Chemical vapor deposition of ruthenium films for metal electrode applications

[0002] 1. Field of the Invention

[0003] The present invention generally relates to the liquid source chemical vapor deposition of metal films as the electrode on capacitors in integrated circuits. More particularly, bis(ethylcyclopentadienyl) ruthenium is used as the liquid source precursor for chemical vapor deposition of ruthenium films at a low temperature in the kinetic-limited temperature regime as the electrode for the application of metal-insulator-metal capacitors.

[0004] 2. Brief Description of the Related Art

[0005] In metal-insulator-metal capacitors, ruthenium is a preferred electrode material for next-generation dynamic random access memories (DRAMs). As the next generation DRAM technology evolves, it will become necessary to use three-dimensional capacitors despite the high-dielectric constant materials employed. The most important characteristics of a capacitor are high dielectric constant and small leakage current even when the electrode film is thin. If the films are being formed for electrodes for DRAM applications, it is advantageous to use a chemical vapor deposition process (CVD).

[0006] Chemical vapor deposition is a broad class of processes using controlled chemical reactions to create layers on wafers and is a key process in the development of ultra-large-scale integrated circuit fabrication. Chemical vapor deposition of thin metal films realizes good step coverage and wafer-to-wafer repeatability on complicated topography. However, the source material for CVD processes must be stable and possess good vaporization properties.

[0007] The use of liquid source metalorganic precursors for the CVD formation of thin films provides a means of repeatably creating these wafers. CVD precursors for advanced DRAM electrodes and dielectrics have traditionally been solid compounds and although soluble in organic solvents such as tetrahydrofuran, solubility is limited. CVD deposition of films can require high vaporization temperatures and residues are left after vaporization and deposition. Barium strontium titanate thin metal films on a substrate with a BST sputtered film using titanyl bis (dipivaloymethanto) (TIO(DPM)) and titantium tetraisopropoxide (TTIP) organometallic solutions as the CVD liquid source material have been formed.

[0008] Traditionally, bis(cyclopentadienyl) ruthenium (Ru(Cp)₂) is used to deposit thin ruthenium films on substrates. However, this precursor is a solid at room temperature possessing the concomitant problems of limited solubility in tetrahydrofuran, lower deposition rate and remaining residue after vaporization which could cause particle formation, process drift and prevent precursor transport. Additionally, bis(cyclopentadienyl) ruthenium easily forms a ruthenium oxide film on substrates at low temperatures in the kinetic-limited temperature regime for virtually all process conditions.

[0009] It is therefore, advantageous, to use a liquid source for CVD thin film deposition that is a stable liquid at room temperature, yet is less susceptible to oxidation during vaporization and deposition of thin films. Ruthenium films have been deposited on silicon substrates using bis-(ethylcyclopentadienyl) ruthenium. However, these ruthenium films had a large column width and contained large quantities of carbon and hydrogen impurities which resulted in a resistivity greater than that of films deposited using the solid precursor bis-(cyclopentadienyl) ruthenium, although, the resistivity was still sufficiently small to be used as capacitor electrodes.

[0010] Bis-(alkyl-cyclopentadienyl) ruthenium complexes, including Ru(EtCp)₂, have been synthesized and have been used in a process to produce ruthenium-containing films. These pure ruthenium films were deposited on a silicon substrate at 600° C. in a hydrogen atmosphere. Thus, deposition occurred in the mass-transfer limited regime.

[0011] The prior art is deficient in the lack of effective means of depositing pure thin ruthenium films on a substrate using a chemical vapor deposition source that is liquid at room temperature and where deposition temperature occurs in the kinetic-limited temperature range. The present invention fulfills this long-standing need and desire in the art.

[0012] The present invention provides a method of depositing ruthenium films on a substrate via liquid source chemical vapor deposition wherein the source material is liquid at room temperature and utilizes process conditions such that deposition of the ruthenium films occurs at a temperature in the kinetic-limited temperature regime.

[0013] In another embodiment of this invention, there is provided a method of depositing a thin ruthenium film on a substrate by liquid source chemical vapor deposition, using bis-(ethylcyclopentadienyl) ruthenium, by vaporizing the bis-(ethylcyclopentadienyl) ruthenium at a vaporization temperature of about 100-300° C. to form a CVD source material gas and forming a thin ruthenium film on a substrate in a reaction chamber using the CVD source material gas and the oxygen source reactant gas such that the substrate has a temperature of about 100-500° C. and deposition of the ruthenium film occurs in the kinetically limited temperature regime.

[0014] Other and further aspects, features, and advantages of the present invention will be apparent from the following description of the presently preferred embodiments of the invention given for the purpose of disclosure.

[0015] So that the matter in which the above-recited features, advantages and objects of the invention, as well as others which will become clear, are attained and can be understood in detail, more particular descriptions of the invention briefly summarized above may be had by reference to certain embodiments thereof which are illustrated in the appended drawings. These drawings form a part of the specification. It is to be noted, however, that the appended drawings illustrate preferred embodiments of the invention and therefore are not to be considered limiting in their scope.

[0016] FIG. 1 depicts the Thermal Gravimetric Analysis and Differential Scanning Calorimetry data for the Ru(Cp)₂(FIG. 1A) precursor the Ru(EtCp)₂(FIG. 1B) precursors.

[0017] FIG. 2 depicts a schematic of an MLDS dual-vaporizer Ru chamber used for the chemical vapor deposition of ruthenium films. Chemical vapor deposition ruthenium deposition only uses one vaporizer.

[0018] FIG. 3 depicts the Vaporizer Temperature Optimization Curve for the Ru(Cp)₂(FIG. 3A) precursor the Ru(EtCp)₂in octane (FIG. 3B) precursor.

[0019] FIG. 4 depicts the Arrhenius plot of CVD Ru process using Ru(Cp)₂precursor on 200 Å CVD TiN and on 2000 Å thermal oxide (FIG. 4A) and 60 Å PVD Ru seed layer (FIG. 4B) and using Ru(EtCp)₂in octane on 200 Å PVD TiN (FIG. 4C) and 60 Å PVD Ru seed layer (FIG. 4D).

[0020] FIG. 5 depicts the Arrhenius plot of CVD Ru process using pure Ru(EtCp)₂at different Ru and O₂flow rates with constant N₂flow rate of 1300 sccm on a 50 Å PVD Ru seed layer/2 kÅ ThOx substrate. The activation energies are also shown. N₂carrier flow is 450 sccm.

[0021] FIG. 6 depicts the Arrhenius plot of CVD Ru process deposition rate and resistance uniformity using pure Ru(EtCp)₂at flow rates of 50 mgm Ru and 100 sccm O₂(FIG. 6A) and rates of 20 mgm Ru and 40 sccm O₂(FIG. 6B) on a 50 Å PVD Ru seed layer/2 kÅ ThOx substrate. N₂flow is 1300 sccm and N₂carrier flow is 450 sccm.

