- Email: [email protected]
Piezoresistive properties of nanocrystalline silicon thin films deposited on plastic substrates by hot-wire chemical vapor deposition
Thin Solid Films 515 (2007) 7658 – 7661 www.elsevier.com/locate/tsf
Piezoresistive properties of nanocrystalline silicon thin films deposited on plastic substrates by hot-wire chemical vapor deposition P. Alpuim ⁎, M. Andrade, V. Sencadas, M. Ribeiro, S.A. Filonovich, S. Lanceros-Mendez Department of Physics, University of Minho, Campus de Azurém, 4800–058 Guimarães, Portugal Available online 16 January 2007
Abstract The piezoresistive property of n-type and p-type nanocrystalline silicon thin films deposited on plastic (PEN) at a substrate temperature of 150 °C by hot-wire chemical vapor deposition, is studied. The crystalline fraction decreased from 80% to 65% in p-type and from 84% to 62% in n-type films, as the dopant gas-to-silane flow rate ratio was increased from 0.18% to 3–3.5%. N-type films have negative gauge factor (−11 to −16) and p-type films have positive gauge factor (9 to 25). In n-type films the higher gauge factors (in absolute value) were obtained by increasing the doping level whereas in p-type films higher gauge factors were obtained by increasing the crystalline fraction. © 2006 Elsevier B.V. All rights reserved. Keywords: Nanocrystalline silicon; Piezoresistivity; Hot-wire chemical vapor deposition; Flexible electronics
1. Introduction The piezoresistive property of crystalline silicon has long been known to scientists and engineers [1–4] and most straingauge sensors use that property as their operation principle [3–5]. These devices measure strain from the relative resistance change of a Si transducer that is deformed to the strain to be measured. Piezoresistance in c-Si is highly anisotropic and is described by a set of coefficients that are the elements of the rank-4 piezoresistive tensor, relating the change in electrical resistivity in a particular crystallographic direction to the components of the applied stress tensor. Due to the symmetry of the silicon fcc crystal structure the piezoresistive tensor has only three independent elements, namely π11, π12 and π44 [1] where the indexes have the usual meaning in six-component vector notation. Polycrystalline metallic strain-gauge sensors are also common but they are much less sensitive than their semiconductor counterparts [5]. This is because their operation is mainly based on resistance variation due to stress-induced dimensional changes and not on large changes on the resistivity, as is the case with Si (or Ge) sensors. Volume changes due to strain affect ⁎ Corresponding author. Tel.: +351 253 510463; fax: +351 253 510461. E-mail address: [email protected] (P. Alpuim). 0040-6090/$ - see front matter © 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.tsf.2006.11.138
both carrier concentration and carrier mobility in a semiconductor and in this way they lead to resistance changes. But more important, deformation potentials modify the band structure near the edge of the conduction and valence bands leading to changes in carrier effective mass and/or occupancy of the density of states at energies close to the Fermi level. This mechanisms dramatically enhance the piezoresistive effect in Si (also in Ge, C-diamond and other semiconductors) when compared to metals [1,5]. As the Fermi level position approaches the conduction or valence band edges the effect is expected to be enhanced and therefore piezoresistance is more important in doped than in intrinsic Si. However, both the physical origin and the corresponding values of the piezoresistive coefficients are very different in p- and n-type material. Detailed studies of both mechanisms can be found in the literature [1–4]. The piezoresistive effect has also been found in thin crystalline silicon films [6,7]. Due to the semi-crystalline nature of these films the reported effect is rather isotropic and it broadly averages the resistance changes of the randomly oriented crystallites, thus retaining the signal of the largest c-Si piezoresistive coefficient. Moreover, grain boundaries also play a role in piezoresistance of nanocrystalline films and their net effect is to attenuate resistivity changes occurring in the crystalline grains [8,9].
