Biosensor Arrays Based on the Degradation of Thin Polymer Films

Oct 24, 2007 - Xinwei Zheng , Joseph P. Cook , Michael Watkinson , Shoufeng Yang , Ian Douglas , Andrew Rawlinson , Steffi Krause. Faraday Discussions...
0 downloads 0 Views 202KB Size
Download PDF
Anal. Chem. 2007, 79, 8974-8978

Biosensor Arrays Based on the Degradation of Thin Polymer Films Interrogated by Scanning Photoinduced Impedance Microscopy Yinglin Zhou,† Shihong Jiang,† Steffi Krause,*,† and Jean-Noe 1 l Chazalviel‡

School of Engineering and Materials Science, Queen Mary University of London, London, E1 4NS, U.K., and Physique de la Matie` re Condense´ e, EÄ cole Polytechnique, CNRS, 91128 Palaiseau, France

Disposable sensors based on the degradation of thin films as a result of an enzymatic reaction have been developed into efficient enzyme detectors. Film degradation has traditionally been monitored using surface plasmon resonance (SPR), quartz crystal microbalance (QCM), or classical ac impedance measurements. The enzyme detection principle has now been integrated with an array technology derived from a recently developed impedance imaging technique, scanning photoinduced impedance microscopy (SPIM). SPIM is based on photocurrent measurements at field-effect structures. The material under investigation is commonly deposited onto a semiconductor-insulator substrate. In this work, field-effect capacitors were replaced by hydrogenated amorphous silicon (a-Si:H) n-i-p photodiode structures, which have recently been shown to be suitable for SPIM measurements with good lateral resolution. To demonstrate the feasibility of SPIM for the characterization of biosensor arrays, polymer dots of the inert polymer cellulose acetate and an r-chymotrypsin-sensitive poly(ester amide) were deposited onto a-Si:H n-i-p/SiO2 structures and their enzymatic degradation was monitored using a laser scanning setup. Disposable biosensors for the detection of enzymes have been developed recently with the advantage of simplicity of the sensor design and the possibility of designing sensor materials with high specificity.1 The principle of this type of sensor is as follows: A biodegradable material is coated onto a sensor substrate. Due to the direct or indirect reaction of the film with the analyte, e.g., an enzyme, the film degrades. The rate of degradation and in some cases the extent of degradation have been found to depend on the analyte concentration and could therefore be used as the sensor signal. The methods used to monitor the degradation of thin films include electrochemical impedance spectroscopy (EIS),2-5 * Corresponding author. E-mail: [email protected]. † Queen Mary University of London. ‡ EÄ cole Polytechnique. (1) Krause, S.; McNeil, C. J.; Ferna´ndez-Sa´nchez, C.; Sabot, A. Encyclopedia of Sensors; Grimes, C. A., Dickey, E. C., and Pishko, M. V., Eds.; American Scientific Publishers: Stevenson Ranch, CA, 2006; Vol. 9, pp 289-306. (2) McNeil, C. J.; Athey, D.; Ball, M.; Ho, W. O.; Krause, S.; Armstrong, R. D.; Wright, J. D.; Rawson, K. Anal. Chem. 1995, 67, 3928-3935. (3) Saum, A. G. E.; Cumming, R. H.; Rowell, F. J. Biosens. Bioelectron. 1998, 13, 511-518.

