughout this specification, refer to any compounds that suppress the plating rate of copper as compared to baths without such suppressors. Suitable suppressors include polymeric materials, preferably having heteroatom substitution, particularly oxygen linkages. In general, suppressors are typically polyethers, such as, but not limited to, those of the formula
R—O—(CXYCX′Y′O)_nH
wherein R is an aryl, alkyl or alkenyl group containing from 2 to 20 carbons; X, Y, X′, and Y′ are each independently hydrogen, alkyl, such as methyl, ethyl or propyl, aryl such as phenyl, and aralkyl such as benzyl; and n is an integer from 5 to 100,000. It is preferred that one or more of X, Y, X′ and Y′ is hydrogen. More than one suppressor may be used.
Suitable suppressors include, but are not limited to: amines such as ethoxylated amines; polyoxyalkylene amines and alkanol amines; amides; poly-glycol type wetting agents such as polyethylene glycols, polyalkylene glycols and polyoxyalkylene glycols; high molecular weight polyethers; polyethylene oxides such as those having a molecular weight in the range of 1,000 to 100,000; polyoxyalkylene block copolymers; alkylpolyether sulfonates; complexing suppressors such as alkoxylated diamines; and complexing agents for cupric or cuprous ions such as citric acid, edetic acid, tartaric acid, potassium sodium tartrate, acetonitrile, cupreine and pyridine.
Particularly useful suppressors include, but are not limited to: ethyleneoxide/propyleneoxide (“EO/PO”) block or random copolymers; ethoxylated polystyrenated phenol having 12 moles of ethyleneoxide (“EO”), ethoxylated butanol having 5 moles of EO, ethoxylated butanol having 16 moles of EO, ethoxylated butanol having 8 moles of EO, ethoxylated octanol having 12 moles of EO, ethoxylated beta-naphthol having 13 moles of EO, ethoxylated bisphenol A having 10 moles of EO, ethoxylated sulfated bisphenol A having 30 moles of EO and ethoxylated bisphenol A having 8 moles of EO.
In general, the suppressor may be added in any amount that provides sufficient lateral growth of the copper layer. Typically, the amount of suppressor is in the range of 0.001 to 10 g/L, and more typically 0.1 to 2.0 g/L.
Levelers may optionally be added to the present electroplating baths. In one embodiment, a leveler compound is used in the present electroplating baths. Such levelers may be used in a wide range of amounts, such as from 0.01 to 50 ppm or greater. Examples of suitable leveling agents are described and set forth in U.S. Pat. Nos. 3,770,598; 4,374,709; 4,376,685; 4,555,315; 4,673,459; and 6,610,192; and U.S. pat. application Ser. No. 2004/0249177. In general, useful leveling agents include those that contain a substituted amino group such as compounds having R—N—R′, where each R and R′is independently a substituted or unsubstituted alkyl group or a substituted or unsubstituted aryl group. Typically the alkyl groups have from 1 to 6 carbon atoms, more typically from 1 to 4 carbon atoms. Suitable aryl groups include substituted or unsubstituted phenyl or naphthyl. The substituents of the substituted alkyl and aryl groups may be, for example, alkyl, halo and alkoxy. Sulfur-containing leveling agents may also be used.
More specifically, suitable leveling agents include, but are not limited to, 1-(2-hydroxyethyl)-2-imidazolidinethione; 4-mercaptopyridine; 2-mercaptothiazoline; ethylene thiourea; thiourea; alkylated polyalkyleneimine; phenazonium compounds disclosed in U.S. Pat. No. 3,956,084; N-heteroaromatic rings containing polymers; quatemized, acrylic, polymeric amines; polyvinyl carbamates; pyrrolidone; and imidazole. An exemplary leveler is 1-(2-hydroxyethyl)-2-imidazolidinethione, although other suitable levelers may be employed.
Other suitable levelers are reaction products of an amine with an epihalohydrin, and preferably epichlorohydrin. Suitable amines include, but are not limited to, primary, secondary or tertiary amines, cyclic amines, aromatic amines and the like. Exemplary amines include, without limitation, dialkylamines, trialkylamines, arylalylamines, diarylamines, imidazole, triazole, tetrazole, benzimidazole, benzotriazole, piperidine, morpholine, piperazine, pyridine, oxazole, benzoxazole, pyrimidine, quinoline, and isoquinoline. Imidazole is the preferred amine. Such amines may be substituted or unsubstituted. By “substituted”, it is meant that one or more of the hydrogens on the amine are replaced by one or more substituent groups, such as alkyl, aryl, alkoxy, halo, and alkenyl. Other suitable reaction products of amines with epichlorohydrin are those disclosed in U.S. Pat. No. 4,038,161 (Eckles et al.). Such reaction products are generally commercially available, such as from Raschig, or may be prepared by methods known in the art.
When present, the leveling agents are typically used in an amount of 0.5 to 1000 ppm. More typically, the leveling agents are used in an amount of 0.5 to 500 ppm, still more typically from 1 to 250 ppm, and even more typically from 1 to 50 ppm.
The present copper electroplating baths may optionally contain a halide ion, and preferably do contain a halide ion. Chloride and bromide are preferred halide ions, with chloride being more preferred. Mixtures of halide ions may be used. A wide range of halide ion concentrations (if a halide ion is employed) may be suitably utilized, e.g. from 0 (where no halide ion employed) to 100 ppm of halide ion in the plating bath, more preferably from 25 to 75 ppm. Such halides may be added as the corresponding hydrogen halide acid or as any suitable salt.
The electroplating baths may be prepared by combining the source of copper ions, the electrolyte, the disulfide-containing accelerator and any optional components in any order. Typically, the plating baths of the present invention may be used at any temperature from 10° to 65° C. or higher. It is preferred that the temperature of the plating baths is from 10° to 35° C. and more preferably from 15° to 30° C.
The present plating baths are typically agitated during use. Any suitable agitation method may be used with the present invention and such methods are well-known in the art. Suitable agitation methods include, but are not limited to, air sparging, work piece agitation, impingement, rotation and the like. Such methods are known to those skilled in the art.
When the present invention is used to plate an integrated circuit substrate, such as a wafer, the wafer may be rotated such as from 1 to 150 RPM and the plating solution contacts the rotating wafer, such as by pumping or spraying. In the alternative, the wafer need not be rotated where the flow of the plating bath is sufficient to provide the desired metal deposit.
In general, the substrate to be copper plated is contacted with the copper electroplating bath by a suitable means, such as by immersion or by pumping or spraying. The substrate typically functions as the cathode. An anode is added to the copper plating bath and a potential is applied.
In one embodiment, the present invention provides a method of depositing copper including the steps of: a) contacting an electronic device substrate having apertures with a copper electroplating bath including a source of copper ions, an electrolyte, and a sulfur-containing compound; b) depositing a layer of copper in the apertures using a duty cycle including 1) applying a first current density for a first period to electrochemically reduce the disulfide-containing accelerator to a thiol compound at a copper surface; and 2) applying a second current density for a second period; and c) repeating step b) until a desired copper deposit is obtained; wherein the second current density is less than the first current density.
