- Email: [email protected]
Controlled thermal oxidation of nanostructured vanadium thin films
Author’s Accepted Manuscript Controlled thermal oxidation of nanostructured vanadium thin films Paulo Pedrosa, Nicolas Martin, Roland Salut, Mohammad Arab Pour Yazdi, Alain Billard www.elsevier.com
PII: DOI: Reference:
S0167-577X(16)30418-9 http://dx.doi.org/10.1016/j.matlet.2016.03.097 MLBLUE20548
To appear in: Materials Letters Received date: 23 February 2016 Revised date: 18 March 2016 Accepted date: 18 March 2016 Cite this article as: Paulo Pedrosa, Nicolas Martin, Roland Salut, Mohammad Arab Pour Yazdi and Alain Billard, Controlled thermal oxidation of nanostructured vanadium thin films, Materials Letters, http://dx.doi.org/10.1016/j.matlet.2016.03.097 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting galley proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Controlled thermal oxidation of nanostructured vanadium thin films
Paulo Pedrosa*, Nicolas Martin, Roland Salut, Mohammad Arab Pour Yazdi, Alain Billard a
Institut FEMTO-ST, UMR 6174, Université de Bourgogne Franche-Comté, CNRS, 15B, Avenue des
Montboucons, 25030 Besançon Cedex, France
ABSTRACT Pure V thin films were dc sputtered with different pressures (0.4 and 0.6 Pa) and particle incident angles α of 0°, 20° and 85°, by using the GLancing Angle Deposition (GLAD) technique. The sputtered films were characterized regarding their electrical resistivity behaviour in atmospheric pressure and invacuum conditions as a function of temperature (40-550 °C), in order to control the oxidation process. Aiming at comprehending the oxidation behaviour of the samples, extensive morphological and structural studies were performed on the as-deposited and annealed samples. Main results show that, in opposition to annealing in air, the columnar nanostructures are preserved in vacuum conditions, keeping metallic-like electrical properties. Keywords: GLancing Angle Deposition, vanadium oxide, in-vacuum resistivity.
1. INTRODUCTION Vanadium oxides (VOx) such as VO2 and V2O5 in the form of thin films display interesting optical and electrical properties that make them suitable for a wide range of applications [1–3]. The reversible first-order semiconductor-to-metal transition exhibited by the VO2 [2,4–6] and V2O5 phases [1,7] makes them especially interesting for gas sensing purposes [3,8,9]. Furthermore, the sensitivity of these VOx phases are commonly attributed to the presence of oxygen vacancies [2,3,6–8]. Consequently, two main strategies are currently being explored to obtain the thin film materials either by (i) ab initio sputtering of the referred oxides [5,10] or by (ii) subsequent thermal oxidation of pure V films [4,11]. Also, the combination of both strategies is often used [2,6,12]. *
Corresponding author: P. Pedrosa; email address: [email protected]; Institut FEMTO-ST, UMR
6174, Université de Bourgogne Franche-Comté, CNRS, 15B, Avenue des Montboucons, 25030 Besançon Cedex, France.
1
In this work, the authors sputtered pure V thin films using the GLAD technique. It is known that porous nanostructured films display interesting electrical and optical properties, thus making them ideal candidates for sensing applications [13,14]. Hence, the effects of controlled thermal oxidation on the resistivity behaviour of the samples sputtered with normal and inclined columnar structures were investigated in atmospheric pressure and in-vacuum conditions. The aim of the present letter is to assess how the combination of nanostructured features and a controlled formation of the VOx phases contribute to the enhancement of the electrical behaviour of the films.
