AMIDINATE LIGAND CONTAINING CHEMICAL VAPOR DEPOSITION PRECURSORS
[0001 ] This application claims priority under 35 U. S. C. § 119(e) of prior U. S. Provisional Patent Application No. 60/582,944 filed June 25, 2004, which is incorporated in its entirety by reference herein.
Field of the Invention
[0002] The present invention is directed to novel precursors for use in chemical vapor deposition. More particularly, the present invention is directed to amidinate ligand containing precursors such as metalloamidinates as chemical vapor deposition precursors.
Background
[0003] Chemical vapor deposition (CVD) precursors that are both volatile and easily handled are needed for numerous applications. The limited availability of suitable precursors for applications based on metal-, metal-oxide-, and metal-nitride films ( such as flat panel displays, organic light-emitting diodes, low emissivity and solar control coatings, photovoltaic cells, and a variety of transparent, opto-, and high temperature electronic applications) is an obstacle to development of new CVD applications.
[0004] One major problem is the overall lack of suitable precursors for CVD such as atmospheric pressure chemical vapor deposition (APCVD); very few species have the combination of volatility (traditionally obtained through construction of low molecular-weight, fluorine-incorporating molecules) and stability (clean decomposition to M, MxOy, MxNy at temperatures above those required for volatility). The present invention is directed to suitable precursors for chemical vapor deposition such as atmospheric pressure chemical vapor deposition (APCVD) of films of metals, metal oxides, and metal nitrides, in which the physical properties of the precursors (or films) may be controlled by modification of the precursor ligand array.
Summary of the invention
[0005] Metalloamidinates [R1NC(R^NR3JxMLn (R1"3 = alkyl or aryl; x > 1 ; M = main-group- or transition-metal; Ln = ancillary ligand array) seemed particularly well suited for precursors. The amidinate [R1NC(R2)NR3]" ligands are low molecular- weight species, capable of multiple binding modes (monodentate (Ia), chelating (Ib3Ic), bridging (Id, Ie)) that may support a variety of metal fragments and metal oxidation states.
N = fR'NC(R2)NR3]- ( Ie)
/
[0006] Furthermore, many substitution patterns (R1, R2, R3) on the amidinate ligand are accessible from starting materials. This permits systematic modification of steric and electronic effects, as well as incorporation of fluorinated substituents, which has been demonstrated to impart volatility to precursors such as (hexafluoroacetylacetonate)Cu(vinyltrimethylsilane) for Cu- deposition.
[0007] The combined steric and electronic effects of the amidinate substituents (R1"3), the specific metal center employed (M), and the ancillary ligands on the metal (Ln) affect the precursor geometry and stability. Changes in geometry and/or stability affect important precursor properties such as volatility and deposition characteristics. Although it is expected, on the basis of molecular weight differences, that monometallic (Ia, Ib, Ic) metalloamidinates are more volatile than the corresponding bimetallic (Id) or oligomeric (Ie) metalloamidinate, the exact combination of R1"3, M, and Ln required to impart practical volatility for CVD use can vary. Furthermore, precursor stability and deposition characteristics are dependent on the geometry as well as on the steric and electronic influences imparted by specific R1"3, M, and Ln. Experimental and computational evaluation of the relationships between these factors
is necessary for development and optimization of suitable precursors for chemical vapor deposition of metal-, metal-oxide-, and metal-nitride films.
Detailed Description of the Invention
[0008] Lithium- and magnesium amidinates [R1NC(R^NR3]' (2a-j)
are easily derived from treatment of carbodiimides (R
1N=C=NR
3) with lithium- or magnesium-reagents. These THF-soluble ligands are obtained in 60-97% yield.
1H NMR spectroscopy indicates that non-stoichiometric and variable quantities of THF often remain in the lattices of 2a-j, even after prolonged exposure to dynamic vacuum. Both symmetrically and unsymmetrically substituted carbodiimides may be converted to amidinates using this procedure. Amidinates with symmetric N- substitution (2a-g) tend to be more crystalline than do the corresponding unsymmetrically N-substituted amidinates (2h-j), although crystallinity depends on the specific N- and C- substituents employed.