8974 Analytical Chemistry, Vol. 79, No. 23, December 1, 2007

holography,6,7 quartz crystal microbalance (QCM),4,8 and surface plasmon resonance (SPR).9,10 For example, McNeil et al. developed a urea sensor that used the alkaline pH change caused by a reaction between urea and urease to degrade a pH-sensitive polymer, a copolymer of methyl methacrylate and methacrylic acid.2 Degradation of the films was accompanied by an increase in capacitance of up to 4 orders of magnitude and was monitored using EIS.2 Millington and Lowe described a holographic sensor that detected trypsin by degrading a gelatin layer.6 Krause's group detected various enzymes by monitoring the degradation of synthetic polymer films using SPR and combined quartz crystal admittance measurements and impedance spectroscopy.8,10 Two systems, a poly(ester amide) degraded by R-chymotrypsin and a dextran hydrogel degraded by dextranase, proved particularly sensitive.10 Enzyme concentrations as low as 40 pM were detected in less than 30 min. In this work, it will be demonstrated that the degradation of polymer in the presence of enzymes can be monitored using a new transducer technology derived from the impedance imaging technique,scanningphotoinducedimpedancemicroscopy(SPIM).11,12 SPIM is based on photocurrent measurements at field-effect structures. The material under investigation is deposited onto a semiconductor/insulator substrate. A bias is applied between the semiconductor and an electrode immersed in the solution in contact with the material under investigation to produce an inversion layer at the semiconductor/insulator interface. A modulated light beam focused into the space-charge region of the semiconductor generates electron-hole pairs, which separate in the inversion layer, causing a local photocurrent to flow through the structure. It has been shown that the photocurrent can be related to the local impedance of the material under investigation.12 (4) Ho, W. O.; Krause, S.; McNeil, C. J.; Pritchard, J. A.; Armstrong, R. D.; Athey, D.; Rawson, K. Anal. Chem. 1999, 71, 1940-1946. (5) Ferna´ndez-Sa´nchez, C.; McNeil, C. J.; Rawson, K.; Nilsson, O. Anal. Chem. 2004, 76, 5649-556. (6) Millington, A. G.; Lowe, C. R. Anal. Chem. 1995, 67, 4229-4233. (7) Mayes, A. G.; Blyth, J.; Millington, R. B.; Lowe, C. R. J. Mol. Recognit. 1998, 11, 168-174. (8) Sabot, A.; Krause, S. Anal. Chem. 2002, 74, 3304-3311. (9) Ballerstadt, R.; Schultz, S. J. Sens. Actuators, B 1998, 46, 50-55. (10) Summer, C.; Sabot, A.; Turner, K.; Krause, S. Anal. Chem. 2000, 72, 52255232. (11) Krause, S.; Talabani, H.; Xu, M.; Moritz, W.; Griffiths, J. Electrochim. Acta 2002, 47, 2143-2148. (12) Krause, S.; Moritz, W.; Talabani, H.; Xu, M.; Sabot, A.; Ensell, G. Electrochim. Acta 2006, 51, 1423-1430. 10.1021/ac071437t CCC: $37.00

© 2007 American Chemical Society Published on Web 10/24/2007

Since SPIM can be used to measure the local impedance of the material, it should be suitable for imaging arrays of different materials on the same substrate. Therefore, it is suggested that SPIM can be used to interrogate biosensor arrays based on the degradation of polymers. To employ SPIM as a technique for interrogating arrays, sufficient resolution of photocurrent measurements needs to be achieved. Similar to light-addressable potentiometric sensors (LAPS),13,14 the resolution of SPIM is limited by the lateral diffusion of charge carriers in the semiconductor substrate. This problem has been tackled by using materials with a short diffusion length of charge carriers such as amorphous silicon.15 However, this kind of structure showed only small photocurrents. Furthermore, the deposition of an oxide on amorphous silicon resulted in large numbers of interface states and high leakage currents. Recently, it has been shown that field-effect structures can be replaced by n-i-p photodiode structures in amorphous silicon as they provide higher photocurrents than field-effect capacitors in amorphous silicon and a lateral resolution of about 10 µm.16 In these structures, the charge separation that causes a photocurrent to flow occurs in the depletion layer of the photodiode structure rather than in an inversion layer formed at the interface between semiconductor and insulator. Hence, no dc voltage needs to be applied to the structure to obtain a photocurrent thereby simplifying the measurement setup, and a good quality insulator with a small interface state density is not required. The electrical behavior of these structures at different frequencies has been described and modeled using a suitable equivalent circuit previously.16 The model polymers chosen to demonstrate the feasibility of using SPIM as a technique for interrogating biosensor arrays include the inert polymer, cellulose acetate, and the biodegradable polymer, poly(ester amide) based on bis(L-phenylalanine)R,ωalkylene diesters. This system was regarded as an ideal model system as it was easy to coat and because it was shown to be degraded rapidly and reproducibly by the protease R-chymotrypsin.8,10 EXPERIMENTAL SECTION Materials and Sample Preparation. The synthesis of the biodegradable polymer, poly(ester amide), MW 6400, was described previously.10 R-Chymotrypsin, type II, MW 25 000, from bovine pancreas, with an activity of 60 units mg-1 was purchased from Aldrich. Cellulose acetate, MW 52 000, was purchased from Fluka. A 0.1 M Na2SO4 solution was used for measuring the lateral resolution on an n-i-p/SiO2 substrate. A 10 mM phosphate buffer pH 7.3 containing 140 mM NaCl was prepared for the degradation experiment. High-purity water (18 MΩ‚cm) was used to prepare all solutions and clean the substrates. Organic solvents such as chloroform and ethyl lactate were purchased from Aldrich. Hydrogenated amorphous silicon (a-Si:H) n-i-p photodiode structures were deposited by low-power plasma-enhanced chemical vapor deposition (PECVD) onto indium tin oxide (ITO)-coated, (13) Hafeman, D. G.; Parce, J. W.; McConnell, H. M. Science 1988, 240, 11821185. (14) Yoshinobu, T.; Ecken, H.; Poghossian, A.; Lu ¨ th, H.; Iwasaki, H. Scho¨ning, M. J. Sens. Actuators, B 2001, 76, 388-392. (15) Moritz, W.; Yoshinobu, T.; Finger, F.; Krause, S.; Martin-Fernandez, M.; Scho ¨ning, M. J. Sens. Actuators, B 2004, 103, 436-441. (16) Zhou, Y.; Chen, L.; Krause, S.; Chazalviel, J.-N. Anal. Chem. 2007, 79, 62086214.