2. EXPERIMENTAL DETAILS Pure V films were deposited by dc magnetron sputtering inside a 40 L stainless-steel custom-made vacuum chamber. The reactor was equipped with a circular planar and water-cooled magnetron sputtering source, which was evacuated with a turbomolecular pump, backed by a mechanical one, in order to obtain a base pressure of 10-6 Pa. A vanadium target (purity 99.6 at.%, 75 mm diameter) was used. This target was dc sputtered in a pure argon atmosphere, using a constant current I = 200 mA. The incident angle of the particle flux was changed by tilting the substrate holder with an angle taken from the substrate normal: = 0°, 20° and 85° (GLAD technique). The argon flow rate was kept at 5.7 sccm ( = 0° sample) and 2.1 sccm ( = 20° and 85° samples), corresponding to a partial pressure of 6×10-1 Pa (S = 16 L.s-1) and 4×10-1 Pa (S = 8.9 L.s-1), respectively. The substrates, introduced through a 1 L airlock, were glass microscope slides (ISO norm 8037-1, with roughness better than 0.5 nm) and (100) silicon wafers (p-type, ρ = 1–30 Ω.cm). Before each run, all substrates were cleaned with acetone and alcohol, and the target was pre-sputtered in a pure argon atmosphere for 5 min., in order to remove the target surface contamination layer. The target-to-substrate distance was kept at 65 mm in all runs. The substrates were grounded and all depositions were carried out at room temperature. The deposition time was adjusted in order to obtain a thickness close to 500 nm. The thickness of the coatings was assessed using a Tencor Alpha Step IQ profilometer, while the morphological features of the samples were probed by scanning electron microscopy (SEM) at 5 keV in a Dual Beam SEM/FIB FEI Helios 600i microscope. The coatings were also characterized by X-ray diffraction (XRD). Measurements were carried out using a Bruker D8 focus diffractometer with a cobalt X-ray tube (Co λKα =1.78897 Å) in a θ/2θ configuration. Scans were performed with a step of 0.02° per
2
0.2 s and a 2θ angle ranging from 20 to 80°. The resistivity measurements were performed using the fourprobe van der Pauw method in the temperature range of 40–550 °C (1st cycle of 40-100-40 °C with a ramp of 2 °C.min-1 followed by 10 similar cycles with 50 °C increments each until a maximum temperature of 550 °C is reached), for both in-vacuum (10-5 Pa) and atmospheric pressure conditions. The atmospheric pressure measurements were done in a custom-made chamber, which is covered in order to have a dark environment; humidity and cleanness were considered as constant. The error associated to all resistivity measurements was always below 1% and the attachment of the contacts was checked prior to every run (I/V correlation close to 1) to ensure that an ohmic contact was attained (use of gold coated tips). All films were characterized in as-deposited and annealed (after atm. pressure and in-vacuum resistivity tests) conditions.
3. RESULTS AND DISCUSSION The resistivity (ρ) behaviour of the normal incidence and GLAD sputtered films with increasing annealing temperature is shown in Fig. 1. Major differences can be seen between atmospheric pressure and vacuum annealing conditions.
Figure 1. Atmospheric pressure (a1-c1) and in-vacuum (a2-c2) resistivity measurements for the V thin films deposited with particle incidence angles of 0°, 20° and 85°. Coloured arrows represent the resistivity evolution with increasing cycle number.
3
The normal incidence sample (V0°) displays a three-step resistivity behaviour (evidenced by the coloured arrows) when the annealing is performed at atmospheric pressure. Firstly, an increase of the values is evident until 300 °C, followed by a decrease of ρ until 450 °C. Finally, a subsequent increase in the final temperature cycle is observed. In opposition, a two-step behaviour is evidenced when the annealing is performed in vacuum conditions. No resistivity increase (oxidation) from 450 °C to 550 °C occurs due to the low oxygen content in the system. The same trend is observed for the sample deposited with α = 85° (V85°). More importantly, the steeper ρ increase (phase transition) occurring at 68 °C is indicative of VO2 formation [15,16]. On the other hand, the sample deposited with α = 20° (V20°) exhibits only a slight increase of resistivity in the whole temperature range. The overall resistivity behaviour of the samples can be explained by a closer analysis to the morphological and structural features of the samples, Figs. 2 and 3, respectively.
Figure 2. SEM cross-section micrographs of the produced samples before and after annealing. The higher sputtering pressure used in the production of the V0° film leads to a more porous morphology than that of V20°. Furthermore, α = 20° is not sufficient to produce the desired shadowing effect, thus no inclined columns can be seen. Hence, due to its denser columnar features, only V surface oxidation occurs when the V20° sample is annealed in vacuum. This is confirmed by both the very small overall increase of ρ and the presence of a shifted bcc V peak, indicative of a polycrystalline V matrix with oxygen insertions in the structure. Hence, a slightly surface-oxidised and polycrystalline V20° film
4
is obtained after in-vacuum annealing. Following the annealing in air, the V20° film displays a slightly higher ρ increase. Moreover, the columnar features are destroyed due to extensive formation of several VOx phases. Also, the film thickness increases (Fig. 2) due to significant surface oxidation, leading to a noisier ρ signal. No crystalline V phase is observable, thus a severely oxidised and amorphous V phase is obtained in atmospheric pressure conditions. The higher ρ increase (from 6.5×10-7 to 2.2×10-6 Ω.m) and Temperature Coefficient of Resistivity (TCR) decrease (from 1.2×10-3 to -4.9×10-5 K-1), in contrast with lower ρ increase (from 6.0×10-7 to 1.0×10-6 Ω.m) and almost constant TCR (5.4 to 8.6×10-4) in vacuum, can be attributed to the combined effects of V amorphization and VOx formation in air annealing. The amorphous VO2 phase formed is insufficient to promote a steep ρ increase.
