Preferentially orientated E-beam TiN thin films using focused jet of nitrogen gas
aa r X i v : . [ c ond - m a t . m t r l - s c i ] O c t Preferentially orientated E-beam TiN thin films using focused jet of nitrogengas
R. Ramaseshan a) and Feby Jose, S. Rajagopalan and S. Dash Thin Film and Coatings Section, Surface and Nanoscience Division, Materials Science Group,Indira Gandhi Centre for Atomic Research, Kalpakkam 603102, India. (Dated: 20 February 2018)
A modified electron beam evaporator has been used judiciously to synthesize TiN thin films with (111) pre-ferred orientation. This new design involved in creating local plasma by accelerating the secondary electronsemitted from the evaporating ingot by a positively biased semi-cylindrical anode plate kept in the vicinityand a jet of N gas has been focused towards the substrate as a reactive gas. We have observed a preferredorientation (111) with 25 ◦ angle to the surface normal and this was confirmed by pole figure analysis. Thephenomenon of preferred orientation (111) has been explained based on the rate of evaporation. The residualstress by the classical sin ψ technique did not yield any tangible result due to the preferred orientation. Thehardness and modulus measured by nanoindentation technique was around 19.5 GPa and 214 GPa. Thecontinuous multicycle indentation test on these films exhibited a stress relaxation. I. INTRODUCTION
Polycrystalline TiN films, belongs to the family ofrefractory transition metal nitrides exhibit characteris-tics of both covalent compounds and metals, such ashigh melting point, thermodynamic stability, high hard-ness, thermal conductivity which are necessary for themechanical components such as cutting tools, formingtools, etc.
In addition, TiN thin films have been usedfor cosmetic ersatz gold surfaces such as watch bezels,watch bands, wavelength selective transparent opticalfilms, diffusion barrier in integrated circuits, and as en-ergy saving coatings for windows due to its strong in-frared reflection.
Preferred oriented thin films largelyaffect the properties, performance and reliability of thecomponents than the polycrystalline thin films.
There are several techniques reported for synthe-sizing TiN in polycrystalline as well as highly ori-ented films includes reactive evaporation, magnetronsputtering, pulsed laser deposition (PLD) and ca-thodic arc evaporation, etc. Diatomic fcc structureTiN has a favored (200) orientation which is substantiallysupported thermodynamically due to their low surfaceenergy. The ion bombardment of the growing film inthe vacuum based plasma assisted process, the presenceof high-density plasma is identified as one of the most rel-evant deposition parameters and its variation may dras-tically alter the film structure and properties.
Pre-ferred crystalline growth and residual stress represent im-portant structural features of polycrystalline thin filmsthat influence the functionality significantly. An impor-tant question about the interrelation between preferredorientation and residual stress that changes the proper-ties of these thin films has been explained by Shin et al . The purpose of this work is first to examine the effectof the positive bias applied to enhance the formation of a) Electronic mail: [email protected] plasma near the vicinity and the gas feed direction onthe growth of TiN films with preferential orientation.