[0022] FIG. 7 depicts the Arrhenius plot of CVD Ru process deposition rate using pure Ru(EtCp)₂at N₂flow rates of 250 sccm, and 1300 sccm on a 50 Å PVD Ru seed layer/2 kÅ ThOx substrate. Ru flow is 50 mgm and N₂carrier flow is 450 sccm.

[0023] FIG. 8 depicts the Arrhenius plot of CVD Ru process deposition rate and resistivity using pure Ru(EtCp)₂at a high N₂flow rate of 1300 sccm on a 50 Å PVD Ru seed layer/2 kÅ ThOx substrate. Ru flow is 50 mgm, O₂flow rate is 100 sccm and N₂carrier flow is 450 sccm.

[0024] FIG. 9 depicts the X-Ray Diffraction (XRD) pattern of chemical vapor deposition Ru film grown with Ru(EtCp)₂in octane with a deposition temperature of 330° C., substrate of 60 Å physical vapor deposition Ru/2000 Å SiO₂and Ru thickness of 600 Å (FIG. 9A) and with a deposition temperature of 320° C., substrate of 60 Å PVD Ru/2000 Å SiO2 and Ru thickness of 280 Å (FIG. 9B) and with Ru(Cp)₂in tetrahydrofuran with a deposition temperature of 340° C., substrate of 60 Å PVD. Ru/200 Å PVD TiN and Ru thickness of 800 Å (FIG. 9C), with a deposition temperature of 320° C., substrate of 60 Å PVD Ru/200 Å PVD TiN and Ru thickness of 650 Å (FIG. 9D) and, in the kinetic-limited regime, with a deposition temperature of 295° C., substrate of 60 Å PVD Ru/200 Å PVD TiN and Ru thickness of 500 Å (FIG. 9E).

[0025] FIG. 10 depicts the X-Ray Diffraction (XRD) pattern of chemical vapor deposition Ru film grown with Ru(EtCp)₂in octane on a PVD Ru seed substrate on 2000 Å thermal oxide and PVD Ru seed substrate on 200 Å PVD TiN.

[0026] FIG. 11 depicts the effect of O₂flow on CVD Ru orientation, using pure Ru(EtCp)₂on a 50 Å PVD Ru seed layer/2 kÅ ThOx substrate. Temperature is 355° C., N₂flow rate is 250 sccm and N₂carrier flow rate is 450 sccm.

[0027] FIG. 12 depicts the effect of O₂flow on CVD Ru (002) orientation and resistivity using pure Ru(EtCp)₂on a 50 Å PVD Ru seed layer/2 kÅ ThOx substrate. Process conditions are: 355° C., N₂250 sccm, N₂carrier 450 sccm, and 8 torr.

[0028] FIG. 13 depicts SEM Images of the CVD Ru films to compare the surface roughness Ru(EtCp)₂precursor at 343° C. on 200 Å PVD TiN (FIG. 13A), Ru(EtCp)₂precursor at 343° C. on PVD Ru seed/2000 Å Oxide (FIG. 13B), Ru(Cp)₂precursor at 343° C. on PVD Ru seed/200 Å PVD TiN (FIG. 13C), Ru(EtCp)₂precursor at 343° C. on PVD Ru seed/2000 Å Oxide (FIG. 13D), Ru(CP)₂precursor at 320° C. on PVD Ru seed/200 Å PVD TiN (FIG. 13E), Ru(EtCp)₂precursor at 330° C. on PVD Ru seed/ 2000 Å Oxide (FIG. 13F).

[0029] FIG. 14 depicts an AFM 2 μm×2 μm scan of CVD Ru film deposited using Ru(Cp)₂at 343° C. on PVD Ru seed/200 Å PVD TiN and PVD TiN.

[0030] FIG. 15 depicts the comparison of step coverage using Ru(Cp)₂precursor at two different temperatures on 200 Å PVD TiN layer. FIG. 15A: Bottom/Top Ratio ˜30%, Sidewall/Top Ratio ˜40%, Conformality ˜60%. FIG. 15B: Bottom/Top Ratio ˜90%, Sidewall/Top Ratio ˜90%, Conformality ˜80%. FIG. 15C reveals a roughness problem at larger geometry.

[0031] FIG. 16 depicts the comparison of step coverage using Ru(Cp)₂(FIG. 16A) and Ru(EtCp)₂in octane (FIG. 16B) on 200 Å PVD TiN seed layers at 340° C. with 0.17 μm AR 5:1. FIG. 16A: Bottom/Top Ratio ˜35%, Sidewall/Top Ratio ˜30%, Conformality ˜60%.

[0032] FIG. 17 depicts the comparison of step coverage using Ru(EtCp)₂in octane precursor with and without 60 Å Ru seed layer. FIG. 17A: Bottom/Top Ratio ˜30%, Sidewall/Top Ratio ˜30%, Conformality ˜50%. FIG. 17B: Bottom/Top Ratio 90%, Sidewall/Top Ratio ˜90%, Conformality ˜80%. Deposition temperatures were changed slightly due to CVD Ru process substrate sensitivity.

[0033] FIG. 18 depicts the comparison of step coverage using Ru(Cp)₂and Ru(EtCp)₂in octane on 60 Å PVD Ru seed layer. FIG. 18A: Bottom/Top Ratio ˜40%, Sidewall/Top Ratio ˜50%, Conformality ˜80%. FIG. 18B: Bottom/Top Ratio 90%, Sidewall/Top Ratio ˜90%, Conformality ˜80%.

[0034] FIG. 19 depicts the step coverage of a chemical vapor deposition Ru film using pure Ru(EtCp)₂at 355° C. for a 0.15 um trench, AR 6:1. Topfilm thickness 400 Å, side film thickness ˜350 Å, bottom film thickness ˜350 Å, sidewall coverage ˜88%, bottom coverage ˜88%.

[0035] FIG. 20 depicts a demonstration of top electrode application: CVD Ru deposited at 340° C. with Ru(EtCp)₂in octane precursor on 80 Å Ta₂O₅/1500 Å poly-Si substrate. CVD Ru film thickness ˜1500 Å. FIG. 20A: 0.1 um AR 16:1, Bottom/Top Ratio ˜10%, Sidewall/Top Ratio ˜10%, Conformality ˜40%. FIG. 20B: 0.05 um AR 20:1, Bottom/Top Ratio N/A, Sidewall/Top Ratio ˜5%, Conformality ˜25%.

[0036] FIG. 21 depicts SEM Images of CVD Ru films deposited on CVD BST/PVD Pt on, patterned cup structures.

[0037] FIG. 22 depicts process repeatability for Ru(Cp)₂and Ru(EtCp)₂in octane precursors. FIG. 22A: Repeatability plot for 100-wafer run using Ru(Cp)₂precursor. The wafer temperature was 340° C., and the substrate was 200 Å PVD TiN/Si. FIG. 22B: Repeatability plot for 25-wafer run using Ru(EtCp)₂precursor. The wafer temperature was 330° C., and the substrate was 60 Å PVD Ru/2000 Å thermal SiO₂.

