`
`Page 1 of 16
`
`REDACTED
`
`REDACTED
`
`REDACTED
`
`RE
`DA
`CT
`ED
`
`
`
`EDACTE
`D
`
`REDACTED
`
`DACTE
`D
`
`ACT
`ED
`
`Page 2 of 16
`
`RE
`DA
`CT
`ED
`
`
`
`Page 3 of 16
`
`REDACTED
`
`REDACTED
`
`
`
`
`
`Volume 307, Numbers l—?.. 10 October 1997
`
`Contents of Volume
`
`
`
`
`
`The structure and residual stress in Si containing diamond-like carbon coating
`W.—J. Wu and M.-H. Hon .
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`Transmission electron microscopy study of Si 8-doped GaAs/AlGaAs/InGaAs/GaAs pseudomorphic high electron mobility
`transistor structures
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`1
`
`6
`
`S.l. Molina and T. Walther .
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`10
`
`The peroxide route of the successive ionic layer deposition procedure for synthesizing nanolayers of metal oxides, hydroxides and
`peroxides
`V.P. Tolstoy .
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`Investigation of the plasma polymer deposited from pyrrole
`J. Zhang, M.Z. Wu. T.S. Pu, Z.Y. Zhang, R.P. Jin, Z.S. Tong, D.Z. Zhu, D.X. Cao, F.Y. Zhu and J.Q. Cao .
`
`Effect of sandblasting on adhesion strength of diamond coatings
`B. Zhang and L. Zhou .
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`14
`
`21
`
`Deposition of boron carbide by laser CVD: a comparison with thermodynamic predictions
`J.C. Oliveira and O. Conde .
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`Correlation between microstructure and the optical properties of TiO2 thin films prepared on different substrates
`Y. Leprince—Wang, K. Yu-Zhang, V. Nguyen Van, D. Souche and J. Rivory .
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`29
`
`38
`
`43
`
`A structural approach to gallium phosphate thin solid films
`F. Touitin, P. Armand, A. lbanez. A. Manteghetti and E. Philippot .
`
`.
`
`Preparation and characterization of ZnO:Al films by pulsed laser deposition
`Z.Y. Ning. S.l-l. Cheng, S.B. Ge. Y. Chao, Z.Q. Gang, Y.X. Zhang and Z.G. Liu .
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`50
`
`Characterization of C—N thin films deposited by reactive excimer laser ablation of graphite targets in nitrogen atmosphere
`A.P. Caricato, G. Leggieri, A. Luches. A. Peirone, E. Gyorgy. I.N. Mihailescu, M. Popescu, G. Barucca, P. Mengucci, J.
`Zemek and M. Trchova .
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`54
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`60
`
`C—The synthesis of CeO3+“ - n H30 nanolayers on silicon and fused-quartz surfaces by the successive ionic layer deposition
`technique
`V.P. Tolstoy and A.G. Ehrlich .
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`Multivariate analysis of noise—corrupted PECVD data
`A. von Keudell, A. Annen and V. Dose .
`.
`.
`.
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`65
`
`7]
`
`A kinetic model for photochemical vapor deposition from germane and silane
`M. Tao .
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`Crystallographic and morphological characterization of reactively sputtered Ta. Ta—N and Ta—N—O thin films
`M. Stavrev. D. Fischer‘. C. Wenzel. K. Drescher and N. Mattem .
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`Texture in evaporated Ag thin films and its evolution during encapsulation process
`Y. Zeng, Y.L. Zou and T.L. Alford .
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`79
`
`89
`
`96
`
`Nucleation and growth of Cu thin films on silicon wafers deposited by radio frequency sputtering
`J.—C. Lin and C. Lee .
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`Surface dilational behavior of docosanic acid monolayers spread on the surface of drops of polymer solutions
`R. Wilstneck. J. Reiche and S. Forster .
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`A model of oxide layer growth on Ag+ and Ptl ion implanted nickel anode in aqueous alkaline solution
`l.S. Tashlykov .
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`Effects of polymer substrate surface energy on nucleation and growth of evaporated gold films
`R.L.W. Smithson, D.J. McClure and DF. Evans .
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`100
`
`106
`
`110
`
`Elsevier Science SA.
`
`Page 4 of 16
`
`Page 4 of 16
`
`
`
`REDACTE
`D
`
`REDACTED
`
`Page 5 of 16
`
`REDACTED
`
`
`
`Comparison of adsorption characteristics of methyl orange and oz-naphthol orange molecules onto the cationic Langmuir-—Blodgett
`films
`
`M. Takahashi, K. Kobayashi, K. Takaoka and K. Tajima .