II. EXPERIMENTAL
An electron beam evaporator (in house assembled)was improvised for synthesizing ARE based nitrides. Aschematic of the deposition process enhanced by the dis-charge assembly setup is depicted in Fig.1. The vac-uum chamber was pumped down to a base pressure of1 x 10 − mbar by turbo molecular pump. A TemescalSIMBA 2 power supply with electron beam rastering ca-pability was used to evaporate titanium (4N pure) ingot.Prior to evaporation, the ingots were surface melted in-situ and degassed. During deposition, a constant flowof 5 N pure nitrogen was allowed (100 SCCM) and itwas directed towards the substrate in the vicinity of thedischarge assembly with an angle 25 ◦ to the substratesurface normal. The typical deposition parameters arelisted in Table.I.A stainless steel semi cylindrical plate having 25 mmradius and 50 mm length was placed 20 mm above theedge of the molten source. A positive DC bias of 100 Vwas applied to the cylindrical plate. The secondary elec-trons emanating from the ingot by electron beam im-pact is accelerated towards the semi cylindrical anode.These secondary electrons induce ionization of Ti vaporand nitrogen molecules. Axi-symmetric plasma engulfs Electron Beam Power (W) 250Working Pressure (mbar) 5 x10 − Cylindrical plate potential (V) +100 DCSubstrate biasing voltage (V) -500 DCSource-substrate distance (mm) 100Plasma current (A) 5Deposition Temperature (K) 525TABLE I. Deposition Parameters
SPECIMEN HOLDER 500-1000V25 -100V COPPER HEARTH N e PLASMA VACUUM CHAMBER
FIG. 1. Schematic of ARE-EB-PVD set up the space between the molten source and the substrateregion, which is sustained by the high constant currentsupply of the anode bias. Throughout the deposition aconstant plasma current of 5 A was maintained. The ion-ized species are transported towards a proximally heldsubstrate kept at a negative DC bias of 500 V, (lowcurrent). The advantage of this geometry for plasmacreation is that, it is confined plasma, which enhancesthe purity of the films. The deposition of crystallineTiN films was carried out on micro slide (glass slides)substrate by maintaining the substrate temperature of525 K. The deposition rate of 20 nm/min was achievedwith the above mentioned conditions.Identification of crystalline phases and crystal struc-ture studies were carried out by a STOE diffractometerin the GIXRD mode at glancing angle of 1 ◦ . The polefigure and residual stress mapping of the thin films wereanalyzed by X-ray diffractometer (D8 Advance, M/s.Bruker, Germany) at an incidence angle of 0 . ◦ . Com-positional homogeneity with respect to depth upto sub-strate interface was studied by Secondary Ion Mass Spec-trometry (SIMS) (IMS 4f, M/s. Cameca, France) instru-ment. The primary ion (Cs + ) beam with impact energyof 1.75 keV was used for sputtering with a beam cur-rent of 10 nA. The primary ion beam was rastered overan area of 200 x 200 µ m and the secondary species werecollected from a central region of 60 µ m diameter. TheSIMS crater depth and thickness of the films were mea-sured with a surface profiler (DEKTAK 6M, M/s. Veeco,USA).Surface morphology of TiN films was studied by using SIS AFM in contact mode. The hardness studies werecarried out using nano-indentation system (Open Plat-form, M/s. CSM, Switzerland) using a Berkovich inden-ter, according to the ISO 14577-1 standard with a loadingduration to peak load in 30 s and unload at the same rateas loading. The hardness and Youngs modulus were cal-culated from the load-displacement curve with the helpof Oliver and Pharr formalism . We have used continu-ous multi-cycle mode (CMC) technique with a load rangeof 1 - 20 mN with a progressive load increment of 1 mNat the same place (20 cycles) to observe any change inthe hardness. The unloading was allowed up to 10% ofthe corresponding applied load to maintain the contactbetween tip and surface. III. RESULTS AND DISCUSSION
In the present deposition set up, enhanced reactivityand high adatom energy imparted by localized plasmaand substrate bias ensures a complete reaction betweenTi and N ions and the formation of TiN films. By vary-ing the constant current supply to the anode plate andevaporation rate, different deposition rates can be real-ized. However, when we increased the deposition rateby increasing the titanium evaporation rate alone by ad-justing the electron beam power, both Ti and TiN wereformed together. So, we have synthesized TiN thin filmsusing the parameters mentioned in the experimental pro-cedure. The GIXRD pattern of TiN films deposited onthe micro slides by ARE based EB-PVD is shown inFig.2. This diffraction pattern matches exactly with thereported JCPDS-ICDD data file, albeit systematic shiftof all Bragg reflections to higher