[0038] FIG. 23 depicts the effects of using a clamp ring for edge exclusion during CVD Ru deposition. FIG. 23A shows the positioning of the clamp ring; FIG. 23B is a thickness profile by alpha step of the CVD Ru film; FIG. 23C shows the SEM of the cross section of the wafer edge where the clamp ring is positioned; and FIG. 23D shows the presence of particles at the wafer's edge under the clamp ring.

[0039] FIG. 24 depicts the dependence of sheet resistance on film thickness.

[0040] In one embodiment, the present invention provides a method of depositing ruthenium films on a substrate via liquid source chemical vapor deposition wherein the source material is liquid at room temperature and utilizes process conditions such that deposition of the ruthenium films occurs at a temperature in the kinetic-limited temperature regime. A representative example of the liquid source is bis-(ethylcyclopentadienyl) ruthenium. Representative examples of the substrate are thermal oxide, titanium nitride, titanium aluminum nitride, tantalum pentoxide, tantalum nitride, tantalum, barium strontium titanate, strontium oxide, ruthenium oxide, silicon nitride, tungsten nitride, lead zirconium titanate (PZT), strontium bismuth tantalate (SBT), and silicon dioxide. Additionally, the substrate wafers have a first film or seed layer deposited via vapor deposition, for example, physical vapor deposition or chemical vapor deposition. Representative examples of the seed layer are ruthenium, iridium, platinum, titanium nitride, titanium aluminum nitride, tantalum pentoxide, ruthenium oxide and titanium silicide. Also provided are process conditions comprising a ruthenium vaporization temperature of about 100-300° C. and a substrate temperature of about 100-500° C.

[0041] Another embodiment of the present invention provides a method of depositing a thin ruthenium film on a substrate by liquid source chemical vapor deposition, using bis-(ethylcyclopentadienyl) ruthenium as the liquid source. This method comprises vaporizing the bis-(ethylcyclopentadienyl) ruthenium at a vaporization temperature of about 100-300° C. to form a chemical vapor deposition source material gas, providing an oxygen source reactant gas and forming a thin ruthenium film on a substrate in a reaction chamber using the CVD source material gas and the oxygen source reactant gas such that the substrate has a temperature of about 100-500° C. and deposition of the ruthenium film occurs in the kinetically limited temperature regime. Representative examples of the substrate are thermal oxide, titanium nitride, titanium aluminum nitride, tantalum pentoxide, tantalum nitride, tantalum, barium strontium titanate, strontium oxide, ruthenium oxide, silicon nitride, tungsten nitride, lead zirconium titanate, strontium bismuth tantalate, and silicon dioxide. Additionally, the substrate wafers have a first film or seed layer deposited via vapor deposition, for example, physical vapor deposition or chemical vapor deposition. Representative examples of the seed layer are ruthenium, iridium, platinum, titanium nitride, titanium aluminum nitride, tantalum pentoxide, ruthenium oxide and titanium silicide.

[0042] The following definitions are given for the purpose of understanding the present invention. Any terms not expressly defined herein should be given their clear and ordinary meaning in the art.

[0043] As used herein the term, “seed layer” shall refer to a layer of material used to facilitate the growth of another layer in order to have desired properties; e.g., high nucleation density, low roughness and specific orientation.

[0044] As used herein the term, “kinetically-limited temperature regime” shall refer to the range of deposition temperatures over which the deposition rate of a chemical vapor deposition film is limited by the kinetics of the chemical reactions at the substrate surface, typically characterized by a strong dependence of deposition rate on temperature.

[0045] As used herein the term, “mass-transfer limited regime” shall refer to the range of deposition temperatures over which the deposition rate of a CVD film is limited by the flux of chemical reactants to the substrate surface, characterized by a strong dependence of deposition rate on chemical flow rates and independent of deposition temperature.

[0046] Process conditions for CVD deposition of ruthenium films are selected so that the ruthenium films are formed at lower temperatures in the-kinetically limited temperature regime thus yielding pure ruthenium films possessing low oxygen content and low resistivity. Optionally, the substrate has a seed layer which controls the CVD ruthenium film roughness, improves the adhesion to the substrate and improves the step coverage and conformality on patterned structures.

[0047] In determining the optimum deposition conditions, the CVD ruthenium films are deposited using a range of process conditions for all precursors. The following film properties are compared: crystalline orientation, thickness WIW uniformity, resistivity, Rs WIW uniformity, surface roughness, step coverage, and conformality on patterned structures. X-ray diffraction (XRD) analysis of crystalline orientation shows that films deposited with the Ru(Cp)₂precursor in the kinetic-limited regime are significantly oxidized to RuO₂, while films deposited with Ru(EtCp)₂, either pure or 1.0 M in octane, in the same regime still contain significant amounts of pure ruthenium. For thickness within-wafer uniformity, there are metrology issues with the XRF and Metapulse tools, Ru(EtCp)₂in octane and Ru(CP)₂are comparable with regard to resistivity and Rs WIW uniformity. SEM images in conjunction with AFM data show that CVD Ru films have high roughness whether deposited with Ru(EtCp)₂or Ru(Cp)₂.

[0048] The use of a PVD Ru seed layer, approximately 50-60 Å thick, reduces surface roughness. Step coverage results show that conformal CVD Ru films in aggressive structures can be grown using Ru(EtCp)₂in octane or pure Ru(EtCp)₂or Ru(Cp) precursors. Both Ru(CP)₂and Ru(EtCp)₂precursors demonstrate 90% step coverage (sidewall/top and bottom/top) and 80% conformality (sidewall thickness variation) for a 0.17 μm trench with a 5:1 aspect ratio (AR). Pure Ru(EtCp)₂precursor demonstrates ˜88% sidewall coverage and ˜88% bottom coverage for a 0.15 μm trench with a 6:1 (AR). From the demonstrable chemical properties, process performance, and film properties, Ru(EtCp)₂is an excellent precursor for CVD Ru process on the basis of high deposition rate, low residue, good step coverage and wafer-to-wafer repeatability, and oxidation resistance in the kinetic-limited temperature regime.

[0049] The following examples are given for the purpose of illustrating various embodiments of the invention and are not meant to limit the-present invention in any fashion.