P. Alpuim et al. / Thin Solid Films 515 (2007) 7658–7661
In this paper, the piezoresistive effect in flexible n- and p-type hydrogenated nanocrystalline silicon thin films, nc-Si:H, deposited on inexpensive plastic substrates by the hot-wire chemical vapor deposition technique (HWCVD) will be studied as a function of crystalline fraction, gas-phase dopant concentration and position of the Fermi level. The parameter used to describe the piezoresistive effect in the present work is the gauge factor, GF, defined as the relative resistance change per unit axial strain: ΔR / R = GF × ε, where R is electrical resistance and ε is axial strain. In this expression the factor (1 + 2ν)ε where ν is the Poisson ratio, which accounts for the resistance change due to geometrical changes of the sample caused by strain, was removed from the second member of the equation since it is very small when compared to the piezoresistive effect. 2. Experimental details 2.1. Film preparation All the films were deposited in an ultra-high vacuum HWCVD chamber equipped with load lock. The pumping system, consisting of a molecular drag pump and a rotary vane pump, allowed to achieve a base pressure lower than 10− 7 Torr. The substrate holder was placed 5 cm above a single tantalum filament (Ø = 0.5 mm, 14 cm long) that was bent to a coil. The filament was resistively heated (Tfil ∼ 1770 °C) for deposition, using a DC power supply (Ifil = 13.5 A) that pyrolitically decomposed the reactant gases at the filament surface. nc-Si:H films were deposited using 95% hydrogen dilution (DH) of reactant gases, defined as the H2 flow rate divided by total gas flow rate × 100%. n- and p- type gas-phase doping was achieved by adding 98% H2-diluted phosphine or diborane, respectively, to the silane–hydrogen gas mixture (see Table 1). Two series of samples were deposited by varying the dopant gas-to-silane flow rate ratio, F = Fdopant / FSiH4. Since P- and Bdoping was performed in the same HWCVD chamber a 2 μm thick sacrificial layer of intrinsic a-Si:H was deposited between the doping series in order to avoid cross contamination. Table 1 Properties of n- and p-type samples prepared by HW-CVD Sample
PH3, B2H6 / SiH4 × 100%
Thickness tf (nm)
σd (Ω− 1 cm− 1)
Ea, σ (meV)
XC
GF
N192 N185 N183 N186 N222 N191 P214 P216 P217 P220 P218 P219 P221
0.15 1.0 1.5 2.0 2.5 3.0 0.18 1.0 1.5 2.0 2.5 3.0 3.5
169 190 174 187 148 183 163 182 186 170 185 178 174
0.57 0.30 1.07 3.50 1.35 2.39 2.72 × 10− 5 0.34 1.46 1.96 3.27 1.44 2.33
61 82 49 36 28 39 580 101 72 61 40 32 45
0.78 0.80 0.77 0.75 † 0.65 0.84 0.81 0.78 0.74 0.72 0.69 0.62
− 10.8 † − 12.9 − 19.0 − 14.3 − 16.0 † 25.3 24.5 15.1 10.3 13.3 9.1
(†) Not measured.
7659
75 μm-thick polyethylene naphthalate (PEN) (maximum working temperature is 155 °C) was used as substrate. Substrate temperature (Tsub) during deposition, measured by a thermocouple embedded in the back side of the substrate holder, was 150 °C. During deposition time (∼ 10 min) no increase in Tsub was detected. Glass substrates were placed side by side with the PEN substrates during all depositions for comparison and to allow for Raman measurements (see next section). 2.2. Film characterization Raman spectroscopy was used to obtain crystalline fraction of the films. In this case films were deposited on glass substrates, since Raman spectrum of PEN substrates contains a sharp pick at 520 cm− 1 which exactly overlaps the TO phonon mode of crystalline Si. The Raman spectra were obtained with a Jobin-Yvon T64000 spectrometer using a 514 nm Ar-laser line. The crystallinity was determined from the amplitudes of crystalline and amorphous bands using the formula XC = Ic / (Ic + yIa), where Ic and Ia are the amplitudes of crystalline and amorphous Raman peaks, respectively, and y = 1.25 is an empirical coefficient expressing the ratio of Raman crosssections for both bands [10]. Film thickness was calculated from the interference fringes in the optical transmission spectra obtained with a Shimadzu UV-3101PC spectrophotometer. Dark conductivity, σd, measurements were performed in the temperature range from 20 °C to 95 °C, during the cooling part of the cycle, between Al evaporated parallel contacts 6 mm wide, 1 mm apart. The activation energy of σd, Ea, was calculated from the slope of an Arrhenius plot of σd as a function of 1000 / T (T is temperature