New Heteroleptic Copper(II) Complexes as MOCVD Precursors

Available online at www.sciencedirect.com

Physics Procedia 46 (2013) 174 – 182

Nineteenth European Conference on Chemical Vapor Deposition, (EUROCVD 19)

New heteroleptic copper(II) complexes as MOCVD precursors V.V. Krisyuk,* S.V. Sysoev, Y.M. Rumyantsev, S.A. Prokhorova, E.V. Maximovskiy, M.L. Kosinova, I.K. Igumenov Nikolaev Institute of Inorganic Chemistry SB RAS, Pr. Lavrentiev 3, Novosibirsk 630090, Russia

Abstract New volatile heteroleptic copper(II) complexes having beta-ketoiminate (O,N) and diketonate (O,O) ligands in one molecule were tested as precursors for LPCVD of copper films. Saturated vapor pressure was measured and compared for new compounds Cu(ki)(hfa) and Cu(dpk)(hfa), where ki = pentane-2-imino-4-onato, hfa = 1,1,1,5,5,5-hexafluoro-pentane2,4-dionato, dpk= 2,2,6,6-tetramethyl-3-iminoheptane-5-onato. The precursors are air stable and non hygroscopic compounds with long shelf life. It was demonstrated that copper metal films can be selectively deposited on metallic surfaces in the presence of hydrogen as a gas-reactant at temperatures of 250, 300, 350 oC and pressure of 20 Torr. Si(100), SiO2 (melted quartz), stainless steel, and Cu, Al, RuO2, Ru and Ta sublayers on Si(100) were tested as substrate materials. Deposited films were analyzed and characterized by X-ray photoelectron spectroscopy (XPS), X-ray diffraction (XRD) and scanning electron microscopy (SEM).

2012 Published The Authors. Published by Elsevier Ltd. © 2013 The© Authors. by Elsevier B.V. Selection and peer-review under responsibility of Organizing Committee of EUROCVD 19. Selection and/or peer-review under responsibility of the Organizing

Committee of EUROCVD 19.

Keywords: volatile copper(II) complexes; heteroleptic complexes; copper films; MOCVD; saturated vapor pressure.

* Corresponding author. E-mail address: [email protected]

1875-3892 © 2013 The Authors. Published by Elsevier B.V. Selection and peer-review under responsibility of Organizing Committee of EUROCVD 19. doi:10.1016/j.phpro.2013.07.065

175

V.V. Krisyuk et al. / Physics Procedia 46 (2013) 174 – 182

1. Introduction Deposition of copper films has been widely studied especially for microelectronics applications due to its low resistivity. Copper films can be deposited by physical methods but MOCVD is intensively studied method due to the possibility of achieving conformal films and of selective deposition. The list of copper MOCVD precursors is well known among which copper(II) β-diketonate derivatives have been the most widely studied [1-3]. The vast majority of the studied volatile copper(II) complexes with β-diketones as precursors refers to homoleptic complexes [1-3]. Homoleptic complexes are easier to prepare and their properties like volatility and thermal stability are more predictable with regard to the ligand structure. The most common representative of this family is Cu(hfa)2, where hfa = 1,1,1,5,5,5-hexafluoro-pentane-2,4-dionato (hexafluoroacetylacetonate), which leads to pure copper films in the 250-380 oC temperature range [4, 5]. Heteroleptic complexes (i.e. chelates containing different β-diketonate ligands) may promise more flexibility in tuning their properties when combining in one molecule, for example, aliphatic ligand, which can give higher thermal stability, and fluorinated ligand, which can provide high volatility. However, reports on the use of them as copper MOCVD precursors are scarce. On the other hand, for some fluorinated heteroleptic βdiketonates of copper(II) a reverse conversion into homoleptic complexes was detected on heating [6-8]. Perhaps, this fact hampered further attempts to obtain and study volatile heteroleptic copper(II) complexes with β-diketonate ligands and their derivatives. Recently, the structure and thermal properties of Cu(ki)(hfa), where ki = pentane-2-imino-4-onato (ketoiminate), were reported [8]. This new compound was shown to be volatile and stable towards thermal conversion into homoleptic complexes in the gas and condensed phase. Thus, these properties allow considering such compounds as promising candidates for MOCVD precursors for the preparation of copper-containing films. In this paper we report a preliminary study of heteroleptic copper(II) complexes, having beta-ketoiminate (O,N) and diketonate (O,O) ligands in one molecule, as precursors for MOCVD copper films. Here we compare volatility of new compounds Cu(ki)(hfa) (1) and Cu(dpk)(hfa) (2), where dpk= 2,2,6,6-tetramethyl-3iminoheptane-5-onato (Figure 1) and discuss some results on deposition of copper films from the precursor 2. In this report we also describe our findings and compare the deposition behavior of the precursor on available metallic and non-metallic substrates. First of all, we wanted to represent our idea and experimental data to readers who know well the subject and can take our results to the consideration and draw conclusions.

