Sensors and Actuators B 145 (2010) 98–103
Contents lists available at ScienceDirect
Sensors and Actuators B: Chemical journal homepage: www.elsevier.com/locate/snb
Nanopore immunosensor for peanut protein Ara h1 Rajdeep Singh, Pranav P. Sharma, Ruth E. Baltus, Ian I. Suni ∗ Department of Chemical and Biomolecular Engineering, Center for Advanced Materials Processing (CAMP), Clarkson University, Potsdam, NY 13699 USA
a r t i c l e
i n f o
Article history: Received 29 May 2009 Received in revised form 14 October 2009 Accepted 16 November 2009 Available online 22 November 2009 Keywords: Biosensors Immunosensors Antibodies Diffusion Impedance Food allergens
a b s t r a c t A nanopore immunosensor is demonstrated for peanut protein Ara h1, a common food allergen. This biosensor is constructed by immobilizing the antibody to peanut protein Ara h1 within Au-coated pores of commercial nanoporous polycarbonate membranes. Peanut protein Ara h1 is detected as the change in the pore conductivity as the pores are partially obscured by antigen binding. Control experiments with the antibody to cockroach protein Bla g1 demonstrate that the observed effects do not arise from nonspecific adsorption. Binding of peanut protein is studied as a function of the pore diameter (15, 30 and 50 nm), and as a function of the concentration of peanut protein. The results show the greatest sensitivity for membranes with the smallest pore diameter, and exhibit mass transfer effects for membranes with larger pore diameters. Approximate calculations of diffusion times for transport of peanut protein Ara h1 through these nanoporous membranes suggest that mass transport effects should not be observed. This discrepancy between experimental observations of mass transfer effects and model predictions suggest that diffusion of peanut protein Ara h1 is anomalously slow in this system. © 2009 Elsevier B.V. All rights reserved.
1. Introduction Commercially available nanoporous alumina and polycarbonate membranes have attracted significant attention as templates for fabricating a variety of different nanomaterials [1,2]. Wellordered nanoporous alumina is typically fabricated by anodization techniques, and these methods have been recently extended to formation of organized nanoporous oxides of Ti and several other metals [3]. Electrodeposition and electroless deposition into nanoporous membranes have been employed to fabricate nanowires and nanotubes from a variety of different metals [4,5], including Au, which is biocompatible. Both nanoporous alumina and nanoporous polycarbonate membranes have been explored for applications to sensors and biosensors [6,7], but polycarbonate membranes have the advantage of superior mechanical stability. We recently reported a novel conductivity biosensor involving immobilization of the glucose–galactose receptor (GGR) protein within the Au-coated pores of nanoporous polycarbonate membranes [8]. Upon glucose binding, this protein undergoes a wide amplitude hinge-bending motion that reduces the protein film thickness, and increases the electrolyte conductivity within the membrane. Here we report a similar nanopore biosensor where the electrolyte conductivity within the membrane instead decreases upon analyte binding. In this case, the antibody to peanut protein Ara h1 is immobilized within the membrane nanopores, and
∗ Corresponding author. Tel.: +1 315 268 4471; fax: +1 315 268 6654. E-mail address:
[email protected] (I.I. Suni). 0925-4005/$ – see front matter © 2009 Elsevier B.V. All rights reserved. doi:10.1016/j.snb.2009.11.039