in K). Room-temperature (RT) dark conductivity was extrapolated from the linear region of the curve that was fit to the data in the Arrhenius plot. The relative content of P in the films was evaluated by TOF-SIMS. The piezoresistive property was studied by quasi-static mechanical tests performed in a Polymer Laboratories MINIMAT machine in the tensile mode at test velocity of dl / dt = 0.02 mm min− 1 on ∼ 35 mm × 10 mm rectangular samples, 0.075 mm thick. The variation of the electrical resistance of the samples with the applied axial strain was calculated from the slope of I–V curves measured in real time with an automated Keithley 487 picoammeter/voltage source and corresponding to several points of the force-deformation curves (see Fig. 1). I–V data points were collected between the same Al parallel contacts used for σd measurements. I–V curves were measured with the electric field applied and the current flowing parallel to the strain axis with applied voltage ranging between − 3 V and + 3 V. 3. Results and discussion Fig. 2 shows the crystalline fraction of p- and n-type films as a function of dopant gas-to-silane flow rate ratio. It can be seen that, in general, as the flow rate of the dopant gas was increased, the crystalline fraction of the doped film decreased, independently of the dopant gas added. XC decreased from 80% to 65% when the phosphine-to-silane flow rate ratio increased from 1%
7660
P. Alpuim et al. / Thin Solid Films 515 (2007) 7658–7661
Fig. 1. Relative resistance change, ΔR/R, as a function of relative length change ΔL/L, for n-type sample N222. The gauge factor, GF, is extracted from the linear part of the graph. Insert shows the load–displacement curve where the region used for electrical resistance measurements is boxed.
Fig. 3. Gauge factor, GF, for p-type films (solid symbols) and for n-type films (open symbols) plotted as a function of dopant gas-to-silane flow rate ratio, F.
to 3% and XC decreased from 84% to 62% when the diboraneto-silane flow rate ratio was increased from 0.18% to 3.5%. The insert in Fig. 2 shows that, as F increased, so did the incorporation of P-atoms in the film Si network and this is arguably the main reason for the decrease in XC: a higher concentration of dopant atoms in the network causes amorphization of the film structure because of the local deformation induced by the impurity atoms. Although SIMS data for Bdoped films is not available at the moment it is also plausible that a similar effect is causing the observed decrease (Fig. 2) of XC in p-type films. Fig. 3 shows the gauge factor, GF, as a function of dopant gasto-silane flow rate ratio. The first important fact to point out is that GF is positive for p-type films and is negative for n-type films. This confirms previous results [6,7] and is consistent with the idea that because the films are semi-crystalline and isotropic, at least within planes parallel to the substrate, they retain the sign of the largest (in absolute value) piezoresistive coefficient of c-Si
which is π44 = 138 × 10− 11 Pa− 1 for p-type and π11 = − 102 × 10− 11 Pa− 1 for n-type material. For example, considering the modulus of elasticity of c-Si in the direction [100], Y100 = 130 GPa [11] one would have, for an n-type single-crystal film to which stress was applied along a [100] direction a GF100 =π11 × Y100 = −132.6. For a p-type monocrystalline film stressed along a [110] direction, with the current and electric field in the same [110] direction one would have [2]: GF110 = 1/2(π11 + π12 + π44) × Y110 = 1/2 × (7 − 1 + 138) × 10− 11 Pa− 1 × 170 GPa = 122.4. These would be extreme values of GF obtainable from n- and p-type monocrystalline Si samples conveniently oriented. Instead, in this work it can be seen (Fig. 3) that in actual polycrystalline films the gauge factor is between 9 and 25 for p-type and between −11 and −19 for n-type films. Besides the effect of random orientation of the crystallites, the damping effect on GF-values caused by grain
Fig. 2. Film crystalline fraction, XC, as a function of dopant gas-to-silane flow rate ratio, F. For p-type films (solid symbols) dopant gas is B2H6 and for n-type films (open symbols) it is PH3.
Fig. 4. Absolute value of gauge factor, GF, as a function of dark-conductivity activation energy, Ea, for p-type (solid symbols) and n-type (open symbols) films.