Figure 3. XRD diffractograms of the sputtered samples before and after annealing. The V0° and V85° films display a similar ρ evolution, confirming their porous morphology. However, the V85° film displays higher ρ values due to its highly porous morphology. Focusing on the nanostructured V85° film, an increase of ρ (from 7.2×10-6 to 1.7×10-5 Ωm) and decrease of the TCR
5
(from 1.1 to -1.7×10-3 K-1) occurs from RT to 300 °C at atmospheric pressure. This indicates a strong oxidation of the film, not only at the surface but also in-depth due to its higher porosity. Afterwards, a semiconductor-like behaviour (decrease of ρ from 1.5×10-5 to 8.9×10-6 Ωm) is visible from 350-450 °C, which may be due to increased crystallisation of the VOx phases formed. Note that no ρ decrease is evidenced in the amorphous V20° film in air. In the last two cycles (450-550 °C), a strong oxidation of the V85° film occurs with ρ increasing one order of magnitude from 1.0×10-5 to 1.3×10-4 Ωm. Around 68 °C, a steeper increase of ρ is noticeable in both V0° and V85° films due to the formation of the VO2 phase. However, this phase may not be pure VO2 since the ρ increase is too low (< 1 order of magnitude). In vacuum conditions the ρ increase in the last temperature cycles is not achieved in both V0° and V85° films. From RT to 350 °C ρ increases from 5.6 to 8.7 ×10-6 Ωm and the TCR (4.7×10-4 K-1) gradually tends to zero due to the decrease of the crystal size of the bcc V phase. Afterwards, ρ decreases until 4.9×10-6 Ωm and the TCR becomes negative (-5.8×10-4 K-1) from 400-550 °C. This is due to the formation of conductive V7O13 and VO0.50 oxide phases in the porous columnar features. Finally, it is possible to say that porous nanostructures may be more easily tailored in terms of their electrical behaviour through controlled oxidation than denser ones. Also, in-vacuum annealing is an effective way to oxidise the pure V films without destroying their nanostructured columnar features. However, no VO2 phase was formed due to the low amount of available oxygen. Hence, further studies on ab initio GLAD nanostructured VOx films is currently underway.
4. CONCLUSIONS The present letter focuses on the study of electrical resistivity vs. temperature cycle of nanostructured vanadium films to be used as gas sensors. The pure vanadium films were nanostructured using the GLAD process and sputtered with incident angles α = 0°, 20° and 85°. In order to promote the formation of the VO2 phase the as-deposited films were annealed in atmospheric pressure and in-vacuum conditions. The controlled oxidation of the films was monitored by in-situ resistivity measurements with the films being morphologically and structurally characterized before and after annealing. The denser V20° films display a steady ρ increase both in air and vacuum conditions. No crystalline VO2 phase is formed, thus no steep ρ increase is present in the last temperature cycle. In opposition, the porous V0° and V85° films exhibit a two-step and three-step behaviour when annealed in vacuum or in air,
6
respectively. No evidences of VO2 formation is found when the annealing is performed in vacuum due to the low available oxygen amount. However, small traces of crystalline VO2 are present when these samples are annealed at the atmospheric pressure, hence the visible steeper ρ increase (although below one order of magnitude) at ~68 °C. The GLAD porous films are more easily tailorable in terms of the envisaged ρ behaviour. Moreover, in-vacuum annealing is effective in tuning the ρ evolution without destroying the nanostructured columns, despite no VO2 phase being attained.
ACKNOWLEDGEMENTS This work was supported by the region of Franche-Comté in cooperation with the Labex ACTION program (contract ANR-11-LABX-01-01). The authors acknowledge the MIMENTO Platform of the RENATECH network.
BIBLIOGRAPHY [1] S. Raja, G. Subramani, D. Bheeman, R. Rajamani, C. Bellan, Structural and optical properties of vacuum evaporated V2O5 thin films, Opt. - Int. J. Light Electron Opt. 127 (2016) 461–464. doi:10.1016/j.ijleo.2015.08.045. [2] Y.-K. Dou, J.-B. Li, M.-S. Cao, D.-Z. Su, F. Rehman, J.-S. Zhang, et al., Oxidizing annealing effects on
VO2
films
with
different
microstructures,
Appl.