[0050] Materials

[0051] Both bis(cyclopentadienyl) ruthenium (Ru(Cp)₂) and bis(ethylcyclopentadienyl) ruthenium (Ru(EtCp)₂) were manufactured and supplied by Advanced Chemical Delivery Systems (ACDS). These precursors are used to deposit CVD Ru films as the electrode for the application of metal-insulator-metal (MIM) capacitors. In general, the CVD Ru film must have the following characteristics: thickness non-uniformity within wafer less than 2% (1σ, 49 pts/9 pts, 15 mmEE); wafer-to-wafer repeatability less than 2% (1σ); film resistivity ρ less than 25 μΩ-cm; and sheet resistance non-uniformity less than 10% (1s, 49 pts, 5 mmEE); and step coverage and sidewall conformality greater than 90%. The film needs to be reasonably smooth so that the capacitor stack is reliable and the film is continuous at thicknesses of 300 Å and below. Table 1 lists the target specifications together with the current process performance and best results for CVD Ru using Ru(EtCp)₂.

[0052] Physical Properties of Ru(Cp)₂and RU(EtCp)₂

[0053] The Ru(Cp)₂precursor is a solid at room temperature and is dissolved in the solvent Tetrahydrofuran (THF) in which the maximum allowable solubility is only 0.12 M. The Ru(EtCp)₂precursor is a liquid at room temperature and is dissolved in octane to 1.0M concentration or used in pure form. Table 2 compares the basic properties of the two precursors.

[0054] Thermal Gravimetric Analysis (TGA) and Differential Scanning Calorimetry (DSC)

[0055] A combination of Thermal Gravimetric Analysis (TGA) and Differential Scanning Calorimetry (DSC) was used to evaluate the precursors Ru(Cp)₂and Ru(EtCp)₂(FIGS. 1A and 1B). These measurements are performed at atmospheric pressure. The thermal gravimetric analysis curve measures the sample mass as a function of temperature. A decrease in sample mass occurs during evaporation, sublimation, or decomposition. The mass of the involatile residue remaining after thermolysis indicates the amount of residue that is left after vaporization. When the precursor has no residue, the process has less potential. for particles and process drifts. The Ru(Cp)₂precursor has a residue of <3 wt % while Ru(EtCp)₂has close to 0 wt % residue.

[0056] Differential Scanning Calorimetry (DSC) curve measures the relative amount of heat flow from the sample with respect to a reference material. The differential scanning calorimetry curve shows the temperatures at which endothermic and exothermic transitions take place. The endothermic behavior indicates sublimation, evaporation or melting while exothermic reactions indicate reactions or decomposition. Thus, from the differential scanning calorimetry curve, the temperature range at which decomposition and sublimation occur are estimated. The information from the curve provides a starting point at which to set the vaporizer temperature. Ru(Cp)₂indicates an endothermic reaction (melting point) at 200° C. while a very broad endothermic peak is observed near 240° C. for Ru(EtCp)₂.

[0057] Ru(EtCp)₂does not have limited solubility in solvents and can be used as a neat (pure) liquid or in a highly concentrated solution with a simple organic solvent, e.g., octane or tetrahydrofuran. As a result, the deposition rate is not limited as it is for the Ru(Cp)₂precursor. Thermal gravimetric analysis data shows that Ru(EtCp)₂leaves virtually no residue in the vaporizer or heated lines which could potentially lead to particles and process drift. Also, because Ru(EtCp)₂is a liquid at room temperature, it is less likely to leave solid residues which can block precursor transport through the liquid lines.

[0058] Hardware and Process Conditions

[0059] All data was collected on an Applied Materials Ru chamber with the following configuration (FIG. 2):

[0060] Chamber: Ru

[0061] Faceplate: Standard

[0062] Heater: NGK HA-12 (95%)

[0063] Vaporizer: MLDS dual vaporizer (No precursor flow through 2^ndvaporizer)

[0064] Frit: Dual, 100 μm pore size

[0065] Tip: 10 mil tapered to 4 mil, 0.1″ insertion length into top flit

[0066] Ampule Push Gas: 65 psi

[0067] The process conditions, unless specified otherwise, for Ru(Cp)₂and Ru(EtCp)₂comparisons were:

[0068] Ru(Cp)₂(0.1 M in tetrahydrofuran)

[0069] Heater Temp. 330-350° C.; Vaporizer Temp. 220° C. Lid/Liner/Nose Temp. 190° C.; Ru Flow 300 mgm; O₂flow 300 sccm; N₂-B Ru 450 sccm; N2-B Pt 250 sccm. (2nd vaporizer); ˜Pressure 8 Torr; and D/R ˜60 Å/min at 340° C. on PVD Ru seed

[0070] Ru(EtCp)₂(1 M in octane):

[0071] Heater Temp. 330-350° C.; Vaporizer Temp. 260° C.; Lid/Liner/Nose Temp. 260° C.; Ru flow 180 mgm; O₂flow 300 sccm; N₂-B 450 sccm (1^stvaporizer); N₂-B 250 sccm (2^ndvaporizer); Pressure 8 Torr; and D/R ˜100 Å/min at 330° C. on PVD Ru seed

[0072] Ru(EtCp)₂(Pure):

[0073] Heater Temp. 330-350° C.; Vaporizer Temp. 260° C.; Lid/Liner/Nose Temp. 220° C.; Ru flow 50 mgm; O₂flow 100 sccm; N₂Flow 1300 sccm (to chamber); N₂-B 450 sccm (1 vaporizer only); Pressure 2 Torr; and D/R ˜450 Å/min at 330° C. on PVD Ru seed

[0074] Vaporizor Temperature Optimization

[0075] In order to evaluate the performance of Ru(Cp)₂and Ru(EtCp)₂on the Applied Materials Ru Captiva tool, vaporizer temperature optimization was performed. The Ru deposition rate was plotted against the vaporizer temperature (FIGS. 3A and 3B).

[0076] Based on these temperature optimization curves, a temperature of 220° C. for the vaporizer using Ru(Cp)₂and 260° C. using Ru(EtCp)₂was selected. These respective temperatures yield near the maximum deposition rate and good wafer-to-wafer repeatability for the respective precursors. The heated lines from the vaporizer to the chamber have the same temperatures. For Ru(Cp)₂, however, the lid, liner, and nose temperatures were lowered, to 190° C. as a precautionary measure due to observed reaction with O₂in the showerhead area at a temperature of 250° C.

[0077] Kinetic study using Ru(Cp)₂(0.1 M in tetrahydrofuran) and Ru(EtCp)₂(1.0 M in octane)

[0078] The deposition temperature was varied for both Ru(Cp)₂and Ru(EtCp)₂precursors to determine the temperature dependence of the deposition rate using Arrhenius plots (log deposition rate vs. 1/T). Arrhenius plots are used to understand the reaction kinetics and to determine the temperature sensitivity of the CVD process. For good step coverage and conformality, the CVD process needs to be operated below the “knee” in the kinetic-limited regime. The “knee” is defined as the transition point between the mass-transfer limited regime and the kinetic-limited regime. Higher activation energies are generally not desirable for a stable metal organic (MOCVD) process due to high temperature sensitivity. The temperature dependence of the CVD Ru process for both Ru(Cp)₂and Ru(EtCp)₂precursors varies on different substrates thereby entailing an Arrhenius study on every substrate used.

