Immittance Electroanalysis in Diagnostics - American Chemical Society

Dec 9, 2014 - Institute of Chemistry, Physical Chemistry Department, Univ. Estadual Paulista (São Paulo State University), Nanobionics Research. Grou...
0 downloads 0 Views 1MB Size
Subscriber access provided by University of Colorado Anschutz Medical Campus Health Sciences Library

Article

Immittance Electroanalysis in diagnostics Amol V. Patil, Flavio C. Bedatty Fernandes, Paulo Roberto Bueno, and Jason J Davis Anal. Chem., Just Accepted Manuscript • DOI: 10.1021/ac503156a • Publication Date (Web): 09 Dec 2014 Downloaded from http://pubs.acs.org on December 15, 2014

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Analytical Chemistry is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 27

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

Immittance Electroanalysis in diagnostics Amol V. Patil†, Flávio C. Bedatty Fernandes‡, Paulo R. Bueno‡*, Jason J. Davis†*

† ‡

Department of Chemistry, University of Oxford, South Parks Road, Oxford OX1 3QZ, UK. Institute of Chemistry, Physical Chemistry Department, Univ. Estadual Paulista (São Paulo State

University), Nanobionics Research Group (www.nanobionics.pro.br) CP 355, 14800-900, Araraquara, São Paulo, Brazil.

Corresponding author Phone: +55 16 3301 9642; Fax: +55 16 3322 2308 Phone: +44 1865 275914 *Email: †[email protected], ‡[email protected]

1

ACS Paragon Plus Environment

Analytical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 27

Abstract

Impedance derived electroanalytical assays are inherently spectroscopic (frequency resolved) and potentially exceedingly sensitive indicators of interfacial change (such as target binding with an appropriate receptor). We introduce here the use of a portfolio of mathematically derived immittance functions and related components capable, from the same raw data sets, of enabling increased assay sensitivity and markedly shorter assay times in comparison to traditional impedance analyses. The methodology, applied herein to faradaic (redox probe amplified) and non-faradaic assays, requires no equivalent circuit analysis or prior assumption of response. Its focus is to optimise analytical potency and to enable the user to select and apply the most frequency-optimized reporter of interfacial change and to, thereafter, run rapid (optimized) analyses at single frequencies.

KEYWORDS: Electrochemical Impedance Spectroscopy, Immunosensors, Biosensor

2

ACS Paragon Plus Environment

Page 3 of 27

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

Introdution It is likely that electrical assays will underpin progress towards increasingly sensitive, miniaturized, multiplexed and portable biosensing configurations. Given that electrical assays can be easily integrated with standard microfabrication and microfluidic formats the resulting electroanalyses can provide a low cost, high throughput and truly convenient multiplexed capability. A variety of voltammetric methods, including linear sweep,1 differential pulse,2 square wave,3 amperometric methods,4 and electrochemical impedance spectroscopy (EIS) have been widely used for biomarker determination. Among them, EIS (faradaic or non-faradaic), which measures the electrical impedance of an interface in AC steady state under a constant DC bias, is uniquely non-destructive and sensitive to interfacial change without amplification or labeling.5-9 During EIS measurements, a relatively small amplitude perturbation (normally 5–20 mV) is superimposed upon a constant DC bias; this mild condition ensures minimal perturbation of the receptive interface (the design and efficacy of which is critical in both sensitive and selective target recruitment). Dependent on whether redox event derived current is associated with the measurement process, EIS can be divided into that which is faradaic (where a signal amplifying redox probe is added to the analytical solution) and that which is non-faradaic (without any redox probe and, in essence, reporting on interfacial capacitance change or dielectric change at the receptor modified surface). In both configurations it is normal to acquire measurement across a broad span of frequencies10-12, although a limited number of studies have reported frequency optimization within otherwise traditional faradaic and non-faradaic measurements by considering the system response13-15. The former (faradaic) is used in the vast majority of reported EIS experiments, where it is associated with both a remarkable sensitivity and inherent practical limitations of requiring high concentrations of redox probe (relative to the analyte) and significant DC bias. In pioneering work by Berggren et al., non-faradaic capacitive immunosensor for the detection of Human chorionic gonadotropin hormone (HCG), and interleukin2, a microbial infection immune factor, were reported some 15 years ago on monoclonal antibody modified Self-Assembly Modified (SAM) gold electrodes.2,3 These assays were reported to be sensitive to targets down to pg/mL levels. In similar work non-faradaic EIS methods have been reported to detect interferon with an astonishing 8 -10 ng/mL level of sensitivity in buffer at thiolated SAM-Ab interfaces. This is some 10 million times greater sensitivity than typically associated with alternative label free immunoassays.2 Rickert and co-workers16 have reported non-faradaic target detection, based on the immobilization of an antibody on a mixed alkane thiol self-assembly monolayer, by measuring capacitance change. Since then, a large number of non-faradaic EIS assays have been reported. During the past decade, focus on an ability to reliably assay protein biomarkers

