Full Paper
`
`Titanium Nitride Grown by Sputtering for
`Contacts on Boron-Doped Diamond
`
`Vincent Mortet,* Omar Elmazria, Wim Deferme, Michael Daenen,
`Jan D’Haen, Andrada Lazea, Anne Morel, Ken Haenen, Marc D’Olieslaeger
`
`Due to its exceptional properties, semiconducting diamond is expected to be used for
`electrically active devices which can be operated in harsh environments. Such devices need
`reliable ohmic contacts that can also stand hostile environments. Titanium nitride (TiN) is a
`chemically stable material with good electrical conductivity. In this work, TiN contacts on
`boron-doped diamond have been made and characterised. TiN films were deposited by
`reactive magnetron sputtering. Boron-doped diamond layers were deposited by plasma
`enhanced chemical vapour deposition. Optimal deposition conditions have been determined
`to obtain TiN films with low resistivity (100 mV cm), high reflectance in the IR region and
`low stress. TiN contacts show ohmic behaviour after annealing at 750 8C.
`
`Introduction
`
`The exceptional properties of diamond and the possibility
`to obtain diamond films at low pressure on different types
`of substrates, make this material a good candidate for a
`large number of novel applications. Diamond also offers
`the possibility to fabricate electrically active devices which
`can be operated at elevated temperatures,
`in hostile
`environments. Diamond can also be used in biomedical
`applications and it fulfils the main requisites for use in
`human implants due to its biocompatibility and its
`chemical stability.
`
`V. Mortet, W. Deferme, M. Daenen, J. D’Haen, A. Lazea, K. Haenen,
`M. D’Olieslaeger
`Institute for Materials Research, Hasselt University,
`Wetenschapspark 1, 3590, Diepenbeek, Belgium
`Fax. þ32 (0)11 26 88 99; E-mail: vincent.mortet@uhasselt.be
`V. Mortet, K. Haenen, J. D’Haen, A. Lazea, M. D’Olieslaeger
`IMEC, Division IMOMEC, Wetenschapspark 1, 3590, Diepenbeek,
`Belgium
`O. Elmazria
`LPMIA, Universite´ H. Poincare´, Nancy I, F-54506, Vandoeuvre-
`les-Nancy Cedex, France
`A. Morel
`ENSAM, boulevard du Ronceray, B.P. 3525, F-49035 Angers Cedex ,
`France
`
`is a hard, dense, refractory
`Titanium nitride (TiN)
`material with high electrical conductivity. TiN has good
`optical properties,
`including an attractive gold-tinged
`appearance when pure, and it has a high reflectance in
`the IR range. TiN is thermodynamically stable in air upto
`600 8C and it is inert to corrosive media. Like diamond, TiN
`is a non-toxic and biocompatible material. TiN meets the
`food and drug administration (FDA) guidelines and it has
`been approved for use in numerous medical/surgical
`devices, including implants. TiN is also widely employed in
`semiconductor manufacturing as a ‘diffusion barrier’ layer
`and it has already been used as ohmic contact on GaN and
`SiC.[1,2]
`The use of TiN in combination with diamond is
`attractive for the construction of ohmic contacts, operating
`either at elevated temperatures, in hostile chemical and
`radiation environments or in biological environment.
`In this paper, an optimisation of TiN thin films grown by
`reactive DC-pulsed magnetron sputtering is reported.
`Structural, mechanical, optical and electrical properties
`of TiN films have been measured. Homoepitaxial bor-
`on-doped diamond layers were grown by plasma
`enhanced chemical vapor deposition and they were
`characterised by Fourier transform photocurrent spectros-
`copy (FTPS).[3] Finally, TiN contacts on boron-doped
`diamond have been made and characterised.
`
`Plasma Process. Polym. 2007, 4, S139–S143
`ß 2007 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
`
`DOI: 10.1002/ppap.200730506
`
`S139
`
`

`
`V. Mortet et al.
