Login| Sign Up| Help| Contact|

Patent Searching and Data


Title:
VOLATILE LIQUID COPPER PRECURSORS FOR THIN FILM APPLICATIONS
Document Type and Number:
WIPO Patent Application WO/2008/085426
Kind Code:
A1
Abstract:
The present invention provides copper +1 compounds for film depositon of copper that retain the low tempertaure processing advantages typical for copper +1 compounds, but which have the thermal stability normally associated with copper +2 compounds. Such compounds are suitable for use as copper precursors for the growth of thin copper films for various applications by Chemical Vapor Deposition (CVD) or Atomic Layer Deposition (ALD).

Inventors:
NORMAN, John Anthony Thomas (709 West Bluff Drive, Encinitas, California, 92024, US)
Application Number:
US2007/026215
Publication Date:
July 17, 2008
Filing Date:
December 21, 2007
Export Citation:
Click for automatic bibliography generation   Help
Assignee:
AIR PRODUCTS AND CHEMICALS, INC. (7201 Hamilton Boulevard, Allentown, Pennsylvania, 18195-1501, US)
NORMAN, John Anthony Thomas (709 West Bluff Drive, Encinitas, California, 92024, US)
International Classes:
C07F7/00; C23C8/00; C07F7/00; C23C8/00
Attorney, Agent or Firm:
ROSSI, Joseph D. (Synnestvedt & Lechner LLP, 1101 Market Street 26th Floo, Philadelphia Pennsylvania, 19107-2950, US)
Download PDF:
Claims:

CLAIMS

1. A metal complex of the structure:

where R 1 , R 2 and R 3 are independently H, alkyl or fluoroalkyl groups, but at least one of R 1 , R 2 and R 3 is fluoroalkyl and M is a monovalent metal and L is a ligand selected from the group consisting of alkene, alkyne, nitrile, isonitrile, cyanate, isocyanate, imine and phosphine, capable of coordinating to the metal M, R 4 is an aliphatic hydrocarbon chain.

2. The metal complex of Claim 1 wherein the R 4 aliphatic hydrocarbon chain is normal or branched C 1-10 .

3. The metal complex of Claim 2 wherein the R 4 aliphatic hydrocarbon chain includes functional groups selected from the group consisting of oxygen, nitrogen, silicon and mixtures thereof.

4. The metal complex of Claim 1 wherein the metal is copper.

5. The metal complex of Claim 1 wherein the fluoroalkyl is C 1-10 .

6. The metal complex of Claim 5 wherein the fluoroalkyl is perfluorinated.

7. The metal complex of Claim 1 in admixture with other liquid molecules that enhance the ALD or CVD process, and the resulting mixtures evaporation for delivery to the deposition chamber by DLL

8. The metal complex of Claim 1 wherein the ligand L is normal or branched C M0 alkenyl.

9. The metal complex of Claim 8 wherein the ligand L is C 2 alkenyl.

10. The metal complex of Claim 1 wherein M is copper, in admixture with a metal complex of Claim 1 wherein M is selected from the group consisting of titanium, tantalum, hafnium, silver, gold, tungsten and nickel, capable of the deposition of copper alloys.

11. The metal complex of Claim 1 of the structure:

12. The metal complex of Claim 1 of the structure:

13. The metal complex of Claim 1 of the structure:

14. The metal complex of Claim 1 of the structure:

15. A thin film deposition process using the metal complexes of Claim 1 in a process selected from the group consisting of CVD, pulsed CVD, PECVD or ALD, where vapors of the metal complexes are contacted with a suitable reagent selected from the group consisting of hydrogen, formic acid, silane, borane, and mixtures thereof, on a heated substrate surface.

16. The thin film process of Claim 15 for the growth of copper films using super critical carbon dioxide and a reducing agent selected from the group consisting of hydrogen, formic acid, borane, silane and mixtures thereof.