3

ACS Paragon Plus Environment

Analytical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 4 of 27

has grown and, with this in mind, both faradaic and non-faradaic EIS assays have shown considerable promise.9,11 In more recent work we have demonstrated that these capabilities can be incorporated into both the analysis of real patient samples and the simultaneous detection of multiple markers.11,17-21 Faradaic EIS assays are almost exclusively based on analysing the parameter of charge transfer resistance (Rct) obtained by fitting the acquired current signal to a Randles equivalent circuit (see SI Figure 1 for details) ,22,23 while non-faradaic analyses most typically use modulus of the impedance (|Z|), double layer capacitance (Cdl) or phase (ϕ) as sampling functions.9,24-27 In both approaches a phenomenological model based either on an equivalent circuit 28 or continuum scheme derived from microscopic Poisson-Nernst-Plank continuum equations (using conventional or alternative approaches for diffusional activity)29-31 can be applied from which physical parameters can be extracted. Though these approaches can be instructive in terms of interfacial modelling, the outputs can be somewhat difficult to compare because of the modelling differences and, from a purely analytical perspective (especially as applied across a range of different interfaces), they are neither necessary nor necessarily informing. Despite the large number of reported impedance works there has been no systematic quantitative consideration of EIS as a generic electrical transfer function (i.e. a generic input/output signal) beyond traditional EIS approaches11,19,32-35 and specifically none which do not require a premodelling of any given interface of interest. We show herein, that, in applying the transfer function concept to EIS datasets it is not only possible to extract analytically potent information but to also automate this without being limited to the confines of a specific physical model. Though applied here to the tracking of specific biorecognition, the methodology is generic in making available more analytical information than is typically available from an EIS experiment. In fact a library of immittance functions (such as  ∗ ,  ∗ ,  ∗ and ∗ , i.e. complex impedance, capacitance, admittance and modulus, respectively) and associated components are available from the same interfacial analyses. These components are present within the same raw EIS transfer function (data). We specifically define herein EIS as a specific form of a transfer function in the case of voltage and current input/output signals. From a modulated potential   =  +    input and a sigmoidal output current response  =  ̅ +   , a whole library of generically termed immittance functions  and their relationships are derivable. These  functions can be evaluated separately as a function of analyte concentration. Subsequently the most sensitive  , at its most responsive frequency, can be selected in generating optimised analytical curves (and, of course,