[0009] In a typical reaction, under air- and moisture-free conditions, 39.6 mmol of diisopropylcarbodiimide was dissolved in 50 ml of anhydrous THF, and the resulting solution cooled to O0C. 4-Fluorophenylmagnesium bromide (36 ml, 1.1 M in THF) was added drop wise to the solution, keeping the temperature below 1O0C during the
addition. After the addition, the flask was allowed to warm to room temperature with stirring. After approximately 3.5 hours of stirring at room temperature, volatiles were removed from the flask and the products dried in vacuo, affording 12.84 grams of a white solid that analyzed (1H NMR) as [(iPr)NC(4-FPh)N(iPr)] [MgBr] [THF]13 (30.6 mmol, 77% yield) (iPr = isopropyl).
[0010] All of the magnesium-amidinates synthesized exhibited multiple (two) isomers in solution, as evidenced by 1-D 1H (and 19F, when possible) NMR spectroscopy, as well as by multidimensional correlation spectroscopy (COSY).
[001 1] Treatment of CuCl2 with two equivalents of
[(iPr)NC(CH3)N(iPr)][Li][THF]o.92 (2a) in THF results in complete consumption of the amidinato-lithium reagent (1H NMR). The resultant metalloamidinate products, however, are difficult to characterize as a result of their insolubility in both nonpolar (toluene) and polar solvents (THF). Similarly, treatment of 2 equivalents of 2a with NiCl2 in THF results in insoluble materials that could not be characterized by 1H NMR spectroscopy, other than observation of the complete consumption of starting 2a.
A eq. + MCl *■ insoluble materials
[001 1] The extreme insolubility of the MX2/2a products is consistent with the formation of oligomeric species, such as Ie.
[0012] By contrast, treatment of NiCl2 with two equivalents of [(iPr)NC(4-F- Ph)N(iPr)]" (2b), in which the central C-methyl group in 2a has been replaced with a sterically more demanding jt?αrα-fluorophenyl group, results in clean formation of a THF- and toluene-soluble green solid. The chemical shift range of the observed H NMR resonances (0-8 ppm) are diagnostic for the formation of diamagnetic species, and the symmetry and integration of the observed resonances are consistent with a rapid interconversion between trigonal bipyramidal and square pyramidal structures. The presence of coordinated THF is clearly evidenced by the chemical shifts of both the α- and β- THF CH2 resonances at 4.02 and 1.34 ppm, respectively (c.f. free THF at 3.58 and 1.73 ppm), with peak integrations revealing the presence of one equivalent
of THF per nickel center, affording a molecular structure of [(iPr)NC(4-F- Ph)N(iPr)]2Ni(THF) (3).
(two isomers)
[0013] In a typical reaction, under air- and moisture-free conditions, 15 ml anhydrous THF was added to a mixture of 46.5 mg of anhydrous NiCl2 (0.359 mmol) and 284 mg (0.715 mmol, 1.99 equivalents) of [(iPr)NC(4-
FPh)N(IPr)][MgBr][THF]1 o- The reaction mixture was agitated at room temperature using an orbital shaker for approximately 2 hours, at which time volatiles were removed in vacuo, and the resulting solids dried under vacuum. The toluene-soluble material was extracted into 10 ml of anhydrous toluene, filtered through a Teflon frit, and dried in vacuo, affording 164 mg of olive-colored powder [0.280 mmol, 79% yield].
[0014] Addition of two equivalents of [(iPr)NC(4-CH3-Ph)C(iPr)]" (2d), in which the/>Ω7"Ω-fluorine-substituent in 2b has been replaced by a methyl group, to NiCl2 in THF also results in complete consumption of starting 2d, and in formation of the monometallic [(iPr)NC(4-CH3-Ph)C(iPr)]2Ni(THF)2 (4).
(two isomers ) (4)
[0015] Incorporation of apαrα-methyl substituent in the C-phenyl group results in the ligation of a second equivalent of THF, resulting in a change in geometry to a 6- coordinate, octahedral Ni-center.
[0016] Addition of two equivalents of the highly congested ø/'t/zo-disubstituted ligand {(iPr)NC[2,6-(CH3)2-Ph]C(iPr)} [MgBr] (2e) to NiCl2 results in a mixture of products, in contrast to the previous />αrø-substituted ligands that reacted cleanly to generate single metalloamidinate species.
( isomers )
+ othe r Products
[0017] A 1H COSY spectrum reveals the existence of at least 4 products from 2e/NiCl2; the 1H NMR spectrum of the major product (-50% of the iPr resonances) is consistent with the bis(amidinate) {(iPr)NC[2,6-(CH3)2-Ph]C(iPr)}2Ni(THF)x (5). The next-most prominent product appears to have undergone a ligand rearrangement, as evidenced by a broad resonance at 4.84 ppm with a cross peak (indicating physical connectivity or proximity) to the multiplet at 4.21 (J = 6.5 Hz), rather than the presence of a septet an intact isopropyl methine.