`
`Experimental Part
`
`The TiN layers were deposited on silicon (100) and fused-silica
`substrates. The substrates were ultrasonically cleaned in trichloro-
`ethylene, acetone and alcohol and they were dried with pure
`nitrogen. The titanium target (10 cm in diameter, 99.99% purity),
`which is held on the water-cooled magnetron cathode, was
`sputtered in a mixture of argon and nitrogen. The argon and
`nitrogen flow rates (FAr and FN, respectively) were controlled by
`two mass flow meters and the total gas flow rate was kept
`constant (50 sccm). The total pressure was controlled with a
`throttling valve situated in front of the turbo-molecular drag
`pump. The target power supply was driven in constant-power
`mode at 250 kHz pulse frequency and 1 600 ns pulse width. The
`distance between the target and the substrate holder can be
`adjusted. The substrates were not heated and their temperature
`was only dependent on the plasma heating.
`Before deposition, the sputtering chamber was evacuated to a
`pressure below 2 10
`6 mbar. The target was cleaned in an argon
`discharge for 10 min and it was pre-sputtered in the same condi-
`tions as the film deposition conditions for additional 10 min.
`During these steps, the substrates are shielded from deposition by
`a shutter. TiN films were deposited under various target power (P),
`total pressure (Pt), target to substrate distance (d) and nitrogen
`ratio (RN¼ FN/(FArþFN)). The investigated deposition conditions
`are summarised in Table 1.
`TiN films were characterised by X-ray diffraction in u–2u scan
`mode with Cu Ka1 radiation and scanning electron microscopy
`(SEM). The films thicknesses were measured by SEM cross-section
`observation. The mechanical stress was calculated from substrate
`curvature measurements using the Stoney formula.[4] The sub-
`strates curvatures were measured by a Dektak3ST profilometer.
`Electrical characterisations of the films were performed using a
`four-points probe. Finally, optical reflectance of TiN films was also
`measured in near IR, visible and ultraviolet range (NIR-Vis-UV).
`Boron-doped diamond layers were grown on (100)
`Ib
`2.5 2.5 0.5 mm3 single crystal diamond samples by plasma
`enhanced chemical vapour deposition (PECVD) in a homemade
`NIRIM type reactor.[5] Before deposition, the vacuum chamber is
`6 mbar with a
`evacuated to a base pressure lower than 10
`turbo-molecular pump. Boron doping is achieved using trimethyl-
`boron (TMB) diluted in hydrogen (200 ppm). The thickness of the
`boron-doped layers was calculated from the mass measurement
`assuming that the density of the epilayer is 3.52. Resistivity of the
`p-type layers was measured using the Van der Pauw resistivity
`measurement method. The incorporation of substitutional boron
`
`in the diamond layer was confirmed using FTPS at liquid
`nitrogen temperature. Ohmic titanium/aluminum inter-digitated
`electrodes with a spacial period of 400 mm were obtained by
`lift-off. Electrical characterisation of the diamond layer and TiN
`contacts were made using circular transmission line model
`measurements (c-TLM).[6]
`
`Results and Discussion
`
`TiN Growth
`
`First, TiN films were deposited at different nitrogen
`low pressure (with d¼ 5 cm, P¼
`concentrations at
`450 W). Films deposited at RN¼ 100% are grey, films
`deposited at lower nitrogen concentration have a copper-
`like colour, whereas films deposited at RN¼ 5% are golden
`which is a particularity of stoichiometric TiN. X-ray dif-
`fraction patterns of these films are reported in Figure 1. All
`films deposited at a nitrogen concentration higher than 30%
`do not exhibit any diffraction peak. At lower nitrogen
`concentration, X-ray diffraction patterns show a peak at 2u
`36.68 related to the (111) TiN peak. The most intense peak
`was obtained at RN¼ 5% and it is shifted to lower 2u. This is
`probably due to a high stress in the layer as one can see from
`the inset of Figure 1 where the variation of the films stress as
`a function of the nitrogen concentration is represented. The
`stress is compressive and maximum (1.3 GPa) at RN¼ 5%.