4

ACS Paragon Plus Environment

Page 5 of 27

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

linear ranges and detection limits). Since some of  functions will be more sensitive to highly specific interfacial changes than others, the advantages of acquiring and processing library of  is considerable and brings with it the added benefits of neither being tied to one sampling function nor having to assume anything about the best function to track. More importantly,  data acquisition can be performed over a markedly smaller frequency range [ranging from: 100.0 – 1.0 Hz, enabling acquisition times of < 3 minutes with optimized sensitivities and Limit of Detection (LoD)]. It is additionally worth noting that the computed behaviour of  are acquired without any presumption or fudge factors, from the raw analytical signal. Herein we report the application of this approach to the highly effective assaying of C-reactive protein (CRP), a clinically important protein biomarker. Inflammation, caused by infection or injury, can lead to an dramatic increase in the CRP level by 1000-fold.36 Recently, studies on CRP conclude that levels exceeding 3.0 mg/L in serum are indicative of risk of diabetes, hypertension and cardiovascular disease.37,38 To select optimum parameter and frequencies from available complex functions ( ∗ ,  ∗ ,  ∗ and  ∗ ), i.e. those associated with high response sensitivity with low background noise were quantitatively analyzed across frequency range spanning six orders of magnitude (from 0.01 Hz to 100 kHz), within both faradaic and non-faradaic formats. Once the most sensitive parameters, which have been frequency optimized, are selected, analytical curves of high diagnostic utility are presented. Experimental

Chemical Reagents – Ethanolamine (98%), 1-ethyl-3-(3-dimethylaminopropyl) carbodiimide (EDC), N-hydroxysuccinimde (NHS), dimethyl sulphoxide (DMSO), and phosphate buffered saline (PBS, 10 mM, pH 7.4) tablets were purchased from Sigma Aldrich while goat anti-human CRP polyclonal antibody and human CRP were purchased from AbD Serotec. Phosphate buffered saline Tween (PBST) was prepared by dissolving PBS tablets (Sigma Aldrich) in ultrapure water with 0.05% v/v Tween-20 added, and then filtered using a 0.22 µm membrane filter. Polyethylene glycol (PEG) containing thiol HS–(CH2)11–(EG)3–OCH2–COOH (PEG thiol) was purchased from Prochimia Surfaces, Poland. 16-Mercaptohexadecanoic acid (16-MHDA) was purchased from Sigma. All other chemicals were of analytical grade. Deionized water (18.2 MΩ/cm, Synergy Ultrapure water system EMD Millipore) was used throughout.

5

ACS Paragon Plus Environment

Analytical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 6 of 27

Apparatus – Electrochemical experiments, cyclic voltammetry and EIS, were conducted with an Autolab Potentiostat equipped with a FRA module using a three-electrode system: either a conventional gold disk working electrodes (1.6 mm diameter, BASi) with a platinum wire counter electrode or gold microelectrode array with six working electrodes (200 µm in diameter) and shared counter electrode (Figure 1), and a silver/silver chloride (Ag|AgCl, filled with 1.0 M KCl, +0.236 versus SHE) reference electrode (CH Instruments). All potentials reported are relative to the Ag|AgCl reference electrode. All EIS responses were recorded over a wide range of frequencies from 0.01 Hz to 100 kHz (at logarithmically spaced values). The error bars are the standard deviations of three successive measurements with the same electrode, with each experiments repeated two to three times. Fabrication of receptor interface – Receptor interfaces were assembled on standard gold disk electrodes or microfabricated gold electrodes arrays. Gold disk electrodes were mechanically polished to flat, mirror-like surfaces (with diamond sprays with three successive sizes (1, 1/4, 1/10 micron) on polishing pads). This was followed by rinsing and ultra-sonic washing with ultrapure water and absolute ethanol. The electrodes were then immersed in hot piranha solution (concentrated H2SO4: 30% H2O2, v/v 3:1 Caution: piranha reacts extremely aggressively with organic materials. Please treat with extreme care!) for approximately 15 min, followed by rinsing and ultra-sonic washing with ultrapure water for 1-2 min. Electrode surface was then electrochemically polished as follows: initially, a number of negative cyclic voltammetry scans (from -1.35 to - 0.35 V) were performed in 0.5M KOH aqueous solution, until stable, smooth curves were obtained. Subsequently, a series of wider range scans, from - 0.35 to 1.5 V, were conducted in 0.5 M H2SO4 at a scan rate of 0.1 V/s, until the height and shape of anodic and cathodic peaks are constant. The gold array microelectrodes were rinsed with ethanol then washed with water prior electrochemical pre-treatment in 0.5 M H2SO4 using cyclic voltammetry (CV) (potential range from 0.2 to 1.6 V at a scan rate of 1 V/s, 100 cycles) until a stable sharp peak (FWHM ~90 mV) was obtained. Such pre-treated electrodes (gold disk electrodes or gold microelectrode arrays) were immersed in an absolute ethanol (HPLC grade) solution of 1 mM PEGylated thiol or 2 mM of 16mercaptohexadecanoic acid (16-MHDA) solution for array microelectrodes and disk electrodes, respectively for 16 h at room temperature to form a self-assembly monolayer. Prior to antibody immobilization, the electrodes were rinsed with absolute ethanol and dried in a flow of nitrogen gas. The terminal carboxyl groups were then activated with 1-Ethyl-3-(36