[0018] The ability to make metal complexes containing two different amidinate ligands greatly expands (beyond the (amidinate)2M complexes) the number of feasible precursor molecules. Addition of one equivalent of amidinate 2b to a cold (-78 - 0° C) suspension OfNiCl2, followed (one hour later) by the addition of one equivalent of the 3,4-difluorophenyl substituted 2c, results in formation of the homoleptic [(iPr)NC(4- F-Ph)N(iPr)]2Ni(THF) (3) and [(iPr)NC(3,4-F2-Ph)N(iPr)]2Ni(THF)x (6), and a small amount of the mixed-amidinate species [(iPr)NC(4-F-Ph)N(iPr)][(iPr)NC(3,4-F2- Ph)N(iPr)]Ni(THF)x (7). Also novel to this invention is the synthesis of mixed (amidinate) 1(amidinate)2M complexes such as 7. The ability to combine different amidinate ligands onto the same metal center may allow additional fine tuning of molecular and macroscopic properties over those dictated by (amidinate)2M species. This also provides proof-of-concept for the generation of mixed ligand complexes such as (for example) (amidinate)M(alkyl), (amidinate)M(alkoxide), and (amidinate)M(β -diketonate) .
2b + 2c + NiCl2 or NiCl2 (DME)
*i P r
[0019] In a typical reaction, 0.46 mmol of 2b dissolved in 15 ml of anhydrous tetrahydrofuran, and 0.46 mmol of 2c was dissolved in 15 ml of THF in a second flask. NiCl2 (0.46 mmol) was suspended in 20 ml of THF in a third flask, and this suspension cooled to -78° C. The solution of 2b was added dropwise (by cannula) to the cooled Ni suspension, and then the -78 0C bath was replaced with a 0° C bath. After one hour of stirring at 0° C, the 2c solution was added to the Ni solution, the 0° C bath removed, and the solution allowed to warm to room temperature and stir for an additional 3 hours. THF was then removed in vacuo, the toluene-solubles extracted into 20 ml of toluene, and magnesium-halides removed by filtration. Toluene was then removed in vacuo, affording a dark green solid. 19F NMR analysis of the toluene- soluble material (THF-<i8) indicates a 46:46:8 mixture of 3:6:7.
[0020] Use of a more soluble Ni-halide source and higher reaction temperatures is a means of affecting the homoleptic:mixed amidinate ratio. Addition of 2b to a homogeneous room-temperature THF solution OfNiCl2(DME) [DME = dimethoxyethane, H3COCH2CH2OCH3], followed (one hour later) by the addition of one equivalent of 2c, affords a 21 :21 :58 mixture of the bis(amidinates) 3 and 6 and the desired mixed-amidinate 7.
[0021 ] In order to help evaluate the factors contributing to amidinate binding and metalloamidinate geometry, the structures of several model metalloamidinates have been predicted using computational techniques. Density Functional Theory (B3LYP) calculations were conducted using the Titan modeling package (Schrόdinger Software). Ground state structures for a series of [R1NC(R2)NR3]2Cu complexes [HNC(H)NH]2Cu (8a), [(nPr)NC(CH3)N(nPr)]2Cu (8b), [(iPr)NC(CH3)N(iPr)]2Cu
(8c), and [(iPr)NC(Ph)N(iPr)]2Cu (8d) have been evaluated, and selected bond lengths, interatomic distances, bond angles, and dihedral angles determined.
[0022] All of the 4-coordinate (amidinate^Cu molecules are expected to adopt approximately square-planar geometries (sum of angles around Cu = 360.00-360.03°), with the Cu atoms clearly exposed, irrespective of the amidinate substituent array employed. Increased steric interactions between the N- and C- substituents serves to expand the R -C-N angle, and significantly widen the R -N-C angles. Also affected by the amidinate substituion is the N-C-N bond angle, which narrows as the substituents become more bulky, effectively serving to push the amidinate C further away from the Cu center, and decreasing Cu-N distances.
[0023] This present invention demonstrates that novel amidinate ligands containing both symmetric- and unsymmetric N- substitution are accessible from readily available starting materials, and that a wide variety of substituents may be successfully be incorporated into the amidinate framework.