`Second, TiN films were deposited at different target’s
`powers and different total pressures at RN¼ 5%. All these
`films show the X-ray diffraction peak of (111) TiN. The
`optimal target power has been found to be 450 W. As one
`can see on Figure 2(a), the stress of the layer is high
`(2 Gpa) at low pressure whatever the target power is, and
`it decreases to nearly no stress for pressures higher than
`(20–30) 10
`3 mbar. We have observed that films depo-
`sited at higher pressure and without stress, are slightly
`less shiny than the films deposited at lower pressure.
`Figure 2(b) shows the resistivity of the TiN films as a func-
`tion of the total pressure and the target power. The resisti-
`vity of the films is low (100 mV cm) at low pressure and
`rises at the threshold pressure of (20–30) 10
`3 mbar.
`Figure 3 shows the reflectivity spectra of TiN films deposited
`
`Table 1. Summary of the investigated deposition conditions and the optimal deposition conditions.
`
`Studied range
`
`Optimal conditions
`
`Base pressure
`
`Target
`
`Gases
`
`Target-to-substrate distance
`
`Target power
`
`Total pressure
`
`S6 mbar
`2 T 10
`Ti (99.99% pure),Ø 10 cm
`
`Argon - Nitrogen
`
`5–13 cm
`
`375–750 W
`S3 mbar
`(4.7–49) T 10
`
`S140
`
`Plasma Process. Polym. 2007, 4, S139–S143
`ß 2007 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
`
`–
`
`–
`
`5% nitrogen in Argon
`
`5 cm
`
`375–450 W
`
`
`(20–30) T 10S3 mbar
`
`DOI: 10.1002/ppap.200730506
`
`

`
` 4.7 ubar
` 16 ubar
` 23 ubar
` 29 ubar
` 36 ubar
` 39 ubar
` 49 ubar
`
`Titanium Nitride Grown by Sputtering for Contacts on . . .
`
`P=450 W, Pt= 5 µbar, d=5 cm
`
`100
`
`Visible
`
`75
`
`50
`
`25
`
`Reflectance (%)
`
`0
`
`20
`40
`60
`80
`N2 concentration (%)
`
`100
`
`1.4
`
`1.2
`
`1.0
`
`0.8
`
`0.6
`
`0.4
`
`0.2
`
`Stress (GPa)
`
`5%
`
`15%
`
`30%
`
`(111) TiN
`
`Si substrate artefact
`
`600
`
`500
`
`400
`
`300
`
`200
`
`100
`
`0
`
`Intensity (cps)
`
`50
`
`60
`
`20
`
`30
`
`40
`2θ (°)
`Figure 1. X-ray diffraction pattern of TiN films obtained at differ-
`ent nitrogen concentrations. Inset: variation of the mechanical
`stress of TiN as a function of the nitrogen concentration in the
`discharge gas.
`
`0
`
`0
`
`500
`
`1000
`1500
`2000
`Wavelenght (nm)
`Figure 3. Variation of the reflectance in UV-visible-NIR range of
`TiN films grown at different total pressures of the deposition
`chamber.
`
`2500
`
`at different pressures. The reflectivity spectra are character-
`istic of a free-electron system in a metal with a reflectivity
`edge at400 nm, due to a screened plasma resonance.[7] One
`can see that the reflectivity in the IR region of TiN films
`deposited at low pressure is higher than those deposited at
`high pressure whatever the target power. The lower
`reflectivity might be due to either surface scattering due
`to the films roughness or different electrical properties of the
`films. TiN films with the lowest resistivity are obtained at
`target powers between 375 and 450 W. Figure 4 show the
`typical SEM pictures of a film deposited at low pressure and
`a film deposited at high pressure. Films deposited at low
`pressure are smooth with very fine grains, whereas films
`deposited at high pressure are rough with large grains.
`
`Properties of films deposited at low pressure and dif-
`ferent target-to-substrate distances (from 5 to 13 cm) have
`been investigated. No significant effects of the target-
`to-substrate distance on the stress, the morphology and
`the reflectivity of the layers have been observed, however
`the distance increase does decrease the crystalline quality
`and the deposition rate.