P. Alpuim et al. / Thin Solid Films 515 (2007) 7658–7661
boundaries is also present [8,9]. The second important fact in Fig. 3 is that while increasing the dopant concentration causes a decrease in GF for p-type films, the opposite is observed (in absolute value of GF) for n-type films. This indicates different origins for the piezoresistive effect in n- and p-type films. Fig. 4 shows the absolute value of the GF as a function of the activation energy, Ea, of σd for n- and p-type samples. While in n-type films the higher absolute value of GF (GF ∼ − 19) are obtained at the lower values of Ea, in p-type films the higher GFs (∼ 25) are obtained at the higher values of Ea, within the range of Ea measured. It is important to notice that all the films studied have σd and Ea-values characteristic of highly doped nc-Si:H. In general, the lower the Ea is, the higher the σd (see Table 1) and for P-doped films also the higher the phosphorous concentration is (see insert in Fig. 2). In n-type films piezoresistivity depends monotonically on σd (Table 1). For borondoped films the GF is higher for higher XC. Since a high XC is correlated with low dopant incorporation (Fig. 2) it can be concluded that in p-type material, contrary to what was observed for n-type material, crystallinity is the main parameter affecting the piezoresistive response of the material. It can thus be concluded that, within the doping range studied in this paper, in n-type films the piezoresistive effect is more favoured by a position of the Fermi level close to the band edge (Ea = 36 meV) than by the crystalline fraction, while in p-type films the GF is higher in less doped films (Ea = 101 meV) but with a higher XC (81%). In order to test the films against fatigue, two samples were submitted to repeated loading–unloading cycles and the gauge factor was re-evaluated during each cycle. In one of the samples GF decreased from 24.5 to 15.0 after 10 cycles and in the other it had a reduction from 25.3 to 14.1 after 16 cycles. Observation of the samples at the microscope after cycling showed that cracks had developed in the films in the region between contacts. Cracks were transverse to the applied strain direction and generally propagated from outside the gap between contacts to inside it. This suggests that in order to use these films as sensors some type of post-deposition patterning or masking during film growth is necessary in order to remove the film from the entire area that is inactive for sensing. In this way the probability to have pinholes and other defects and crack propagation from them will be highly reduced. This will be the next step in the study of shape sensors.
7661
4. Conclusions In this work the piezoresistive property of n- and p-type ncSi:H thin films deposited at Tsub = 150 °C by HWCVD on PEN substrates was demonstrated. N-type films showed a decrease in their resistivity as they were strained axially under tension (GF b 0) whereas p-type films exhibited GF N 0. The highest (absolute) values of GF were 19 for n-type and 25 for p-type films. While the most piezoresistive n-type nc-Si:H films were those with higher dark conductivity, the highest GF was found in p-type films with larger crystalline fraction. Repeated loading and unloading the films under tensile stress caused cracks to develop and the GF to decrease (∼ 10 units) below their original absolute values. This suggests that in order to make devices most of the film area should be etched out to prevent crack nucleation and growth. With such an approach it should be possible to fabricate high-elongation sensors taking full advantage of the specific mechanical properties of the flexible film–substrate system. Acknowledgements The authors thank A. Rolo (U. Minho) for help with the Raman measurements and S. Chiussi (U. Vigo) for the TOFSIMS measurements. One of the authors (S.A. Filonovich) acknowledges Fundação para a Ciência e Tecnologia (FCT) for a Post-doctorate grant (SFRH/BPD/14919/2004). References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11]
C.S. Smith, Phys. Rev. 94 (1954) 42. W.G. Pfann, R.N. Thurston, J. Appl. Phys. 32 (1961) 2008. O.N. Tufte, P.W. Chapman, D. Long, J. Appl. Phys. 33 (1962) 3322. I. Goroff, L. Kleinman, Phys. Rev. 132 (1961) 1080. M. Elwenspoek, R. Wiegerink, Mechanical Microsensors, Springer, New York, 2001, p. 87. P. Alpuim, V. Chu, J.P. Conde, IEEE Sens. J. 2 (2002) 336. S. Nishida, M. Konagai, K. Takahashi, Jpn. J. Appl. Phys. 25 (1986) 17. L. Fang, W.L. Wang, P.D. Ding, K.J. Liao, J. Wang, J. Appl. Phys. 86 (1999) 5185. M. Eickhoff, M. Stutzmann, J. Appl. Phys. 96 (2004) 2878. M. Ledinský, L. Fekete, J. Stuchlík, T. Mates, A. Fejfar, J. Kočka, J. NonCryst. Solids 352 (2006) 1209. J.J. Wortman, R.A. Evans, J. Appl. Phys. 36 (1965) 153.