CF3

(H3C)3C

O O Cu N O

CF3 O O Cu N O

CF3 1 Fig. 1. Structural formulas for Cu(ki)(hfa) (1) and Cu(dpk)(hfa) (2).

(H3C)3C

CF3 2

176

V.V. Krisyuk et al. / Physics Procedia 46 (2013) 174 – 182

2. Material and methods 2.1 Preparation Compound 2 was synthesized from Cu(hfa)2 and Cu(dpk)2 [9] by the same technique described in [8] for 1 and purified by vacuum sublimation. Cu(dpk)(hfa) has been identified by mass spectrometry and X-ray structural study and these data will be published separately. 2.2. Vapor pressure measurements The saturated vapor pressure for compound 1 has been reported in [8]. The saturated vapor pressure for the compound 2 over solid and liquid phases was measured by the flow method in an environment of a dry inert gas (helium). We have systematically used this technique to measure the saturated vapor pressure of many volatile organometallic complexes [10, 11]. Compounds were purified by double vacuum sublimation before measurements. The amount of the substance vaporized on the passage of certain helium volume with the following condensation in a cold zone was determined by weighing. The total error of this method was less than 5% with the accuracy concerning the temperature ± 0.5 oC and the error in the measurement of the flow rate ±2%. The measurements were carried out in quasi-equilibrium conditions. The independence of the vapor pressure on a helium flow rate was the experimental evidence of this statement and was checked for different helium flow rates. The saturated vapor pressure was calculated from the equation: P = Ptotal *n/(n+N), where n is the number of moles of the carried compound, N is the number of the moles of the carrier gas, Ptotal is the total pressure in the system. The calculation is based on the assumption that the substance vaporizes in the monomolecular form. It should be noted that the results obtained both on the loss in mass of the substance and the rise in mass of the condensed substance in the cold zone practically coincided meaning good thermal stability of the compound under long heating. 2.3 Thin films deposition Copper metal films were deposited in a vertical cold walls quartz reactor equipped with gas flow and vacuum control units. Hydrogen was used as a gas-carrier with flow rate of 50 sccm. Deposition temperatures varied within the range from 200 to 350 oC and total pressures were 0.1, 2 and 20 Torr. Si(100), SiO2 (melted quartz), stainless steel and Cu, Al, RuO2, Ru and Ta sublayers on Si(100) were tested as substrate materials. Films composition was studied by XRD with a Shimadzu XRD-7000 diffractometer (Bregg-Brentano geometry, Cu Kα radiation, Ni – filter, 5 – 60° 2ș range, 5 s per point). XRD with standard scanning mode was used to identify predominate phases in the film. Indexing of the diffraction patterns was carried out using data for compounds reported in the PDF database (Powder Diffraction File. Alphabetical Index. Inorganic Phases, JCPDS, 1983, International Centre for Diffraction Data, Pennsylvania, USA). Films morphology was examined by SEM with a JEOL-JSM 6700 F device. Investigation of the film chemical composition by X-ray photoelectron spectroscopy (XPS) was conducted with SPECS spectrometer. The spectrometer was equipped with a source of characteristic X-ray radiation XR50M with double Al/Ag anode, X-ray monochromator FOCUS-500, hemispherical analyzer PHOIBOS-150MCD-9. Monochromatic radiation of Al Kα (hν = 1486.74 eV) were used to record the spectra. Sample was mounted on a sample holder using double sided conductive adhesive tape. The relative concentrations of the elements in the area of the analysis were determined on the basis of the integrated intensities of XPS lines