[0079] The four substrates used for these Arrhenius studies are 2000 Å thermal oxide, 200 Å CVD TiN, 200 Å PVD TiN, and 60 Å PVD Ru seed layer. The CVD Ru process using either Ru(Cp)₂or Ru(EtCp)₂precursor is sensitive to substrate, so deposition rate and film properties vary on different substrates. Initially, thermal oxide substrates were used as substrates, but adhesion problems (peeling) and poor film quality and appearance were observed. For the CVD Ru(Cp)₂process, CVD TiN and PVD TiN substrates improved the film adhesion and appearance. However, for CVD Ru(EtCp)₂, the haze and surface roughness were found to be very high on the TiN substrates. The use of a PVD Ru seed layer, approximately 60 Å thick, dramatically improved the haze of the CVD Ru(EtCp)₂films.

[0080] FIG. 4A shows the temperature dependence of the deposition rate for Ru(Cp)₂deposited on PVD TiN and thermal oxide substrates. The CVD Ru process for this Arrhenius plot uses slightly different conditions (e.g. 2 Torr pressure, O₂flow 500 sccm). The deposition rate dependence was similar on thermal oxide and CVD TiN. The activation energy is about 2-4 eV for Ru(Cp)₂on TiN and thermal oxide substrates, and the “knee” was close to 300° C.

[0081] The Arrhenius plot of Ru(Cp)₂on a PVD Ru seed layer is shown in FIG. 4B. The estimated activation energy of ˜0.7 eV was approximately four times lower than the activation energy of Ru(Cp)₂on TiN. A lower activation energy improves the process stability in the kinetic-limited regime. The temperature dependence of film resistivity is also plotted in FIG. 4B. The resistivity increases from about 20 μΩ-cm to 60 μΩ-cm when the temperature is reduced from 330° C. to 308° C. The sudden increase in resistivity is likely due to the formation of RuO₂at the lower temperatures as indicated in XRD analysis. There appears to be a transition point around 320° C. in which RuO₂is preferentially formed.

[0082] The Arrhenius plot of Ru(EtCp)₂on PVD TiN shown in FIG. 4C has an activation energy of ˜5 eV which is higher than Ru(Cp)₂. Also, the curve and “knee” are shifted to slightly higher temperatures than the Ru(Cp)₂precursor. The higher activation energy is perhaps caused by the higher deposition rates achieved using the Ru(EtCp)₂precursor.

[0083] The Arrhenius plot and resistivity dependence for the Ru(EtCp)₂precursor on PVD Ru seed substrates is shown in FIG. 4D. The curve was shifted to higher temperatures relative to the Arrhenius plot for Ru(Cp)₂on PVD Ru seed substrates. The estimated activation energy for Ru(EtCp)₂over the observed temperature range was about 3 eV, approximately four times higher than the activation energy for Ru(Cp)₂on PVD Ru seed.

[0084] The resistivity increases gradually to 40 μΩ-cm as the deposition temperature was lowered, in contrast to the abrupt increase in resistivity observed with the Ru(Cp)₂precursor. In this case, the decrease in resisitvity may be caused by several factors: smaller grain size, decrease in film thickness, and some initial RuO₂formation. However, Ru(EtCp)₂appears to form less RuO₂than Ru(Cp)₂in the kinetic-limited regime.

[0085] The study of Arrhenius behavior of CVD Ru(CP)₂and CVD Ru(EtCp)₂shows both processes to be sensitive to substrate. For both Ru(Cp)₂and Ru(EtCp)₂precursors, the activation energy was lower when deposited on a PVD Ru seed layer rather than on other substrates such as TiN and oxide. In addition, the deposition rates on PVD Ru seed layers are higher than on TiN, BST, or TanOx (latter two results not shown) in the kinetic-limited regime. The lower activation energies and higher deposition rates on PVD Ru highlight the importance of using a PVD Ru seed layer as a substrate for CVD Ru growth.

[0086] The Ru(EtCp)₂process has a higher activation energy and higher “knee” temperature than Ru(Cp)₂. However, at lower temperatures (T<320° C.), the wafers deposited with Ru(Cp)₂form primarily the RuO₂phase. Thus, below 320° C., the activation energy is mainly controlled by the formation of RuO₂rather than ruthenium making it difficult to compare the activation energies for Ru(EtCp)₂and Ru(Cp)₂.

[0087] The plots of resistivity temperature dependence show resistivity increasing as temperature decreases for both precursors. The increase in resistivity is more dramatic for Ru(Cp)₂. The higher resistivity for Ru(Cp)₂is probably due to the formation of RuO₂, which has a higher bulk resistivity (45 μΩ-cm) than pure Ru (7 μΩ-cm).

[0088] Ruthenium Film Properties

[0089] The CVD Ruthenium film properties and process performance are summarized below in Table 3.

[0090] Kinetic Study Using Pure Ru(EtCp)₂

[0091] Using a pure Ru(EtCp)₂precursor lowers the activation energy to 0.8-0.9 eV. Activation energy is also dependent on the rate of Ru flow. Using a CVD Arrhenius plot, comparing the Ru flow rates of 20, 50 and 80 mgm with concomitant O₂flows of 40, 160 and 100 sccm and a constant N₂flow on deposition rates on a 50 Å PVD Ru/2 kÅ ThOx substrate indicates that best conditions occur at Ru flow 50 mgm with an activation energy of 0.79 eV (FIG. 5). The higher activation energies at low temperatures for high Ru flow may possibly be due to a lack of N₂flow. At an Ru flow rate of 50 mgm, uniformity increases rapidly as the temperature decreases to the optimal 335° C. Under these conditions Ru CVD yields a silver Ru film with a resistivity of 22 μΩ-cm (FIG. 6A). At an Ru flow of 20 mgm on the same substrate, uniformity increases dramatically below 345° C., a result of increased temperature sensitivity, and resistivity remains fairly constant at temperatures below 335° C., (FIG. 6B); under these process conditions the heater has a 4° C. max.-min. variation for setpoints from 300-350° C.).

[0092] Increasing the N₂flow increases the deposition rate of Ru films on a 50 Å PVD Ru/2 kÅ ThOx substrate and lowers the activation energy in the kinetic-limited regime (FIG. 7), whereas a higher N₂flow reduces the deposition rate in the mass transport regime. Additionally, at high N₂flow, 1300 sccm, and 50 mgm Ru flow on the same substrate, resistivity increases at lower deposition temperatures (FIG. 8). Generally, increasing the N₂flow rate improves the Ru CVD process, making it more stable and less temperature dependent. It is therefore possible to tune the nitrogen and oxygen flow rates and deposition temperature to achieve the desired Ru film.