Piezoresistive properties of nanocrystalline silicon thin films deposited on plastic substrates by hot-wire chemical vapor deposition P. Alpuim ⁎, M. Andrade, V. Sencadas, M. Ribeiro, S.A. Filonovich, S. Lanceros-Mendez Department of Physics, University of Minho, Campus de Azurém, 4800–058 Guimarães, Portugal Available online 16 January 2007
Abstract The piezoresistive property of n-type and p-type nanocrystalline silicon thin films deposited on plastic (PEN) at a substrate temperature of 150 °C by hot-wire chemical vapor deposition, is studied. The crystalline fraction decreased from 80% to 65% in p-type and from 84% to 62% in n-type films, as the dopant gas-to-silane flow rate ratio was increased from 0.18% to 3–3.5%. N-type films have negative gauge factor (−11 to −16) and p-type films have positive gauge factor (9 to 25). In n-type films the higher gauge factors (in absolute value) were obtained by increasing the doping level whereas in p-type films higher gauge factors were obtained by increasing the crystalline fraction. © 2006 Elsevier B.V. All rights reserved. Keywords: Nanocrystalline silicon; Piezoresistivity; Hot-wire chemical vapor deposition; Flexible electronics
1. Introduction The piezoresistive property of crystalline silicon has long been known to scientists and engineers [1–4] and most straingauge sensors use that property as their operation principle [3–5]. These devices measure strain from the relative resistance change of a Si transducer that is deformed to the strain to be measured. Piezoresistance in c-Si is highly anisotropic and is described by a set of coefficients that are the elements of the rank-4 piezoresistive tensor, relating the change in electrical resistivity in a particular crystallographic direction to the components of the applied stress tensor. Due to the symmetry of the silicon fcc crystal structure the piezoresistive tensor has only three independent elements, namely π11, π12 and π44 [1] where the indexes have the usual meaning in six-component vector notation. Polycrystalline metallic strain-gauge sensors are also common but they are much less sensitive than their semiconductor counterparts [5]. This is because their operation is mainly based on resistance variation due to stress-induced dimensional changes and not on large changes on the resistivity, as is the case with Si (or Ge) sensors. Volume changes due to strain affect ⁎ Corresponding author. Tel.: +351 253 510463; fax: +351 253 510461. E-mail address: [email protected] (P. Alpuim). 0040-6090/$ - see front matter © 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.tsf.2006.11.138
both carrier concentration and carrier mobility in a semiconductor and in this way they lead to resistance changes. But more important, deformation potentials modify the band structure near the edge of the conduction and valence bands leading to changes in carrier effective mass and/or occupancy of the density of states at energies close to the Fermi level. This mechanisms dramatically enhance the piezoresistive effect in Si (also in Ge, C-diamond and other semiconductors) when compared to metals [1,5]. As the Fermi level position approaches the conduction or valence band edges the effect is expected to be enhanced and therefore piezoresistance is more important in doped than in intrinsic Si. However, both the physical origin and the corresponding values of the piezoresistive coefficients are very different in p- and n-type material. Detailed studies of both mechanisms can be found in the literature [1–4]. The piezoresistive effect has also been found in thin crystalline silicon films [6,7]. Due to the semi-crystalline nature of these films the reported effect is rather isotropic and it broadly averages the resistance changes of the randomly oriented crystallites, thus retaining the signal of the largest c-Si piezoresistive coefficient. Moreover, grain boundaries also play a role in piezoresistance of nanocrystalline films and their net effect is to attenuate resistivity changes occurring in the crystalline grains [8,9].