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`274
`
`280
`
`vii
`
`Self-assembled multilayer formation of an aromatic bifunctional molecule via selective ionic interaction
`V. Patil, K.S. Mayya and M. Sastry .
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`Photovoltaic properties of indium selenide thin films prepared by van der Waals epitaxy
`J.F. Sanchez-Royo, A. Segura, 0. Lang, C. Peuenkofer, W. Jaegermarm, A. Chevy and L. Roa .
`
`Sol—ge1 prepared In2O3 thin films
`A. Gurlo, M. Ivanovskaya, A. Pfau, U. Weimar and W. Gtipel
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`,-
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`283
`
`288
`
`The study of the antigen—antibody reaction by fluorescence method in LB films for immunosensor
`G.K. Chudinova, A.V. Chudinov, V.V. Savransky and A.M. Prokhorov .
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`294
`
`Structural properties of a-Si
`N,:H films grown by plasma enhanced chemical vapour deposition by SiH. + NI-[3 + H2 gas mixtures
`F. Giorgis. C.F. Pirri and E. Tresso .
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`Nucleation of strontium titanate films grown by PLD on silicon: a kinetic model
`R. Castro-Rodriguez, E. Vasco, F. Leccabue, B.E. Watts, M. Zapata-Torres and A.l. Oliva .
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`.
`
`.
`
`.
`
`.
`.
`
`.
`
`.
`
`.
`.
`
`.
`
`.
`
`.
`.
`
`.
`
`.
`
`.
`.
`
`298
`
`306
`
`311
`313
`
`Author Index of Volume 307 .
`Subject Index of Volume 307 .
`
`.
`.
`
`.
`.
`
`.
`.
`
`.
`.
`
`.
`.
`
`.
`.
`
`.
`.
`
`.
`.
`
`.
`.
`
`.
`.
`
`.
`.
`
`.
`.
`
`.
`.
`
`.
`.
`
`.
`.
`
`.
`.
`
`.
`.
`
`.
`.
`
`.
`.
`
`.
`.
`
`.
`.
`
`.
`.
`
`.
`.
`
`.
`.
`
`.
`.
`
`.
`.
`
`.
`.
`
`.
`.
`
`.
`.
`
`.
`.
`
`.
`.
`
`.
`.
`
`.
`.
`
`.
`.
`
`.
`.
`
`.
`.
`
`.
`.
`
`.
`.
`
`.
`.
`
`.
`.
`
`.
`.
`
`.
`.
`
`.
`.
`
`.
`.
`
`.
`.
`
`.
`.
`
`.
`.
`
`.
`.
`
`.
`.
`
`.
`.
`
`.
`.
`
`.
`.
`
`.
`.
`
`.
`.
`
`.
`.
`
`.
`.
`
`The table of contents of Thin Solid Films is included in ESTOC—Elsevier Science Tables of Contents service—which can be
`accessed on the World Wide Web at the following URL addresses:
`http://www.elsevier.nl/locate/estoc or http://www.elsevier.com/locate/estoc
`
`describing our requirements is available from the publisher upon request.
`
`The publisher encourages the submission of articles in electronic form thus saving time and avoiding rekeying errors. A leaflet
`
`Page 6 of 16
`
`Page 6 of 16
`
`REDACTED
`
`
`
`
`
`Thin Solid Films 307 (1997) 79-88
`
`the
`
`fl
`
`rid]
`
`Crystallographic and morphological characterization of reactively
`sputtered Ta, Ta—N and Ta—N—O thin films
`
`Momtchil Stavrev "°' , Dirk Fischer ", Christian Wenzel ", Kurt Drescher 3, Norbert Mattem "
`
`' Dresden University of Technology, Semiconductor and Micrcsystems Technology laboratory, Dresden 01062, Germany
`b Institute of Solid State and Materials Research Helmholtzstr. 20, Dresden 01069, Germany
`
`Received 29 January 1997; accepted 21 May 1997
`
`
`
`Abstract
`
`This paper concentrates on the deposition of Ta, Ta—N and Ta—N—O thin films by r.f. magnetron sputtering in Ar/N,/O2 gas
`mixtures. The film properties and their suitability as diffusion barriers and protective coatings in silicon devices were characterized using
`four-point probe measurements, Auger electron spectroscopy, Rutherford backscaxtering, glancing angle X-ray diffractometry, atomic
`force microscopy and scanning electron microscopy. With the addition of N, to the gas mixture a transition from tetragonal Ta to
`b.c.c.-Ta(N) was detected,
`leading to the nanocrystalline metastable b.c.c.-Ta(N) phase with approximately 20 at.% interstitially
`incorporated nitrogen. Increasing the nitrogen flow above a critical value, an abmpt transition between metal-sputtering to nitride-sputter-
`ing mode was observed, resulting in a sharp increase in the N:Ta atomic ratio slightly above the stoichiometric value for the TaN phase.