Surf.
Sci.
345
(2015)
232–237.
doi:10.1016/j.apsusc.2015.03.044. [3] K. Schneider, M. Lubecka, A. Czapla, VOx Thin Films for Gas Sensor Applications, Procedia Eng. 120 (2015) 1153–1157. doi:10.1016/j.proeng.2015.08.1009. [4] Y. Ji, G.A. Niklasson, C.G. Granqvist, M. Boman, Thermochromic VO 2 films by thermal oxidation of
vanadium
in
SO2,
Sol.
Energy
Mater.
Sol.
Cells.
144
(2016)
713–716.
doi:10.1016/j.solmat.2015.10.012. [5] D. Zhang, M. Zhu, Y. Liu, K. Yang, G. Liang, Z. Zheng, et al., High performance VO 2 thin films growth by DC magnetron sputtering at low temperature for smart energy efficient window application, J. Alloys Compd. 659 (2016) 198–202. doi:10.1016/j.jallcom.2015.11.047. [6] J.Y. Suh, R. Lopez, L.C. Feldman, R.F. Haglund, Semiconductor to metal phase transition in the nucleation and growth of VO2 nanoparticles and thin films, J. Appl. Phys. 96 (2004) 1209. doi:10.1063/1.1762995. [7] R. Mustafa Öksüzoğlu, P. Bilgiç, M. Yıldırım, O. Deniz, Influence of post-annealing on electrical, structural and optical properties of vanadium oxide thin films, Opt. Laser Technol. 48 (2013) 102–109. doi:10.1016/j.optlastec.2012.10.001.
7
[8] A. Simo, B. Mwakikunga, B.T. Sone, B. Julies, R. Madjoe, M. Maaza, VO 2 nanostructures based chemiresistors for low power energy consumption hydrogen sensing, Int. J. Hydrogen Energy. 39 (2014) 8147–8157. doi:10.1016/j.ijhydene.2014.03.037. [9] J. Huotari, R. Bjorklund, J. Lappalainen, A. Lloyd Spetz, Pulsed Laser Deposited Nanostructured Vanadium Oxide Thin Films Characterized as Ammonia Sensors, Sensors Actuators B Chem. 217 (2015) 22–29. doi:10.1016/j.snb.2015.02.089. [10] X. Wei, S. Li, J. Gou, X. Dong, X. Yang, W. Li, et al., Preparation and characteristics of vanadium oxide thin films by controlling the sputtering voltage, Opt. Mater. (Amst). 36 (2014) 1419–1423. doi:10.1016/j.optmat.2014.02.021. [11] T.C.-K. Yang, Y.-L. Yang, R.-C. Juang, T.-W. Chiu, C.-C. Chen, The novel preparation method of high-performance thermochromic vanadium dioxide thin films by thermal oxidation of vanadiumstainless steel co-sputtered films, Vacuum. 121 (2015) 310–316. doi:10.1016/j.vacuum.2015.05.001. [12] Y.-Y. Su, X.-W. Cheng, J.-B. Li, Y.-K. Dou, F. Rehman, D.-Z. Su, et al., Evolution of microstructure in vanadium oxide bolometer film during annealing process, Appl. Surf. Sci. 357 (2015) 887–891. doi:10.1016/j.apsusc.2015.09.068. [13] A. Barranco, A. Borras, A.R. Gonzalez-Elipe, A. Palmero, Perspectives on oblique angle deposition of
thin
films:
From
fundamentals
to
devices,
Prog.
Mater.
Sci.
76
(2016)
59–153.
doi:10.1016/j.pmatsci.2015.06.003. [14] M.M. Hawkeye, M.T. Taschuk, M.J. Brett, Glancing Angle Deposition of Thin Films, John Wiley & Sons, Ltd, Chichester, UK, 2014. doi:10.1002/9781118847510. [15] A. Zylbersztejn, N.F. Mott, Metal-insulator transition in vanadium dioxide, Phys. Rev. B. 11 (1975) 4383–4395. doi:10.1103/PhysRevB.11.4383. [16] R.E. Marvel, R.R. Harl, V. Craciun, B.R. Rogers, R.F. Haglund, Influence of deposition process and substrate on the phase transition of vanadium dioxide thin films, Acta Mater. 91 (2015) 217–226. doi:10.1016/j.actamat.2015.03.009.
Highlights:
*pure vanadium thin films were GLAD deposited with incident angles of 0, 20 and 85°. *the films were annealed in atmospheric pressure (air) and vacuum conditions. *the resistivity evolution was monitored with increasing annealing temperatures. *metallic-like electrical properties and columnar features are preserved in vacuum. *VO2 phase is formed in air for α=0 and 85° samples.