ACS Paragon Plus Environment

Page 7 of 27

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

dimethylaminopropyl) carbodiimide (0.4 M) and N-Hydroxysuccinimide (0.1 M) in ultrapure water for 40 min at room temperature, and then reacted with 1µM antibody receptor in PBS solution for 10 h at 4oC (16-MHDA modified electrode were immersed for an 1 h in the Anti-CRP solution) .11,19 The final step involved immersing the probes in 1 M ethanol amine (pH ~8.5) to deactivate any extra activated carboxylic groups. Impedance analysis and cyclic voltammetry measurements were carried out after each of the biosensor fabrication/ modification steps in order to ensure an appropriated receptive surface construction. The observed significant increase in Rct at this point is fully consistent with an increase in interfacial steric bulk which is associated with successful antibody immobilization.11 The resistance of the interface thereafter responds in manner that correlates to the amount of target protein binding as expected. In the absence of a redox probe (non-faradaic), stepwise surface modification is reported through observed capacitative decrease. Therefore, EIS and CV present a useful means of characterizing the stepwise fabrication of captive surface. The cyclic voltammograms report a typical electrochemical behaviour as surface is modified. It is clear that charge transfer is predictably inhibited after 16-MHDA formation. Additionally with EIS measurement it is possible to obtain the Nyquist plots, in which the “semi-circular” behaviour at low sampling frequency helps in elucidating charge transfer restrictions imposed either sterically or electrostatically as interface is modified. Predictably, there are sharp increases in Rct as the receptor layer is fabricated. Rct specifically increases from less than 18 ± 0.8 kΩ/cm2 to 5.7 ± 0.6 MΩ MΩ/cm2 after the formation of the 16-MHDA and further upwards on antibody immobilization. Antibody modified electrode surfaces were preserved in 10 mM PBS at 4oC before further experiments. Prior to any EIS measurement (at a specific concentration of CRP) the antibody modified gold electrode was rinsed with PBS, incubated in solution of CRP (at the required specific concentration) for 30 minutes, followed by another rinse in PBS. A typical acquisition of impedance data over the full frequency range takes about an hour; the acquisition of a 5 point calibration curve, with triplicate repeats at each concentration, thus requires approximately 7 hours of analytical time. Faradaic/Non-Faradaic Electrochemical Impedance Measurements and Cyclic Voltammetry – The measurements were recorded in a KCl (0.1M) supporting electrolyte containing 1 mM [Fe(CN)6]3−/4− as the standard redox probe. For all the methods, CV was performed at a scan rate of 100 mV s−1 between -0.20 V and 0.6 V relative to Ag|AgCl. EIS measurements were conducted in a frequency range of 0.01 Hz to 100 kHz with RMS amplitude of 3 mV (or 10 mV peak to peak) around 0.21V relative to Ag|AgCl. Non-Faradaic measurements were carried out under identical parameters in absence of ferricyanide and at 0.0V relative to Ag/AgCl. Standard faradaic impedance analysis,

7

ACS Paragon Plus Environment

Analytical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 8 of 27

utilizing Rct were carried out by fitting acquired raw impedance data to an equivalent circuit (see SI 1). Data Analysis – The analysed Immittance functions are impedance ( ∗ ), capacitance ( ∗ ), modulus ( ∗ ) and admittance ( ∗ ) with their phasorial relationships:  ∗ = 1/ ∗

(1a)