`which was found to exhibit f.c.c. structure. With the addition of oxygen at fixed nitrogen flow the films tend to grow in an amorphous
`state. Due to the lack of short-circuit diffusion paths, the as-deposited amorphous Ta(N,O) films are considered as excellent candidates for
`ultra-thin diffusion barriers and protection layers in future Cu-metalliaed ULSI devices. ©1997 Elsevier Science S.A.
`
`Keywords: Tantalum; PVD; Amorphous materials; Diffusion barrier
`
`1. Introduction
`
`In the past, Ta and Ta-based compounds have been
`investigated as thin film resistor materials with low tem-
`perature coefficient of resistivity [1] and as stable contact
`materials to Si [2]. More recently, the suitability as diffu-
`sion or drift barriers between metal layers (Al, Cu. etc.)
`and semiconductors (Si, GaAs) or dielectrics (SiO2, poly-
`imide) has been examined and a superior thermal stability
`has been reported [3—12].
`Concerning the future Cu-metallized interconnection
`systems, as the line widths diminish, the barrier layer can
`significantly increase the net line resistance. Hence, the
`barrier has to be very thin to avoid increasing net line
`resistance for a given cross-sectional area of the conductor,
`but it has to be stable enough to provide acceptable barrier
`properties [4—7]. The use of amorphous or amorphous-like
`diffusion barriers, free of extended defects, can signifi-
`
`‘ Corresponding author. Fax: +49 351 4637172;
`stavrev@ehmgw1.et.tu-dresden.de
`
`e-mail:
`
`0040-6090/97/$17.00 0 1997 Elsevier Science S.A. All rights reserved.
`PII S0040-6090(97)00319-2
`Page 7 of 16
`
`cantly improve the thermal stability and reliability of the
`Cu-based contact and interconnect systems [3].
`Sputter-deposited Ta—N alloys have been known as an
`attractive class of materials because of their high chemical
`and mechanical stability and good conductivity. As has
`been reported earlier,
`the electrical properties,
`the stoi-
`chiometry, the crystallographic structure and the morphol-
`ogy of these metallic compounds depend on the sputtering
`conditions, substrate type, pre-treatment and film thick-
`ness. In order to optimize the Ta—N film properties many
`researchers have investigated the role of the reactive gas
`component, either as nitrogen partial pressure or nitrogen
`partial flow rate [l,4,13-18]. The reported sequence of the
`phases produced by increasing the nitrogen content [l,l3-
`15] is consistent with the equilibrium binary phase diagram
`[19] and can be summarized as follows: (i) equilibrium
`body-centered cubic at-Ta phase or metastable tetragonal
`[3-Ta, which are known to have a low solubility for N (5 S
`at.%) at room temperature [I9]; (ii) h.c.p.-Ta,N existing
`over a range between 22 and 35 at.% N [1,13,15,19]; (iii)
`f.c.c.-TaN phase [l,l3,l4]; and (iv) various nitrogen-rich
`compounds (Ta5N5, Ta4N,, Ta3N5), at which the nitrogen
`concentration levels off [l3,l5,19].
`
`Page 7 of 16
`
`
`
`80
`
`M. Sravrev er a1. / Thin Solid Films 307 (1997) 79-88
`
`On the contrary, other researchers have observed in the
`initial stage of N2 addition a partial or complete transfor-
`mation of B-Ta to oz-Ta phase prior to the Ta2N or TaN
`formation [16—18]. Aita et al. [18] observed that after such
`transformation, a change from b.c.c. Ta(N) to f.c.c.-TaN
`occurred abruptly without deposition of Ta,N, when the
`reactive gas component was increased by only 0.3%. But
`no sufficient fundamental research on the deposition condi-
`tions leading to this transfomration and on the film compo-
`sition and microstructure is present in the literature.
`In the Ta—O binary system, up to approximately 30
`at.% of oxygen could be interstitially dissolved in the
`b.c.c.-Ta lattice prior to conversion into amorphous or
`polycrystalline Ta20, phase [1,l6]. As an impurity in
`polycrystalline Ta films, 0 is believed to increase the
`effectiveness of the diffusion barriers by decorating the
`extended defects such as grain boundaries, thereby block-
`ing the active paths for grain boundary diffusion [12].
`Furthermore, with the growth of tantalum nitrides and
`oxides a reduction in grain size was observed, leading to
`nanocrystalline or amorphous-like structure [l,5,6,l3].