8
PII: DOI: Reference:
S0167-577X(16)30418-9 http://dx.doi.org/10.1016/j.matlet.2016.03.097 MLBLUE20548
To appear in: Materials Letters Received date: 23 February 2016 Revised date: 18 March 2016 Accepted date: 18 March 2016 Cite this article as: Paulo Pedrosa, Nicolas Martin, Roland Salut, Mohammad Arab Pour Yazdi and Alain Billard, Controlled thermal oxidation of nanostructured vanadium thin films, Materials Letters, http://dx.doi.org/10.1016/j.matlet.2016.03.097 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting galley proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Controlled thermal oxidation of nanostructured vanadium thin films
Paulo Pedrosa*, Nicolas Martin, Roland Salut, Mohammad Arab Pour Yazdi, Alain Billard a
Institut FEMTO-ST, UMR 6174, Université de Bourgogne Franche-Comté, CNRS, 15B, Avenue des
Montboucons, 25030 Besançon Cedex, France
ABSTRACT Pure V thin films were dc sputtered with different pressures (0.4 and 0.6 Pa) and particle incident angles α of 0°, 20° and 85°, by using the GLancing Angle Deposition (GLAD) technique. The sputtered films were characterized regarding their electrical resistivity behaviour in atmospheric pressure and invacuum conditions as a function of temperature (40-550 °C), in order to control the oxidation process. Aiming at comprehending the oxidation behaviour of the samples, extensive morphological and structural studies were performed on the as-deposited and annealed samples. Main results show that, in opposition to annealing in air, the columnar nanostructures are preserved in vacuum conditions, keeping metallic-like electrical properties. Keywords: GLancing Angle Deposition, vanadium oxide, in-vacuum resistivity.
1. INTRODUCTION Vanadium oxides (VOx) such as VO2 and V2O5 in the form of thin films display interesting optical and electrical properties that make them suitable for a wide range of applications [1–3]. The reversible first-order semiconductor-to-metal transition exhibited by the VO2 [2,4–6] and V2O5 phases [1,7] makes them especially interesting for gas sensing purposes [3,8,9]. Furthermore, the sensitivity of these VOx phases are commonly attributed to the presence of oxygen vacancies [2,3,6–8]. Consequently, two main strategies are currently being explored to obtain the thin film materials either by (i) ab initio sputtering of the referred oxides [5,10] or by (ii) subsequent thermal oxidation of pure V films [4,11]. Also, the combination of both strategies is often used [2,6,12]. *
Corresponding author: P. Pedrosa; email address: [email protected]; Institut FEMTO-ST, UMR
6174, Université de Bourgogne Franche-Comté, CNRS, 15B, Avenue des Montboucons, 25030 Besançon Cedex, France.
1
In this work, the authors sputtered pure V thin films using the GLAD technique. It is known that porous nanostructured films display interesting electrical and optical properties, thus making them ideal candidates for sensing applications [13,14]. Hence, the effects of controlled thermal oxidation on the resistivity behaviour of the samples sputtered with normal and inclined columnar structures were investigated in atmospheric pressure and in-vacuum conditions. The aim of the present letter is to assess how the combination of nanostructured features and a controlled formation of the VOx phases contribute to the enhancement of the electrical behaviour of the films.