177

V.V. Krisyuk et al. / Physics Procedia 46 (2013) 174 – 182

taking into account the photoionization cross sections of the corresponding terms [12]. The calibration of the scale of binding energies was done by the line Cu2p3/2 (E = 932.67 eV). The spectra were recorded both before and after etching by a focused Ar+ ion beam with energy of 2.4 keV at a current density of ~ 10 mkȺ/ɫɦ2. These conditions refer to the etching rate of the thin film SiO2/Si of about ~ 0.5 nm/min. 3. Results and discussion 3.1 General remarks and comparison of volatility In our previous work we introduced new type of copper(II) heteroleptic complexes and reported thermal properties of the compound 1 [8]. It was shown that the formation of stable heteroleptic copper(II) complexes is possible when the compound molecule contains both ketoiminate and diketonate ligands. The reported compound has a molecular structure and can exist in two polymorphic modifications. The Cu(ki)(hfa) compound is volatile and thermally stable; it is easily sublimed in vacuum without decomposition into homoleptic complexes. The saturated vapor pressure for both polymorphs takes intermediate values between those for Cu(ki)2 and Cu(hfa)2. Encouraged by those results we synthesized another new compound of this type: Cu(dpk)(hfa). The main advantages of the compound 2 are its low melting point and higher volatility. All prepared complexes are air-stable and non hygroscopic substances. These compounds can be recrystallized from toluene, sublimated in vacuum or in an inert gas flow without decomposition into homoleptic complexes. The temperature dependence of saturated vapor pressure for 2 and Cu(hfa)2 are shown in Figure 2 for comparison. Saturated vapor pressure of Cu(dpk)(hfa) was measured over solid and liquid phases. It is seen that under sublimation saturated vapor pressure for 2 is close to that one of Cu(hfa)2.

0.0 -0.5 log(P(Torr))

-1.0 -1.5

o

MP=71 ɋ

Cu(hfa)2

-2.0 -2.5 -3.0 -3.5 2.6

2.7

2.8

2.9

3.0

3.1

3.2

3.3

3.4

3.5

1000/T, K Fig. 2. Temperature dependences of saturated vapor pressure for 2 (liquid and solid) in comparison with Cu(hfa)2 [1] (solid).