P. Alpuim et al. / Thin Solid Films 515 (2007) 7658–7661
In this paper, the piezoresistive effect in flexible n- and p-type hydrogenated nanocrystalline silicon thin films, nc-Si:H, deposited on inexpensive plastic substrates by the hot-wire chemical vapor deposition technique (HWCVD) will be studied as a function of crystalline fraction, gas-phase dopant concentration and position of the Fermi level. The parameter used to describe the piezoresistive effect in the present work is the gauge factor, GF, defined as the relative resistance change per unit axial strain: ΔR / R = GF × ε, where R is electrical resistance and ε is axial strain. In this expression the factor (1 + 2ν)ε where ν is the Poisson ratio, which accounts for the resistance change due to geometrical changes of the sample caused by strain, was removed from the second member of the equation since it is very small when compared to the piezoresistive effect. 2. Experimental details 2.1. Film preparation All the films were deposited in an ultra-high vacuum HWCVD chamber equipped with load lock. The pumping system, consisting of a molecular drag pump and a rotary vane pump, allowed to achieve a base pressure lower than 10− 7 Torr. The substrate holder was placed 5 cm above a single tantalum filament (Ø = 0.5 mm, 14 cm long) that was bent to a coil. The filament was resistively heated (Tfil ∼ 1770 °C) for deposition, using a DC power supply (Ifil = 13.5 A) that pyrolitically decomposed the reactant gases at the filament surface. nc-Si:H films were deposited using 95% hydrogen dilution (DH) of reactant gases, defined as the H2 flow rate divided by total gas flow rate × 100%. n- and p- type gas-phase doping was achieved by adding 98% H2-diluted phosphine or diborane, respectively, to the silane–hydrogen gas mixture (see Table 1). Two series of samples were deposited by varying the dopant gas-to-silane flow rate ratio, F = Fdopant / FSiH4. Since P- and Bdoping was performed in the same HWCVD chamber a 2 μm thick sacrificial layer of intrinsic a-Si:H was deposited between the doping series in order to avoid cross contamination. Table 1 Properties of n- and p-type samples prepared by HW-CVD Sample
PH3, B2H6 / SiH4 × 100%
Thickness tf (nm)
σd (Ω− 1 cm− 1)
Ea, σ (meV)
XC
GF
N192 N185 N183 N186 N222 N191 P214 P216 P217 P220 P218 P219 P221
0.15 1.0 1.5 2.0 2.5 3.0 0.18 1.0 1.5 2.0 2.5 3.0 3.5
169 190 174 187 148 183 163 182 186 170 185 178 174
0.57 0.30 1.07 3.50 1.35 2.39 2.72 × 10− 5 0.34 1.46 1.96 3.27 1.44 2.33
61 82 49 36 28 39 580 101 72 61 40 32 45
0.78 0.80 0.77 0.75 † 0.65 0.84 0.81 0.78 0.74 0.72 0.69 0.62
− 10.8 † − 12.9 − 19.0 − 14.3 − 16.0 † 25.3 24.5 15.1 10.3 13.3 9.1
(†) Not measured.
7659
75 μm-thick polyethylene naphthalate (PEN) (maximum working temperature is 155 °C) was used as substrate. Substrate temperature (Tsub) during deposition, measured by a thermocouple embedded in the back side of the substrate holder, was 150 °C. During deposition time (∼ 10 min) no increase in Tsub was detected. Glass substrates were placed side by side with the PEN substrates during all depositions for comparison and to allow for Raman measurements (see next section). 2.2. Film characterization Raman spectroscopy was used to obtain crystalline fraction of the films. In this case films were deposited on glass substrates, since Raman spectrum of PEN substrates contains a sharp pick at 520 cm− 1 which exactly overlaps the TO phonon mode of crystalline Si. The Raman spectra were obtained with a Jobin-Yvon T64000 spectrometer using a 514 nm Ar-laser line. The crystallinity was determined from the amplitudes of crystalline and amorphous bands using the formula XC = Ic / (Ic + yIa), where Ic and Ia are the amplitudes of crystalline and amorphous Raman peaks, respectively, and y = 1.25 is an empirical coefficient expressing the ratio of Raman crosssections for both bands [10]. Film thickness was calculated from the interference fringes in the optical transmission spectra obtained with a Shimadzu UV-3101PC spectrophotometer. Dark conductivity, σd, measurements were performed in the temperature range from 20 °C to 95 °C, during the cooling part of the cycle, between Al evaporated parallel contacts 6 mm wide, 1 mm apart. The activation energy of σd, Ea, was calculated from the slope of an Arrhenius plot of σd as a function of 1000 / T (T is temperature