`To clarify the influence of N and 0 addition on the
`diffusion barrier behaviour,
`it is necessary not only to
`determine the composition and crystalline structure of the
`thin films but also the film morphology in terms of grain
`size, grain orientation, degree of amorphization, etc. Since
`there are some discrepancies in the literature on how the
`nitrogen and oxygen influence the film properties,
`this
`paper contributes to the general understanding of the con-
`ditions for formation of different Ta, Ta—N and Ta—N—O
`
`phases. Furthermore, it concentrates on the reactive sput-
`tering of conductive diffusion barriers with a special em-
`phasis on the crystallography and the existence of
`nanocrystalline or amorphous phases in the Ta—N binary
`system and partially in the Ta—N—O ternary system.
`
`2. Experimental details
`
`Both substrate cleaning and film deposition were per-
`fonned on a five-chamber-cluster-tool including load-lock,
`dealer, Ta-PVD module and inductively coupled plasma
`(ICP) soft etch module. The base pressure in the PVD
`chamber was 3 X 10" Pa. The Ar, N2 and 02 gas flows
`were controlled within 10.1 sccm by mass flow con-
`trollers, which guarantee reproducible deposition condi-
`tions. The gas purity was 99.9999%. The base vacuum and
`the Ar/N2/O2 gas mixture were analyzed by quadrupole
`mass spectrometer, mounted in a separate throttled UHV
`chamber, which was differentially pumped down to 10"
`Pa. With this configuration a reduction in total and partial
`pressure in the spectrometer chamber occurs. This enables
`gas monitoring during sputtering and guarantees high sig-
`nal-to-noise ratio.
`
`For this study, 100 mm Q Si(100) wafers were used. A
`standard RCA clean was performed prior to loading them
`
`Page 8 of 16
`
`into the load-lock. After soft-etching at 200 W in Ar
`plasma and without breaking vacuum, 10 to 100 nm thin
`Ta-based films were deposited by r.f. magnetron sputtering
`at 1 kW forward power from a 332 m E Ta target
`(99.9S%). The target-to-substrate distance was 50 mm.
`While the Ar flow was kept at 5 sccm, the nitrogen flow
`<15": and oxygen flow (D02 were varied between 0-5 sccm
`and 0-10 sccm respectively, resulting in process pressures
`between 0.22 and 0.3 Pa. For determination of the deposi-
`tion rates and thickness non-uniforrnities the films were
`
`patterned by lift-off technique and measured by surface
`profilometry. For characterization of the film microstruc-
`ture 1.5 pm thick multilayers consisting of several single
`films were sputtered under simultaneous variation of the
`nitrogen flow from 0 to 5 sccm with the rate of 0.1
`sccm/rnin.
`The Ta, Ta—N and Ta—N—O films were analyzed by
`four-point probe sheet resistance measurements and laser
`profilometry for characterizing the resistivity and intrinsic
`film stress, respectively. The fihn composition was deter-
`mined by Auger electron spectroscopy (ABS) in combina-
`tion with Ar sputter etching using elemental sensitivity
`values from Ref. [20]. Additionally, the film composition
`was
`characterized by 3.34 MeV ‘I-le"' Rutherford
`backscattering spectrometry (RBS). The RBS measure-
`ments were quantified by RUMP simulation. For determi-
`nation of the nitrogen content a calibration using the
`procedure described by Herring [21] was performed. X-ray
`diffraction (XRD) patterns of 100 nm Ta, Ta—N and
`Ta—N—O films were recorded at a HZG4 diffractometer
`
`equipped with Siemens rotating Cu anode, thin film sup-
`plement and secondary graphite monochromator. To en-
`hance the sensitivity for X-rays, the measurements were
`performed in parallel beam geometry at a constant incident
`angle of a:= 2°. The registered angle range was 20=
`20.0—l00.0° with a step size A2@= 0.05° and a measur-
`ing time of 20 s per step. To analyze the phase composi-
`tion of the thick Ta—N multilayers, XRD patterns at differ-
`
`3a.
`
`5..
`
`-0...
`
`[Pa] 3
`PartialPressurepm»
`
`
`J.
`
`'
`
`high-nu
`maul-sputtering
`mode
`
`0.0
`0.5
`1.0
`1,5
`2.0
`2.5
`3.0
`3.5
`4.0
`4.5
`5.0
`
`Nitrogen Flow Rate (DN: [sccm]
`
`Fifi. l. Nitrogen partial pressure p"; in the quadrupole chamber VS-
`“i“'°8¢|1 flow <15": Parameters: Ar flow £15,“: 5 sccm; forward power! 1
`kW; total pressure: 0.22. . .0.3 Pa.