The duty cycle may be repeated at a variety of different frequencies. For example, the duty cycle may be repeated up to multiple times per second or may take multiple seconds to perform one duty cycle. The particular duty cycle chosen will depend upon the size of the aperture to be copper plated, the particular copper electroplating bath used and the level of impurities desired. Suitable duty cycle frequencies are from 0.05 to 10 Hz (or cycles per second) or even higher frequencies may be used, such as up to 100 Hz. In the manufacture of integrated circuits having wide metal lines, a suitable duty cycle has a frequency of 0.1 to 10 Hz, more typically from 0.1 to 5 Hz and still more typically from 0.1 to 2 Hz, although higher or lower frequencies may suitably be used.
A wide variety of current densities may be used for the first current density. Suitable first current densities are from 1 to 100 mA/cm²although higher or lower current densities may be used. More typically, the first current density is from 5 to 100 mA/cm², and still more typically from 15 to 90 mA/cm². A particularly suitable range of first current densities is from 40 to 85 mA/cm^{2 .}A wide variety of current densities may be used for the second current density, provided that the second current density is less than the first current density. Exemplary second current densities are from 1 to 50 mA/cm², although higher or lower current densities may be used. More typically, the second current density is from 1 to 35 mA/cm², still more typically from 2 to 25 mA/cm², and even more typically from 5 to 10 mA/cm².
While not intending to be bound by theory, it is believed that the first period of high current density reduces the disulfide-containing accelerator to one or more thiol compounds. Such thiol compounds may contain one or more thiol groups. It is believed that the disulfide-containing accelerator is electrochemically reduced at the freshly growing copper surface to form the thiol compound. Such thiol compounds are believed to adsorb on the copper surface during the relatively high current density first period. In one embodiment, the first period is performed for a time of 0.1 msec. to 10 sec., more typically from 0.1 msec. to 5 sec., and still more typically from 0.1 msec. to 1 sec. Further without wishing to be bound by theory, it is believed that the longer the period of relatively low current density, the greater the amount of total impurities incorporated into the copper deposit. In one theory, but not the only theory, such period of relatively low current density allows the copper surface to recrystallize to incorporate any organic material on the copper surface. Thus, the amount of total impurities incorporated into the copper deposited can be controlled by the choice of second current density and by the time period the substrate is subjected to this current density.
In the present process, the range of amounts of impurities incorporated in the copper layer as deposited, that is before annealing) may be from 1 to 500 ppm for each impurity, such as chloride, sulfur, carbon, oxygen and nitrogen. The total amount of impurities before annealing may be up to a couple of thousand ppm. . Typically, such total impurities are in the range of from 1 to 500 ppm, more typically from 1 to 300 ppm. The impurity levels are determined by Secondary Ion Mass Spectrometry (“SIMS”), which provides a value of ion concentration per unit area, as compared to an ion implanted standard. The average impurity values are obtained by summing the ppm values from the SIMS analysis for each data point for each impurity and then dividing by the total number of data points for the depth (in nm) of the copper layer evaluated. The average impurity levels throughout the depth of the copper deposit are much lower than the individual values. For example, an impurity level of chloride ion by SIMS analysis may show a maximum value of 200 ppm for a given unit area, where the average chloride ion impurity level may only be 5 ppm for the entire copper deposit. In one embodiment, the range of average total impurity level is from 1 to 500 ppm, and more typically from 1 to 300 ppm.
In integrated circuit manufacture, copper layers are typically annealed. During such annealing step, certain impurities, such as sulfur and oxygen, are typically reduced. Copper layers deposited according to the present invention, following annealing, typically have average total impurities in the range of 1 to 500 ppm, more typically 1 to 300 ppm, and still more typically from 1 to 250 ppm. In one embodiment, the average total impurity level following annealing is from 1 to 100 ppm.
After the desired copper deposit is obtained, an optional further plating step may be employed to smooth the surface of the deposit. Such optional plating step includes applying a current density for a third period. A further optional resting step may be included. No current is applied during the resting portion of the step. In one embodiment, the third current density is in the range of 20 to 90 mA/cm²and typically 30 to 60 mA/cm². In another embodiment, additional plating steps are performed, such additional steps may include cycling the plating on and off to smooth the surface of the copper deposit.
The present invention is useful for depositing copper as well as copper alloys such as, but not limited to, copper-silver, copper-tin and tin-copper-silver. The present invention is expected to be beneficial in the deposition of metals other than copper, such as silver and tin.
An advantage of the present invention is that it provides for the tailoring of doping (impurity) levels in a metal layer, particularly a copper layer, to balance electromigration performance and void stress migration performance. High purity levels (i.e. low doping levels) are advantageous from electromigration performance. However, the incorporation of certain levels of impurities may be beneficial for void stress migration control where small vias land on a wide line.
A further advantage of the present invention is that a single metal plating bath may be used to provide an electronic device having a first metal layer having a first purity and a second metal layer having a second purity, where the purities of the two metal layers are different. In this way, a metal layer can be deposited having a desired level of total impurities needed for a specific purpose, such as for control of void stress migration. Accordingly, the present invention provides an electronic device including a first layer of metal and a second layer of metal, wherein the first layer of metal includes total impurities in the range of up to 10 ppm and the second layer of metal includes total impurities in the range of 10 to 100 ppm. In one embodiment, the first and second metal layers are copper. For example, in an integrated circuit, the first layer of metal may be a via layer, a small line (i.e., a line having a width of ≦1 μm), or a mixture of these and the second layer of metal may be a wide line.