Figure 1. a-Si:H n-i-p photodiode structure for localized impedance measurements.

500 µm thick glass substrates. The n-type layer was doped with 0.5% phosphorus and the p-type layer with 0.1% boron. Some 60 nm of silicon dioxide was deposited on top of the p-type layer by reactive ion sputtering. Prior to film deposition, the a-Si:H n-ip/SiO2 substrates (Figure 1) were cleaned by acetone, 2-propanol, and ultrapure water, and finally dried with a stream of nitrogen. Film Preparation. To obtain a model system for measuring the lateral resolution, cellulose acetate (CA) films were spin-coated from a 5% solution of the polymer in ethyl lactate at a speed of 800 rpm for 5 min. Immediately after spin-coating, a part of the cellulose acetate film was peeled off in order to create areas with different impedance. Afterward, the samples were heated on a hot plate at 200 °C for 5 min to improve the adhesion of the films. Microscope images of the samples showed that sharp polymer edges were produced in this manner.16 For the degradation experiments, arrays of four polymer dots were formed by depositing 0.1 µL drops of solutions of 50 mg/ mL poly(ester amide) in chloroform and 5% cellulose acetate in ethyl lactate onto the n-i-p/SiO2 substrate. The arrays were left to dry overnight. The thickness of the films was measured by a Dektak3ST surface profiler. Experimental Setup for SPIM. The experimental setup for photocurrent measurements consisted of an EG&G 7260 lock-in amplifier, a photodiode laser module (405 nm), a pinhole spatial filter, an LD Plan-Neofluar 40×/0.6 microscope objective with correction ring (Zeiss) to compensate for the spherical aberration when focusing through glass, and a cell mounted onto an XYZ positioning system (Newport Ltd., Newbury, U.K.) (Figure 2). The use of a laser diode module with the short wavelength of 405 nm had the advantage that most of the light was absorbed by the thin silicon layer, i.e., effects caused by the reflection of light from the front of the sample such as broadening of the focus or changes in the photocurrent caused by different reflectivities of coated and uncoated areas rather than by differences in impedance were excluded. The intensity of the laser photodiode was modulated sinusoidally at 10 kHz using the oscillator output of the lock-in amplifier. This frequency was chosen as it provides the maximum photocurrent16 and therefore the best signal-to-noise ratio for photocurrent measurements. The control software was written inhouse using LabView.12 The samples were mounted into a cell with the polymer coating facing the electrolyte solution. A carbon electrode served as the gate electrode, while the contact to the semiconductor was made on the ITO using silver paint. The carbon electrode and ITO were directly connected to the current input of the lock-in amplifier as no dc voltage needed to be applied Analytical Chemistry, Vol. 79, No. 23, December 1, 2007

8975

Figure 2. Experimental setup for SPIM measurements.

