- Email: [email protected]
Study on the applicability of polytetrafluoroethylene–silver composite thin films as sensor material
Applied Surface Science 278 (2013) 117–121
Contents lists available at SciVerse ScienceDirect
Applied Surface Science journal homepage: www.elsevier.com/locate/apsusc
Study on the applicability of polytetrafluoroethylene–silver composite thin films as sensor material Tomi Smausz a,∗ , Gabriella Kecskeméti a , Tamás Csizmadia a , Ferenc Benedek b , Béla Hopp a a b
Department of Optics and Quantum Electronics, University of Szeged, H-6720 Szeged, Dóm tér 9., Hungary Institute for Engineering and Materials Science, University of Szeged, H-6720 Szeged, Tisza Lajos Krt. 103., Hungary
a r t i c l e
i n f o
Article history: Received 15 June 2012 Received in revised form 6 November 2012 Accepted 8 January 2013 Available online 23 January 2013 Keywords: Pulsed laser deposition Composite layer Silver PTFE Fluctuation-enhanced sensing Cholesterol
a b s t r a c t A study on applicability of conductive high specific surface PTFE/Ag composite layers as active electrodes of a non-enzymatic cholesterol sensor is presented. The composite layers were prepared on one of the two neighboring electrode of a printed circuit board by pulsed laser deposition technique where targets composed of silver and PTFE were ablated by an ArF excimer laser. Cholesterol was dissolved in 0.1 M NaOH in different concentrations in 0–5 mM range. A drop of cholesterol covered the two electrodes and a constant current of 10 A was forced through the sample while the voltage between the electrodes was measured by means of a high resolution A/D converter with 1 kHz sampling rate for 5 s periods. Instead of the time-averaged signal monitoring we applied the Fluctuation-Enhanced Sensing (FES) method which is based on the analysis of the stochastic component of the signal. The power spectral density of the fluctuation was found to be dependent on the cholesterol concentration of the samples. Principal Component Analysis method was used for quantifying the difference between the recorded spectra. A tendentious variation of the spectral properties as the function of the cholesterol concentration was observed. The results indicate that the FES technique combined with high specific surface composite electrodes may be a useful tool for cholesterol detection. © 2013 Elsevier B.V. All rights reserved.
1. Introduction Polymer-metal composites are good candidates for sensor materials. Their electrical, optical and dielectric behavior is strongly influenced by the metal content, the size and shape of the particles [1–3]. One of the promising basic materials for composite preparation is the Teflon (polytetrafluoroethylene, PTFE) which has good mechanical, thermal and chemical stability [4,5]. PTFE thin films can be prepared by various techniques depending on the used substrate and quality requirements. The pulsed laser deposition (PLD) allows the deposition of stoichiometric Teflon thin films with compact structure or sponge-like morphology depending on the deposition parameters and post-treatment [6–8]. Recent studies showed that PTFE/silver composite structures deposited by PLD using PTFE/Ag targets have a rough morphology with increased specific surface attributed to the deposition of PTFE grains and show improved conductive and wetting properties due to the Ag content [9]. Conductive layers with high specific surface can find applications in the field of biosensors, too. Amperometric and potentiometric devices often aiming the detection of analytes with
∗ Corresponding author. Tel.: +36 62 544657; fax: +36 62 544658. E-mail address: [email protected] (T. Smausz). 0169-4332/$ – see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.apsusc.2013.01.051
clinical importance (e.g. urea, glucose, cholesterol) are mainly based on incorporation of enzymes into the active electrodes [10–14]. The enzyme-based sensors can have good selectivity and high sensitivity, however, the enzyme immobilization process is the most difficult step of the production process. In the last few years several attempts were made for fabricating enzyme-free biosensors. Wang et al. produced an amperometric glucose biosensor based on the modification of functional nickel hexacyanoferrate nanoparticles [15]. Lee and Park used a macroporous gold electrode with incorporated Pt [16], while Li et al. built porous tubular silver nanoparticles [17] for detecting cholesterol. In these amperometric measurements charges involved in the electrocathalitic oxidation of cholesterol at the electrode surface were detected, therefore the increased specific surface of the electrode is of high importance. Most of the detection techniques are based on the measurement of the time-averaged value of the sensor signal, however, also the stochastic component (noise) can serve important information on the detector’s ambience. For example, in case when detector signal is the result of the desorption of the analyte at the sensing surface the dynamics of the adsorption–desorption and the diffusion properties can serve as a “fingerprint” for a given analyte. In the Fluctuation-Enhanced Sensing (FES) the low amplitude timevarying components are amplified and statistically analyzed to find the “fingerprints” of the analytes [18,19]. While for simultaneous detection of multiple components a number of different sensors
118
T. Smausz et al. / Applied Surface Science 278 (2013) 117–121
have to be used, with use of proper analysis, pattern database and pattern recognition tools FES can allow their detection using one single sensor [20,21]. The FES method was introduced and mainly used for increasing the selectivity of gas sensors, its usefulness in liquid phase was demonstrated for detection of bacteria [22]. In latter case a voltage fluctuation between two electrodes caused by ions emitted by phage infected bacteria was detected. The aim of present work is a preliminary study on the applicability of our laser deposited PTFE/Ag composite layers as sensing electrodes for the non-enzymatic detection of cholesterol. Differently from conventional amperometric methods, we intended to study whether the cholesterol concentration can influence the noise of the detected signal which is substantial from the point of view of fluctuation-enhanced sensing.