17. The thin film process of Claim 2 to grow a copper film onto a diffusion barrier material whereby the metal complex selectively deposits copper onto the diffusion barrier material rather than onto copper, wherein the copper deposition process becomes self limiting upon achieving a certain copper film thickness and is essentially 'self healing' in that if barrier metal becomes exposed through the copper film, fresh copper will deposit there to restore the continuous copper film.

18. The thin film process of Claim 17 wherein the diffusion barrier material is ruthenium.

19. A thin film deposition process using the metal complexes of Claim 11 in a process selected from the group consisting of CVD, pulsed CVD, PECVD or ALD, where vapors of the metal complexes are contacted with a suitable reagent selected from the group consisting of hydrogen, formic acid, silane, borane, and mixtures thereof, on a heated substrate surface.

20. A thin film deposition process using the metal complexes of Claim 13 in a process selected from the group consisting of CVD, pulsed CVD 1 PECVD or ALD, where vapors of the metal complexes are contacted with a suitable reagent selected from the group consisting of hydrogen, formic acid, silane, borane, and mixtures thereof, on a heated substrate surface.

Description:

TITLE OF THE INVENTION: VOLATILE LIQUID COPPER PRECURSORS FOR THIN FILM APPLICATIONS

CROSS REFERENCE TO RELATED APPLICATION

[0001] This patent application claims the benefit of US Provisional Patent Application Serial Number 60/878,027 filed 28 December 2006.

BACKGROUND OF THE INVENTION [0002] In the semiconductor industry there continues to be an ongoing interest in the development of volatile copper precursor compounds for the growth of thin copper films for various interconnect applications by Chemical Vapor Deposition (CVD) or Atomic Layer Deposition (ALD). A relevant example of the prior art is found in US patent application US2006/0145142and US 7205422

BRIEF SUMMARY OF THE INVENTION

[0003] These new compounds have very high volatility combined with thermal stability, but they are still chemically reaxtive a low temperature towards growing a copper film when reacted with a reducing agent like hydrogen, formic acid, silane, borane or mixtures thereof. So, they are highly suited to ALD or CVD copper. In addition, being so thermally stable they are easy to use in a standard bubbler for deilevery, or they can be used in DLI mode. Thermal stability also means they can purified by distillation.

BRIEF DESCRIPTION OF THE DRAWINGS

[0004] Figure 1 is the TGA/DSC, respectively, for copper precursor Cu-1 [0005] Figure 2 is the TGA/DSC for copper precursor Cu-2. [0006] Figure 3 is a graph comparing the vapor pressures of Cu-1 and Cu-2 [0007] Figure 4 shows the molecular structure of Cu-3 as determined by X-ray Crystallography.

[0008] Figure 5 is the TGA for Cu-3 (here named Kl 4).

[0009] Figure 6 is an Arrenhius Graph of the several experimental deposition runs of Example 9.

DETAILED DESCRIPTION OF THE INVENTION

[0010] The present invention is copper +1 compounds that retain the low tempertaure processing advantages typical for copper +1 compounds, but the compounds of the present invention have the thermal stability normally associated with copper +2 compounds, which always have higher process temperatures. Also, the compounds of the present invention are easy to handle for the process engineer, using simple bubbler delivery. In the semiconductor industry, there continues to be an ongoing interest in the development of volatile copper precursor compounds for the growth of thin copper films for various interconnect applications by Chemical Vapor Deposition (CVD) or Atomic Layer Deposition (ALD).

[0011] For copper precursors to be useful for these processes, it is preferred that they are liquid at ALD or CVD source temperatures, highly volatile, thermally stable, and yet, chemically reactive towards reducing agents, to yield pure copper films. It is well known that incorporation of the element fluorine into the chemical structure of a copper precursor will increase its volatility. This is especially highly advantageous if the precursor is thermally stable, because it can then be safely heated to a high vapor pressure without fear

of the precursor decomposing. A high deposition rate of copper is then achieved, when this high vapor pressure of precursor is flowed through the CVD or ALD reactor and reacted with a reducing agent. High thermal stability also offers other processing advantages, one of which is the ability to simply heat the precursor in a container, and then, bubble a carrier gas through it to achieve a steady and constant stream of precursor vapor into the reactor.