`
`Page 8 of 16
`
`RED
`ACT
`ED
`
`
`
`/.2"?ammm...ammm...an4%l.|n.0«z..<HA55Aamnzv
`
`
`
`
`oar...GL95A50divq?_E\E5MA50036A50035nEo8v9mmmouamaflmaE=_S:22=owo.=_Z.mmo.=m.b_>cu_woM.82:oE»onoQ$6:NO.>6=«Z.Boc._<EE:5...
`
`
`
`
`
`MandmadIone:own93.Nonnouzua.W,I3..8..2,...I3%2:onmZIa._.UII«E3..IEm~«.2o..m218...Hms;84Cl3..Iasmaconnzua.WoiIw:_m._Imen9.8onnznfl.mweas.8:3:8n5;.o3nzua.FRdI.33...:S:as.oNnzua.WR...IIEdiR.23o2mZ13.
`SandIEd3Im:3..ogmZuah%«no85vmodV«mo:mm:newoomup.
`
`
`88
`
`.uo.:.m3E82AIv
`
`cm.25¢55OIZIE.EaZluh.5..«o8:32..a5nobuEE_._w
`
`
`
`
`
`
`M._2..."...
`
`Page 9 of 16
`
`Page 9 of 16
`
`
`
`RED
`ACT
`ED
`
`REDACTED
`
`Page 10 of 16
`
`
`
`M. Staurev et al. / Thin Solid Films 307 (1997) 79-88
`
`83
`
` Nitrogen Flow
`
`5 seem
`
`out
`
`3.5 SOCIT1
`
`3 seem
`
`2.5 seem
`
`2 seem
`
`1.5 SCCITI
`
`1 seem
`
`0 seem
`
`
`
`Intensity(a.u.)
`
`550
`
`500
`
`450
`
`400
`
`0) 01O
`
`008
`
`250
`
`200
`
`150
`
`100
`
`50
`
`0
`
`o
`
`10
`
`20
`
`so 40 so so 70
`
`so so 1oo
`
`Fig. 4. XRD patterns of 100 nm Ta—N thin films sputtered in Ar/N, gas mixture.
`
`28 [deg]
`
`Table 1 gives a summaiy of the main properties of the
`investigated Ta—N—O films.
`
`3.2. Composition of the Ta, Ta—N and Ta—N-0 thin films
`
`The nitrogen-to-tantalum (N:Ta) atomic ratio in the
`films as determined by RBS and AES is presented in Table
`1. As measured by both techniques, there is only a slight
`increase in the nitrogen concentration between 0 and 2.5
`sccm, resulting in about 20 at.% of N for ID": = 2.5 sccm.
`By further addition of nitrogen, an abrupt increase of the
`N:Ta-ratio is observed. Neither by RBS nor by ABS,
`N:Ta—ratios typical for the fonnation of the Ta2N phase
`were found. According to ABS, for <15", =3.5 sccm a
`nearly stoichiometric N:Ta-ratio was detemiined, suggest-
`ing the deposition of a TaN phase. Above 3.5 sccm N2
`flow a saturation of the N:Ta atomic ratio at 1.1 occurs.
`
`On the other hand, using RBS a much higher N:Ta-ratio of
`about 1.6 is estimated. In order to exclude non-uniformity
`effects, ABS depth profiling was performed. The results,
`plotted in Fig. 3 for 45": = 2.5 sccm and 47”, - 3.5 sccm,
`show uniform distribution of the nitrogen throughout both
`
`films. As further impurities small amounts of 0, C and Ar
`(total impurity concentration 5 5 at.%) were found in the
`Ta—N films.
`
`In the case of Ta—N—O films, sputtered at <PA,:<15N2:¢oz
`= 5:2.5:2, the predominant impurity dissolved in the films
`was found to be nitrogen (about 17 at.%). Furthennore,
`approximately 3 at.% oxygen were measured by AES.
`
`Table 2
`
`Summary of the XRD results on Ta and Ta—N thin films
`
`Nitrogen
`flow, <17"1
`(sccm)
`
`Crystalline
`structure/
`phase
`
`0
`1
`
`1.5
`2
`2.5
`3
`3.5
`5
`
`predominantly tetragonal Ta
`tetragonal Ta +
`b.c.c.-Ta
`b.c.c.-Ta(N)
`b.c.c.-Ta(N)
`b.c.c.-Ta(N)
`f.c.c.-TaN
`f.c.c.-TaN
`f.e.c.-TaN + amorphous
`
`Lattice
`constants.