The target rotated with 10 rpm was ablated with pulses of an ArF excimer laser ( = 193 nm, FWHM = 20 ns) having 8 J/cm2 fluence at an area of 0.8 mm2 , the repetition rate was 10 Hz. The relatively low deposition rate does not allow the formation of sandwich structure. Layers were deposited with 7500 and 12,500 pulses, respectively to obtain composite electrodes with different roughness. During deposition the substrates (sample board or glass plate for electron microscopic investigations) were placed onto a 150 ◦ C heatable holder. The base pressure of the PLD chamber was about 2 × 10−3 Pa and the distance between the target and the substrate was 4 cm. The morphology of the layers was studied with a Hitachi S4700 scanning electron microscope, while their elemental composition was analyzed with the energy-dispersive X-ray spectrometer (EDX) of the electron microscope.
2. Experimental
2.2. The fluctuation-enhanced sensing
2.1. Thin film deposition
Cholesterol was dissolved in 0.1 M NaOH solvent containing 2 v% Triton X-100 (Sigma) in different concentrations: 0.5, 1, 2 and 3.5 and 5 mM, respectively. (The normal cholesterol concentration is below 5 mM in human blood.) A drop of solution was placed onto the sample board to overlap the two adjacent electrodes, one being covered with the PTFE/Ag composite layer, as shown in Fig. 2. Reference measurements on untreated Au electrodes were also carried out. A constant current of 10 A was drawn through the circuit and the U(t) voltage between the two adjacent electrodes was measured with a sampling rate of 1000 Hz for three consecutive periods of
PTFE/Ag composite layers were prepared by pulsed laser deposition method onto one of the electrodes of printed circuit sample boards containing two pairs of 2 mm × 2 mm gold plated electrodes as shown in Fig. 1. The experimental parameters were chosen based on the results of our earlier studies [8,9]. The rotating disk form targets were composed of two sectors of circles: 1/6 part PTFE (Goodfellow, grain size 6–9 m compacted at 520 MPa pressure) and 5/6 part Ag (Goodfellow, 2 mm thickness, purity 99.95%).
Fig. 1. Set-up used for composite layer deposition (a) and photograph of a sample board with one covered electrode (b).
Cholesterol Solution Gold plated Deposited layer electrode
a
U(t)
V A
n
samples
S(f)
2
y=S(f)/U=
y1 ...
b
Measurements on
FFT Power Spectral Density
yn
Principal Component Analysis
SCORES
PLOTTING (PC-1, PC-2)
Fig. 2. Scheme of the measurement procedure of FES method: obtaining individual noise spectrum for each sample (a) and comparison of spectra by means of PCA (b).
T. Smausz et al. / Applied Surface Science 278 (2013) 117–121
119
5 s. A home-made A/D converter allowed the digital recording of the time-varying component (>0.2 Hz) of the U(t) voltage with a resolution of 38 nV (least significant bit). The S(f) power spectral density was obtained by FFT and y(f) = S(f)/U= 2 function (corrected spectrum) was calculated, where U= is the constant component of the U(t) signal. For a quantitative comparison of the y(f) functions obtained at different cholesterol concentrations we used the Principal Component Analysis (PCA) method, one of the pattern recognition tools used in sensor technology [23,24] and it was performed using the MATLAB software. The PCA transforms the array formed of rows consisting of PSD data into PC array, where the most of the information about the similarity/difference of the spectra is preserved into the first few columns. 3. Results and discussions 3.1. Thin film characterization Our earlier study [8] demonstrated the PFFE is mainly deposited in form of grains and larger clusters as also shown on the electron microscopic image (Fig. 3a) of the layers. According to elemental composition maps (Fig. 3b and c) recorded with the EDX spectrometer the silver has a more uniform distribution, the darker areas in Fig. 3c can be attributed to the shielding effect of the PTFE grains, since their size is larger than the detection depth (∼1 m) of the EDX. Due to the higher ablation rate of the Teflon, the silver content of the deposit is lower than that of the target [9]. Moreover, as the consequence of the alternate ablation of PTFE and silver the Teflon grains are covered with silver forming a conductive film on their surface. The applied experimental setup results in layers with the following properties: approximately 20 wt% silver content, rough surface with an average thickness of 3 and 5 m for the applied 7500 and 12,500 pulses, respectively. The thicker layer showed an increased roughness as compared to the thinner one. The void content of the layers is under the rough surface profile (measurable with a stylus profiler) is around 48–50%. 3.2. The fluctuation-enhanced sensing measurements Preliminary measurements showed that after few minutes exposure to the electric current the PTFE/silver electrode turned into brown from its initial silvery color and the spectra became concentration-independent, therefore each sample board was used for one 3 × 5 s measurement set, only. According to the sampling frequency and acquisition time the FFT spectra were obtained in 0–500 Hz ranges with 0.2 Hz data interval. On samples deposited with 12,500 laser pulses the U= constant components were in 1.2–1.4 V range, however, these values, the standard deviation and the ratio of these did not show any tendentious dependence on the cholesterol concentration. This may be attributed to the small variation of the conductive and morphologic properties from layer to layer. Differently from these, the y(f) spectra of the voltage fluctuation strongly depended on cholesterol concentration in the 0.6–100 Hz range (Fig. 4), which range was further used for the Principal Component Analysis. The power spectral density of the fluctuation measured at the presence of the conductive pure solvent is the cumulative result of the dynamics of charge carrier motion in the liquid, the sensor material and the charge exchange at the electrode surface. The presence of cholesterol and the dynamics of its electro-oxidation at the surface of the electrode resulted in a decrease of the detected noise at the studied range. An intensive peak appeared at 50 Hz corresponding to the mains frequency, which was cut off before the further data processing. Among the data of the PC (SCORES) matrix given by PCA analysis the first two columns (PC-1 and PC-2) have the highest significance.