[0012] This is in contrast to the well know liquid fluorinated copper precursor Cu(hfac)(tmvs) (J. A. T. Norman, B. A. Muratore, P. N. Dyer, D. A. Roberts, A.K. Hochberg and L. H. Dubois, Materials Science and Engineering, B17, 87, (1993).), which, due to its thermal instability under vaporization conditions of heat and vacuum, must be mixed with stabilizing excess (tmvs) and the resulting solution vaporized by Direct Liquid Injection (DLI). In some applications, DLI can also be used for thermally stable liquid precursors, as an alternative to bubbling, and may even be preferred due, to the precision of vaporization control that it offers. Another advantage of thermal stability is that the precursor will only metallize the heated substrate surface and not decompose in other parts of the reactor. Once the precursor reacts with the reducing agent, it deposits a copper film and releases volatile by-products, which are pumped away as vapors.

[0013] By contrast, copper precursors of the Cu(hfac)(tmvs) type metallize a substrate by a thermal disproportionation reaction, which releases Cu +2 (hfac) 2 as one of its volatile by- products. This reaction is shown below in Formula 1. The Cu +2 (hfac) 2 compound is a solid at room temperature, which accumulates in cool regions of the reactor, so special provisions must be made to manage this solid material. In addition, the disproportionation dictates that 50% of the original copper precursor is lost as this by-product. In contrast, the thermally stable fluorinated copper precursors of the present invention do not disproportionate, so both the effective 50% loss of precursor and accumulation of copper *2 by-product are avoided. Additionally, the copper *1 complexes of the present invention are

sufficiently chemically reactive that they can grow a copper film by reaction with a reducing agent below 200 0 C, despite their outstanding thermal stability.

[0014] This low processing temperature regime is especially attractive, since it permits an overall low 'thermal budget' for growing copper films onto silicon based electronic devices, such as central processing units, CPUs, which helps avoid their overheating during processing. Another advantage of the volatile low mellting fluorinated thermally stable copper precursors of the present invention is that they can be readily purified by vacuum distillation or sublimation. Other liquid fluorinated copper precursors of the type Cu(hfac)(tmvs) do not readily lend themselves to vacuum distillation since they tend to thermally degrade during the process.

Cu(hfac)(tmvs) Cu(hfac)2 tmvs

Formula 1 : disproportionation of Cu(hfac)(tmvs)

[0015] The use of fluorinated copper precursors for the deposition of copper onto tantalum based diffusion barrier materials has proven problematic, due to the tantalum producing unwanted side reactions with the fluorine of the precursor. This leads to poor adhesion of the copper to the underlying tantalum. However, it is now widely accepted that ruthenium metal will be used as an intermediate 'glue' layer between copper and tantalum base diffusion barrier materials, such as tantalum nitride, including when both the tantalum nitride and ruthenium are deposited by ALD. Since ruthenium is not expected to

react with fluorine in the way that tantalum does, the use of fluorinated copper precursors for depositing copper films onto ruthenium is not anticipated to be problematic. Indeed, this has been confirmed by recent publications from Japan (including H. Kim, Y. Kojima, H. Sato, N. Yoshii, S. Hosaka, Y. Shimogaki, Jpn. J. Appld. Phy, VoI 45, No.8, L233-235, (2006).) describing the growth of high purity fluorine free copper onto ruthenium using the fluorinated copper precursor Cu(hfac)(tmvs) and confirming that unwanted side reactions do not occur and that pure highly adhering copper is yielded.