178

V.V. Krisyuk et al. / Physics Procedia 46 (2013) 174 – 182

3.2 Copper film deposition The growth rate of Cu films at low pressures (0.1 and 2 Torr) appeared to be very slow and consequently long deposition times were needed to obtain satisfactory films thickness. So, more details are discussed here for Cu films prepared at 20 Torr with deposition time not exceeding 1 h. The slowest growth rate of Cu films was observed on Si(100) and quartz substrates. In contrast, selective deposition of metal Cu was visible on Cu, Al, RuO2, Ru, Ta sublayers but not on the open Si areas. To prepare Cu films on Si and quartz we used substrates with predeposited by PVD Cu layers (~3-6 nm). In this case Cu films were prepared selectively on PVD spots and with good growth rate (100-200 nm/h). Figure 3 displays the SEM micrographs of copper films deposited on Cu/Si(100) substrates at 250, 300 and 350 oC. It is seen that the film deposited at 250 oC consists of randomly distributed grains. At 300 oC these grains grow denser. At 350 oC the coalescence of grains increased largely and the film became less porous. As films thickness increased with temperature they were deposited in the surface kinetic regime at least below 350 oC. Very similar morphology of the copper films was observed on other metallic surfaces (Figure 4). For copper films deposited at 350 oC on Al/Si(100) sublayers and stainless steel substrates we noticed yellow interface layer in the places which were less exposed to the precursor vapor (where a sample holder touched the substrate). Moreover, all surface of the stainless steel substrate became yellow and no free Cu film was visible above deposition temperatures of 250 and 300 oC. So, for stainless steel and Al/Si(100) substrates we suspect that definite interface alloy formed in these cases due to intensive diffusion of Cu into the substrate material at that deposition temperatures. Study of such interfaces will be a subject of the separate work. Figure 5 shows the XRD pattern of copper film prepared on Al/Si(100) substrate. The scanning mode used did not allow detecting reflections for copper alloys, although peaks for main phases were wide and shifted. Cu films prepared on Ru/Si(100) substrates had the same appearance and morphology as that ones on other metallic sublayers, but films prepared on RuO2/Si(100) substrates have quite special morphology. The microstructure of the coating on initially RuO2 sublayer is illustrated by SEM micrograph of Figure 6. The copper film deposited at 350 oC is not compact and very porous in this case. It is well seen that islands of fused copper grains are evenly distributed over the surface of fine-grained material which is a metallic ruthenium resulted from its oxide reduction in hydrogen atmosphere. Copper grains also formed in the pores on free Si surfaces. The formation of pores in the film can be explained by the release of water vapor from the film under the deposition conditions. 3.3 XPS study of films composition XPS data of copper film prepared at 350 oC on all substrate have shown that given films include Cu, C and O. Carbon and oxygen are mainly localized on the surface, and their concentration decreases sharply after 2 minutes of the ion etching. For example for the film prepared on Al/Si(100) substrate, Al was detected at the interface film/substrate after 20 minute ion etching. Because of low ionization cross section the intensity of Si2p and Si2s lines did not allow analyzing the state of silicon. The relative atomic concentrations of Cu, C, O and Al are shown in Table I. After 20 minutes of the ion etching at the interface area the shape of the spectrum Cu2p and the value of the Auger parameter Į of 1851.4 eV correspond to the metallic state of copper. The position of the Al2p line is 73.0 eV, indicating aluminum in the metallic state. For comparison, aluminum in Al2O3 is characterized by the value of the binding energy in the region of 74.2-74.6 eV [22-24]. Regardless of the fact that Al and Cu were in the metallic state after 20 minutes etching, O1s line intensity indicated high oxygen concentration in the analyzed zone: the atomic ratio [O]/[Cu] = 0.9, [O]/[Al]= 1.8. This fact can be explained by the presence of SiO2 covering silicon substrate under Al sublayer. This is confirmed by the high value of the binding energy of the O1s, equal to 532.4 eV.

179

V.V. Krisyuk et al. / Physics Procedia 46 (2013) 174 – 182

a)

a)

b)

b)

c)

c)

Fig. 3. Plain-view SEM images of copper film deposited on Cu/Si substrates a) at 250 oC; b) at 300 oC; c) at 350 oC.

Fig. 4. Plain-view SEM images of copper film deposited at 350 oC on a) stainless steel substrate; b) Ta/Si substrate; c) Al/Si substrate.

180

V.V. Krisyuk et al. / Physics Procedia 46 (2013) 174 – 182

For comparison, the binding energy for O1s line in Al2O3 is 531.0-531.2 eV and in SiO2 it is 533 eV [2224]. For all studied films no signal of fluorine was detected.

Table I. Results of XPS study of the copper film prepared at 350 oC on Al/Si substrate.