binding of peanut protein Ara h1 partially obscures the nanopores. Peanut protein Ara h1 is a common food allergen [9,10], and is often consumed inadvertently, thus presenting a significant public health hazard [11]. The reported biosensor for this peanut protein may have significant advantages in speed and simplicity over current methods for detecting peanut protein, which are based on either enzyme-linked immunosorbent assays (ELISA) or lateral flow immunoassays [9,12–14]. 2. Materials and methods 2.1. Materials Polycarbonate track-etched membranes with nominal pore size 15, 30 and 50 nm (Catalog numbers 110601, 110602 and 110603) were purchased from SPI Supplies. 11-Mercaptoundecanoic acid (11-MUA) was purchased from Aldrich; N-(3-dimethylaminopropyl)-N -ethylcarbodiimide hydrochloride (EDC), potassium dihydrogen phosphate and dipotassium hydrogen phosphate were purchased from Sigma; N-hydroxysulfosuccinimide sodium salt (NHSS) was purchased from Pierce Biotechnology; peanut protein Ara h1 (monomeric), rabbit polyclonal antibody to Ara h1, and rabbit polyclonal antibody to cockroach protein Bla g1 were purchased from Indoor Biotechnologies; formaldehyde, ammonium hydroxide, sulfuric acid, nitric acid and methanol were purchased from J.T. Baker; tin (II) chloride, sodium sulfite, silver nitrate, potassium chloride and calcium chloride were purchased from Sigma Aldrich; and trifluoroacetic acid was purchased from Superclo. Double distilled deionized water was freshly prepared
R. Singh et al. / Sensors and Actuators B 145 (2010) 98–103
99
and used for all experiments. Au plating solution (Oromerse Part B) was purchased from Technic. All reagents were used as received. 2.2. Au deposition onto and within nanoporous membrane Au deposition involved a multi-step procedure of activation and galvanic/electroless Au deposition [4]. First, Sn2+ ions were adsorbed onto the membrane by immersion for 45 min. into 0.026 M SnCl2 and 0.07 M triflouroacetic acid with 1:1 ratio of water: methanol, and rinsed afterwards to remove excess SnCl2 . Next, the membrane was immersed for 2 min. into 0.029 M AgNO3 saturated with NH3 to reduce Ag+ to neutral Ag. The membrane was then rinsed and refrigerated for 30 min. in methanol. Next, a continuous layer of Au was deposited by immersion for 3 h into the Au plating solution. Prior to use, the Au plating solution (Oromerse Part B) was diluted by 40×, and the pH was reduced to 10 by dropwise addition of 0.5 M H2 SO4 . Diluting the Au plating bath and reducing its pH slows down Au deposition. Following Au deposition, the membrane was immersed for 12 h in 25 vol% HNO3 to remove residual Sn and Ag. 2.3. Membrane characterization Several uncoated polycarbonate membranes were imaged in a JEOL model 7400F field emission scanning electron microscope (SEM) to estimate the true pore diameter. This yielded average () true pore diameters of 18 (2), 27 (2), and 50 (2) nm for the membranes of nominal pore diameter 15, 30 and 50 nm, respectively. The hydrodynamic pore diameter was also determined by measuring pressure drop as a function of flow rate through the membrane, as described elsewhere [15]. This yielded hydrodynamic diameters of 17 (3), 30 (3), and 53 (2) nm for the membranes of nominal pore diameter 15, 30 and 50 nm, respectively. Thus the pore diameters determined from hydrodynamic measurements and from SEM imaging are both quite close to the nominal pore diameter given by the manufacturer. For this reason the nominal pore diameter was used for quantitative analysis below. Several Au-coated polycarbonate membranes were analyzed by SEM to determine the average Au coating thickness. Membrane pieces were attached to an Al stub using carbon epoxy. For each membrane, approximately 5–6 areas were imaged and the average thickness was determined. Averaging these measurements over three membranes yielded an Au thickness of 3 nm. 2.4. Protein immobilization within nanopores After Au nanotube formation, the membrane was immersed into phosphate buffer solution (PBS) for 2 h to enhance wetting. The membrane was then immersed into 1.0 M 11-MUA and 50 mM PBS at pH 10 to form a carboxylate terminated self-assembled monolayer (SAM) [16]. The carboxylate groups were then activated in 75 mM EDC, 15 mM NHSS and 50 mM PBS at pH 7.3. Several drops (about 180 l) of PBS solution containing 0.057 nM antibody to peanut protein Ara h1, or antibody to cockroach protein Bla g1, were placed on one side of the membrane, and antibody immobilization was allowed to proceed for 1 h. This results in antibody immobilization through amide bond formation [17]. Unbound antibody was removed by gently rinsing the membrane in PBS solution.