angle. This systematicshift in angle corresponds to the lattice contraction thatis attributed to compressive residual stress in the film.The lattice constant of this film calculated from the high-est peak of the x-ray diffraction is 0.41955 nm, whereasthe reference value is 0.42410 nm. The high ∆d% (1.64)indicating there is a large compressive stress in the filmcompared to standard TiN. Such residual strain are ex-pected in reactively deposited films due to (i) the thermalmismatch between thin film phases and the substrate, (ii)the lattice mismatch between the two, (iii) point defectsdue to non stoichiometry, (iv) preferred orientations ifany and (v) presence of dislocations. However, the pres-ence of compressive residual stress in the film enhancesits life by hardening the surface and controlling the crackgrowth.Although the relative diffracted intensities are notquantitative in GIXRD mode, these films can be inferredto have ideally random polycrystalline nature as learnedfrom the relative intensities comparison given in the Ta-ble.II with slight prominence to (111) and (200) orien-tations. The texture coefficient of observed peaks arecalculated using the γ ∗ hkl = I hkl / P I hkl , where I hkl is theintensity of the specific peak and P I hkl is sum of in-tensities of all observed TiN peaks. The atomic packing Miller Indices Intensity (JCPDS) Normalized Intensity d-reference (nm) d-measured (nm) ∆d% γ ∗ hkl (111) 72 84.3 0.24491 0.24214 1.13 30(200) 100 100.0 0.21207 0.20865 1.61 36(220) 45 56.6 0.14996 0.14750 1.64 20(311) 19 19.6 0.12789 0.12642 1.15 7(222) 12 15.0 0.12244 0.12107 1.12 5TABLE II. Relative intensity, d-spacing and texture coefficient of observed peaks I n t e n s it y ( a . u . ) ( ) ( ) ( ) ( )( ) FIG. 2. GIXRD profile of ARE-EB-PVD-TiN film (1,-1,-1)(1,-1,1) (-1,-1,-1)(-1,-1,1)
FIG. 3. (Color online) Pole figure of (111) plane with 25 ◦ off in the ψ with respect to surface normal and (111) planegenerated by CaRine density of (200) plane (4 atoms/ a ) is higher than (111)plane (2.3 atoms/ a ) and are higher than all other millerindex planes. In general, the (200) planes of rock salttype structure have the lowest surface energy whereas(111) planes have the lowest strain energy. When therate of deposition is low, the adatoms can get enoughtime to diffuse and react. In this process, they releaseself-reaction energy and reach the lowest surface energy.In reactive electron beam evaporation Ti ions are prefer-entially excited and the substrate bias leading to self-ion bombardment further assists atomic rearrangement.The films with slight preference to orientation of highatomic density planes will pose higher resistance againstoxidation. The texture analysis by LEPTOS software (-1,0,-1)(1,0,-1)(1,-1,0)(0,-1,1) (-1,0,1)(0,-1,-1)(-1,-1,0) FIG. 4. (Color online) Pole figure of (220) plane and (220)plane generated by CaRine has shown a clear preferential orientation of (111) and(220) planes. Fig.3 in particular shows a huge texturewith (111) planes inclined about 25 ◦ from the surfacenormal and its corresponding simulated pattern of (111).A similar preferred orientation was observed by some re-searchers, where the Ion beam assisted deposition andmagnetron sputtering has been used to get the same withan angle. Similarly, fig.4 shows the (220) oriented planes alsogrown at the expense of (200) planes in this system alongwith its simulated pattern of (220). In this depositionsystem a tube for N gas focused at an angle 25 ◦ to thesurface normal. Such configuration enhanced the forma-tion of (111) planes as well as the tilt of this plane tosurface normal along with the low rate of deposition.The classical sin ψ technique was used to determinethe internal stress present in the coatings, in whichan inter-planar spacing ”d” serves as an internal straingauge. The preferential orientation or huge texture ofthis TiN thin film makes the bi-axial residual stress mea-surement difficult. Very few ψ orientations will allow toget some diffraction signal. Therefore, the stress evalua-tion will be only based on a few measurement points andmakes the evaluated residual stress suspicious. Evalua-tion of residual stress in this case was based on the (422)reflection at around 2 θ = 125 ◦ . Fig.5 shows the intensitydistribution of the (422) Miller plane 2 θ vs ψ . Texture isclearly observed and hence the reduced number of acces-sible ψ positions for the stress evaluation.SIMS depth profile of this TiN film is shown in Fig.6.The secondary ion intensities (CsTi + and CsN + ) aremarkedly different due to dissimilar secondary ion sensi-tivity factors of the species. The quasi molecular ions of S c a n O r d e r FIG. 5. (Color online) Residual stress mapping Angle (2 θ ) vs ψ I n