Etching time, min

[Si]/[Cu]

[C]/[Cu]

[O]/[Cu]

Concentration, At. % Cu

C

O

0 min

0

2.3

2.0

19

44

37

2 min

0

0.068

0.056

89

6

5

5 min

0

0.050

0.044

91.4

4.0

4.6

8 min

0

0.039

0.059

91.2

3.5

5.3

20 min

0

0.043

0.88

[Al]/[Cu] = 0.48

Al (111)

Cu (111)

Cu (200)

30

35

40

45

50

55

60

2θ, degr. Fig. 5. Diffraction pattern of Cu film deposited on Al sublayer at 350 oC. Peaks of Al and Cu are labeled. All other peaks refer to the Si substrate.

V.V. Krisyuk et al. / Physics Procedia 46 (2013) 174 – 182

181

Fig. 6. Plain-view SEM images of copper film deposited on RuO2/Si substrate substrates at 350oC.

4. Conclusion At this stage, we can conclude that the deposition of copper films at a sufficient rate occurs from Cu(dpk)(hfa) precursor on the metal surfaces in the presence of hydrogen. Thus, we may assume that the main chemical reaction, producing copper metal, in this case is the reduction reaction of the precursor molecules by hydrogen which is well adsorbed on the metal surfaces. This reaction plays a key role at the initial stage of the film growth while further growth of the film occurs already on a copper surface. Our results demonstrated that this type of heteroleptic complexes can be used as promising precursors to prepare CVD copper-containing films. The precursor Cu(dpk)(hfa) is air and moisture stable. Low melting point of this compound may provide using it in a bubbler or other liquid-precursor delivery systems but additional tests of thermal stability of the liquid phase are of legitimate interest. This precursor seems to be good alternative for Cu(hfa)2 and many other copper(II) precursors.

Acknowledgement We are indebted to Dr. I.V. Korolkov (NIIC SB RAS, Novosibirsk, Russia) for XRD analysis and to Dr. V.V. Kaichev (IC SB RAS, Novosibirsk, Russia) for XPS analysis.

References [1] M.J. Hampden-Smith, T.T. Kodas (Eds), The Chemistry of Metal CVD, VCH, Weinheim, 1994, pp. 178-238. [2] V.N. Vertoprakhov, S.A. Krupoder, Preparation of thin copper films from the vapour phase of volatile copper(I) and copper(II) derivatives by the CVD method, Russ. Chem. Rev. 69(12) (2000) 1057-1082.