EXAMPLE 1-10

A copper plating bath was prepared by combining copper sulfate (40 g/L of copper ion), sulfuric acid (10 g/L), hydrochloric acid (50 mg/L of chloride ion), a disulfide-containing sulfonic acid accelerator (10 mL/L), an EO/PO copolymer suppressor (5 mL/L), a leveler (3 mL/L) that is a reaction product of an epoxide and imidazole and water.
Wafers were plated by immersing them individually in the copper plating bath with rotation to cause net mass transport to the wafer surface. Different first and second current densities were used for each wafer. In each case, a rectangular pulsed waveform having a 75% duty cycle was used. Copper was deposited to approximately 1 μm. After deposition, the wafers were removed from the plating bath, rinsed and dried. The copper deposits were then analyzed by Secondary Ion Mass Spectrometry (“SIMS”) for total impurity levels and found to contain oxygen, nitrogen, chlorine, sulfur and carbon as impurities. The approximate total amount of impurities (C, N, O, S, Cl) in each deposit is reported in the following table.

High Current Low Current Average Total

Density Density Frequency Impurity Level

Example (mA/cm²) (mA/cm²) (Hz) (ppm)

1 40 5 1 11.0

2 40 5 0.5 14.2

3 40 5 0.25 13.6

4 55 5 1 26.4

5 55 5 0.5 21.0

6 55 5 0.25 20.7

7 65 5 0.25 26.0

8 65 5 0.1 24.4

9 85 5 0.25 38.3

10 85 5 0.1 35.5

EXAMPLE 11—Comparative

The plating bath of Examples 1-10 was used to deposit copper on a wafer using a DC waveform. The average (C, N, O, S, Cl) impurity level by SIMS analysis was <5 ppm.