Fig. 3. Electron microscopic image showing the rough morphology of a composite layer deposited with 12,500 pulses (a) and the distribution of the fluorine (b) and silver (c) atoms on the corresponding area.
120
T. Smausz et al. / Applied Surface Science 278 (2013) 117–121
able to detect multiple components based on the signal supplied by the sensor. 4. Conclusions
Fig. 4. Corrected spectra of voltage fluctuation recorded for different cholesterol concentrations.
When plotting (PC-1, PC-2) data points, those corresponding to similar spectra are closely grouped, while the highest is the difference in spectra the larger is the separation distance between the data points in the PC-1–PC-2 plane. As Fig. 5 shows, the measurement points for the composite electrodes at different cholesterol concentrations are distributed along a line in an order corresponding to the cholesterol concentration. The closely grouped points for to the uncoated Au electrode do not show such an obvious distribution. The results of PCA analysis on samples produced with 7500 laser pulses also showed a concentration dependence, however the separation of the data points on the (PC-1, PC-2) plot was about ten times smaller as compared to previous measurements and was not significantly larger than for those corresponding to the reference Au electrodes. This indicates that when increasing the roughness of the composite electrodes the spectral difference of the recorded voltage fluctuation also increased. The presence of other electro-active (interfering) compounds would probably further influences the PSD spectra. Although the sensors often do not produce a linear response, i.e. the spectrum of a multi-compound sample is not a linear superposition of the elementary characteristic spectra (due to the interaction between the compounds) [25] a proper pattern recognition method can be
Fig. 5. Result of PCA analysis represented in PC-1–PC-2 plane: in contrast to the data points corresponding to the spectra measured on uncoated Au electrodes (empty symbols) the distribution for the composite electrodes (solid symbols) show an obvious tendency (arrow) as the function of the cholesterol concentration.
PTFE/Ag composite layers with high specific surface were produced by pulsed laser deposition method and used as active electrodes in fluctuation-enhanced detection method. The results showed that when driving a constant current through the electrodes covered with cholesterol solutions the spectra of the voltage fluctuation measured between the electrodes is concentration dependent. The Principal Component Analysis of the spectra showed a tendentious displacement of the corresponding data points for different cholesterol concentrations as compared to the solvent. The separation of the data points depended on the morphology of the layers: the rougher structure with higher specific surface resulted in more significant separation as the function of the cholesterol concentration. Although, there are different aspects to be studied in detail (optimization of experimental parameters for increasing the spectral differentiation, cross-selectivity with interfering chemical compounds) these preliminary results indicate that the FES technique combined with high specific surface composite electrode may be a useful tool for enzymatic-free cholesterol detection. Acknowledgements The study was funded by the National Development Agency of Hungary with financial support from the Research and Technology Innovation Fund (OTKA-CNK-78549 and OTKA-K-67818). References [1] S. Nambiar, J.T.W. Yeow, Conductive polymer-based sensors for biomedical applications, Biosensors and Bioelectronics 26 (2011) 1825–1832. [2] M.V. Fuke, A. Vijayan, M. Kulkarni, R. Hawaldar, R.C. Aiyer, Evaluation of copolyaniline nanocomposite thin films as humidity sensor, Talanta 76 (2008) 1035–1040. [3] A. Choudhury, Polyaniline/silver nanocomposites: dielectric properties and ethanol vapour sensitivity, Sensors and Actuators B 138 (2009) 318–325. [4] M. Nebel, S. Neugebauer, H. Kiesele, W. Schuhmann, Local reactivity of diamond-like carbon modified PTFE membranes used in SO2 sensors, Electrochimica Acta 55 (2010) 7923–7928. [5] M. Wienecke, M.