in K). Room-temperature (RT) dark conductivity was extrapolated from the linear region of the curve that was fit to the data in the Arrhenius plot. The relative content of P in the films was evaluated by TOF-SIMS. The piezoresistive property was studied by quasi-static mechanical tests performed in a Polymer Laboratories MINIMAT machine in the tensile mode at test velocity of dl / dt = 0.02 mm min− 1 on ∼ 35 mm × 10 mm rectangular samples, 0.075 mm thick. The variation of the electrical resistance of the samples with the applied axial strain was calculated from the slope of I–V curves measured in real time with an automated Keithley 487 picoammeter/voltage source and corresponding to several points of the force-deformation curves (see Fig. 1). I–V data points were collected between the same Al parallel contacts used for σd measurements. I–V curves were measured with the electric field applied and the current flowing parallel to the strain axis with applied voltage ranging between − 3 V and + 3 V. 3. Results and discussion Fig. 2 shows the crystalline fraction of p- and n-type films as a function of dopant gas-to-silane flow rate ratio. It can be seen that, in general, as the flow rate of the dopant gas was increased, the crystalline fraction of the doped film decreased, independently of the dopant gas added. XC decreased from 80% to 65% when the phosphine-to-silane flow rate ratio increased from 1%
7660
P. Alpuim et al. / Thin Solid Films 515 (2007) 7658–7661
Fig. 1. Relative resistance change, ΔR/R, as a function of relative length change ΔL/L, for n-type sample N222. The gauge factor, GF, is extracted from the linear part of the graph. Insert shows the load–displacement curve where the region used for electrical resistance measurements is boxed.
Fig. 3. Gauge factor, GF, for p-type films (solid symbols) and for n-type films (open symbols) plotted as a function of dopant gas-to-silane flow rate ratio, F.
to 3% and XC decreased from 84% to 62% when the diboraneto-silane flow rate ratio was increased from 0.18% to 3.5%. The insert in Fig. 2 shows that, as F increased, so did the incorporation of P-atoms in the film Si network and this is arguably the main reason for the decrease in XC: a higher concentration of dopant atoms in the network causes amorphization of the film structure because of the local deformation induced by the impurity atoms. Although SIMS data for Bdoped films is not available at the moment it is also plausible that a similar effect is causing the observed decrease (Fig. 2) of XC in p-type films. Fig. 3 shows the gauge factor, GF, as a function of dopant gasto-silane flow rate ratio. The first important fact to point out is that GF is positive for p-type films and is negative for n-type films. This confirms previous results [6,7] and is consistent with the idea that because the films are semi-crystalline and isotropic, at least within planes parallel to the substrate, they retain the sign of the largest (in absolute value) piezoresistive coefficient of c-Si
which is π44 = 138 × 10− 11 Pa− 1 for p-type and π11 = − 102 × 10− 11 Pa− 1 for n-type material. For example, considering the modulus of elasticity of c-Si in the direction [100], Y100 = 130 GPa [11] one would have, for an n-type single-crystal film to which stress was applied along a [100] direction a GF100 =π11 × Y100 = −132.6. For a p-type monocrystalline film stressed along a [110] direction, with the current and electric field in the same [110] direction one would have [2]: GF110 = 1/2(π11 + π12 + π44) × Y110 = 1/2 × (7 − 1 + 138) × 10− 11 Pa− 1 × 170 GPa = 122.4. These would be extreme values of GF obtainable from n- and p-type monocrystalline Si samples conveniently oriented. Instead, in this work it can be seen (Fig. 3) that in actual polycrystalline films the gauge factor is between 9 and 25 for p-type and between −11 and −19 for n-type films. Besides the effect of random orientation of the crystallites, the damping effect on GF-values caused by grain
Fig. 2. Film crystalline fraction, XC, as a function of dopant gas-to-silane flow rate ratio, F. For p-type films (solid symbols) dopant gas is B2H6 and for n-type films (open symbols) it is PH3.
Fig. 4. Absolute value of gauge factor, GF, as a function of dark-conductivity activation energy, Ea, for p-type (solid symbols) and n-type (open symbols) films.