2. EXPERIMENTAL DETAILS Pure V films were deposited by dc magnetron sputtering inside a 40 L stainless-steel custom-made vacuum chamber. The reactor was equipped with a circular planar and water-cooled magnetron sputtering source, which was evacuated with a turbomolecular pump, backed by a mechanical one, in order to obtain a base pressure of 10-6 Pa. A vanadium target (purity 99.6 at.%, 75 mm diameter) was used. This target was dc sputtered in a pure argon atmosphere, using a constant current I = 200 mA. The incident angle of the particle flux was changed by tilting the substrate holder with an angle taken from the substrate normal: = 0°, 20° and 85° (GLAD technique). The argon flow rate was kept at 5.7 sccm ( = 0° sample) and 2.1 sccm ( = 20° and 85° samples), corresponding to a partial pressure of 6×10-1 Pa (S = 16 L.s-1) and 4×10-1 Pa (S = 8.9 L.s-1), respectively. The substrates, introduced through a 1 L airlock, were glass microscope slides (ISO norm 8037-1, with roughness better than 0.5 nm) and (100) silicon wafers (p-type, ρ = 1–30 Ω.cm). Before each run, all substrates were cleaned with acetone and alcohol, and the target was pre-sputtered in a pure argon atmosphere for 5 min., in order to remove the target surface contamination layer. The target-to-substrate distance was kept at 65 mm in all runs. The substrates were grounded and all depositions were carried out at room temperature. The deposition time was adjusted in order to obtain a thickness close to 500 nm. The thickness of the coatings was assessed using a Tencor Alpha Step IQ profilometer, while the morphological features of the samples were probed by scanning electron microscopy (SEM) at 5 keV in a Dual Beam SEM/FIB FEI Helios 600i microscope. The coatings were also characterized by X-ray diffraction (XRD). Measurements were carried out using a Bruker D8 focus diffractometer with a cobalt X-ray tube (Co λKα =1.78897 Å) in a θ/2θ configuration. Scans were performed with a step of 0.02° per
2
0.2 s and a 2θ angle ranging from 20 to 80°. The resistivity measurements were performed using the fourprobe van der Pauw method in the temperature range of 40–550 °C (1st cycle of 40-100-40 °C with a ramp of 2 °C.min-1 followed by 10 similar cycles with 50 °C increments each until a maximum temperature of 550 °C is reached), for both in-vacuum (10-5 Pa) and atmospheric pressure conditions. The atmospheric pressure measurements were done in a custom-made chamber, which is covered in order to have a dark environment; humidity and cleanness were considered as constant. The error associated to all resistivity measurements was always below 1% and the attachment of the contacts was checked prior to every run (I/V correlation close to 1) to ensure that an ohmic contact was attained (use of gold coated tips). All films were characterized in as-deposited and annealed (after atm. pressure and in-vacuum resistivity tests) conditions.
3. RESULTS AND DISCUSSION The resistivity (ρ) behaviour of the normal incidence and GLAD sputtered films with increasing annealing temperature is shown in Fig. 1. Major differences can be seen between atmospheric pressure and vacuum annealing conditions.
Figure 1. Atmospheric pressure (a1-c1) and in-vacuum (a2-c2) resistivity measurements for the V thin films deposited with particle incidence angles of 0°, 20° and 85°. Coloured arrows represent the resistivity evolution with increasing cycle number.
3
The normal incidence sample (V0°) displays a three-step resistivity behaviour (evidenced by the coloured arrows) when the annealing is performed at atmospheric pressure. Firstly, an increase of the values is evident until 300 °C, followed by a decrease of ρ until 450 °C. Finally, a subsequent increase in the final temperature cycle is observed. In opposition, a two-step behaviour is evidenced when the annealing is performed in vacuum conditions. No resistivity increase (oxidation) from 450 °C to 550 °C occurs due to the low oxygen content in the system. The same trend is observed for the sample deposited with α = 85° (V85°). More importantly, the steeper ρ increase (phase transition) occurring at 68 °C is indicative of VO2 formation [15,16]. On the other hand, the sample deposited with α = 20° (V20°) exhibits only a slight increase of resistivity in the whole temperature range. The overall resistivity behaviour of the samples can be explained by a closer analysis to the morphological and structural features of the samples, Figs. 2 and 3, respectively.
Figure 2. SEM cross-section micrographs of the produced samples before and after annealing. The higher sputtering pressure used in the production of the V0° film leads to a more porous morphology than that of V20°. Furthermore, α = 20° is not sufficient to produce the desired shadowing effect, thus no inclined columns can be seen. Hence, due to its denser columnar features, only V surface oxidation occurs when the V20° sample is annealed in vacuum. This is confirmed by both the very small overall increase of ρ and the presence of a shifted bcc V peak, indicative of a polycrystalline V matrix with oxygen insertions in the structure. Hence, a slightly surface-oxidised and polycrystalline V20° film
4
is obtained after in-vacuum annealing. Following the annealing in air, the V20° film displays a slightly higher ρ increase. Moreover, the columnar features are destroyed due to extensive formation of several VOx phases. Also, the film thickness increases (Fig. 2) due to significant surface oxidation, leading to a noisier ρ signal. No crystalline V phase is observable, thus a severely oxidised and amorphous V phase is obtained in atmospheric pressure conditions. The higher ρ increase (from 6.5×10-7 to 2.2×10-6 Ω.m) and Temperature Coefficient of Resistivity (TCR) decrease (from 1.2×10-3 to -4.9×10-5 K-1), in contrast with lower ρ increase (from 6.0×10-7 to 1.0×10-6 Ω.m) and almost constant TCR (5.4 to 8.6×10-4) in vacuum, can be attributed to the combined effects of V amorphization and VOx formation in air annealing. The amorphous VO2 phase formed is insufficient to promote a steep ρ increase.