[0016] The precursors described in the present invention are the fluorinated analogues of the non-fluorinated copper precursors described in US 2006/0145142. The precursors comprise the class of metal complexes of the structure:

where R 1 , R 2 and R 3 are independently H, alkyl or fluoroalkyl groups, but at least one of R 1 ,

R 2 and R 3 is fluoroalkyl and M is a monovalent metal and L is a ligand, capable of coordinating to the metal M, R 4 is an aliphatic hydrocarbon chain. [0017] Preferably, the R 4 aliphatic hydrocarbon chain is normal or branched C M0 .

Alternately, the R 4 aliphatic hydrocarbon chain includes functional groups selected from the group consisting of oxygen, nitrogen, silicon and mixtures thereof.

[0018] Preferably, the metal is copper.

[0019] Preferably, the fluoroalkyl is C M0 . More preferably, the fluoroalkyl is perfluorinated.

[0020] Preferably, the ligand L is selected from the group consisting of alkene, alkyne, nitrile, isonitrile, cyanate, isocyanate, imine and phosphine. More preferably, the ligand L is normal or branched C M O alkenyl. Most preferably, the ligand L is C 2 alkenyl. [0021] In a further embodiment, the metal complex is in admixture with other liquid molecules that enhance the ALD or CVD process, and the resulting mixtures evaporation for delivery to the deposition chamber by DLL

[0022] In yet another embodiment, the metal complex wherein M is copper, is in admixture with a metal complex wherein M is selected from the group consisting of titanium, tantalum, hafnium, silver, gold, tungsten and nickel, capable of the deposition of copper alloys.

[0023] Unexpectedly, these molecules are low melting solids at room temperature, are exceptionally thermally stable, are highly volatile and are highly reactive towards reducing agents, such as hydrogen, formic acid, silane, borane or mixtures thereof at low temperatures. The unexpected combined properties of high thermal stability and high volatility of these new compounds is readily demonstrated by comparing the TGA/DSC scans of these compounds compared to their unfluorinated counterparts. In these experiments a small quantity of a precursor is placed in a microbalance pan, which is then steadily heated at a fixed temperature ramp rate under a steady fixed flow of nitrogen. [0024] As the precursor heats up it begins to vaporize, as manifest by a steady weight loss. This process continues until either complete evaporation occurs or an involatile residue remains. The former case indicates no thermal degradation to have occurred during evaporation, whereas the latter indicates some decomposition to have taken place. A greater amount of residue indicates a greater degree of decomposition. During this heating process, any exothermic events (often decomposition) or endothermic events (often phase changes such as melting or heat loss by evaporative cooling) are detected by the Differential Scanning Calorimeter (DSC). Thus, a precursor, which remains perfectly

thermally stable during evaporation, shows a smooth curve for 100% weight loss down to almost 0% residue (the experimental error of TGS is + or - 0.5%). During this evaporation, the DSC curve is also smooth, as it registers only the endotherm of evaporative cooling of the sample. This is the case for the fluorinated copper precursor Cu-1 of the present invention, whose TGA/DSC and chemical structure are shown in Figure 1 and Formula 2, below, respectively.

Formula 2; The chemical structure of copper precursor Cu-1

[0025] Figure 2 and Formula 3, below, respectively, show the TGA/DSC and chemical structure of the homologous unfluorinated precursor Cu-2. By comparing Formula 2 with Formula 3, it can be seen that the only difference between Cu-1 and Cu-2 is that Cu-1 has fluorinated CF 3 groups on the ketoimine anion portion of the complex in place of the corresponding unfluorinated methyl groups of Cu-2. However, the TGA/DSC performance of the copper precursor Cu-1 is far superior to that of Cu-2 as seen by comparing the Figures 1 and 2. In Figure 1 , the involatile residue for Cu-1 is only 2.3%, with the final evaporation temperature being ~225°C, whereas Figure 2 shows an involatile residue for Cu-2 of ~ 30%, with a final evaporation temperature of ~240°C. Additionally, Figure 2 also

shows an endotherm of melting at -73 0 C, whereas Figure 1 shows an endotherm of melting at 41 0 C, which is an additional advantage, since lower melting points are desirable for CVD and ALD precursors, because then only a relatively low temperature is needed to maintain the molecule in the liquid state.