182

V.V. Krisyuk et al. / Physics Procedia 46 (2013) 174 – 182

[3] A.C. Jones, M.L. Hitchman (Eds), Chemical Vapour Deposition: Precursors. Processes and Applications, RSC, 2009, pp. 331-337 [4] D. Temple, A. Reisman, Chemical Vapor Deposition of Copper from Copper (II) Hexafluoroacetylacetonate, J. Electrochem. Soc., 136 (11) (1989) 3525-3529. [5] A.E. Kaloyelos, A. Feng, J. Garhart, K.C. Brooks, S.K. Ghosh, A.N. Saxena, F. Luehers, J. Electron. Mater., Low-temperature metal-organic chemical vapor deposition (LTMOCVD) of device-quality copper films for microelectronic applications 19 (3) (1990) 271-276. [6] P.A. Stabnikov, I.A. Baidina, K.V. Kradenov, S.A. Gromilov, I.K. Igumenov, Izv. Sib. Otd. Akad. Nauk SSSR, Ser. Khim. Nauk (Engl. Transl.), 5 (1989) 21-28. [7] S.A. Gromilov, I.A. Baidina, Regularities of Crystal Structures of Cu(II) ȕ-Diketonates, J. Struct. Chem. 45 (2004) 1031-1081. [8] V.V. Krisyuk, I.A. Baidina, I.V. Korolkov, P.P. Semyannikov, P.A. Stabnikov, S.V. Trubin, A.E. Turgambaeva, New volatile heteroleptic complex of copper(II): Comparison of two polymorphs, Polyhedron, 49(1) (2013) 1-6. [9] P.A. Stabnikov, G.I. Zharkova, I.A. Baidina, S.V. Tkachev, V.V. Krisyuk, I.K. Igumenov, Synthesis and properties of a new ketoiminate derivative of the 2,2,6,6-tetramethylheptane-3,5-dionate ligand to prepare volatile CVD precursors, Polyhedron 26 (2007) 4445–4450 [10] S.V. Sysoev, N.B. Morozova, G.I. Zharkova, I.K. Igumenov, P.P. Semyannikov, V.M. Grankin, Vapour Pressure and Thermoanalytical Study of Diethyldithiocarbamates of Platinum Metals, Therm. Anal. Calorim. 53 (1998) 87-96. [11] G.I. Zharkova, S.V. Sysoev, P.A. Stabnikov, V.A. Logvinenko, I.K. Igumenov, Vapor pressure and crystal lattice energy of volatile palladium(II) ȕ-iminoketonates, J. Therm. Anal. Calorim. 103 (2011) 381-385. [12] J.H. Scofield, J. Electron Spectrosc. Relat. Phenom. Hartree-Slater subshell photoionization cross-sections at 1254 and 1487 eV, 8 (1976) 129-137. [13] J. Batista, A. Pintar, D. Mandrino, M. Jenko, XPS and TPR examinations of Ȗ-alumina-supported Pd-Cu catalysts, V. Martin, Appl. Catal. A. 206 (2001) 113-124. [14] V.I. Bukhtiyarov, V.V. Kaichev, I.P. Prosvirin, X-ray photoelectron spectroscopy as a tool for in-situ study of the mechanisms of heterogeneous catalytic reactions, Top. Catal. 32 (2005) 3-15. [15] N.S. McIntyre, M.G. Cook, X-ray photoelectron studies on some oxides and hydroxides of cobalt, nickel, and copper, Anal. Chem. 47 (1975) 2208-2213. [16] J.C. Otamiri, S.L.T. Andersson, A. Andersson, Ammoxidation of toluene by YBa2Cu3O6+x and copper oxides: Activity and XPS studies, Appl. Catal. 65 (1990) 159-174. [17] S. Poulston, P.M. Parlett, P. Stone, M. Bowker, Surface Oxidation and Reduction of CuO and Cu2O Studied Using XPS and XAES, Surf. Interface Anal. 24 (1996) 811-820. [18] M. Richter, M.J.G. Fait, R. Eckelt, M. Scneider, J. Radnik, D. Heidemann, R. Fricke, Gas-phase carbonylation of methanol to dimethyl carbonate on chloride-free Cu-precipitated zeolite Y at normal pressure, J. Catal. 245 (2007) 11-24. [19] B.R. Strohmeier, D.E. Leyden, R.S. Field, D.M. Hercules, Surface spectroscopic characterization of Cu/Al2O3 catalysts, J. Catal. 94 (1985) 514-530. [20] A. Wöllner, F. Lange, H. Schmelz, H. Knözinger, Characterization of mixed copper-manganese oxides supported on titania catalysts for selective oxidation of ammonia, Appl. Catal. A. 94 (1993) 181-203. [21] G. Moretti, Auger parameter and wagner plot in the characterization of chemical states: initial and final state effects, J. Electron Spectrosc. Relat. Phenom. 76 (1995) 365-370. [22] J. van den Brand, P.C. Snijders, W.G. Sloof, H. Terryn, J.H.W. de Wit, Acid−Base Characterization of Aluminum Oxide Surfaces with XPS, J. Phys. Chem. B, 108 (2004) 6017-6024. [23] J. Mendialdua, R. Casanova, F. Rueda, A. Rodriguez, J. Quinones, L. Alarcon, E. Escalante, P. Hoffmann, I. Taebi, L. Jalowiecki, X-ray photoelectron spectroscopy studies of laterite standard reference material, J. Mol. Catal. A, 228 (2005) 151-162. [24] N. Kosova, E. Devyatkina, A. Slobodyuk, V. Kaichev, Surface chemistry study of LiCoO2 coated with alumina, Solid State Ionics, 179 (2008) 1745-1749.