Figure 3. XRD diffractograms of the sputtered samples before and after annealing. The V0° and V85° films display a similar ρ evolution, confirming their porous morphology. However, the V85° film displays higher ρ values due to its highly porous morphology. Focusing on the nanostructured V85° film, an increase of ρ (from 7.2×10-6 to 1.7×10-5 Ωm) and decrease of the TCR
5
(from 1.1 to -1.7×10-3 K-1) occurs from RT to 300 °C at atmospheric pressure. This indicates a strong oxidation of the film, not only at the surface but also in-depth due to its higher porosity. Afterwards, a semiconductor-like behaviour (decrease of ρ from 1.5×10-5 to 8.9×10-6 Ωm) is visible from 350-450 °C, which may be due to increased crystallisation of the VOx phases formed. Note that no ρ decrease is evidenced in the amorphous V20° film in air. In the last two cycles (450-550 °C), a strong oxidation of the V85° film occurs with ρ increasing one order of magnitude from 1.0×10-5 to 1.3×10-4 Ωm. Around 68 °C, a steeper increase of ρ is noticeable in both V0° and V85° films due to the formation of the VO2 phase. However, this phase may not be pure VO2 since the ρ increase is too low (< 1 order of magnitude). In vacuum conditions the ρ increase in the last temperature cycles is not achieved in both V0° and V85° films. From RT to 350 °C ρ increases from 5.6 to 8.7 ×10-6 Ωm and the TCR (4.7×10-4 K-1) gradually tends to zero due to the decrease of the crystal size of the bcc V phase. Afterwards, ρ decreases until 4.9×10-6 Ωm and the TCR becomes negative (-5.8×10-4 K-1) from 400-550 °C. This is due to the formation of conductive V7O13 and VO0.50 oxide phases in the porous columnar features. Finally, it is possible to say that porous nanostructures may be more easily tailored in terms of their electrical behaviour through controlled oxidation than denser ones. Also, in-vacuum annealing is an effective way to oxidise the pure V films without destroying their nanostructured columnar features. However, no VO2 phase was formed due to the low amount of available oxygen. Hence, further studies on ab initio GLAD nanostructured VOx films is currently underway.
4. CONCLUSIONS The present letter focuses on the study of electrical resistivity vs. temperature cycle of nanostructured vanadium films to be used as gas sensors. The pure vanadium films were nanostructured using the GLAD process and sputtered with incident angles α = 0°, 20° and 85°. In order to promote the formation of the VO2 phase the as-deposited films were annealed in atmospheric pressure and in-vacuum conditions. The controlled oxidation of the films was monitored by in-situ resistivity measurements with the films being morphologically and structurally characterized before and after annealing. The denser V20° films display a steady ρ increase both in air and vacuum conditions. No crystalline VO2 phase is formed, thus no steep ρ increase is present in the last temperature cycle. In opposition, the porous V0° and V85° films exhibit a two-step and three-step behaviour when annealed in vacuum or in air,
6
respectively. No evidences of VO2 formation is found when the annealing is performed in vacuum due to the low available oxygen amount. However, small traces of crystalline VO2 are present when these samples are annealed at the atmospheric pressure, hence the visible steeper ρ increase (although below one order of magnitude) at ~68 °C. The GLAD porous films are more easily tailorable in terms of the envisaged ρ behaviour. Moreover, in-vacuum annealing is effective in tuning the ρ evolution without destroying the nanostructured columns, despite no VO2 phase being attained.
ACKNOWLEDGEMENTS This work was supported by the region of Franche-Comté in cooperation with the Labex ACTION program (contract ANR-11-LABX-01-01). The authors acknowledge the MIMENTO Platform of the RENATECH network.
BIBLIOGRAPHY [1] S. Raja, G. Subramani, D. Bheeman, R. Rajamani, C. Bellan, Structural and optical properties of vacuum evaporated V2O5 thin films, Opt. - Int. J. Light Electron Opt. 127 (2016) 461–464. doi:10.1016/j.ijleo.2015.08.045. [2] Y.-K. Dou, J.-B. Li, M.-S. Cao, D.-Z. Su, F. Rehman, J.-S. Zhang, et al., Oxidizing annealing effects on
VO2
films
with
different
microstructures,
Appl.
Surf.
Sci.