-C. Bunescu, M. Pietrzak, K. Deistung, P. Fedtke, PTFE membrane electrodes with increased sensitivity for gas sensor applications, Synthetic Metals 138 (2003) 165–171. [6] S.T. Li, E. Arenholz, J. Heitz, D. Bauerle, Pulsed-laser deposition of crystalline Teflon (PTFE) films, Applied Surface Science 125 (1998) 17–22. [7] G.B. Blanchet, S.I. Shah, Deposition of polytetrafluoroethylene films by laser ablation, Applied Physics Letters 62 (1993) 1026–1028. [8] T. Smausz, B. Hopp, N. Kresz, Pulsed laser deposition of compact high adhesion PTFE thin films, Journal of Physics D: Applied Physics 35 (2002) 1859–1863. [9] G. Kecskeméti, B. Hopp, T. Smausz, Zs. Tóth, G. Szabó, Production of porous PTFE-Ag composite thin films by pulsed lased deposition, Applied Surface Science 258 (2012) 7982–7988. [10] J.M.C.S. Magalhães, A.A.S.C. Machado, Urea potentiometric biosensor based on urease immobilized on chitosan membranes, Talanta 47 (1998) 183–191. [11] R. Nenkova, D. Ivanova, J. Vladimirova, T. Godjevargova, New amperometric glucose biosensor based on cross-linking of glucose oxidase on silica gel/multiwalled carbon nanotubes/polyacrylonitrile nanocomposite film, Sensors and Actuators B 148 (1) (2010) 59–65. [12] P.-C. Nien, P.-Y. Chen, K.-C. Ho, Fabricating an amperometric cholesterol biosensor by a covalent linkage between poly(3-thiopheneacetic acid) and cholesterol oxidase, Sensors 9 (2009) 1794–1806. [13] S. Saha, S.K. Arya, S.P. Singh, K. Sreenivas, B.D. Malhotra, V. Gupta, Nanoporous cerium oxide thin film for glucose biosensor, Biosensors and Bioelectronics 24 (2009) 2040–2045. [14] M. Guo, J. Chen, J. Li, L. Nie, S. Yao, Carbon nanotubes-based amperometric cholesterol biosensor fabricated through layer-by-layer technique, Electroanalysis 16 (23) (2004) 1992–1998. [15] X. Wang, Y. Zhang, C.E. Banks, Q. Chen, X. Ji, Non-enzymatic amperometric glucose biosensor based on nickel hexacyanoferrate nanoparticle film modified electrodes, Colloids and Surfaces B 78 (2) (2010) 363–366. [16] Y.-J. Lee, J.-Y. Park, Nonenzymatic free-cholesterol detection via a modified highly sensitive macroporous gold electrode with platinum nanoparticles, Biosensors and Bioelectronics 26 (2010) 1353–1358.
T. Smausz et al. / Applied Surface Science 278 (2013) 117–121 [17] Y. Li, H. Bai, Q. Liu, J. Bao, M. Han, Z. Dai, A nonenzymatic cholesterol sensor constructed by using porous tubular silver nanoparticles, Biosensors and Bioelectronics 25 (2010) 2356–2360. [18] G. Schmera, C. Kwan, P.M. Ajayan, R. Vajtai, L.B. Kish, Fluctuation-enhanced sensing: status and perspectives, IEEE Sensors Journal 8 (2008) 714–719. [19] L.B. Kish, J. Smulko, P. Heszler, C.-G. Granqvist, On the sensitivity selectivity, sensory information and optimal size of resistive chemical sensors, Nanotechnology Perceptions 3 (2007) 43–52. [20] C. Kwan, B. Ayhan, G. Chen, J. Wang, B. Ji, C.-I. Chang, A novel approach for spectral unmixing, classification, and concentration estimation of chemical and biological agents, IEEE Transactions on Geoscience and Remote Sensing 44 (2006) 409–419. [21] G. Schmera, J.M. Smulko, L.B. Kish, P. Heszler, C.-G. Granqvist, Advanced agent identification with fluctuation-enhanced sensing, IEEE Sensors Journal 8 (2008) 706–713.
121
[22] J.R. Biard, L.B. Kish, Enhancing the sensitivity of the SEPTIC bacterium sensing method by concentrating the phage-infected bacteria via DC electrical current, Fluctuation and Noise Letters 5 (2005) 153–158. [23] H. Haspel, R. Ionescu, P. Heszler, A. Kukovecz, Z. Konya, Z. Gingl, J. Maklin, T. Mustonen, N. Halonen, K. Kordas, R. Vajtai, P.M. Ajayan, Fluctuation enhanced gas sensing on functionalized carbon nanotube thin films, Physica Status Solidi (B) 245 (2008) 2339–2342. [24] D. Molnar, P. Heszler, R. Mingesz, Z. Gingl, A. Kukovecz, Z. Konya, H. Haspel, M. Mohl, A. Sapi, I. Kiricsi, K. Kordas, J. Maklin, N. Halonen, G. Toth, H. Moilanen, S. Roth, R. Vajtai, P. Ajayan, Y. Pouillon, A. Rubio, Increasing chemical selectivity of carbon nanotube-based sensors by fluctuation-enhanced sensing, Fluctuation and Noise Letters 9 (2010) 277–287. [25] J.L. Solis, G. Seeton, L.B. Kish, Agent-induced excess noise in commercial chemical sensors, Proceedings of the SPIE 5115 (2003), http://dx.doi.org/10.1117/12.497153.