Fig. 1. Schematic of the electrochemical cell used for impedance measurements, from Ref. [8].
cell contains valves at the bottom to remove the solution, allowing for rapid, complete changeover between different samples [8]. The test cell was then repeatedly filled with solutions of increasing peanut protein Ara h1 concentration, and the conductivity measured between the two Au electrodes. One Au electrode served as the working electrode, while the other served as both the counter and working electrodes. The impedance at 1 kHz was recorded every 5 min. using a Princeton Applied Research (PAR) model 273 potentiostat coupled with a Solartron 1250 frequency response analyzer (FRA) at a probe voltage of 10 mV. At 1.0 kHz the imaginary component of the impedance was negligible for all studies, so only the real component is reported. 3. Results and discussion 3.1. Biosensor principle of operation Figs. 2–4 show the increase in the real component of the impedance (ZRe ) at 1.0 kHz with increasing peanut protein concentration for membranes of nominal pore diameter 15, 30 and 50 nm. In all cases, ZRe increases as peanut protein Ara h1 is introduced. In most electrochemical systems, the high frequency impedance corresponds to the solution phase resistance [18], so ZRe in Figs. 2–4 measures the electrolyte resistance between the two Au electrodes. These results demonstrate that the pore conductivity decreases as the peanut protein concentration increases, since the effective pore diameter is reduced upon binding of peanut protein to its surfaceimmobilized antibody. The antigen concentrations studied here (0.04–0.12 g/ml) are considerably lower than the range of detec-
2.5. Impedance measurements The Au- and antibody-coated membrane was placed between two Au electrodes (SPI Supplies) in a virgin Teflon test cell, as shown in Fig. 1 [8]. In this cell, the membrane was mounted between two Teflon blocks using Chemraz O-rings, with the cell volume of about 1800 l divided equally on both sides of the membrane. This
Fig. 2. ZRe at 1 kHz for membranes with 15 nm pore diameter in 0 g/ml (), 0.04 g/ml (), 0.08 g/ml () and 0.12 g/ml (X) peanut protein Ara h1 in 50 mM PBS buffer at pH 7.3.
100
R. Singh et al. / Sensors and Actuators B 145 (2010) 98–103
Fig. 3. ZRe at 1 kHz for membranes with 30 nm pore diameter in 0 g/ml (), 0.04 g/ml (), 0.08 g/ml () and 0.12 g/ml (X) peanut protein Ara h1 in 50 mM PBS buffer at pH 7.3.
tion limits (1–100 mg/kg) suggested as being necessary for food allergens [19]. Using the density of water, this range of detection limits translates to about 1–100 g/ml. The method reported here for conductivity measurements is used for fuel cell membrane characterization, where the real component of the impedance at 1 kHz is often employed as a method for estimating the ionic conductivity [20,21]. Because significant electrochemistry does not occur at either Au electrode, one Au electrode can serve as the working electrode, and the other as both counter and reference electrode, even though this is not a true reference electrode. Two-electrode conductivity measurements have similarly been employed to estimate the ionic conductivity of fuel cell membranes [20,22]. The impedance measured between the two Au electrodes is [8]: ZT = ZA + ZM + ZB
(1)
where ZA and ZB are the electrolyte impedance (resistance) on either side of the membrane, and ZM is the electrolyte impedance (resistance) through the membrane pores. For the current cell design, the Au electrode positions are not fixed, so ZA and ZB are poorly reproduced when the cell is dissembled and reassembled to change membranes. For this reason, the magnitude of ZRe varies substantially in Figs. 2–4 for successive experiments with membranes of different pore diameter. Although ZRe varies substantially between membranes of different pore diameter, the impedance changes in Figs. 2–4 upon introduction of peanut protein Ara h1 are reproducible to within about 15%. For this reason, only the change in ZRe in Figs. 2–4 upon
Fig. 5. Variation in impedance change with increasing peanut protein Ara h1 concentration for membranes with 15 nm (), 30 nm () and 50 nm () pore diameter in 50 mM PBS buffer at pH 7.3.