 ∗ =  ∗

(1b)

 ∗ =  ∗

(1c)

where  = √−1 and  = 2!" As is evident through the above stated phasorial relations,  ∗ and  ∗ are related to impedance ( ∗ ) and capacitance ∗ . All  are complex functions so that individually divisible into constituent real (') and imaginary ('') components, the modulus as well as their ratios (and inverted ratios), each of which can be separately analysed as a reporter of specific interfacial change. Thus there are sixteen parameters in total for any given sampling frequency. The responsiveness (R), of each function was systematically evaluated across the frequency range as the concentration of the analyte (i.e. CRP) was varied. The Relative Response (RR) was defined as; %

%

%

%

##$ % = #$ − #' /#' ∗ 100 %

(2) %

where #' is the initial value of  function in absence of analyte and #$ the value of the  function after exposure to a specific target concentration (n) at the same frequency f. Thus an optimal analytical function can be defined where the relative response (RR) is maximal at a particular frequency across the sampled concentrations. A home written MATLAB R2013b algorithm was used to calculate and convert the original electrical transfer function data into phasorial related immittance functions, determine the optimal frequency at which the response of  is monotonic, linear (with R2>0.95). However only those functions that showed favourable response (in terms of monotonic behaviour, linear relationship between signal variation and range of the target i.e. the linearity & sensitivity as compared to Rct in the case of faradaic assays) were further analysed while the rest were discarded. Limits of Detection

8

ACS Paragon Plus Environment

Page 9 of 27

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

(LoD) were calculated, based upon the standard deviation (SD) of the blank response & slope of the analytical curve, as 3.3*SD. Thus for faradaic EIS analyses, direct comparisons between selected immittance functions (with the highest sensitivity & linearity) and charge transfer resistance (Rct - frequency independent and determined from Randles equivalent circuit) were made. For the non-faradaic format, analyses were restricted to immittance functions.

RESULTS AND DISCUSSION Currently a number of turbidimetric, nephelometric or ELISA based assays of CRP (acute phase protein which is cardiac marker and indicator of general trauma/infection) are available. These are, however, commonly insensitive, prone to false negatives and/or time consuming.39,40 We sought here to build upon previous impedance based analysis of this protein11,20,41,42 in applying the aforementioned immittance function analysis and refining both assay sensitivity and speed, with and without the use of a solution based redox probe. In order to map out the efficacy/applicability of the computed phasorial  at any particular frequency (and Rct calculated from a traditional equivalent circuit approach) and additionally the effects of the interfacial chemistry on sensitivity and assay reproducibility, faradaic EIS studies of CRP were run at two different SAM interfaces (PEG-thiol & 16-MHDA). Once having obtained the relative response (RR) of all of the  complex functions and subparameters (real & imaginary components, modulus and ratios thereof, some which are shown herein in Figure 2) by scanning through the frequency range, in presence of specific concentration of analyte, it is facile to determine the optimal sampling frequency for all wherein RR is maximal. Taking the spectral response of Z'' as an example (Figure 3), it is evident that RR varies across the frequency range with a large apparent change in magnitude, i.e. high sensitivity, in response to changes in concentration of analyte (CRP) within a frequency range of 1 to 20 Hz. By computationally calculating the averaged RR response (for three repetitions at each concentration value) across the concentration range (0.01-100 nM), and fitting a linear response (R2 > 0.95) if monotonic, then it is possible to determine the optimal frequency for each of the four complex  with eight components generating a total of sixteen sensitive components at which the Sensitivity (S) (defined as the slope of the linear fit) is maximal (Table 1). Thus one can compare the performance and the optimal frequency of each of  parameter in order to isolate that which is most 9