`no (nm)/
`co (nm)
`
`0.53/ 1 .006
`0.53/ 1.006
`0.334
`0.335
`0.337
`0.342
`0.436
`0.435
`0.435
`
`Mean
`crystallite
`size,
`D (nm)
`
`90
`80
`10
`5.5
`4
`3
`6
`5
`5
`
`Page 11 of 16
`
`Page 11 of 16
`
`
`
`M. Slaureu et al. / Thin Solid Film: 307 (1997) 79-88
`
`3.3. Crystallographic and morphological characterization
`of thin Ta—based films
`
`Fig. 4 shows the X-ray diffraction patterns of 100 nm
`Ta—N films grown in various Ar/N, mixtures on Si(100)
`
`substrates. The variations in position, intensity and shape
`of the reflections indicate changes in the phase composi-
`tion of the Ta—N films. The analysis of the angle depen-
`dence of the width at half maximum points to grain size
`effects. The calculation of the mean crystallite size, D
`
`
`
`
`.....,,......n»_e.....esm.m.m..:.....ami.at.§..._._eE3....e.wa..%eh...L....._....mrw....K$.LE.“
`
`
`
`
`..4...l.....,3»)
`
`
`
`
`0.‘3...............2...fl..»..5.....a.u.....:
`
`Fig. 5. AFM images of 100 nm Ta-N thin films sputtered in Ar ‘-
`/N gum’:
`scan size: 500 X 500 nmz; Note an differences
`Z ranlgetflt On
`seem: (0 3.5 socm
`
`Page 12 of 16
`
`'..
`
`gt
`
`t,.__.n.,ew__
`..~dfi
`. 4%.. '~
`;()3
`-
`-.
`-
`s'1oo.
`I(
`) (a) 0 sccm. (b) 1 sccm. (c) 1.5 sccm. (d) 2.5 Sccm 9
`
`..
`..£.,..
`.._
`...
`
`.
`
`.....a............
`...u..2...~..H
`
`
`....m?%...M..%...
`
`Page 12 of 16
`
`
`
`
`M. Staureu er al. / Thin Solid Films 307 (I997) 79-88
`
`85
`
`500
`
`‘5°
`400
`
`350
`
`amorphous Ta(N,O)
`
`300
`
`250
`
`
`
`100 50
`
`from the full width at half maximum (FWHM) after
`correction for instrumental contribution using the Scherrer
`equation [24]:
`A
`D FWHM - cos0
`
`(2)
`
`with A, X-ray wave length and 9, diffraction angle, gives
`values in the nanometer range. These values along with the
`determined phase composition and the calculated lattice
`constants are given in Table 2.
`Pure Ta films sputtered in Ar plasma grow predomi-
`nantly in the tetragonal B-phase and are apparently poly-
`crystalline as determined by XRD (Fig. 4). In Fig. 5a, an
`AFM image of the film surface is plotted, showing grains
`with circular cross-section and low surface roughness
`(RMS '4 0.54 nm).
`The addition of only 1 sccm N, to the sputtering gas
`results in phase mixture of tetragonal B-Ta and b.c.c.-Ta.
`According to the AFM image (Fig. 5b) this film consists
`not only of circular grains like the [%-Ta film, but also of
`long grains parallel to the surface.
`At 4?": = 1.5 sccm, a b.c.c. phase is formed with lattice
`constant ac, = 0.335 nm, which is slightly higher than the
`value reported for the b.c.c.-Ta phase [JCPDS-ICDD 4-
`788]. As shown in Fig. 5c, this films consist predominantly
`of long. randomly oriented grains attributable to the b.c.c.
`phase. The calculated mean crystallite size is about 6 nm
`and the RMS roughness about 0.75 nm.
`By further addition of N, up to the threshold flow <1>,';2,
`the b.c.c. crystalline structure is preserved. Additionally, a
`peak broadening and a slight shift in peak position towards
`smaller 20-angles become obvious,
`the latter revealing
`further increase of the lattice constant an. The calculated
`lattice constant for the phase sputtered at <15N2 = 2.5 sccm
`is 4% higher than that of pure b.c.c.-Ta and suggests the
`formation of a metastable b.c.c.-Ta(N) phase. This lattice
`constant ao= 0.342 nm is even higher than the lattice
`constant of
`the b.c.c.-TaN0_,-phase
`(a0 = 0.337 nm)
`[JCPDS-ICDD 25-1278]. According to the calculations,
`this film is nanocrystalline with crystallites about 3 nm in
`size. The surface, examined by AFM (Fig. 5d). appears to
`be nearly structureless and very smooth (RMS = 0.3 nm).