introduction of peanut protein Ara h1 is significant. Fig. 5 allows for easier comparison of the impedance change that occurs for each membrane with each step increase in the concentration of peanut protein Ara h1. The results in Figs. 2–4 demonstrate the essential elements of an immunsensor, since the current sensor employs antibodies immobilized within the membrane nanopores, and changes in ZRe correlate to changes in the concentration of peanut protein Ara h1. However, as will be discussed in detail below, detailed analysis indicates that for some conditions, the sensor response is limited by mass transport within the nanopores. 3.2. Test for non-specific adsorption One should also consider whether or not the results of Figs. 2–4 can be ascribed to non-specific adsorption. For example, peanut protein Ara h1 might bind to the internal nanopore surface through gaps within the antibody film. To test for non-specific adsorption, control experiments were performed using membranes with 30 and 50 nm pores within which the rabbit polyclonal antibody to cockroach protein Bla g1 was immobilized. The change in the real component of the impedance at 1 kHz ranged from 7–10 cm2 , and was independent of the concentration of Ara h1. These values are considerably lower than the impedance changes reported in Figs. 3 and 4, demonstrating that these results arise largely from specific biomolecular recognition of the peanut protein antigen by its surface-immobilized antibody. 3.3. Effect of peanut protein concentration and nominal pore diameter The results of Figs. 2–5 show interesting trends with pore size and antigen concentration. Upon introduction of the lowest concentration studied (0.04 g/ml), the largest change in pore conductivity is observed for the smallest pores (15 nm). However, exposing this membrane to further increase in the antigen concentration has little effect on the pore conductivity. On the other hand, for 30 and 50 nm pores, the impedance continues to rise as the antigen concentration is increased. The results for the 15 nm pores can be explained by proposing that the pore mouth becomes obscured at low concentrations of peanut protein, preventing further protein diffusion into the membrane. Assuming complete and uniform film coverage within the nanopores, the effective pore diameter (dpore ) following antigen binding is: dpore = dnominal − 2tAu − 2tantibody − 2tantigen
Fig. 4. ZRe at 1 kHz for membranes with 50 nm pore diameter in 0 g/ml (), 0.04 g/ml (), 0.08 g/ml () and 0.12 g/ml (X) peanut protein Ara h1 in 50 mM PBS buffer at pH 7.3.
(2)
where dnominal is the nominal pore diameter, and tAu , tantibody , and tantigen are the film thickness for Au, antibody, and antigen,
R. Singh et al. / Sensors and Actuators B 145 (2010) 98–103
respectively. Recent studies of protein binding to an antibody film estimated the average thickness of both the antibody and antigen layers as 4 nm [23]. For the 15 nm pores, Eq. (2) then yields a negative value for the effective pore diameter. This supports the argument above that for low peanut protein concentrations, the pore opening becomes smaller than the protein diameter, so diffusion is blocked, even though the measured impedance remains finite. One might consider whether the size of bound peanut protein Ara h1 is consistent with the impedance change measured for the largest pore size (50 nm). Assuming a uniform cross-sectional area along the pore length, the change in resistance through the membrane (ZRe ) is [8]: ZRe =
4L A
1 2 dpore
−
1 (dpore + 2tantigen )
2
101
Table 1 Calculated pore diffusion coefficients and diffusion times for peanut protein Ara h1. Nominal pore size (nm)
dpore (nm)
D∞ (cm2 /s)
Dp (cm2 /s)
(s)
15 30 50
– 8 28
– 0.13 0.04
6.33 × 10−7 6.33 × 10−7 6.33 × 10−7
– 4.69 × 10−7 5.89 × 10−7
– 0.767 0.614
Table 2 Calculated pore diffusion coefficients and diffusion times for glucose. Nominal pore size (nm)
dpore (nm)
D∞ (cm2 /s)
Dp (cm2 /s)
(s)
15 30 50
3 18 38
0.24 0.04 0.02
6.8 × 10−6 6.8 × 10−6 6.8 × 10−6
3.54 × 10−9 6.23 × 10−9 6.53 × 10−9
0.102 0.058 0.055
(3) [25,26]:
where A is the membrane area exposed to the electrolyte, is the nanopore density, is the electrolyte conductivity, and L is the pore length (membrane thickness). For previous studies of glucose sensing [8], time-dependent effects were not observed, suggesting no axial variations in the effective pore diameter. However, the time dependence of ZRe in Figs. 3 and 4 suggests axial variations in dpore upon introduction of peanut protein Ara h1. Although Eq. (3) could be rewritten in differential form, this is not warranted given the approximate nature of our analysis. From the results in Fig. 4 for the highest concentration of Ara h1 and the longest time, the observed value of ZRe corresponds to an antigen film thickness (tanalyte ) of 1.4 nm. This value is of the correct order of magnitude, but is considerably smaller than the diameter of peanut protein Ara h1. Back-of-the-envelope calculations show that this discrepancy likely arises from incomplete surface coverage of the antigen within the nanopores. This may arise from incomplete surface coverage by the antibody film, incomplete saturation with antigen, or inactivation of a substantial fraction of the antibody upon surface immobilization [24].