Figure 3. Photocurrent line scans across the edge of the cellulose acetate film at 10 kHz using different microscope objectives for focusing the laser: open symbols, 50×/0.5 microscope objective (Olympus, LMPlanFl); closed symbols, LD Plan-Neofluar 40×/0.6 microscope objective with correction ring (Zeiss). Linear regressions of the steepest regions of the line scans are shown including their intercepts with horizontal lines indicating a constant photocurrent on uncoated and coated parts of the sample.

to the structure (see the Introduction).16 During degradation experiments, the solution was stirred continuously using a magnetic stirrer. RESULTS AND DISCUSSION Resolution of SPIM Obtained by n-i-p Photodiode Structures. In this work, SPIM measurements were carried out on n-i-p photodiode structures in amorphous silicon. As shown in a previous investigation,16 these structures provide higher photocurrents and eliminate the necessity of a high-quality insulator with a small interface state density compared to fieldeffect capacitors in amorphous silicon. Amorphous hydrogenated silicon photodiode structures were characterized by focusing a 405 nm, modulated laser beam through the glass substrate and the ITO layer onto the a-Si:H layer. A patterned cellulose acetate film served as the model system for assessing the resolution of photocurrent measurements. To investigate the resolution of photocurrent measurements in this arrangement, the focused 405 nm laser was scanned from the uncoated to the coated part of the sample (Figure 3). Previously, a resolution of 10 µm was obtained by scanning the 8976

Analytical Chemistry, Vol. 79, No. 23, December 1, 2007

laser focused through a 50×/0.5 microscope objective across a cellulose acetate edge (Figure 3, open symbols).16 It was assumed that the resolution was mainly limited by the spherical aberration caused by focusing the laser through a solid glass substrate. In this work, the 50×/0.5 microscope objective was replaced with a 40×/0.6 microscope objective with correction ring that allowed correcting for spherical aberration. A comparison of the photocurrent line scans obtained using both objectives is shown in Figure 3. The resolution was determined from the intercepts of linear fits of the steepest regions of the photocurrent line scans with horizontal lines indicating a constant photocurrent on the uncoated and coated parts of the sample. Correcting for spherical aberration resulted in an improvement of the resolution from 10 to 7 µm (Figure 3, closed symbols). As it has been shown previously that a focus of 1.3 µm can easily be achieved at this wavelength,12 this was somewhat surprising. It is assumed that scattering of light caused by the roughness of the substrate may have contributed to the broadening of the focused laser spot. However, the resolution of 7 µm was deemed sufficient for the characterization of arrays of polymer dots. Biosensor Array Based on the Degradation of Thin Films. To demonstrate the suitability of SPIM for the interrogation of biosensor arrays, dots of poly(ester amide), a material which has been shown to degrade rapidly and reproducibly in the presence of R-chymotrypsin,8,10 and an inert material, cellulose acetate, were deposited on a-Si:H n-i-p/SiO2 substrates. As shown in Figure 4A, an array consisting of two dots of cellulose acetate and two dots of poly(ester amide) on an a-Si:H n-i-p/SiO2 substrate is clearly visible in the photocurrent image measured in the absence of R-chymotrypsin in the buffer solution. After adding 1 µM R-chymotrypsin to the buffer solution, photocurrent images were taken every 10 min to observe the degradation of the films. Each image took 2 min to record. The two dots corresponding to poly(ester amide) were found to degrade gradually with time. Most of the poly(ester amide) in the center of the dot had completely degraded after 1 h (Figure 4B). However it took much longer to degrade the edges of the films. After 3 h, the films of poly(ester amide) were completely degraded (Figure 4C). As expected, the cellulose acetate dots were not affected by the presence of R-chymotrypsin.

Figure 5. Thickness profile of a poly(ester amide) dot on an a-Si:H n-i-p/SiO2 substrate obtained by Dektak3ST surface profiler.

Figure 6. Effect of film thickness on rate of degradation (Rchymotrypsin concentration, 1 µM; film thickness in the range from 1 to 12 µm). The reciprocal current i normalized with the initial current i0 vs time relationship is shown. Arrows indicate the addition of enzyme to the buffer solution.

Figure 4. Photocurrent images of an array of different polymer dots on n-i-p substrate before (A), 50 min (B), and 3 h (C) after adding R-chymotrypsin (1 µM) into pH 7.3 buffer solution (bottom left and top right, cellulose acetate; top left and bottom right, poly(ester amide)).