t e n s it y ( a . u ) FIG. 6. (Color online) Dynamic SIMS depth profile of TiNfilm the analyte species with cesium markedly reduce the ma-trix effect observed in the secondary ion intensity. The el-emental analyte intensities and its ratio ( I CsT i + / I CsN + )are constant up to a depth of 1.8 µ m (film thickness) sig-nifying the stoichiometric homogeneity across the depthof the film up to the substrate interface. The intensitiesof the profiles are matching with TiN standard. How-ever, we cannot calculate the composition quantitatively.Yet the relative intensities of the sample, the standard and their uniformity with respect to depth confirm thesample stoichiometry and its homogeneity. This is signif-icant advantage over many other coating processes wherea nitrogen gradient is generated which can lead to sub-stoichiometry and mechanical property gradations. Theintensity of oxygen and carbon impurities are very low inthese thin films. This stems from the localized and highpurity selective plasma used in this deposition.A widely used nanoindentation technique has beenused to determine the mechanical properties of e-beamgrown TiN thin films. These films have uniform surfacemorphology with RMS roughness of ≈ µ m. The lowsurface roughness is the resultant of low deposition ratewhich in turn ensured the stoichiometry throughout thefilm. A typical load - displacement curve obtained for thisTiN film is shown in fig.7. The nanoindentation hardnessusually depends on the surface roughness, chemical stateof surface layer and the size of the indenter. Here, rough-ness of this film is small enough and it is much lowerthan the depth penetrated by the indenter. The maxi-mum penetration depth was maintained well within the10% of the TiN film thickness. A typical value of inden-tation hardness and modulus is measured as 19.5 GPaand 214 GPa, respectively. L o a d ( m N ) FIG. 7. A typical nanoindentation profile of TiN film
Continuous Multi Cycle (CMC) indentation is a tech-nique that can provide the fatigue information of the thinfilms. The load dependent indentation behavior was in-vestigated through CMC testing on the TiN films coatedon micro slides (hardness ≈ The reloading pathsdo not necessarily overlap with the unloading path ofthe previous loading cycle resulting in hysteresis loops.The reason for this can be ascribed to the chemistryand microstructure of the material led to Bauschinger L o a d ( m N ) H a r dn e ss ( G P a ) FIG. 8. (Color online) CMC loading profile of TiN film effect. Figure 8 shows the continuous loading-unloadingindentation profiles with hardness corresponding to pen-etration depth. Initially, the hardness increases due tothe indentation contact in the elastic region and con-stant hardness range follows. This exhibit the inden-tation is in the plastic zone (indentation depth range,150 to 225 nm). The hardness should not vary withdepth beyond this range since these films are quite hard.But, at higher indentation depths the residual stress re-lieving makes the hardness to decrease rapidly. It isunderstood from the GIXRD profile systematic shift tohigher angle supports the existence of compressive resid-ual stress in this film. Usually residual stress is relievedthrough the formation of crack on the films. But, it isclear from the fig.8 that there is no pop-in or pop-out wasobserved in the typical indentation profile. This signifiesthat there is no fracture of films can be expected dur-ing this relaxation process. Usually, compressive resid-ual stress makes the films much harder than the stressrelieved films. IV. CONCLUSION
Stoichiometric single phase fcc TiN films have beensuccessfully synthesized after modification of the e-beamevaporation assembly using an anode plate in the vicin-ity of the heated titanium ingot to ignite dense plasma,which enhances the reactive deposition. This reactiongets activated further by substrate bias which causes theself-ion bombardment. This plasma enhances ion yieldand ion molecule interactions to deposit films at rela-tively lower temperature. We have obtained a high qual-ity polycrystalline TiN films of grain size ∼
150 nm.A N jet focusing induced the preferential orientationin the same direction. Also systematic lattice contrac-tion indicated the presence of compressive residual stresswhich is beneficial for wear resistance applications. Dy- namic SIMS analysis revealed compositional uniformityacross the depth and negligible low Z (O, C) impurities.Nano-indentation studies on these films resulted hardnessvalues close to 19.5 GPa. V. ACKNOWLEDGEMENT
F.J would like to acknowledge Mr.A.K.Balamuruganfor the discussion regarding the SIMS studies. We wouldlike to thank the UGC-DAE-CSR facility at Kalpakkamfor the GIXRD.