Contents lists available at SciVerse ScienceDirect
Applied Surface Science journal homepage: www.elsevier.com/locate/apsusc
Study on the applicability of polytetrafluoroethylene–silver composite thin films as sensor material Tomi Smausz a,∗ , Gabriella Kecskeméti a , Tamás Csizmadia a , Ferenc Benedek b , Béla Hopp a a b
Department of Optics and Quantum Electronics, University of Szeged, H-6720 Szeged, Dóm tér 9., Hungary Institute for Engineering and Materials Science, University of Szeged, H-6720 Szeged, Tisza Lajos Krt. 103., Hungary
a r t i c l e
i n f o
Article history: Received 15 June 2012 Received in revised form 6 November 2012 Accepted 8 January 2013 Available online 23 January 2013 Keywords: Pulsed laser deposition Composite layer Silver PTFE Fluctuation-enhanced sensing Cholesterol
a b s t r a c t A study on applicability of conductive high specific surface PTFE/Ag composite layers as active electrodes of a non-enzymatic cholesterol sensor is presented. The composite layers were prepared on one of the two neighboring electrode of a printed circuit board by pulsed laser deposition technique where targets composed of silver and PTFE were ablated by an ArF excimer laser. Cholesterol was dissolved in 0.1 M NaOH in different concentrations in 0–5 mM range. A drop of cholesterol covered the two electrodes and a constant current of 10 A was forced through the sample while the voltage between the electrodes was measured by means of a high resolution A/D converter with 1 kHz sampling rate for 5 s periods. Instead of the time-averaged signal monitoring we applied the Fluctuation-Enhanced Sensing (FES) method which is based on the analysis of the stochastic component of the signal. The power spectral density of the fluctuation was found to be dependent on the cholesterol concentration of the samples. Principal Component Analysis method was used for quantifying the difference between the recorded spectra. A tendentious variation of the spectral properties as the function of the cholesterol concentration was observed. The results indicate that the FES technique combined with high specific surface composite electrodes may be a useful tool for cholesterol detection. © 2013 Elsevier B.V. All rights reserved.
1. Introduction Polymer-metal composites are good candidates for sensor materials. Their electrical, optical and dielectric behavior is strongly influenced by the metal content, the size and shape of the particles [1–3]. One of the promising basic materials for composite preparation is the Teflon (polytetrafluoroethylene, PTFE) which has good mechanical, thermal and chemical stability [4,5]. PTFE thin films can be prepared by various techniques depending on the used substrate and quality requirements. The pulsed laser deposition (PLD) allows the deposition of stoichiometric Teflon thin films with compact structure or sponge-like morphology depending on the deposition parameters and post-treatment [6–8]. Recent studies showed that PTFE/silver composite structures deposited by PLD using PTFE/Ag targets have a rough morphology with increased specific surface attributed to the deposition of PTFE grains and show improved conductive and wetting properties due to the Ag content [9]. Conductive layers with high specific surface can find applications in the field of biosensors, too. Amperometric and potentiometric devices often aiming the detection of analytes with
∗ Corresponding author. Tel.: +36 62 544657; fax: +36 62 544658. E-mail address: [email protected] (T. Smausz). 0169-4332/$ – see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.apsusc.2013.01.051
clinical importance (e.g. urea, glucose, cholesterol) are mainly based on incorporation of enzymes into the active electrodes [10–14]. The enzyme-based sensors can have good selectivity and high sensitivity, however, the enzyme immobilization process is the most difficult step of the production process. In the last few years several attempts were made for fabricating enzyme-free biosensors. Wang et al. produced an amperometric glucose biosensor based on the modification of functional nickel hexacyanoferrate nanoparticles [15]. Lee and Park used a macroporous gold electrode with incorporated Pt [16], while Li et al. built porous tubular silver nanoparticles [17] for detecting cholesterol. In these amperometric measurements charges involved in the electrocathalitic oxidation of cholesterol at the electrode surface were detected, therefore the increased specific surface of the electrode is of high importance. Most of the detection techniques are based on the measurement of the time-averaged value of the sensor signal, however, also the stochastic component (noise) can serve important information on the detector’s ambience. For example, in case when detector signal is the result of the desorption of the analyte at the sensing surface the dynamics of the adsorption–desorption and the diffusion properties can serve as a “fingerprint” for a given analyte. In the Fluctuation-Enhanced Sensing (FES) the low amplitude timevarying components are amplified and statistically analyzed to find the “fingerprints” of the analytes [18,19]. While for simultaneous detection of multiple components a number of different sensors
118
T. Smausz et al. / Applied Surface Science 278 (2013) 117–121
have to be used, with use of proper analysis, pattern database and pattern recognition tools FES can allow their detection using one single sensor [20,21]. The FES method was introduced and mainly used for increasing the selectivity of gas sensors, its usefulness in liquid phase was demonstrated for detection of bacteria [22]. In latter case a voltage fluctuation between two electrodes caused by ions emitted by phage infected bacteria was detected. The aim of present work is a preliminary study on the applicability of our laser deposited PTFE/Ag composite layers as sensing electrodes for the non-enzymatic detection of cholesterol. Differently from conventional amperometric methods, we intended to study whether the cholesterol concentration can influence the noise of the detected signal which is substantial from the point of view of fluctuation-enhanced sensing.