3.4. Effects of mass transport The results shown in Figs. 3 and 4 differ quantitatively and qualitatively from those shown in Fig. 2. Quantitatively, the introduction of 0.04 g/ml of peanut protein Ara h1 causes smaller changes in pore conductivity for membranes with 30 and 50 nm pores than in membranes with 15 nm pores, since the fractional reduction in the pore cross-sectional area is smaller. Qualitatively, unlike the results in Fig. 2, the results in Figs. 3 and 4 exhibit a discernable upward slope following each successive addition of peanut protein Ara h1. This indicates likely mass transfer effects on sensor performance. Since the pore mouth in the 30 and 50 nm pore membranes remains relatively open, further antigen diffusion into the nanopore is possible, resulting in a gradual increase in the impedance with time. Further evidence for this argument is that similar results for ZRe previously reported for glucose detection do not exhibit a discernable upward slope for membranes with a pore diameter of 30 nm [8]. One can test the argument that detection of peanut protein Ara h1 is at least partially rate-limited by diffusion in 30 and 50 nm pore membranes by estimating protein diffusion times through these nanoporous membranes, and by comparing these to estimates for glucose diffusion. Since proteins have bulk diffusion coefficients about one order of magnitude lower than that of glucose, mass transfer limitation is much more likely to be observed for proteins. Estimates of the intrapore diffusion coefficients for both analytes within the membrane can be estimated using the Rankin equation
Dp = 1 − 2.104 + 2.0903 − 0.9485 D∞
(4)
where D∞ is the bulk diffusion coefficient and is the dimensionless solute size: =
dsolute dpore
(5)
The bulk diffusion coefficient for glucose has been reported as 6.8 × 10−6 cm2 /s [27]. For peanut protein Ara h1, the bulk diffusion coefficient can be estimated from Ref. [28]:
−8
D∞ = 8.34 × 10
T
M 1/3
cm2 /s
(6)
where T is the absolute temperature, is the solvent viscosity (cP), and M is the molecular weight. From the known molecular weight (63.5 kDa) of peanut protein Ara h1, this yields a bulk diffusion coefficient (D∞ ) of about 6.33 × 10−7 cm2 /s. For peanut protein Ara h1, values for can be calculated by determining the pore diameter (dpore ) from Eq. (1) above. On the other hand, glucose binding reduces the thickness of the glucose galactose receptor (GGR) protein film [8], so the last term is neglected for the glucose calculations. In addition, the Au film thickness (tAu ) is neglected, since the Au film in those experiments was not continuous. The GGR Q26C protein film thickness on Au (trecognition ) is taken to be 6 nm [29]. The intrapore diffusion coefficients (Dp ) determined using Eqs. (4)–(6) are summarized in Tables 1 and 2 for peanut protein Ara h1 and glucose. One can then estimate the time ( ) required for diffusion through the nanoporous polycarbonate membrane as: =
L2 Dp
(7)
Since the membrane thickness is about 6 m, the diffusion times ( ) for glucose and peanut protein Ara h1 through the membrane can be calculated, as summarized in Tables 1 and 2. As expected, the diffusion time of glucose through the 6 m thick polycarbonate membrane is rapid, with diffusion times considerably less than 1 s for all pore sizes. This short time is consistent with experimental observations that showed no mass transfer effects for glucose sensing [8]. The effective diffusion coefficient of peanut protein Ara h1 is about one order of magnitude less than that of glucose, so the diffusion time for peanut protein through the membrane is still quite rapid, on the order of 1 s. This is surprising given the experimental evidence from Figs. 3 and 4 that this nanopore biosensor exhibits time lags of the order of minutes to tens of minutes. This suggests that mass transfer of peanut protein Ara h1 in this system is anomalously slow.