VI. REFERENCES DongHau Kuo and KwonWen Hung, ”Kinetics and microstruc-ture of TiN coatings by CVD”, Surf. Coat. Technol, , 150,(2001). J. Wagner, C. Motterer, M.Penoy, C. Michotte, W.Wallgramand M. Kathrein, ”The effect of deposition temperature on mi-crostructure and properties of thermal CVD TiN coatings”, Int.J. Refract. Met., , 120, (2008). C.-H.Ma, J.-H.Huang and Haydn Chen, ”Nanohardness ofnanocrysgalline TiN thin films”, Surf. Coat. Technol., , 3868,(2006) N. X. Randall and A. Harris,”Nanoindentation as a tool for char-acterising the mechanical properties of tribological transfer films”Wear, , 196, (2000). J.M. Lackner,W. Waldhauser, R.Berghauser, R.Ebner, B. Majorand T.Schoberl, ”Structural, mechanical and tribological investi-gations of pulsed laser deposited titanium nitride coatings”, ThinSolid Films, − , 195-202, (2004). Allan Mathews, ”Titanium nitride PVD coating Technology”,Surf. Eng., , 93, (1985). M. Yustea, R. Escobar Galindoa, S. Carvalhob, J.M. Albellaa, O.Sncheza, ”Improving the visible transmittance of low-e titaniumnitride based coatings for solar thermal applications”, Appl. Surf.Sci., , 1784, (2011). Elias Penilla, Francisco G. Prez-Gutirrez, Wyatt Duvall,Guillermo Aguilar, Junlan Wang, ”Optical properties of superstoichiometric
T iN x thin films”, Thin Solid Films, , 272,(2012). D. Deniz and J.M.E.Harper, ”Fiber textures of titanium nitrideand hafnium nitride thin films deposited by off-normal incidencemagnetron sputtering”, J. Appl. Phys., , 063519, (2008). G. Abadias and Y. Y. Tse, ”Diffraction stress analysis in fiber-textured TiN thin films grown by ion-beam sputtering: Applica-tion to (001) and mixed (001)+(111) texture”, J. Appl. Phys., , 2414, (2004). G. Hakansson, L Hultman, J. E. Sundgren, J. E. Greene andW. D. Munz, ”Microstructures of TiN films grown by variousphysial vapour deposition techniques”, Surf. Coat. Technol, ,51, (1991). H. Mori and H.Yoshihara, ”Synthesis of Titanium Nitride byenhanced ARE process ” Japan. J. Appl. Phys., , 575, (1978). Y.M.Chen, G..P.Yu, Y.J.Huang, ”Role of process parameters inthe texture evolution of TiN films deposited by hollow cathodedischarge ion plating”, Surf. Coat. Technol., , 156, (2001). Alicja Krella and Andrzej Czyzniewski, ”Cavitation erosion re-sistance of nanocrystalline TiN coating deposited on stainlesssteel”, Wear, , 963, (2008). S.Mahieu, P. Ghekiere, D. Depla, R. De Gryse, ”Biaxial align-ment in sputter deposited thin films”, Thin Solid Films, ,1229, (2006). A. P. Ehiasarian, A. Vetushka, Y. Aranda Gonzalvo, G. Sa fran,L. Szekely and P. B. Barna,”Influence of high power impulsemagnetron sputtering plasma ionization on the microstructureof TiN thin films”, J. Appl. Phys., , 104314, (2011). Feby Jose, R. Ramaseshan, S. Tripura Sundari, S. Dash, A. K.Tyagi, M. S. R. N. Kiran, and U. Ramamurty,”Nanomechanicaland optical properties of highly a-axis oriented AlN films”, Appl.Phys. Letts, , 254102, (2012). G. Abadias, Y. Y. Tse, and Ph. Gurin, V. Pelosin, ”Interdepen-dence between stress, preferred orientation and surface morphol-ogy of nanocrystalline TiN thin films