The target rotated with 10 rpm was ablated with pulses of an ArF excimer laser ( = 193 nm, FWHM = 20 ns) having 8 J/cm2 fluence at an area of 0.8 mm2 , the repetition rate was 10 Hz. The relatively low deposition rate does not allow the formation of sandwich structure. Layers were deposited with 7500 and 12,500 pulses, respectively to obtain composite electrodes with different roughness. During deposition the substrates (sample board or glass plate for electron microscopic investigations) were placed onto a 150 ◦ C heatable holder. The base pressure of the PLD chamber was about 2 × 10−3 Pa and the distance between the target and the substrate was 4 cm. The morphology of the layers was studied with a Hitachi S4700 scanning electron microscope, while their elemental composition was analyzed with the energy-dispersive X-ray spectrometer (EDX) of the electron microscope.
2. Experimental
2.2. The fluctuation-enhanced sensing
2.1. Thin film deposition
Cholesterol was dissolved in 0.1 M NaOH solvent containing 2 v% Triton X-100 (Sigma) in different concentrations: 0.5, 1, 2 and 3.5 and 5 mM, respectively. (The normal cholesterol concentration is below 5 mM in human blood.) A drop of solution was placed onto the sample board to overlap the two adjacent electrodes, one being covered with the PTFE/Ag composite layer, as shown in Fig. 2. Reference measurements on untreated Au electrodes were also carried out. A constant current of 10 A was drawn through the circuit and the U(t) voltage between the two adjacent electrodes was measured with a sampling rate of 1000 Hz for three consecutive periods of
PTFE/Ag composite layers were prepared by pulsed laser deposition method onto one of the electrodes of printed circuit sample boards containing two pairs of 2 mm × 2 mm gold plated electrodes as shown in Fig. 1. The experimental parameters were chosen based on the results of our earlier studies [8,9]. The rotating disk form targets were composed of two sectors of circles: 1/6 part PTFE (Goodfellow, grain size 6–9 m compacted at 520 MPa pressure) and 5/6 part Ag (Goodfellow, 2 mm thickness, purity 99.95%).
Fig. 1. Set-up used for composite layer deposition (a) and photograph of a sample board with one covered electrode (b).
Cholesterol Solution Gold plated Deposited layer electrode
a
U(t)
V A
n
samples
S(f)
2
y=S(f)/U=
y1 ...
b
Measurements on
FFT Power Spectral Density
yn
Principal Component Analysis
SCORES
PLOTTING (PC-1, PC-2)
Fig. 2. Scheme of the measurement procedure of FES method: obtaining individual noise spectrum for each sample (a) and comparison of spectra by means of PCA (b).
T. Smausz et al. / Applied Surface Science 278 (2013) 117–121
119
5 s. A home-made A/D converter allowed the digital recording of the time-varying component (>0.2 Hz) of the U(t) voltage with a resolution of 38 nV (least significant bit). The S(f) power spectral density was obtained by FFT and y(f) = S(f)/U= 2 function (corrected spectrum) was calculated, where U= is the constant component of the U(t) signal. For a quantitative comparison of the y(f) functions obtained at different cholesterol concentrations we used the Principal Component Analysis (PCA) method, one of the pattern recognition tools used in sensor technology [23,24] and it was performed using the MATLAB software. The PCA transforms the array formed of rows consisting of PSD data into PC array, where the most of the information about the similarity/difference of the spectra is preserved into the first few columns. 3. Results and discussions 3.1. Thin film characterization Our earlier study [8] demonstrated the PFFE is mainly deposited in form of grains and larger clusters as also shown on the electron microscopic image (Fig. 3a) of the layers. According to elemental composition maps (Fig. 3b and c) recorded with the EDX spectrometer the silver has a more uniform distribution, the darker areas in Fig. 3c can be attributed to the shielding effect of the PTFE grains, since their size is larger than the detection depth (∼1 m) of the EDX. Due to the higher ablation rate of the Teflon, the silver content of the deposit is lower than that of the target [9]. Moreover, as the consequence of the alternate ablation of PTFE and silver the Teflon grains are covered with silver forming a conductive film on their surface. The applied experimental setup results in layers with the following properties: approximately 20 wt% silver content, rough surface with an average thickness of 3 and 5 m for the applied 7500 and 12,500 pulses, respectively. The thicker layer showed an increased roughness as compared to the thinner one. The void content of the layers is under the rough surface profile (measurable with a stylus profiler) is around 48–50%. 3.2. The fluctuation-enhanced sensing measurements Preliminary measurements showed that after few minutes exposure to the electric current the PTFE/silver electrode turned into brown from its initial silvery color and the spectra became concentration-independent, therefore each sample board was used for one 3 × 5 s measurement set, only. According to the sampling frequency and acquisition time the FFT spectra were obtained in 0–500 Hz ranges with 0.2 Hz data interval. On samples deposited with 12,500 laser pulses the U= constant components were in 1.2–1.4 V range, however, these values, the standard deviation and the ratio of these did not show any tendentious dependence on the cholesterol concentration. This may be attributed to the small variation of the conductive and morphologic properties from layer to layer. Differently from these, the y(f) spectra of the voltage fluctuation strongly depended on cholesterol concentration in the 0.6–100 Hz range (Fig. 4), which range was further used for the Principal Component Analysis. The power spectral density of the fluctuation measured at the presence of the conductive pure solvent is the cumulative result of the dynamics of charge carrier motion in the liquid, the sensor material and the charge exchange at the electrode surface. The presence of cholesterol and the dynamics of its electro-oxidation at the surface of the electrode resulted in a decrease of the detected noise at the studied range. An intensive peak appeared at 50 Hz corresponding to the mains frequency, which was cut off before the further data processing. Among the data of the PC (SCORES) matrix given by PCA analysis the first two columns (PC-1 and PC-2) have the highest significance.