102
R. Singh et al. / Sensors and Actuators B 145 (2010) 98–103
Anomalously low effective diffusion coefficients have been reported in a variety of systems, but these have generally been systems of high tortuosity. Surprisingly, only a few authors have measured quantitative diffusion coefficients in systems such as those studied here, with straight, well-defined pores of no or minimal tortuosity [30–33]. For small inorganic cations in nanoporous alumina, some diffusion coefficients obtained from radiotracer experiments were anomalously low [30,31]. However, such phenomena were observed only at very low ionic strength, and this behavior was easily attributed to electrostatic interactions with charged pore walls [31]. Results at higher ionic strength (>0.1 M) yielded diffusion coefficients close to bulk values [31]. The diffusion coefficient for Ru(NH3 )6 3+ in aligned multiwall carbon nanotube membranes was also quite near its bulk value [32]. The only report of truly anomalous diffusion coefficients in straight, well-defined pores was for organic species, including caffeine, methyl orange, and malachite green oxalate, in nanoporous alumina membranes [33]. These authors reported diffusion coefficients about two orders of magnitude less than the expected bulk values. They also note the paucity of quantitative diffusion coefficient measurements in nanoporous membranes [33]. Although diffusion coefficients of peanut protein Ara h1 are not measured here, the results of Figs. 3 and 4, along with the approximate mass transfer calculations given above, suggest anomalously low diffusion coefficients here as well. One possible explanation is repeated collisions between peanut protein Ara h1 and the antibody-coated pore walls during which non-specific interactions occur that slow protein transport. Such adsorptive interactions have been reported to explain anomalously low diffusion coefficients for protein molecules within a porous polymer scaffold [34]. Mass transfer resistance external to the membrane was also considered to explain the anomalously low protein diffusion rates observed here. Because the test cell solution was not stirred, a stagnant boundary layer might contribute to the overall resistance to protein transport into and through the membrane. An analysis developed by Wakeham and Mason can be used to estimate the contribution of boundary layer resistance for this experimental system, which has a pore length much greater than the pore diameter (>100×) and a low pore density. This analysis indicates that boundary layer resistances can increase the effective path length for diffusion by a factor of 0.410 dpore /L [35]. Thus any boundary layer effects external to the membrane should contribute less than 1% of the mass transfer resistance within the membrane. The best pore size for application of the nanopore conductivity biosensor reported here depends on the trade-offs between several considerations. On the one hand, the greatest sensitivity is obtained for the smallest practical pore size, as illustrated here for the 15 nm diameter pores. On the other hand, development of a biosensor with a signal that correlates to the concentration may require a larger pore size, since here the pores appear to be effectively blocked at the lowest concentration of peanut protein Ara h1 in the 15 nm pores. This suggests that measurements to quantify the peanut protein concentration might require larger pore diameters. 4. Conclusions A nanopore immunosensor is demonstrated for peanut protein Ara h1, a common food allergen. This biosensor is constructed by immobilizing the antibody to peanut protein Ara h1 within Au-coated pores of commercial nanoporous polycarbonate membranes. Peanut protein Ara h1 is detected as the change in the pore conductivity as the pore is obscured by antigen binding. Binding of peanut protein is studied in membranes with pore diameters of 15, 30 and 50 nm. For the 15 nm pores, the largest change in pore conductivity is observed for the lowest concentration (0.04 g/ml) studied of peanut protein Ara h1. However, protein binding rapidly