`In the optimal deposition conditions (P¼ 375–450 W,
`Pt¼ 20 mbar, RN¼ 5%, d¼ 5 cm),
`low stress and high
`reflectivity in the IR range, low resistivity TiN films with a
`(111) crystalline orientation are obtained. The chemical
`stability of TiN films has been tested. TiN films were let in
`Aqua Regia and in an Al etchant solution (H3PO4/HNO3/
`H2O at 60:7:10). TiN films are slowly etched in Aqua Regia
`
`Resistivity (uΩ cm)
`
`600
`
`500
`
`400
`
`300
`
`200
`
`00
`
`RN=5 %, d=5 cm
`
` 375W
` 450W
` 600W
` 750W
`
`RN=5 %, d=5 cm
`
`(b)
`
` 375W
` 450W
` 600W
` 750W
`
`(a)
`
`2.0
`
`1.5
`
`1.0
`
`0.5
`
`0.0
`
`Stress (GPa)
`
`01
`
`50
`
`0
`
`10
`
`40
`
`50
`
`0
`
`10
`
`40
`
`20
`30
`30
`20
`Pressure (µbar)
`Pressure (µbar)
`Figure 2. Variation of the mechanical stress (a) and the resistivity (b) as a function of the total pressure in the deposition chamber.
`
`Plasma Process. Polym. 2007, 4, S139–S143
`ß 2007 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
`
`www.plasma-polymers.org
`
`S141
`
`

`
`V. Mortet et al.
`
`come from the quality of the samples
`surface before deposition and/or the
`cleaning process before deposition.
`Figure 5 shows the FTPS spectra of a
`boron-doped diamond layer obtained
`using a B/C ratio of 4 ppm. The
`spectrum exhibits a clear photoionisa-
`tion threshold at 0.37 eV and a peak at
`0.347 eV. The photocurrent
`signal
`shows two series of equidistant minima
`starting at 0.30 and 0.35 eV with a period
`of 165 meV. All these results are a clear
`signature of boron incorporation in
`the diamond layer.[8] The resistivity of
`the diamond layer decreases from 14 to
`0.8 V cm as the TMB concentration
`increases.
`Electrical characterisation of TiN contacts on boron-
`doped diamond layer were made using circular transmis-
`sion line model measurements. The circular contacts were
`obtained by lift-off. Prior to TiN deposition, the surface of
`the doped diamond layers were oxidised in hot H2SO4 and
`KNO3 solution to remove non-diamond carbon. Figure 6
`shows the I–V curves of a boron-doped diamond layer (18
`ppm) with as deposited and annealed (at 450 and 750 8C)
`TiN contacts. As deposited and 450 8C annealed contacts
`are highly resistive, while contacts annealed at 750 8C
`show ohmic contact behaviour. The specific contact
`resistance (rc) could only be determined for the diamond
`layers grown with a B/C 26 ppm: rc 10
`2 V cm2. This
`result is one order of magnitude higher than Ti/Pt/Au
`contacts.[9] This might be due to a low dopant concentra-
`tion[9] or/and a different reactivity of TiN and Ti to form
`ohmic contact.
`
`Figure 4. SEM images of TiN films’ surface deposited at low and high pressure (P¼ 450 W,
`RN¼ 5% and d¼ 5 cm).
`
`(10 nm/h) and they are not etched in the Al etching
`solution while they are etched in hot H2O2.