[0093] X-Ray Diffraction Data of RU(Cp)₂(0.1 M in tetrahydrofuran) and Ru(EtCp)₂(1.0 M in octane)

[0094] FIG. 9A shows an X-Ray diffraction plot of films grown with Ru(EtCp)₂precursor in octane with a deposition temperature of 330° C., in the kinetic-limited regime. This film has a relatively strong Ru(100) orientation as well as Ru(002) and Ru(101) peaks. Although no peaks for RuO₂were present, the resistivity of this film is greater than 30 μΩ-cm and RBS analysis shows that the film has 30±10% oxygen. This film was shiny and metallic in appearance. At 320° C., the X-Ray diffraction pattern (FIG. 9B) shows some evidence of RuO₂formation with weak RuO₂(110) and RuO₂(002) peaks, but the dominant phase was Ru.

[0095] CVD Ru films grown with the Ru(Cp)₂precursor in tetrahydrofuran have different crystalline orientation than films grown with Ru(EtCp)₂. FIG. 9C shows an X-Ray Diffraction pattern of CVD Ru grown with Ru(Cp)₂at 340° C., in the mass-transfer limited regime. This film has a relatively strong Ru(002) peak as well as Ru(101) and Ru(100) peaks. At 320° C., some RuO₂phase was mixed with the Ru phases (FIG. 9D). In the kinetic-limited regime at 295° C., the film was almost completely RuO₂(FIG. 9E).

[0096] The extent of oxidation of the films grown with Ru(Cp)₂in the kinetic-limited temperature regime appears to be greater than the oxidation of films grown with Ru(EtCp)₂. In fact, it has not been possible to grow films with mostly pure Ru phase at low temperatures using Ru(Cp)₂using a wide range of gas flows and pressures. On the other hand, the O₂flow can be varied using the, Ru(EtCp)₂precursor to form mostly pure Ru phase or mostly RuO₂phase. In this way the Ru(EtCp)₂precursor has the advantage of tuning the film Ru/O composition ratio under low-temperature conditions where conformal films can be deposited.

[0097] The orientation of CVD Ru films depends on the substrate even when a PVD Ru, seed is used. FIG. 10 shows the X-Ray Diffraction patterns of CVD Ru using Ru(EtCp)₂on two slightly different substrates each having a ˜60 Å PVD Ru seed layer. One wafer had 2000 Å thermal oxide under the PVD Ru layer while the other had 200 Å PVD TiN. The CVD Ru film with the Ru seed on thermal oxide has a very strong Ru(101) peak while the film with Ru seed on TiN had only weak Ru peaks. This result emphasizes the significance of the substrate on the film properties even when a thin PVD Ru seed layer was used.

[0098] X-Ray Diffraction Data of Ru Films Grown with Pure Ru(EtCp)₂

[0099] As with the Ru(EtCp)₂precursor in octane, pure Ru(EtCp)₂precursor can be used to tune the Ru/O film composition ratio to control the crystalline orientation of the Ru films being deposited. The best Ru(002) orientation occurs at the lowest O₂flow, FIG. 11 shows the effect of O₂flow on CVD Ru orientation on a 50 Å PVD Ru/2 kÅ ThOx substrate at 355° C. RuO₂formation begins to occur at O₂flow 1200 sccm as evidenced by the abrupt appearance of RuO₂(101) peaks and, to a lesser extent, RuO₂peaks at 200 and 110.

[0100] An examination of the effect of O₂flow on CVD Ru (002) orientation at 355° C. using process conditions of Ru 50 mgm, N₂250 sccm and 8 torr of pressure, again indicates an abrupt increase in the relative intensity of the 002 peak at O₂flow 1200 sccm (FIG. 12). The concomitant increase in resistivity to greater than 30 μΩ-cm suggests the appearance of RuO₂in the deposited film. Below ˜900 sccm O₂flow, the film is silver and has a high Ru content which becomes a more brown/purple hue as the RuO₂content increases with an increase in the O₂flow rate.

[0101] Within-Wafer Film Thickness Uniformity

[0102] Measuring the within-wafer thickness uniformity of CVD Ru by XRF has been limited due to a background signal that varies from the center to the edge of the wafer. Uncorrected WIW thickness uniformity values by XRF ate typically 10-15% 1σ. The WIW thickness uniformity for blank substrates (which should measure zero thickness everywhere) is also very high. It is necessary to zero the background signal at all points on the wafer surface to ensure that reliable WIW measurements are made.

[0103] The Rudolph Metapulse is another tool that may be capable of measuring within-wafer Ru thickness. Initial data using the Ru(Cp)₂precursor on TiN/Si substrates indicates the WIW thickness uniformity is as low as 2% 1σ. However, a discrepancy exists between the WIW thickness uniformity numbers and the XRF measurement. Possibly the Metapulse measurement is affected by poor adhesion between the CVD Ru and TiN layers. A second set of CVD Ru films grown on 60 Å. PVD Ru/2 kÅ SiO₂yielded very weak signals. In this instance the oxide thickness could be reducing the film reflectivity. Metapulse measurements using SiO₂substrates less than 1000 Å thick may correct the problem.

[0104] Within-Wafer Sheet Resistance Uniformity

[0105] Sheet resistance was measured by the KLA-Tencor 4-point probe analyzer. The within-wafer sheet resistance uniformity was taken from 49 points measured across the wafer using a 5 mm edge exclusion. A 5 mm edge exclusion was used to accommodate the PVD seed layer or adhesion layer which has a 3 mm edge exclusion from the clamp ring. The MRS requirement for WIW Rs uniformity (49pt) was 10% 1σ. The CVD Ru films using either precursor were able to achieve WIW Rs uniformity below 10% 1σ using the PVD Ru seed layer. Table 3 above shows the dramatic improvement in Rs uniformity by using a PVD Ru seed layer. For Ru(Cp)₂, the Rs uniformity was reduced from 30% to 5% 1σ.

[0106] Dependence of Sheet Resistance on Film Thickness and Deposition Temperature

[0107] The CVD Ru film resistivity (ρ) increases with decreasing deposition temperature. Typical ρ vs T data is shown on the Arrhenius plots for the Ru(Cp)₂(FIG. 4B) and Ru(EtCp)₂(FIG. 4D) precursors. The trend was not linear but was instead sigmoidal; a sharp increase in film resistivity is seen near the transition from mass-transfer to kinetic-limited temperature regime. This increase in resistivity was most likely the result of increasing O incorporation in the Ru film with decreasing deposition temperatures. XRD patterns confirm the emergence of RuO₂peaks at lower process temperatures. The extent of oxidation in the kinetic-limited temperature regime is greater for films grown with the Ru(Cp)₂precursor relative to the Ru(EtCp)₂in octane precursor, as evidenced by greater intensities of RuO₂peaks and higher film resistivity (FIG. 9E).