Formula 3: chemical structure of copper precursor Cu-2

[0026] A further exceptional property of the copper precursor Cu-1 , when compared to the copper precursor Cu-2, is the greater volatility of the former, which is quantified in Figure 3, below, which plots vapor pressure versus temperature for these two precursors. This shows that for a given temperature, Cu-1 is about eight times more volatile than Cu-2. In practical terms, this represents a great advantage, because higher volatility means it is easier to flow a greater vapor pressure of precursor through an ALD or CVD reactor, which can yield faster cycle times, greater copper growth rates and overall lower temperatures for the evaporation of precursor and heated vapor delivery lines to the reactor. This performance is only possible by the unexpected combination of thermal stability and volatility of Cu-1.

[0027] Below in Formula 4 is shown the chemical structure of the precursor Cu-3, which has only one CF 3 group (in Formula 4, "Me" symbolizes a methyl radical). Therefore, the

properties of the copper precursor Cu-3 are in between those of Cu-1 and Cu-2, achieving a higher vapor pressure and thermal stability, when compared to Cu-2 by adding only one CF 3 group. See Figure 5.

Formula 4: Cu-3

[0028] The chemical structures for the copper precursors, Cu-4 (Formula 5), Cu-5 (Formula 6) and Cu-6 (Formula 7), are listed below, where it can be seen that they have an (-NMe-) link from the (-NCH 2 CH 2 -) to silicon group versus the analogous (-O-) link in Cu-1 , Cu-2 and Cu-3. Note that Cu-4 bears two CF 3 groups, whereas Cu-6 has only one and Cu-5 has only unfluorinated methyl groups. Thus, Cu-6 is more volatile and more stable than Cu-5.

Formula 5: Cu-4

Formula 6: Cu-5

Formula 7: Cu-6

[0029] The partially fluorinated metal complexes are used for metal deposition, such as copper deposition comprising a thin film deposition process using the metal complexes in a process selected from the group consisting of CVD, pulsed CVD, PECVD or ALD, where vapors of the metal complexes are contacted with a suitable reagent selected from the group consisting of hydrogen, formic acid, silane, borane, and mixtures thereof, on a heated substrate surface.

[0030] Preferably, the growth of copper films uses super critical carbon dioxide and a reducing agent selected from the group consisting of hydrogen, formic acid, borane, silane and mixtures thereof.

[0031] More preferably, the process is used to grow a copper film onto a diffusion barrier material whereby the metal complex selectively deposits copper onto the diffusion barrier material rather than onto copper by selection of appropriate process conditions, wherein the copper deposition process becomes self limiting upon achieving a certain copper film thickness and is essentially 'self healing' in that if barrier metal becomes exposed through the copper film, fresh copper will deposit there to restore the continuous copper film. Preferably, the diffusion barrier material is ruthenium.

Experimental [0032] The chemical syntheses below first shows the preparation of the fluorinated ligand CF 3 C(O)CH 2 C(NCH 2 CH 2 θSiMe 2 (C 2 H 3 ))CF 3 from which Cu-1 is prepared (in this Specification, where "Me" appears, it represents a methyl radical).

Example 1 Synthesis of CF 3 C(O)CH 2 C(NCH 2 CH 2 θSiMe 2 (C 2 H 3 ))CF 3

[0033] 45g of O,N-bis(dimethylvinylsilyl)ethanolamine were dissolved in 100 ml of dry tetrahydrofuran under an atmosphere of nitrogen and cooled to OC. 41g of 1 ,1 ,1 ,5,5,5- hexafluoro-2,4-pentanedione in 10 ml of tetrahydrofuran were then added with stirring over a 1 hour period. The solvent was then removed by vacuum at room temperature and the remaining crude product (90% yield) was then vacuum distilled at 70 0 C to give the finished product.