`
`Homoepitaxial Boron-Doped Diamond Growth and
`Electrical Characterisation
`
`Doped diamond layers were grown in a mixture of 1% of
`methane diluted in hydrogen at a total pressure of
`110 mbar, a microwave power of 500 W and a substrate
`temperature of 1 100 8C. The B/C ratio in the plasma was
`adjusted from 4 to 32 ppm. During deposition, the total
`gases mass flow rate was kept constant at 500 sccm. The
`layers morphology has been observed by optical micros-
`copy. The diamond layers show pits and non-epitaxial
`crystallites on their surfaces. The number of those defects
`varies a lot from one sample to another. Those defects can
`
`Not annealed &
`Annealed @ 450°C
`
`Annealed @ 750°C
`
`10
`
`02468
`
`-2
`
`-4
`
`-6
`
`Voltage (V)
`
`10-1
`
`10-2
`
`10-3
`
`Photocurrent (a.u.)
`
`10-4
`0.3
`
`0.4
`0.5
`0.6
`0.7
`Photon energy (eV)
`Figure 5. FTPS spectra of boron-doped diamond layer (B/C¼
`4 ppm).
`
`0.8 0.9
`
`1
`
`-8
`
`-1.0x10-3
`
`0.0
`5.0x10-4
`-5.0x10-4
`Current (A)
`I–V characteristics of TiN contacts on boron-doped
`Figure 6.
`diamond layer (B/C¼ 18 ppm) after different annealing treat-
`ments.
`
`1.0x10-3
`
`S142
`
`Plasma Process. Polym. 2007, 4, S139–S143
`ß 2007 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
`
`DOI: 10.1002/ppap.200730506
`
`

`
`Titanium Nitride Grown by Sputtering for Contacts on . . .
`
`Conclusion
`
`TiN films deposited by reactive magnetron sputtering have
`been studied and characterised. Smooth and conductive
`TiN films with low stress were obtained in optimal
`deposition conditions.
`It has been observed that the
`properties of the TiN films are strongly dependent on the
`nitrogen concentration and the total pressure. TiN contacts
`were deposited using photolithography and lift-off tech-
`niques onto boron-doped diamond obtained by PECVD.
`Experimental results show that ohmic contacts can be
`formed after annealing at temperatures 750 8C. The first
`results show that TiN conctacts can be formed on
`boron-doped diamond and they can be used for electronic
`applications in harsh environments or
`in biological
`environments on a p-type diamond semiconductor.
`
`Acknowledgements: This work was supported by the IWT-
`SBO-project no. 030219 ‘CVD Diamond, a novel multifunctional
`material for high temperature electronics, high power/high
`frequency electronics and bioelectronics’ and the EU FP6 Marie
`
`Curie RTN ‘‘DRIVE’’, MRTN-CT-2004-512224. K.H. is a Postdoctoral
`Fellow of the Research Foundation – Flanders (FWO-Vlaanderen).
`
`Received: September 19, 2006; Revised: November 8, 2006;
`Accepted: November 23, 2006; DOI: 10.1002/ppap.200730506
`
`Keywords: films; titanium nitride; characterization; sputtering;
`X-ray
`
`[1] C. A. Dimitriadis, Th. Karakostas, S. Logothetidis, G. Kamarinos,
`J. Brini, G. Nouet, Solid-State Electron. 1999, 43, 1969.
`[2] B. Pe´cz, Appl. Surf. Sci. 2001, 184, 287.
`[3] M. Vanecek, A. Poruba, Appl. Phys. Lett. 2002, 80, 719.
`[4] G. G. Stoney, Proc. R. Soc. 1909, A82, 172.
`[5] Koji Kobasi, ‘‘Diamond Films – Chemical Vapor Deposition for
`Oriented and Heteroepitaxial Growth’’, Elsevier, Amsterdam
`2005,p. 17.
`[6] G. K. Reeves, Solid-State Electron. 1980, 23, 487.
`[7] A. Tarniowy, R. Mania, M. Rekas, Thin solid Films 1997, 311, 93.
`[8] R. Kravets, Ph.D. Thesis, Czech Technical University, Prague,
`2005.
`[9] Y. G. Chen, M. Ogura, S. Yamasaki, H. Okushi, Diamond Relat.
`Mater. 2004, 13, 2121.
`
`Plasma Process. Polym. 2007, 4, S139–S143
`ß 2007 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
`
`www.plasma-polymers.org
`
`S143