`Above the critical value 45,32, at CDNZ = 3 sccm the
`crystallographic structure changes abruptly to f.c.c.-TaN
`according to the XRD measurements. The AFM image
`(Fig. 5e) shows larger grains as compared to 3 or b.c.c.-Ta,
`accompanied by an increase in the surface roughness
`(RMS= 1.29 nm). This trend is even much more pro-
`nounced for the f.c.c.-TaN film sputtered at <15”: = 3.5
`sccm (see Fig. 5f). The further addition of nitrogen up to 5
`sccm leads to the formation of a phase mixture consisting
`of f.c.c.-TaN and amorphous phase, evidenced by the
`changed shape of the first reflection. The peak consists of
`a superposition of the (111)-TaN-reflection and a diffuse
`maximum (see Fig. 4).
`The addition of O2 to the Ar/N2 gas mixture at the
`
`Page 13 of 16
`
`Intensity(a.u.) 200
`
`150
`
`20
`
`30
`
`40
`
`50
`
`60
`
`70
`
`B0
`
`90
`
`100
`
`26 [deg]
`
`Fig. 6. XRD pattern of 100 um thin b.c.c.-Ta(N,O) film sputtered at
`d5A,:¢,,z:<Poz = 5:2.5:2.
`
`ratio ¢,,_,:<I>N2:45o2 = 5:2.5:2 produces an XRD pattern
`(Fig. 6) with only two very broad peaks comparable to the
`(110) and (211) peaks of the b.c.c. Ta(N) phase at Q”: =
`2.5 sccm. The peak shape and the disappearing of the
`(200) and (220) peaks indicate lattice distortion and/or
`growth of an amorphous Ta(N,O) thin film. The shift in
`peak position towards smaller 20-angles reveals a further
`increase of the interatomic distances. AFM measurements
`
`showed that the surface of the 100 nm thin amorphous
`Ta(N,O) films was very smooth (RMS = 0.28 nm).
`
`3.4. Investigation of thick Ta / Ta—N multilayers
`
`As shown above the phase composition of the Ta—N
`films changes from predominantly tetragonal
`to b.c.c.-
`Ta(N) and finally to f.c.c-TaN by variation of the nitrogen
`flow. This offers the possibility to prepare gradual multi-
`layers with a defined sequence under variation of the
`nitrogen flow from 0 to 5 sccm. Fig. 7 shows the cross-
`section of an 1.5 um thick multilayer sputtered under
`variation of the (DMZ. Three different zones could be
`distinguished. In the first Zone A, the Ta film consists of
`vertically oriented close-packed columns. With further ad-
`dition of nitrogen to the gas mixture the formation of a
`
`chose "*9" o3eu_3
`‘
`tunzrmvno
`
`
`Fig. 7. Cross-sectional SEM image of cleaved Ta/Ta—N multilayer.
`
`Page 13 of 16
`
`
`
`86
`
`".~
`3.
`
`5200
`E‘WC
`
`9E
`
`29 [deg]
`
`Fig. 8. XRD patterns of 1.5 thick multilayer at different angles of
`incidence a from 1 to 20".
`
`second Zone B becomes obvious. In this zone, the film
`
`fracture results in tilted patterns, which could be associated
`with the elongated grains of the b.c.c.-Ta(N) phase. The
`density of these patterns increases with increasing film
`thickness and hence nitrogen flow. At a thickness of
`approximately 1 pm and 45”‘ =1 Q3: the abrupt formation
`of a third Zone C can be observed. The Ta—N layers in
`Zone C exhibit an apparent polycrystalline microstructure,
`i.e., vertical columns with mean diameter about 50 nm.
`This is consistent with the grain size estimated by AFM on
`the 100 nm f.c.c.-TaN films. The conformation of the
`
`phase sequence is given in Fig. 8, where the XRD patterns
`obtained at different incident angles are plotted. For (I = 1°
`only the reflections of the f.c.c.-TaN phase are visible. At
`a= 1.5° (t=100 nm), small contributions of the b.c.c-
`Ta(N) are additionally observed, which become enhanced
`with increasing incident angle. As in the case of single
`Ta—N films the formation of the Ta2N is not observed. At
`the penetration depth of zz 0.5 ,u.m (at -= 6°), the multi-
`layer consists of several hundred nanometers surface layer
`of f.c.c.-TaN with b.c.c.-Ta(N) underneath. The further
`increase of a (a2 10°) shows additional reflections of
`
`tetragonal Ta in the XRD pattern, which indicates the
`existence of the B—Ta phase at the bottom of the multi-
`layer.
`
`4. Discussion
`
`The commonly used tetragonal or B-Ta thin films sput-
`tered in Ar plasma show typical columnar microstructure
`(Zone A in Fig. 7). Despite of the high chemical stability
`towards Cu and Si,
`the nature of the film defects,
`like
`voids and grain boundaries, could limit the performance
`and suitability of these films. Especially in applications,
`where ultra-thin (S 20 nm) diffusion barriers are needed,
`the use of polycrystalline L3-Ta films is expected to lead to
`short-circuit diffusion of Cu.