obscures the pore mouth, blocking diffusion of antigen into the nanopores. For the 30 and 50 nm pores, a more modest initial change in pore conductivity is observed, but the pore conductivity gradually increases with time as the antigen diffuses into the nanopores. The observed transient response following step increases in the concentration of peanut protein is inconsistent with approximate calculations of protein diffusion times through the nanoporous membranes, suggesting that diffusion of peanut protein Ara h1 is anomalously slow in this system. Acknowledgements This research was supported by US Army grant W911NF-05-10339. References [1] A. Huczko, Template-based synthesis of nanomaterials, Appl. Phys. A 70 (2000) 365–376. [2] G. Schmid, Materials in nanoporous alumina, J. Mater. Chem. 12 (2002) 1231–1238. [3] J.M. Macak, H. Tsuchiya, A. Ghicov, K. Yasuda, R. Hahn, S. Bauer, P. Schmuki, TiO2 nanotubes: self-organized electrochemical formation, properties and applications, Curr. Opin. Solid State Mater. Sci. 11 (2007) 3–18. [4] V.P. Menon, C.R. Martin, Fabrication and evaluation of nanoelectrode ensembles, Anal. Chem. 87 (1995) 1920–1928. [5] D. Routkevich, A.A. Tager, J. Haruyama, D. Almawlawi, M. Moskovits, J.M. Xu, Nonlithographic nano-wire arrays: fabrication, physics, and device applications, IEEE Trans. Elect. Dev. 43 (1992) 1646–1658. [6] M. Wirtz, M. Parker, Y. Kobayashi, C.R. Martin, Template-synthesized nanotubes for chemical separations and analysis, Chem. Eur. J. 8 (2002) 3572–3578. [7] P. Kohli, M. Wirtz, C.R. Martin, Nanotube membrane based biosensors, Electroanalysis 16 (2004) 9–18. [8] A. Tripathi, J. Wang, L.A. Luck, I.I. Suni, Nanobiosensor design utilizing a periplasmic E. coli receptor protein immobilized within Au/polycarbonate nanopores, Anal. Chem. 79 (2007) 1266–1270. [9] A. Pomes, R.M. Helm, G.A. Bannon, A.W. Burks, A. Tsay, M.D. Chapman, Monitoring peanut allergen in food products by measuring Ara h1, J. Allergy Clin. Immunol. 111 (2003) 640–645. [10] G.W. Palmer, D.A. Dibbern, A.W. Burks, G.A. Bannon, S.A. Bock, H.S. Porterfield, R.A. McDermott, S.C. Dreskin, Comparative potency of Ara h1 and Ara h2 in immunochemical and functional assays of allergenicity, Clin. Immunol. 115 (2005) 302–312. [11] T.J. Furlong, J. DeSimone, H. Sicherer, Peanut and tree nut allergic reactions in restaurants and other food establishments, J. Allergy Clin. Immunol. 108 (2001) 867–870. [12] M.L. Nogueira, R. McDonald, C. Westphal, Can commercial peanut assay kits detect peanut allergens? J. AOAC Int. 87 (2004) 1480–1484. [13] D.A. Schmitt, H. Cheng, S.J. Maleki, A.W. Burks, Competitive inhibition ELISA for quantification of Ara h1 and Ara h2, the major allergens of peanuts, J. AOAC Int. 87 (2004) 1492–1497. [14] M. Kiening, E.R. Niessner, S. Baumgartner, R. Krska, M. Bremer, V. Tomkies, P. Reece, C. Danks, U. Immer, M.G. Weller, Sandwich immunoassays for the determination of peanut and hazelnut traces in foods, J. Agric. Food Chem. 53 (2005) 3321–3327. [15] J. Shao, R.E. Baltus, Effect of solute concentration on hindered diffusion in porous membranes, AIChE J. 46 (2000) 1307–1316. [16] F. Schreiber, Structure and growth of self-assembling monolayers, Prog. Surf. Sci. 65 (2000) 151–257. [17] B. Johnsson, S. Lofas, G. Lindquist, Immobilization of proteins to a carboxymethyldextran-modified gold surface for biospecific interaction analysis in surface plasmon resonance sensors, Anal. Biochem. 198 (1991) 268–277. [18] A.J. Bard, L.R. Faulkner, Electrochemical Methods, 2nd ed., John Wiley & Sons, 2001. [19] R. Krska, E. Welzig, S. Baumgartner, Immunoanalytical detection of allergenic proteins in food, Anal. Bioanal. Chem. 378 (2004) 63–65. [20] P. Zapata, P. Basak, J.C. Meredith, High-throughput screening of ionic conductivity in polymer membranes, Electrochim. Acta 54 (2009) 3899–3900. [21] Z. Siroma, R. Kakitsubo, N. Fujiwara, T. Ioroi, S.I. Yamazaki, K. Yasuda, Depression of proton conductivity in recast Nafion film measured on flat substrate, J. Power Sources 189 (2009) 994–998. [22] M.A. Navarra, A. Fernicola, S. Panero, B. Scrosati, Composite gel-type proton membranes, J. Electrochem. Soc. 153 (2006) A1284–A1289. [23] T.A. Bendikov, A. Rabinkov, T. Karakouz, A. Vaskevich, I. Rubinstein, Biological sensing and interface design in gold island film based localized plasmon transducers, Anal. Chem. 80 (2008) 7487–7498. [24] Y. Jung, J.Y. Jeong, B.H. Chung, Recent advances in immobilization methods of antibodies on solid supports, Aanlyst 133 (2008) 697–701. [25] E.M. Renkin, Filtration, diffusion, and molecular sieving through porous cellulose membranes, J. Gen. Physiol. 38 (1954) 225–243.