EXAMPLE 12—Comparative

The plating bath of Examples 1-10 was used to deposit a copper film on a wafer using a with a current density of 7 mA/cm²for the first 100 nm of copper deposited, followed by 40 mA/cm²for approximately the next 900 nm of copper deposited. The non-annealed copper deposit (approximately 1000 nm thick) was then analyzed for impurity levels using SIMS. The results are shown in FIG. 1 and illustrate the principal impurities: carbon, sulfur, chlorine, nitrogen, and oxygen. The plotted concentration values are not average values but are instead actual data points. The increasing levels of impurities observed near 0 nm of copper deposit depth arose from surface contamination from the additives in the plating bath. The high level oxygen impurity at depths of >800 nm arose from the TaO liner used to fabricate the silicon test wafer. These data show, for example, a maximum concentration of approximately 6 ppm of chloride ion per unit area in the region of 400-700 nm depth. The average total impurity level of this copper deposit was quite low, i.e. , 5 ppm. The average total impurity level was not controlled using this process.

EXAMPLE 13 —Comparative

The plating bath of Examples 1-10 was used to deposit a copper film on a wafer using a constant current density of 7 mA/cm²for the entire depth of the copper deposit (approximately 1000 nm). The non-annealed copper deposit was then analyzed for impurity levels using SIMS. The results are shown in FIG. 2 and illustrate the principal impurities: carbon, sulfur, chlorine, nitrogen, and oxygen. The plotted concentration values are not average values but are instead actual data points. The increasing level of oxygen at depths of >850 nm arose from the TaO liner used to fabricate the silicon test wafer. The very high levels of impurities observed between 300 and 450 nm depth arose from a natural surface recrystallization phenomenon that results in the incorporation of surface adsorbates from the plating bath and exposure of a fresh surface of copper atoms. This natural cycle of accumulation of surface adsorbates followed by recrystallization of the surface layers can be repeated indefinitely if plating is continued at low current density. However, when such recrystallization occurs is not predictable. Accordingly, the average total impurity level cannot be controlled using such natural recrystallization process.

EXAMPLE 14

The plating bath of Examples 1-10 was used to deposit a copper film on a wafer using alternating current densities of 5 mA/cm²(for 100 nm of copper deposit) and 60 mA/cm²(for 22.5 nm) repeated four times. The final 410 nm of copper deposit was plated at 5 mA/cm². The non-annealed copper deposit was then analyzed for impurity levels using SIMS. The results are shown in FIG. 3 and illustrate the principal impurities: carbon, sulfur, chlorine, nitrogen, and oxygen. The plotted concentration values are not average values but are instead actual data points. The high level of oxygen at depths of >950 nm arose from the TaO liner used to fabricate the silicon test wafer. The very high levels of other impurities observed between 300 and 950 nm depth arose from induced surface recrystallization caused by enrichment with surface adsorbed thiols that are produced during brief applications for a high current density plating pulse (1 sec, 60 mA/cm²). Pulsed waveforms of the present invention can be used, therefore, to incorporate very large levels of impurities compared with a conventional DC plating process.

Claims

1. A method of depositing copper comprising the steps of: a) contacting an electronic device substrate having an aperture with a copper electroplating bath comprising a source of copper ions, an electrolyte, and a disulfide-containing accelerator; b) depositing a layer of copper in the aperture using a duty cycle comprising 1) applying a first current density for a first period to electrochemically reduce the disulfide-containing accelerator to a thiol compound at a copper surface; and 2) applying a second current density for a second period; and c) repeating step b) until a desired copper deposit is obtained; wherein the second current density is less than the first current density.

2. The method of claim 1 wherein the copper deposit comprises from 1 to 500 ppm of average total impurities after annealing of the copper.

3. The method of claim 2 wherein the impurities comprise one or more of carbon, oxygen, nitrogen, sulfur and chloride.

4. The method of claim 1 wherein the first period is up to 5 seconds.

5. The method of claim 4 wherein the duty cycle has a frequency of 0.1 to 10 Hz.

6. The method of claim 1 wherein the first current density is in the range of 10 to 100 mA/cm².

7. The method of claim 1 wherein the second current density is in the range of 1 to 20 mA/cm².

8. The method of claim 1 wherein the duty cycle has a ratio of step 1) to step 2) of 1:1 to 10:1.

9. An electronic device comprising a first layer of metal and a second layer of metal, wherein the first layer of metal comprises average total impurities in the range of up to 10 ppm and the second layer of metal comprises average total impurities in the range of 10 to 100 ppm.

10. The electronic device of claim 10 wherein the first and second metal layers are copper.