`
`.
`
`In Ar/N2 gas mixtures, the effect of an abrupt increase
`
`Page 14 of 16
`
`M. Staurev et al. / Thin Solid Films 307 (I997) 79-88
`
`
`
`of the reactive gas partial pressure above the critical flow
`rate C15,}: - 2.6 sccm indicates the change of target surface
`composition due to transition between sputtering from
`metallic target and target being covered by reaction prod-
`ucts, presumably TaN.
`The behaviour of the partial pressure, deposition rate
`and resistivity for IDNI 5 45,}:
`is typical for the metallic
`mode. In this mode the majority of the neutral and ionized
`reactive species are gettered by the condensation area,
`which includes the substrate surface. Concurrently,
`the
`substrate is not only condensation surface but also the
`reaction surface needed to produce stable or metastable
`Ta—N compounds. The main parameter for the formation
`of compounds is the ratio of the impact number of the N
`atoms to that of the Ta atoms. As determined by RBS,
`AES and XRD,
`the formation of the metastable Ta(N)
`
`phase only and the absence of the Ta,N and TaN phases in
`the metallic mode are not typical for reactively-sputtered
`Ta—N films [1,l3,l4]. Similar observations were reported
`in Ref. [16]. The small target-to-substrate distance (5 cm)
`and/or the large ratio between the target area and the
`substrate area are supposed to be responsible for this
`behavior. In this case, the deposition of the Ta,N and the
`TaN phases in the metallic mode could be suppressed, and
`a direct transition from b.c.c.-Ta(N) to f.c.c-TaN deposited
`in the compound-sputtering mode above (15132 can occur.
`Additionally, the film stoichiometry, structure and mor-
`phology, which depend on the nitrogen and tantalum flows
`towards the substrate, the sticking probability especially
`for nitrogen and the surface mobility, could differ drasti-
`cally from equilibrium phases. Due to the existence of film
`defects like voids, dislocations and impurity incorporation
`the formation of a nitrogen-stabilized cubic structure could
`be favored.
`
`For 45,,‘ < 2.5 sccm, the broad XRD peaks along with
`the determined nitrogen content, indicate continuous inter-
`stitial incorporation of nitrogen atoms into an expanded
`b.c.c.-Ta(N) lattice (Table 2). The AFM and SEM analyses
`of these films indicate the growth of randomly oriented
`elongated grains and according to XRD a decrease in mean
`grain size.
`The compound obtained at <15", = 2.5 sccm can be
`described as metastable b.c.c.-Ta(N) phase with nearly 20
`at.% nitrogen. The interstitial incorporation of nitrogen at
`room temperature exceeds that reported for thermal equi-
`librium conditions [19] and is presumably due to the
`non-equilibrium nature of the reactive deposition in the
`metal-sputtering mode. The broadening of the peaks in the
`XRD pattern are mainly due to smaller crystallite size.
`Since the XRD gives an estimation of the mean crystallite
`size only, the features observed by AFM (Fig. 5d) Could
`consist of several crystallites. In Ref. [18], the formation of
`9}, b-C-0--Ta(N) phase with similar lattice constant (00 = 339
`A) have been reported. Due to its nanocrystalline structure
`(mean crystallite size about 3 nm) in the as-deposited state.
`this metastable b.c.c.-Ta(N) phase could be observed as
`
`Page 14 of 16
`
`
`
`M. Staureu at al. / Thin Solid Films 307 (1997) 79-88
`
`87
`
`preferred candidate for diffusion barrier between 01 and Si
`or interlevel dielectrics.
`
`The introduction of oxygen to the sputtering mixture at
`the ratio ¢A,:<15N1:d>o1 = 5:2.5:2 forces the deposition of
`amorphous Ta(N,O). The further shift of the peak maxi-
`mum in the XRD pattern as compared to b.c.c.-Ta(N) may
`be due to additional interstitial oxygen incorporation. Ap-
`parently, this small amount of oxygen (3 at.%) and the
`already incorporated nitrogen (17 at.%) are sufficient to
`lead to a considerable distortion of the b.c.c. lattice and the
`
`formation of a novel metastable amorphous Ta(N,O) mate-
`rial. Due to the lack of short-circuit diffusion paths, 50 nm
`Ta(N,O) thin films have been found to be excellent diffu-
`sion barriers between Cu and Si and remain intact even
`
`after armealing at 600°C for 1 h in N,/H, ambient [6].
`The incorporation of 0 and N into the Ta films can
`additio