Fig. 3. Electron microscopic image showing the rough morphology of a composite layer deposited with 12,500 pulses (a) and the distribution of the fluorine (b) and silver (c) atoms on the corresponding area.
120
T. Smausz et al. / Applied Surface Science 278 (2013) 117–121
able to detect multiple components based on the signal supplied by the sensor. 4. Conclusions
Fig. 4. Corrected spectra of voltage fluctuation recorded for different cholesterol concentrations.
When plotting (PC-1, PC-2) data points, those corresponding to similar spectra are closely grouped, while the highest is the difference in spectra the larger is the separation distance between the data points in the PC-1–PC-2 plane. As Fig. 5 shows, the measurement points for the composite electrodes at different cholesterol concentrations are distributed along a line in an order corresponding to the cholesterol concentration. The closely grouped points for to the uncoated Au electrode do not show such an obvious distribution. The results of PCA analysis on samples produced with 7500 laser pulses also showed a concentration dependence, however the separation of the data points on the (PC-1, PC-2) plot was about ten times smaller as compared to previous measurements and was not significantly larger than for those corresponding to the reference Au electrodes. This indicates that when increasing the roughness of the composite electrodes the spectral difference of the recorded voltage fluctuation also increased. The presence of other electro-active (interfering) compounds would probably further influences the PSD spectra. Although the sensors often do not produce a linear response, i.e. the spectrum of a multi-compound sample is not a linear superposition of the elementary characteristic spectra (due to the interaction between the compounds) [25] a proper pattern recognition method can be
Fig. 5. Result of PCA analysis represented in PC-1–PC-2 plane: in contrast to the data points corresponding to the spectra measured on uncoated Au electrodes (empty symbols) the distribution for the composite electrodes (solid symbols) show an obvious tendency (arrow) as the function of the cholesterol concentration.
PTFE/Ag composite layers with high specific surface were produced by pulsed laser deposition method and used as active electrodes in fluctuation-enhanced detection method. The results showed that when driving a constant current through the electrodes covered with cholesterol solutions the spectra of the voltage fluctuation measured between the electrodes is concentration dependent. The Principal Component Analysis of the spectra showed a tendentious displacement of the corresponding data points for different cholesterol concentrations as compared to the solvent. The separation of the data points depended on the morphology of the layers: the rougher structure with higher specific surface resulted in more significant separation as the function of the cholesterol concentration. Although, there are different aspects to be studied in detail (optimization of experimental parameters for increasing the spectral differentiation, cross-selectivity with interfering chemical compounds) these preliminary results indicate that the FES technique combined with high specific surface composite electrode may be a useful tool for enzymatic-free cholesterol detection. Acknowledgements The study was funded by the National Development Agency of Hungary with financial support from the Research and Technology Innovation Fund (OTKA-CNK-78549 and OTKA-K-67818). References [1] S. Nambiar, J.T.W. Yeow, Conductive polymer-based sensors for biomedical applications, Biosensors and Bioelectronics 26 (2011) 1825–1832. [2] M.V. Fuke, A. Vijayan, M. Kulkarni, R. Hawaldar, R.C. Aiyer, Evaluation of copolyaniline nanocomposite thin films as humidity sensor, Talanta 76 (2008) 1035–1040. [3] A. Choudhury, Polyaniline/silver nanocomposites: dielectric properties and ethanol vapour sensitivity, Sensors and Actuators B 138 (2009) 318–325. [4] M. Nebel, S. Neugebauer, H. Kiesele, W. Schuhmann, Local reactivity of diamond-like carbon modified PTFE membranes used in SO2 sensors, Electrochimica Acta 55 (2010) 7923–7928. [5] M. Wienecke, M.