P. Alpuim et al. / Thin Solid Films 515 (2007) 7658–7661
boundaries is also present [8,9]. The second important fact in Fig. 3 is that while increasing the dopant concentration causes a decrease in GF for p-type films, the opposite is observed (in absolute value of GF) for n-type films. This indicates different origins for the piezoresistive effect in n- and p-type films. Fig. 4 shows the absolute value of the GF as a function of the activation energy, Ea, of σd for n- and p-type samples. While in n-type films the higher absolute value of GF (GF ∼ − 19) are obtained at the lower values of Ea, in p-type films the higher GFs (∼ 25) are obtained at the higher values of Ea, within the range of Ea measured. It is important to notice that all the films studied have σd and Ea-values characteristic of highly doped nc-Si:H. In general, the lower the Ea is, the higher the σd (see Table 1) and for P-doped films also the higher the phosphorous concentration is (see insert in Fig. 2). In n-type films piezoresistivity depends monotonically on σd (Table 1). For borondoped films the GF is higher for higher XC. Since a high XC is correlated with low dopant incorporation (Fig. 2) it can be concluded that in p-type material, contrary to what was observed for n-type material, crystallinity is the main parameter affecting the piezoresistive response of the material. It can thus be concluded that, within the doping range studied in this paper, in n-type films the piezoresistive effect is more favoured by a position of the Fermi level close to the band edge (Ea = 36 meV) than by the crystalline fraction, while in p-type films the GF is higher in less doped films (Ea = 101 meV) but with a higher XC (81%). In order to test the films against fatigue, two samples were submitted to repeated loading–unloading cycles and the gauge factor was re-evaluated during each cycle. In one of the samples GF decreased from 24.5 to 15.0 after 10 cycles and in the other it had a reduction from 25.3 to 14.1 after 16 cycles. Observation of the samples at the microscope after cycling showed that cracks had developed in the films in the region between contacts. Cracks were transverse to the applied strain direction and generally propagated from outside the gap between contacts to inside it. This suggests that in order to use these films as sensors some type of post-deposition patterning or masking during film growth is necessary in order to remove the film from the entire area that is inactive for sensing. In this way the probability to have pinholes and other defects and crack propagation from them will be highly reduced. This will be the next step in the study of shape sensors.
7661
4. Conclusions In this work the piezoresistive property of n- and p-type ncSi:H thin films deposited at Tsub = 150 °C by HWCVD on PEN substrates was demonstrated. N-type films showed a decrease in their resistivity as they were strained axially under tension (GF b 0) whereas p-type films exhibited GF N 0. The highest (absolute) values of GF were 19 for n-type and 25 for p-type films. While the most piezoresistive n-type nc-Si:H films were those with higher dark conductivity, the highest GF was found in p-type films with larger crystalline fraction. Repeated loading and unloading the films under tensile stress caused cracks to develop and the GF to decrease (∼ 10 units) below their original absolute values. This suggests that in order to make devices most of the film area should be etched out to prevent crack nucleation and growth. With such an approach it should be possible to fabricate high-elongation sensors taking full advantage of the specific mechanical properties of the flexible film–substrate system. Acknowledgements The authors thank A. Rolo (U. Minho) for help with the Raman measurements and S. Chiussi (U. Vigo) for the TOFSIMS measurements. One of the authors (S.A. Filonovich) acknowledges Fundação para a Ciência e Tecnologia (FCT) for a Post-doctorate grant (SFRH/BPD/14919/2004). References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11]
C.S. Smith, Phys. Rev. 94 (1954) 42. W.G. Pfann, R.N. Thurston, J. Appl. Phys. 32 (1961) 2008. O.N. Tufte, P.W. Chapman, D. Long, J. Appl. Phys. 33 (1962) 3322. I. Goroff, L. Kleinman, Phys. Rev. 132 (1961) 1080. M. Elwenspoek, R. Wiegerink, Mechanical Microsensors, Springer, New York, 2001, p. 87. P. Alpuim, V. Chu, J.P. Conde, IEEE Sens. J. 2 (2002) 336. S. Nishida, M. Konagai, K. Takahashi, Jpn. J. Appl. Phys. 25 (1986) 17. L. Fang, W.L. Wang, P.D. Ding, K.J. Liao, J. Wang, J. Appl. Phys. 86 (1999) 5185. M. Eickhoff, M. Stutzmann, J. Appl. Phys. 96 (2004) 2878. M. Ledinský, L. Fekete, J. Stuchlík, T. Mates, A. Fejfar, J. Kočka, J. NonCryst. Solids 352 (2006) 1209. J.J. Wortman, R.A. Evans, J. Appl. Phys. 36 (1965) 153.