[0108] Surface Roughness Comparison

[0109] The surface roughness and morphology of Ru films deposited by CVD varies depending on the wafer temperature and substrate. CVD Ru films were deposited on PVD TiN as well as on PVD Ru seed layers with PYD TiN. SEM images were taken of the ruthenium films to compare the surface morphologies. CVD Ru films deposited directly onto PVD TiN using Ru(EtCp)₂in octane precursor had very rough surfaces (FIG. 13A). Using the same process conditions, the surface morphology was improved by using a thin PVD Ru seed layer, approximately 60 Å in thickness below the CVD Ru film (FIG. 13B). Both films were deposited using Ru(EtCp)₂in octane precursor at a wafer temperature of 343° C.

[0110] With the Ru(Cp)₂precursor, the surface roughness of the CVD Ru film was slightly better on PVD TiN than on the PVD Ru seed. FIG. 14 shows the AFM 2 μm×2 μm scans comparing CVD Ru on the PVD Ru seed with the PVD TiN substrate. Both films were deposited using the same 343° C. Ru(Cp)₂process. Even though the RMS values cannot be directly compared because CVD Ru thickness was 50% greater for the sample on PVD Ru than the one on PVD TiN, the RMS roughness value was still 100% greater.

[0111] Comparing the surface morphology at 343° C. using a PVD Ru seed layer, the film deposited with Ru(Cp)₂looks smoother than the films deposited with Ru(EtCp)₂(FIGS. 13C and 13D). At a wafer temperature of 343° C., the process was mass-transfer limited for both precursors. In Table 4, AFM results show the roughness values to be fairly high for both films. The RMS value is 121.9 Å for a 1430 Å film deposited with the Ru(EtCp)₂precursor and 62 Å for a 710 Å film deposited with the Ru(Cp)₂precursor.

[0112] FIGS. 13E and 13F compare the two films at their respective temperatures in the kinetic-limited regime. At a wafer temperature of 320° C., the film deposited with Ru(Cp)₂has raised and depressed regions which are detectable at lower magnifications, 2,000 to 10,000 times. This overlying macroscopic pattern causes visible haze on the wafer. FIG. 13F shows that the film deposited using Ru(EtCp)₂at 331° C. had small grains with sharp, high aspect ratio features but no overlying pattern. The RMS value for a 540 Å film deposited with Ru(EtCp)₂was 66 Å.

[0113] Overall, the CVI) Ru films deposited using either Ru(EtCp)₂or Ru(Cp)₂as precursor appear to be quite rough. The average roughness RMS values were about 10% of the total film thickness for each of the conditions. The surface roughness does not seem to change dramatically with temperature. However, surface morphology was affected by the wafer temperature and the grain size tends to increase with temperature. For Ru(EtCp)₂precursor, the surface roughness improves by using a PVD Ru seed layer relative to a PVD TiN layer. The opposite trend was observed for the Ru(Cp)₂precursor.

[0114] Although surface roughness is an important consideration with the CVD Ru films, the actual roughness requirement has not been defined. It is not clear which precursor, Ru(EtCp)₂or Ru(Cp)₂, produces smoother films. The substrate seems to be a strong variable that affects surface roughness and morphology. Additionally, process parameters such as pressure, carrier gas flow, and spacing, also can affect surface roughness.

[0115] Step Coverage

[0116] Ru(Cp)₂and Ru(EtCp)₂in octane precursor performance for step coverage and conformality were compared (FIGS. 15-20). The comparisons include a) deposition temperature, b) type of Ru precursor, and c) substrate (PVD Ru seed layer vs. TiN). Conformality is defined as the thinnest part of the test structure sidewall divided by the thickest part of the sidewall. Step coverage refers to sidewall (sidewall thickness/top thickness) and bottom coverage (bottom thickness/top thickness). Average thicknesses were used for the step coverage calculation.

[0117] For CVD processes, lower temperatures in the kinetic-limited regime were typically required to achieve good step coverage and conformality. FIG. 15 shows films deposited on PVD TiN using Ru(Cp)₂at two different wafer temperatures, 343° C. (FIG. 15A) and 320° C. (FIG. 15B). At 343° C. the step coverage for a 0.17 μm structure with AR 5:1 is ˜40% on the sidewall and ˜30% on the bottom while the conformality is ˜60%. The SEM cross-section images show better step coverage for the film deposited at 320° C., below the “knee” in the kinetic-limited regime. Step coverage (sidewall and bottom) for the 320° C. process is ˜90%, and conformality is ˜80% for a 0.17 μm structure at an aspect ratio (AR) of 5:1. However, an SEM of a structure with larger geometry (0.28 μm AR 3:1) reveals that the film is very rough and not continuous in some areas.

[0118] FIG. 16 compares the step coverage of Ru(Cp)₂(FIG. 16A) and Ru(EtCp)₂in octane (FIG. 16B) films at 343° C. on PVD TiN substrates. The SEM images reveal that the Ru(EtCp)₂in octane film at 340° C. is extremely rough with faceted and discontinuous grains.

[0119] By depositing CVD Ru onto a PVD Ru seed layer, the film roughness and step coverage can be greatly reduced. FIG. 17 compares CVD Ru films with (FIG. 17B) and without (FIG. 17A) a PVD Ru seed layer using Ru(EtCp)₂in octane. Because the CVD Ru process is slightly substrate dependent, the film without the PVD Ru seed layer is deposited at 340° C.; however, the two films are deposited in similar temperature regimes. The conformality and step coverage were much better with the PVD Ru seed layer, and the surface was smoother. The PVD Ru seed layer is a good template substrate for CVD Ru growth on patterned structures.

[0120] FIG. 18 shows SEM crosssections of CVD Ru films deposited onto ˜60 Å PVD Ru seed layers. The film deposited using the Ru(EtCp)₂in octane (FIG. 18B) precursor demonstrates much better step coverage with ˜90% sidewall and bottom coverage for a feature size, of 0.15 μm and AR 6:1. The Ru(Cp)₂(FIG. 18A) film only showed 50% sidewall coverage and 40% bottom coverage for the same structure. Both films have about 80% conformality that is higher than the films deposited without the PVD seed layer.

[0121] FIG. 19 shows an SEM crosssection of a CVD Ru film using pure Ru(EtCp)₂on a 50 Å PYD,Ru/2 k Å substrate at 355° C. for a feature size of 0.15 1 μm and AR 6:1. Rs film uniformity is ˜8% under these conditions. Top film thickness is 400 Å and side film thickness is 350 Å with both sidewall and bottom coverage at ˜88%.

[0122] FIG. 20 shows the step coverage and conformality of CVD Ru deposited directly on a Ta₂O₅/poly-Si substrate without a PVD Ru seed layer at 340° C. The CVD Ru can not be deposited at the best known process temperature of 330° C. used for PVD Ru seed layers. The step coverage and conformality were quite good considering the narrow spacing and severe aspect ratios. The CVD Ru surface was smooth, even in the absence of PVD Ru seed, Thus, CVD Ru can be used as a top electrode for TanOx MIM applications even for very high aspect ratios.