1 H NMR: (500 MHz 1 C 6 D 6 ): δ = 0.13 (s, 6H), δ = 2.9 (m, 2H), δ = 3.1 (m, 2H), δ = 5.7 (dd, 1H), δ = 5.75 (S, 1H), δ = 5.9 (dd, 1H), δ = 6.05 (dd, 1H), δ = (10.65 (bs, 1 H). 19 F NMR: (500 MHz, C 6 D 6 ): δ = -67.1 (s, 3F), δ = -76.8 (s, 3F).

Example 2

Synthesis of Cu-1

[0034] 9.6g of CF 3 C(O)CH 2 C(NCH 2 CH 2 θSiMe 2 (C 2 H 3 ))CF 3 from the above synthesis was slowly added to OJg of sodium hydride, stirring in 100ml of dry deoxygenated tetrahydrofuran under an atmosphere of nitrogen. After 20 minutes, there was no further evolution of hydrogen, and the faintly turbid solution was filtered, before slowly adding 3.2g of cuprous chloride in 10 ml of tetrahydrofuran under an atmosphere of nitrogen at 0 0 C. The resulting yellow green mixture was then allowed to warm to room temperature and stirred overnight. The solvent was then removed by vacuum, and 200 ml of dry deoxygenated hexane added with stirring for 30 minutes. The resulting mixture was then filtered, and the hexane removed by vacuum to yield 8.Og of crude product, as a green oil, which was then vacuum distilled to give a yellow final product.

1 H NMR: (500 MHz, C 6 D 6 ): δ = 0.08 (bs, 6H), δ = 3.1(bs, 1H), δ = 3.5(bm, 3H), δ = 3.9(d, 1 H), δ = 4.0(d, 1 H), δ = 6.02(s, 1 H). 19 F NMR: (500 MHz, C 6 D 6 ): δ = -63.05(s, 3F), δ = -73.85(S, 3F).

Example 3 Synthesis of the ligand CF 3 C(O)CH 2 C(NCH 2 CH 2 OSiMe 2 C 2 H 3 )Me from which to make Cu-3

[0035] 6.1g (0.1 moles) of ethanolamine were added to 15.4 g (0.1 moles) of 1 ,1,1- trifluoro-2,4-pentanedione in 100 ml tetrahydrofuran, containing 2Og of anhydrous sodium sulfate, and stirred overnight. The sodium sulfate was then removed, and the solvent vacuum distilled off to give CF 3 C(O)CH 2 C(NCH 2 CH 2 OH)Me as a waxy solid. 9.55g (0.05 moles) of this product was then dissolved in 20 ml of dry tetrahydrofuran and slowly added to 1.2 g (0.05 moles) of sodium hydride in 100 ml of dry tetrafuran under an atmosphere of nitrogen. When the evolution of hydrogen ceased, the mixture was stirred for an additional 1 hour, after which 6.Og (0.05 moles) of chlorodimethylvinylsilane were slowly added, and the resultant mixture was stirred overnight. Filtration of the sodium chloride by-product, followed by removal of the solvent by vacuum, yielded the liquid product CF 3 C(O)CH 2 C(NCH 2 CH 2 OSiMe 2 C 2 H 3 )Me.

Example 4 Synthesis of Cu-3

[0036] 11 -6g (0.01 moles) of CF 3 C(O)CH 2 C(NCH 2 CH 2 OSiMe 2 (C 2 H 3 ))Me were slowly added to 1.05g (0.043 moles.10% excess) of sodium hydride in 250 ml of dry tetrahydrofuran under an atmosphere of nitrogen. When the evolution of hydrogen ceased (1.5 hrs), the reaction was stirred for an additional hour. The resulting solution was then added to 4.3g (0.43 moles, 10% excess) of cuprous chloride in 20 ml of dry tetrahydrofuran cooled to 0 0 C under an atmosphere of argon, and this mixture was then stirred overnight. The solvent was then removed by vacuum, and the resulting mixture stirred with 500ml of dry hexane under an atmosphere of argon. This mixture was then filtered, the filter cake

washed with 2 X 200 ml of further hexane. The hexane was then removed by vacuum to yield the copper precursor Cu-3 as off-white crystals. [0037] Yield = 10.4g (79%).