From the above experiments, it was apparent that the polymer dots formed by drop-coating were not uniform. The polymer thickness in the center part of the dots was much thinner than that at the edge of the dots. This was also confirmed by the thickness measurement of poly(ester amide) films obtained using a Dektak3ST surface profiler (Figure 5). The thickness of the edge of the poly(ester amide) dots reached up to 12 µm, whereas the polymer was about 1 µm thick in the center. Therefore, it took much longer to degrade the edge of the films. To confirm that the different degradation times at the different positions of the poly(ester amide) dots were caused by the

thickness of the film and not by different degradation rates, the photocurrent was monitored at different positions on the poly(ester amide) dots during degradation at a single enzyme concentration. As thickness changes should be directly proportional to changes in film impedance, reciprocal photocurrents normalized with the initial photocurrent were plotted. The slope of the linear part of the degradation curves, which is equivalent to the degradation rate, was found be independent from film thickness (Figure 6). This is consistent with results obtained by SPR using the same poly(ester amide)10 and can be an important factor for batch-to-batch reproducibility of a disposable sensor array. Rather than recording full images of the entire sample, which would have been very time-consuming, a single image was recorded initially and up to six appropriate positions on the sample were chosen from that image. The time dependent measurements shown in Figures 6 and 7 were carried out as four or five point scans, each of which took 8 or 10 s to record. The experimental setup allows for six-point scans to be recorded within less than 1 s. However, faster recording times were deemed unnecessary as 10 s was short compared to the degradation times measured. The change of the reciprocal photocurrent with time with the laser focused onto the center of poly(ester amide) dots was measured at different concentrations of the enzyme R-chymotrypsin (Figure 7). The samples were left in the buffer solution for 10 min. The poly(ester amide) dots were found to be stable in the buffer. After 10 min, different concentrations of enzyme were added into the buffer solution. The higher the concentration of R-chymotrypsin added, the faster the reciprocal photocurrent decreased and the faster the films degraded. The noise of the Analytical Chemistry, Vol. 79, No. 23, December 1, 2007

8977

Figure 7. Change of the reciprocal photocurrent i normalized with the initial photocurrent i0 at 10 kHz of poly(ester amide)-coated n-i-p substrate with time after adding the different concentrations of R-chymotrypsin into pH 7.3 buffer solution: (a) 100 pM; (b) 500 pM; (c) 2 nM; (d) 10 nM; (e) 50 nM; (f) 200 nM.

Figure 8. Relationship between the rate of the change of the normalized reciprocal photocurrent at 10 kHz of poly(ester amide)coated n-i-p substrate and the concentration of R-chymotrypsin on a double-logarithmic scale. Error bars represent the 95% confidence interval derived from five measurements at each enzyme concentration.

signal measured was estimated to be about 5% of the total change in photocurrent for complete degradation (see curves d-f in Figure 7). After an initial period, the degradation was roughly linear with time (Figure 7). The gradients of the linear portions of the curves, which were equivalent to the rates of degradation, were used to produce a calibration curve (Figure 8). Using the rate of change as a measure of enzyme concentration allowed detection of concentrations of 100 pM R-chymotrypsin in less than 40 min. There is a good linear relationship between the logarithm

8978

Analytical Chemistry, Vol. 79, No. 23, December 1, 2007

of the rate of change of the normalized reciprocal photocurrent and the logarithm of the concentration of the enzyme in a range of more than 3 orders of magnitude of the concentration (between 100 pM and 200 nM) (Figure 8). Error bars at each enzyme concentration were calculated from five measurements carried out in parallel using the multipoint scan described above. The response of the system is slightly slower than that measured with SPR previously.10 An induction period was observed before the maximum rate of degradation could be detected by SPIM (Figure 7). The smaller the concentration the longer was the induction period. This behavior can be ascribed to the rough surface of the polymer layer in the present investigation. It is assumed that initially degradation occurred at rough features and edges on the polymer film which were more readily attacked by the enzyme but contributed little to the overall film impedance. Similar behavior was observed when comparing changes in the impedance with changes in the mass during degradation of the same polymer on a gold-coated quartz crystal.8 While the mass of the film decreased instantly upon addition of R-chymotrypsin, a delay was observed before the film capacitance changed significantly.8 CONCLUSIONS It was shown that a-Si:H n-i-p/insulator structures can be used to carry out SPIM measurements with a lateral resolution of about 7 µm. This resolution was deemed sufficient for the characterization of arrays of polymer dots. It was demonstrated that SPIM is suitable for interrogating polymer-based biosensor arrays. R-Chymotrypsin concentrations in a range from 0.1 to 200 nM were detected using poly(ester amide) dots deposited onto a-Si:H n-i-p/insulator structures. It is envisaged that this technology will be used for interrogating biosensor arrays that are capable of detecting different enzymes simultaneously. ACKNOWLEDGMENT The authors thank the EU for providing a Marie-Curie International Incoming Fellowship (MIF1-CT-2005-007808) and EPSRC for providing an Advanced Fellowship and project funding.

Received for review July 6, 2007. Accepted September 10, 2007. AC071437T