[0123] The SEM images in FIG. 21 indicate that CVD Ru films do not grow uniformly on CVD BST films. CVD Ru was deposited on 300 Å BST/PVD Pt on 0.2 μm wide cup structures with AR 1.5:1. As with the deposition on Ta₂O₅, the process temperature had to be increased to 340° C. From the SEM photos, it appears that the CVD Ru growth is not controlled well on the BST surface. Large crystallites of ruthenium are seen randomly spread across the wafer inside and outside of the cup structures. In some areas, the ruthenium seems to be depositing along the sidewalls but the film looks rough.

[0124] In summary, conformality and step coverage improve at lower temperatures, especially in the kinetic-limited temperature regime. Lower deposition temperatures also reduce surface roughness. Both Ru(Cp)₂and Ru(EtCp)₂produce CVD Ru films with good step coverage and conformality in the kinetic-limited temperature regime; however, the deposition rate is higher using Ru(EtCp)₂in comparable temperature regimes resulting in greater wafer throughput. Additionally, using a PVD Ru seed layer reduces surface roughness and appears to improve nucleation of CVD Ru films deposited with Ru(EtCp)₂.

[0125] Process Condition Repeatability

[0126] The wafer-to-wafer thickness for both the Ru(Cp)₂(FIG. 22A) and Ru(EtCp)₂(FIG. 22B) precursors meet the C&F goal of less than 2% 1 sigma. However, the run with Ru(EtCp)₂has a process drift, with the average thickness increasing 10-15 Å over 25 wafers. As this is only a 25 wafer run, longer run is needed to determine whether this trend of increasing thickness is an anomaly.

[0127] Edge Exclusion Ring

[0128] An aluminum clamp ring for edge exclusion of CVD Ru deposition was tested on a total of 8 wafers using the BKM CVD Ru process. Edge exclusion of CVD Ru is a common requirement and may be accomplished by means other than a clamp ring such as is well known by those having ordinary skill in this art. It is noted that the clamp ring was not centered on the wafer (FIG. 23A). This pattern was present for all 8 wafers. A centering mechanism would make the clamp ring concentric to the wafer. The film edge was reasonably well-defined at the top half of the wafer, suggesting a good contact to the wafer surface (FIG. 23B). The edge exclusion was primarily opposite the notch and extended approximately over the top half of the wafer only, as indicated by optical microscope observations of the film edge (FIG. 23C). On the very first wafer, particles were observed under the clamp ring at the top half of the wafer (FIG. 23D). The number of particles was reduced after several wafers were run.

[0129] The effects of the clamp ring on the CVD Ru process are identified below in TABLE 5. The clamp ring was allowed to touch the wafer for 60 seconds before the start of the deposition. Heater-showerhead spacing was 400 mil.

[0130] data taken with Ru(EtCp)₂precursor at 330° C. using PVD Ru seed/TiN/Si substrates; data with clamp ring Ru9117; data without clamp ring Ru9124. *Note: high WIW uniformity due to varying background signal (wafer center to edge) in XRF. WIW uniformity is typically less than 5% 1_ from Metapulse (not measured on both sets of samples).

[0131] The biggest impact of the clamp ring was on the within-wafer Rs uniformity. The clamp ring may be affecting the wafer temperature near the edge and thus alter the film, The WIW thickness uniformity was similar; however, the current XRF measurement for WIW uniformity was not accurate. In addition, the deposition rate was slightly lower, and the film resistivity was slightly higher.

[0132] Metrology

[0133] One possible method of measuring CVD Ru thickness is using the sheet resistance measurement of the 4-point probe. FIG. 24 shows how sheet resistance varies with film thickness. If the film resistivity was assumed to be constant over this thickness range, then this is essentially a 1/thickness vs. thickness curve. With a proper calibration, the film thickness could be determined from the measured sheet resistance.

[0134] The within-wafer thickness uniformity of CVD Ru films was difficult to measure by XRF due to a large background signal from the silicon substrate that varies from wafer center to edge. While it may be possible to correct for this problem using proper calibration procedures, an alternative method is the Rudolph Metapulse tool, which measures laser-induced sound wave pulses in the film to obtain film thickness.

[0135] Ru(CP)₂vs. Ru(EtCp)₂precursors: Summary

[0136] The Ru(Cp)₂precursor has a limited solubility (0.12M in THF) which limits the CVD deposition rate. The current deposition rate for the 343° C. process using Ru(Cp)₂is about 70 Å/min in the mass-transfer-limited regime. In contrast, the Ru(EtCp)₂precursor, which is a liquid at room temperature and is currently used at 1M in octane, has a current deposition rate of greater than 300 Å/min in the mass-transfer-limited regime and greater than 100 Å/min for the BKM process in the kinetic-limited regime.

[0137] The residue from the TGA curve is zero for Ru(EtCp)₂, while it is about 2% for Ru(Cp)₂. Finite residues are not desirable because residues can result in particles and/or process drift.

[0138] The step coverage using the Ru(EtCp)₂precursor with a PVD Ru seed layer was close to 90% with the BKM process. The step coverage on TanOx substrates using Ru(EtCp)₂for top electrode applications has excellent filling capabilities to greater than 10:1 AR.

[0139] The wafer-to-wafer thickness repeatability using Ru(EtCp)₂meets the MRS alpha-exit requirements of less than 2% 1σ.

[0140] Compared with Ru(Cp)₂, Ru(EtCp)₂has a wider process window for tuning the O/Ru film ratio. Ru(EtCp)₂is less susceptible to oxidation during MOCVD Ru deposition at low temperatures. A low-oxygen Ru film with low film resistivity can be formed in the kinetic-limited temperature regime using certain process conditions. By changing process parameters, e.g., higher O₂flow, it is also possible to form high-resistivity RuO₂films with Ru(EtCp)₂. This is in contrast with Ru(Cp)₂which has a greater tendency to form high-resistivity RuO₂films in the kinetic-limited regime for nearly all process conditions. See Table 1 for a summary of CVD Ru film deposition using Ru(EtCp)₂as the precursor as compared with Alpha Exit Specifications. Table 6 provides a summary of the processing parameter ranges for forming an Ru film using using Ru(EtCp)₂as the precursor for metalorganic chemical vapor deposition (MOCVD) while Table 7 provides specific process parameters for optimal deposition of Ru films.

[0141] Any patents or publications mentioned in this specification are indicative of the levels of those skilled in the art to which the invention pertains. These patents and publications are herein incorporated by reference to the same extent as if each individual publication was specifically and individually indicated to be incorporated by reference.

[0142] One skilled in the art will readily appreciate that the present invention is well adapted to carry out the objects and obtain the ends and advantages mentioned, as well as those inherent therein. It will be apparent to those skilled in the art that various modifications and variations can be made in practicing the present invention without departing from the spirit or scope of the invention. Changes therein and other uses will occur to those skilled in the art which are encompassed within the spirit of the invention as defined by the scope of the claims.