deposited by dual ion beamsputtering”, J. Appl. Phys., , 113519, (2006). S.Mahieu, P.Ghekiere, G. De Winter, S. Heirwegh, D.Depla,R.De Gryse, O.I. Lebedev, G. Van Tendeloo,”Mechanism of pref-erential orientation in sputter deposited titanium nitride andyttria-stabilized zirconia layers”, J. Cryst. Growth, , 100,(2005). C-S. Shin, D.Gall, Y-W.Kim, N.Hellgren, I.Petrov and E.Greene,”Development of preferred orientation in polycrystalline NaCl-structure d-TaN layers grown by reactive magnetron sputtering:Role of low-energy ion surface interactions”, J. Appl. Phys., ,5084, (2003). Hongchen Wu and Andre Andres, ”Measurements of the asym-metric dynamic sheath around a pulse biased sphere immersed inflowing metal plasma”, Plasma Sources Sci. Technol. , 035030,(2008). W.C. Oliver and G.M. Pharr, ”Measurement of hardness andelastic modulus by instrumented indentation: Advances in un-derstanding and refinements to methodology”, J. Matter. Res., , 3, (2004). Powder diffraction file JCPDS No: 38-1420. XU Xuan-qian, YE Hui and ZOU Tong,”Characterization ofDC magnetron sputtering deposited thin films of TiN forSBN/MgO/TiN/Si structural waveguide”, JZUS A, ( ), 472,(2006). Dongdong Wang and Takeo Oki, ”The Morphology and Orien-tation of Cr-N films deposited by Reactive Ion Plating”, Thin Solid Films, , 219, (1990). Shiqin Xiao, Cristian P. Lungu and Osamu Takai, ”Comparisonof TiN deposition by rf magnetron sputtering and electron beamsustained arc ion plating”, Thin Solid Films, , 173, (1998). Jia-Hong Huang, Cheng-Hsin M, Haydn Chen, ”Effect of Ti in-terlayer on the residual stress and texture development of TiNthin films”, Surf. Coat. Technol, , 5937, (2006). M. Birkholz, C. Genzel and T. Jung, ”X-ray diffraction study onresidual stress and preferred orientation in thin titanium filmssubjected to a high ion flux during deposition”, J. Appl. Phys., , 7202, (2004). P. K. Ajikumar, Ananthi Sankaran, M. Kamruddin, R. Nithya, P.Shankar, S. Dash, A. K. Tyagi and Baldev Raj, ”Morphology andgrowth aspects of Cr(N) phases on gas nitridation of electroplatedchromium on AISI 316 LN stainless steel”, Surf. Coat. Technol, , 102, (2006). Feby Jose, R. Ramaseshan, A.K. Balamurugan, S. Dash, A.K.Tyagi and Baldev Raj, ”Continuous multi cycle nanoindentationstudies on compositionally graded
T i − x Al x N multilayer thinfilms”, Mater. Sci. Engg. A, , 6438, (2011). A.C.Fischer-Cripps, ”Critical review of analysis and interpreta-tion of nanoindentation test data”, Surf. Coat. Technol, ,4153, (2006). T. Saraswati, T.Sritharan, S. Mhaisalkar, C.D. Breach andF.Wulff, ”Cyclic loading as an extended nanoindentation tech-nique”, Mater. Sci. Engg. A, , 14, (2006). Bharat Bhushan and Xiaodong Li, ”Nanomechanical characteri-sation of solid surfaces and thin films”, Int. Mater. Rev., , 125,(2003). Trevor F. Page and Sarah V. Hainsworth,”Using nanoindentationtechniques for the characterization of coated systems: a critique”,Surf. Coat. Technol. , 201, (1993). Y-H.Lee, K.Takashima and D.Kwon, ”Micromechanical analysison residual stress-induced nanoindentation depth shifts in DLCfilms”, Scr. Mater.,50