[0024] Addition of two equivalents of the least sterically hindered ligand, 2a, to CuCl2 or NiCl2 in THF affords highly insoluble material, consistent with the formation of oligomeric- or polymeric materials. The N-isopropyl and C-methyl group combination may not be sufficiently bulky to promote a chelating geometry, thus leading to bridging amidinate ligands and oligomeric products.
[0025] Increasing the steric bulk of the ligand array, by replacing the C-methyl group with a substituted phenyl group, results in soluble metalloamidinates that may be characterized by NMR spectroscopy, and that appear to adopt monometallic structures. Smaller changes in the ligand array allow "tuning" of the molecular properties of the metalloamidinates; replacing mepαrø-fiuorine substituent with a methyl group results in additional equivalents of THF bound the to metal center, and thus a shift from 5- to 6-coordinate nickel.
[0026] Further increasing the steric bulk of the ligand, to the 2,6-dimethylphenyl substituted 2e, results in dramatic changes in ligand reactivity, affording both
(amidinate)2Ni compounds, species which have undergone ligand activation/rearrangement, and additional uncharacterized products.
[0027] Also novel to this invention is the synthesis of mixed
(amidinate) (amidinate) M complexes such as 7. The ability to combine different amidinate ligands onto the same metal center allows additional "fine tuning" of molecular and macroscopic properties over those dictated by (amidinate)2M species.
This also provides proof-of-concept for the generation of mixed ligand complexes such as (for example) (amidinate)M(alkyl), (amidinate)M(alkoxide), and
(amidinate)M(β-diketonate).
[0028] The ability to model representative metalloamidinates allows detailed analysis of the steric environments of potential precursors, and may be used to help optimize precursor properties. The relationship between amidinate steric bulk and calculated [R NC(R )NR ]2Cu distances and angles is consistent with the net effect of increased movemement of the N- subsitutents (R1 and R3) into the open portion of the molecule located between the amidinate ligands, and even out of the plane of the amidinate ligand, with increasing substituent bulk. These results suggest that incorporation increasingly bulky amidinate substiuents many be a way to enforce the chelating, as opposed to either monodentate or undesired bridging, amidinate geometry.
[0029] These results also suggest that, for a monometallic metalloamidinate, increasing steric bulk may be a way to increase unfavorable amidinate-amidinate interactions, leading to distortion from an electronically favored square-planar geometry (Ib) towards an electronically destabilized tetrahedral geometry (Ic). Electronic destabilization (or stabilization) of a metal center through amidinate substituent modification may be a means of controlling precursor thermal stability or deposition characteristics.
[0030] The preferred principle and method of operation may be extended in several ways without losing the advantages of the invention. The invention may be extended to metalloamidinates with amidinate substiruents not listed explicitly above, but that would be obvious to those skilled in the art. The amidinate ligands may be
coordinated to metals other than Cu and Ni; this invention covers metalloamidinates of main-group- and transition metals for use as CVD precursors, including but not limited to nickel, copper, zinc, titanium, silver, molybdenum, tungsten, tantalum, tin, aluminum, gallium, and indium. This invention also includes metalloamidinate precursors in which one- or more than two amidinate fragments are bound to a metal center. When more than one amidinate fragment is bound to a metal center, the amidinates may have the same or different substituents. The metal fragment to which one or more amidinate ligand is bound may have additional coordinated (ancillary) ligands; these ligands may include alkyls, aryls, hydrides, alkoxides, acetylacetonates and other β-diketonates, amines, phosphines, alkenes, alkynes, allyls, cyclopentadienyls, carbonyls, nitriles, halides, oxides, imides, nitrides, and tetrahydrofurans. Amidinate-ancillary ligand interactions may also be used to influence precursor properties
[0031 ] This invention also covers metalloamidinates made by methods other than addition of lithium- or magnesium-amidinates to metal halides. Alternate possible methods of synthesizing metalloamidinates include addition of carbodiimides (R1N=C=NR3) or amidines (R1NHC(R^NR3) to metal-alkyls.
[0032] This invention covers the use of metalloamidinates as precursors for the deposition of metal-, metal-oxide-, and metal-nitride films of main-group- or transition-metals. Techniques used to deposit these films include, but are not limited to pyrolytic atmospheric pressure CVD, low-pressure CVD, plasma assisted CVD, solution spray CVD, and powder spray CVD.
[0033] While the present invention has been described with respect to particular embodiments thereof, it is apparent that numerous other forms and modifications of the invention will be obvious to those skilled in the art. The appended claims and this invention generally should be construed to cover all such obvious forms and modifications that are within the true spirit and scope of the present invention.