[0038] The structure of Cu-3 was confirmed by X-ray crystallography, as shown below in Figure 4. Its TGA curve is shown in Figure 5 showing only a low involatile residue at 2.7%. [0039] It is noted that with only one CF 3 group, Cu-3 still has excellent evaporation characteristics with only 2.7wt% residue. Thus, its stability is similar to that of Cu-1 , but it is less volatile, since it does not possess two CF 3 groups, as in Cu-1.

Example 5

Synthesis of the ligand CF 3 (O)CH 2 C(NCH 2 CH 2 NMeSiMe 2 (C 2 H 3 ))CF3 from which to make Cu-4

[0040] 20.8g (0.1 moles) of hexafluoroacetylacetone were slowly added to 2.4g of sodium hydride stirred in 200ml of dry terahydrofuran under an atmosphere of nitrogen. When the hydrogen evolution ceased, the mixture was stirred for an additional hour. 12.Og of chlorodimethylvinylsilane were then added, and the resulting mixture refluxed overnight.

The mixture was then filtered, and the solvent removed by vacuum to leave behind

CF 3 C(O)CH 2 C(OSiMe 2 (C 2 H 3 ))CF 3 .

[0041] To this product, 7.4g of N-methylethylenediamine were added with stirring under nitrogen. This mixture was warmed and stirred for an additional 2 hours. The by-product, tetramethyldivinyldisiloxane, was removed by vacuum distillation to leave behind the

PrOdUCt CF 3 C(O)CH 2 C(NCH 2 CH 2 NHMe)CF 3

[0042] 2.64g (0.01 moles) of CF 3 C(O)CH 2 C(NCH 2 CH 2 NHMe)CF 3 in 20ml of dry tetrahydrofuran were then slowly added to 0.24g (0.01 moles) of sodium hydride stirring in 20 ml of dry tetrahydrofuran under an atmosphere of nitrogen. When the hydrogen evolution ceased, the mixture was stirred for 1 additional hour. To this solution was then

added 1.2g of chlorodimethylvinylsilane, and the resulting mixture was stirred overnight. The mixture was then filtered, and the solvent removed by vacuum to leave behind the ligand CF 3 (O)CH 2 C(NCH 2 CH 2 NMeSiMe 2 (C 2 H 3 ))CF 3 .

Example 6

Synthesis of Cu-4

[0043] 0.348g (0.001 moles) of CF 3 (O)CH 2 C(NCH 2 CH 2 NMeSiMe 2 (C 2 H3))CF 3 in 5 ml dry tetrahydrofuran were then added to 0.024g (0.001 moles) of sodium hydride stirring in 5 ml of dry terahydrofuran under an atmosphere of nitrogen. When the evolution of hydrogen ceased, the mixture was stirred an additional hour, and then, it was added to 0.1 g (0.001 moles) of cuprous chloride stirring in 5 ml of dry tetrahydrofuran at O 0 C under an atmosphere of nitrogen. This mixture was then allowed to stir overnight. The solvent was then removed, 50 ml of dry deoxygenated hexane was added, and the resulting mixture was filtered. The solvent was then removed to yield the copper precursor Cu-4.