R. Singh et al. / Sensors and Actuators B 145 (2010) 98–103 [26] J.L. Anderson, J.A. Quinn, Restricted transport in small pores, Biophys. J. 14 (1974) 130–150. [27] N. Mogi, E. Sugai, Y. Fuse, T. Funazukuri, Infinite dilution binary diffusion coefficients for six sugars at 0.1 MPa and temperatures from 273.2 to 353.2 K, J. Chem. Eng. Data 52 (2007) 40–43. [28] M.E. Young, P.A. Carroad, R.L. Bell, Estimation of diffusion coefficients of proteins, Biotechnol. Bioeng. 22 (1980) 947–955. [29] I. Sokolov, V. Subba-Rao, L.A. Luck, Change in rigidity in the activated form of the glucose/galactose receptor from Escherichia coli: a phenomenon that will be key to the development of biosensors, Biophys. J. 90 (2006) 1055–1063. [30] E.A. Bluhm, E. Bauer, R.M. Chamberlin, K.D. Abney, J.S. Young, G.D. Jarvinen, Surface effects on cation transport across porous alumina membranes, Langmuir 15 (1999) 8668–8672. [31] E.A. Bluhm, N.C. Schroeder, E. Bauer, J.N. Fife, R.M. Chamberlin, K.D. Abney, J.S. Young, G.D. Jarvinen, Surface effects on metal ion transport across porous alumina membranes. 2. Trivalent cations: Am, Tb, Eu, and Fe, Langmuir 16 (2000) 7056–7060. [32] B.J Hinds, N. Chopra, T. Rantell, R. Andrews, V. Gavalas, L.G. Bachas, Aligned multiwalled carbon nanotube membranes, Science 303 (2004) 62–65. [33] X. Jiang, N. Mishra, J.N. Turner, M.G. Spencer, Diffusivity of sub-1000 Da molecules in 40 nm silicon-based alumina pores, Microfluid. Nanofluid. 5 (2008) 695–701. [34] K. Whang, T.K. Goldstick, K.E. Healy, A biodegradable polymer scaffold for delivery of osteotropic factors, Biomaterials 21 (2000) 2545– 2551. [35] W.A. Wakeham, E.A. Mason, Diffusion through multiperforate laminae, Ind. Eng. Chem. Fundam. 18 (1979) 301–305.
103
Biographies Rajdeep Singh is currently a MS student in the Department of Chemical and Biomolecular Engineering at Clarkson University. He received his bachelor’s degree in Chemical Engineering in 2005 from the National Institute of Technology in Jalandhar, Punjab, Inda. Pranav P. Sharma is currently a MS student in the Department of Chemical and Biomolecular Engineering at Clarkson University. He received his bachelor’s degree in Chemical Engineering in 2007 from the National Institute of Technology in Warangal, Andhra Pradesh, India. Ruth E. Baltus received her PhD in Chemical Engineering in 1982 from Carnegie Mellon University and joined the faculty in the Department of Chemical and Biomolecular Engineering at Clarkson University in 1983. She has served as department Chair since 2007. Her research interests involve macromolecular and particle transport in nanoporous and microporous systems, characterization of porous membranes and characterization of room temperature ionic liquids. Ian I. Suni received his PhD in Chemistry in 1992 from Harvard University, where he studied with William Klemperer. He did postdoctoral research with Edmund Seebauer in the Department of Chemical and Biomolecular Enginering at the University of Illinois from 1991 to 1993. He has been a Professor in the Department of Chemical and Biomolecular Engineering at Clarkson University since 1993. His research focuses on the application of electrochemistry and electrochemical engineering to the development and understanding of biosensors, nanomaterials, and new methods for thin film growth and dissolution.