-C. Bunescu, M. Pietrzak, K. Deistung, P. Fedtke, PTFE membrane electrodes with increased sensitivity for gas sensor applications, Synthetic Metals 138 (2003) 165–171. [6] S.T. Li, E. Arenholz, J. Heitz, D. Bauerle, Pulsed-laser deposition of crystalline Teflon (PTFE) films, Applied Surface Science 125 (1998) 17–22. [7] G.B. Blanchet, S.I. Shah, Deposition of polytetrafluoroethylene films by laser ablation, Applied Physics Letters 62 (1993) 1026–1028. [8] T. Smausz, B. Hopp, N. Kresz, Pulsed laser deposition of compact high adhesion PTFE thin films, Journal of Physics D: Applied Physics 35 (2002) 1859–1863. [9] G. Kecskeméti, B. Hopp, T. Smausz, Zs. Tóth, G. Szabó, Production of porous PTFE-Ag composite thin films by pulsed lased deposition, Applied Surface Science 258 (2012) 7982–7988. [10] J.M.C.S. Magalhães, A.A.S.C. Machado, Urea potentiometric biosensor based on urease immobilized on chitosan membranes, Talanta 47 (1998) 183–191. [11] R. Nenkova, D. Ivanova, J. Vladimirova, T. Godjevargova, New amperometric glucose biosensor based on cross-linking of glucose oxidase on silica gel/multiwalled carbon nanotubes/polyacrylonitrile nanocomposite film, Sensors and Actuators B 148 (1) (2010) 59–65. [12] P.-C. Nien, P.-Y. Chen, K.-C. Ho, Fabricating an amperometric cholesterol biosensor by a covalent linkage between poly(3-thiopheneacetic acid) and cholesterol oxidase, Sensors 9 (2009) 1794–1806. [13] S. Saha, S.K. Arya, S.P. Singh, K. Sreenivas, B.D. Malhotra, V. Gupta, Nanoporous cerium oxide thin film for glucose biosensor, Biosensors and Bioelectronics 24 (2009) 2040–2045. [14] M. Guo, J. Chen, J. Li, L. Nie, S. Yao, Carbon nanotubes-based amperometric cholesterol biosensor fabricated through layer-by-layer technique, Electroanalysis 16 (23) (2004) 1992–1998. [15] X. Wang, Y. Zhang, C.E. Banks, Q. Chen, X. Ji, Non-enzymatic amperometric glucose biosensor based on nickel hexacyanoferrate nanoparticle film modified electrodes, Colloids and Surfaces B 78 (2) (2010) 363–366. [16] Y.-J. Lee, J.-Y. Park, Nonenzymatic free-cholesterol detection via a modified highly sensitive macroporous gold electrode with platinum nanoparticles, Biosensors and Bioelectronics 26 (2010) 1353–1358.
T. Smausz et al. / Applied Surface Science 278 (2013) 117–121 [17] Y. Li, H. Bai, Q. Liu, J. Bao, M. Han, Z. Dai, A nonenzymatic cholesterol sensor constructed by using porous tubular silver nanoparticles, Biosensors and Bioelectronics 25 (2010) 2356–2360. [18] G. Schmera, C. Kwan, P.M. Ajayan, R. Vajtai, L.B. Kish, Fluctuation-enhanced sensing: status and perspectives, IEEE Sensors Journal 8 (2008) 714–719. [19] L.B. Kish, J. Smulko, P. Heszler, C.-G. Granqvist, On the sensitivity selectivity, sensory information and optimal size of resistive chemical sensors, Nanotechnology Perceptions 3 (2007) 43–52. [20] C. Kwan, B. Ayhan, G. Chen, J. Wang, B. Ji, C.-I. Chang, A novel approach for spectral unmixing, classification, and concentration estimation of chemical and biological agents, IEEE Transactions on Geoscience and Remote Sensing 44 (2006) 409–419. [21] G. Schmera, J.M. Smulko, L.B. Kish, P. Heszler, C.-G. Granqvist, Advanced agent identification with fluctuation-enhanced sensing, IEEE Sensors Journal 8 (2008) 706–713.
121
[22] J.R. Biard, L.B. Kish, Enhancing the sensitivity of the SEPTIC bacterium sensing method by concentrating the phage-infected bacteria via DC electrical current, Fluctuation and Noise Letters 5 (2005) 153–158. [23] H. Haspel, R. Ionescu, P. Heszler, A. Kukovecz, Z. Konya, Z. Gingl, J. Maklin, T. Mustonen, N. Halonen, K. Kordas, R. Vajtai, P.M. Ajayan, Fluctuation enhanced gas sensing on functionalized carbon nanotube thin films, Physica Status Solidi (B) 245 (2008) 2339–2342. [24] D. Molnar, P. Heszler, R. Mingesz, Z. Gingl, A. Kukovecz, Z. Konya, H. Haspel, M. Mohl, A. Sapi, I. Kiricsi, K. Kordas, J. Maklin, N. Halonen, G. Toth, H. Moilanen, S. Roth, R. Vajtai, P. Ajayan, Y. Pouillon, A. Rubio, Increasing chemical selectivity of carbon nanotube-based sensors by fluctuation-enhanced sensing, Fluctuation and Noise Letters 9 (2010) 277–287. [25] J.L. Solis, G. Seeton, L.B. Kish, Agent-induced excess noise in commercial chemical sensors, Proceedings of the SPIE 5115 (2003), http://dx.doi.org/10.1117/12.497153.