Example 7 Synthesis of the ligand CF 3 (O)CH 2 C(NCH 2 CH 2 NMeSiMe 2 (C 2 H 3 ))Me from which to make Cu-6

[0044] 7.4g of N-methylethylenediamine (0.1 moles) were added to 15.4 g (0.1 moles) of 1 ,1 ,1 -trifluoro-2,4-pentanedione in 100 ml tetrahydrofuran containing 2Og of anhydrous sodium sulfate and stirred overnight. The sodium sulfate was then removed, and the solvent vacuum distilled off to give CF 3 C(O)CH 2 C(NCH 2 CH 2 NMeH)Me. 10.5g (0.05 moles) of this product was then dissolved in 20 ml of dry tetrahydrofuran and slowly added to 1.2 g (0.05 moles) of sodium hydride in 100 ml of dry tetrahdrofuran under an atmosphere of nitrogen. When the evolution of hydrogen ceased, the mixture was stirred for an additional 1 hour, after which 6.Og (0.05 moles) of chlorodimethylvinylsilane were slowly added, and

the resultant mixture was stirred overnight. Filtration of the sodium chloride by-product, followed by removal of the solvent by vacuum, yielded the product CF 3 C(O)CH 2 C(NCH 2 CH 2 NMeHSiMe 2 C 2 H 3 )Me.

Example 8

Synthesis of Cu-6

[0045] 2.94g (0.01 moles) of CFSC(O)CH 2 C(NCH 2 CH 2 NMeHSiMe 2 C 2 H 3 )Me in 50 ml of dry tetrahydrofuran were added to 0.24 g of sodium hydride in 50 ml of dry tetrahydrofuran under an atmosphere of nitrogen. The mixture was stirred for one additional hour, after the evolution of hydrogen ceased. This solution was then slowly added to 1.Og (0.001 moles) of cuprous chloride stirred in 10ml of dry tetrahydrofuran at O 0 C. This mixture was then stirred overnight. The solvent was then removed by vacuum distillation, 200 ml of dry deoxygenated hexane added, the mixture stirred for 30 minutes at room temperature, and then filtered. The solvent was then removed from the filtrate to yield the precursor Cu-6.

[0046] The synthesis for the ligand used to prepare Cu-5 and the synthesis of Cu-5 are described in detail in the published application US2006/0145142.

Example 8 Examples of copper ALD deposition

[0047] Cu- 1 was used at a source temperature of 75 0 C using nitrogen as carrier gas. Its vapors were pulsed through an ALD reactor, alternating with hydrogen reagent gas, over ruthenium coated silicon wafer samples at between 20 and 500 mTorr pressure. Three experiments were run at 200 0 C, 225°C and 250 0 C for 1800, 600 and 700 cycles

respectively resulting in the deposition of copper films, as confirmed by Auger spectroscopy.

Example 9 Chemical Vapor Deposition (CVD) of copper from Cu-1

[0048] A 25g sampleof Cu-1 in a stainless steel bubbler was fitted to an experimental hotwall CVD reactor, and CVD copper films were grown onto ruthenium substrates using formic acid vapor as a reducing gas added by vapor draw from a quartz bubbler. The CVD chamber pressure was maintained at 2.0 Torr, and 25 seem of helium carrier gas was flowed through the bubbler. Each run was of 30 minutes duration at wafer temperatures of 150, 200 and 250 0 C, and for each of these wafer temperatures, the Cu-1 bubbler held at 57, 64 or 75°C. The formic acid vapor was metered through a needle valve and determined to be an average of 160 seem for each run, as determined by weight loss of formic acid from its container for each run. Table 1 , below, shows all the runs with their respective average CVD copper film thickness grown, shown in units of Angstroms. Copper film thickness was determined by etching a hole in the copper film down to the underlying ruthenium (not etched) using 50% nitric acid, washing with water, drying, then measuring the resulting copper step by stylus profilometry. These results are also shown in the Figure 6 Arrhenius graph, below, which indicates a linear relationship between the logarithum of copper growth rate in Angstoms per minute versus the reciprocal of absolute temperature, along with a higher overall growth rate of copper with increasing source temperature, as typically seen for CVD metallization processes. The copper metal deposited was confirmed by electron dispersive X-ray (EDX) analysis.

Table 1