345
(2015)
232–237.
doi:10.1016/j.apsusc.2015.03.044. [3] K. Schneider, M. Lubecka, A. Czapla, VOx Thin Films for Gas Sensor Applications, Procedia Eng. 120 (2015) 1153–1157. doi:10.1016/j.proeng.2015.08.1009. [4] Y. Ji, G.A. Niklasson, C.G. Granqvist, M. Boman, Thermochromic VO 2 films by thermal oxidation of
vanadium
in
SO2,
Sol.
Energy
Mater.
Sol.
Cells.
144
(2016)
713–716.
doi:10.1016/j.solmat.2015.10.012. [5] D. Zhang, M. Zhu, Y. Liu, K. Yang, G. Liang, Z. Zheng, et al., High performance VO 2 thin films growth by DC magnetron sputtering at low temperature for smart energy efficient window application, J. Alloys Compd. 659 (2016) 198–202. doi:10.1016/j.jallcom.2015.11.047. [6] J.Y. Suh, R. Lopez, L.C. Feldman, R.F. Haglund, Semiconductor to metal phase transition in the nucleation and growth of VO2 nanoparticles and thin films, J. Appl. Phys. 96 (2004) 1209. doi:10.1063/1.1762995. [7] R. Mustafa Öksüzoğlu, P. Bilgiç, M. Yıldırım, O. Deniz, Influence of post-annealing on electrical, structural and optical properties of vanadium oxide thin films, Opt. Laser Technol. 48 (2013) 102–109. doi:10.1016/j.optlastec.2012.10.001.
7
[8] A. Simo, B. Mwakikunga, B.T. Sone, B. Julies, R. Madjoe, M. Maaza, VO 2 nanostructures based chemiresistors for low power energy consumption hydrogen sensing, Int. J. Hydrogen Energy. 39 (2014) 8147–8157. doi:10.1016/j.ijhydene.2014.03.037. [9] J. Huotari, R. Bjorklund, J. Lappalainen, A. Lloyd Spetz, Pulsed Laser Deposited Nanostructured Vanadium Oxide Thin Films Characterized as Ammonia Sensors, Sensors Actuators B Chem. 217 (2015) 22–29. doi:10.1016/j.snb.2015.02.089. [10] X. Wei, S. Li, J. Gou, X. Dong, X. Yang, W. Li, et al., Preparation and characteristics of vanadium oxide thin films by controlling the sputtering voltage, Opt. Mater. (Amst). 36 (2014) 1419–1423. doi:10.1016/j.optmat.2014.02.021. [11] T.C.-K. Yang, Y.-L. Yang, R.-C. Juang, T.-W. Chiu, C.-C. Chen, The novel preparation method of high-performance thermochromic vanadium dioxide thin films by thermal oxidation of vanadiumstainless steel co-sputtered films, Vacuum. 121 (2015) 310–316. doi:10.1016/j.vacuum.2015.05.001. [12] Y.-Y. Su, X.-W. Cheng, J.-B. Li, Y.-K. Dou, F. Rehman, D.-Z. Su, et al., Evolution of microstructure in vanadium oxide bolometer film during annealing process, Appl. Surf. Sci. 357 (2015) 887–891. doi:10.1016/j.apsusc.2015.09.068. [13] A. Barranco, A. Borras, A.R. Gonzalez-Elipe, A. Palmero, Perspectives on oblique angle deposition of
thin
films:
From
fundamentals
to
devices,
Prog.
Mater.
Sci.
76
(2016)
59–153.
doi:10.1016/j.pmatsci.2015.06.003. [14] M.M. Hawkeye, M.T. Taschuk, M.J. Brett, Glancing Angle Deposition of Thin Films, John Wiley & Sons, Ltd, Chichester, UK, 2014. doi:10.1002/9781118847510. [15] A. Zylbersztejn, N.F. Mott, Metal-insulator transition in vanadium dioxide, Phys. Rev. B. 11 (1975) 4383–4395. doi:10.1103/PhysRevB.11.4383. [16] R.E. Marvel, R.R. Harl, V. Craciun, B.R. Rogers, R.F. Haglund, Influence of deposition process and substrate on the phase transition of vanadium dioxide thin films, Acta Mater. 91 (2015) 217–226. doi:10.1016/j.actamat.2015.03.009.
Highlights:
*pure vanadium thin films were GLAD deposited with incident angles of 0, 20 and 85°. *the films were annealed in atmospheric pressure (air) and vacuum conditions. *the resistivity evolution was monitored with increasing annealing temperatures. *metallic-like electrical properties and columnar features are preserved in vacuum. *VO2 phase is formed in air for α=0 and 85° samples.
8