ACS Paragon Plus Environment

Analytical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 10 of 27

responsive. For CRP responses at PEG-thiol-antiCRP interfaces, numerical calculations show that |Z| and |M| (sensitivity of 264) Z’’ and M’ (sensitivity of 245) are the most sensitive  parameters at 12.6 Hz and LoDs of approximately 8 pM, values that compare favourably with Rct (sensitivity 108 and LoD 6 pM, Figure 4). An equivalent faradaic data analysis at 16-MHDA modified gold surfaces shows that |Z| and the Z′′/ Z′ ratio are significantly more sensitive than Rct (Figure 5). The small experiment-to-experiment variance of these observations, as indicated by the error bar magnitudes, is an indication of the stability of the interfacial construct and analytical reproducibility (Figures 4 & 5). It is also evident that, irrespective of underlying surface modification chemistry (PEG-Thiol or 16-MHDA), one or more  parameter has a response whose sensitivity is larger than the conventionally used Rct. It is also clear, based upon results from multiple (4) electrodes, that parameter sensitivity depends on both surface chemistry and analyte (for example while |Z| is most sensitive across both PEG-Thiol & 16-MHDA, outperforming Rct, the Z''/ Z' ratio is sensitive at 16MHDA films but relatively unresponsive on PEG-Thiol (as shown by comparing Figures 4 & 5). Further analysis of the generated data sets shows that for most of the  parameters there exists a range of frequencies, around the optimal frequency, for which the sensitivity is largely constant. Z'', for example, has a broadly constant sensitivity to CRP (PEG-thiol modified gold) across a frequency range of 4.8 - 20.4 Hz; (with an average value of 215.88 ± 29 compared against the maximum value 245 at 12.6 Hz) (Figure 6)*. The existence of a broader optimal frequency range enables the user to scan over a limited region of frequency without having to “dial in” an exact value of optimal frequency and yet be assured of a response whose sensitivity is equivalent or superior to Rct. A detailed analysis of optimal frequency ranges for some of the ’s parameters, in faradaic assay, is presented in Figure 7. It is additionally noteworthy that, for all faradaic analyses carried out across both interfacial chemistries,  optimal response/sensitivity was found to be confined to the 0.01200 Hz frequency range. By limiting the sampling frequency range (initially 0.01-10 MHz to 0.1 to 200 Hz) acquisition & subsequent an analysis time is reduced from 1 hour to less than 10 minutes. To summarize, once an optimal frequency region is known, for a given target and interface, acquisition and analysis time can be dramatically reduced while achieving superior assay sensitivity and LoD values that are equivalent or better than those obtained with traditional analysis (based on using Rct as the analytical transducer signal).

*

Note that the parameters with optimal response (in terms of sensitivity) were resolved at frequencies in excess of 10 Hz. Diffusion effect contributions become significant at frequencies below 1 Hz (See SI figure 2) and do not contribute.

10

ACS Paragon Plus Environment

Page 11 of 27

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

Extension of Immittance function analysis to non-faradaic CRP assays The strength of non-faradaic EIS analysis lies in the absence of a high doping concentration of redox probe and a potential perturbing initial bias voltage; however there is often a concomitant drawback of low sensitivity (the result of lack of signal amplification due to redox probes/labels present in faradaic EIS). There is, then, considerable value in optimizing non-faradaic  parameters for sensitivity. The analytical curves (Figure 8-a) show that the most sensitive functions all span a three orders of magnitude linear range (0.5 pM to 50 nM). It should be noted that, although the sensitivities demonstrated by non-faradaic parameters are significantly smaller than observed when a redox probe is present, the lower background noise (defined as 3.3 times the standard deviation of the baseline) leads to significantly smaller LoD values (52.4 ± 42.1pM to 619 ± 234 pM). By applying the above mentioned analytical methodology, one can readily find the optimal regime in which one or more  parameters has maximal sensitivity and minimal LoD. Such analysis of non-faradaic CRP data herein showed that each of the investigated  displayed an optimum frequency at the lower end of the sampled range (0.05 to 0.6 Hz). As with the analogous faradaic data, optimized frequency zones are apparent wherein optimal sensitivity is retained. The most sensitive parameters analysed here, across 3 electrodes and 9 repetitions in total, were consistently found to be Z'', M', C'', Z'', |Z| and Y' (Figure 8-b).

CONCLUSIONS EIS methods possess an innate ability to sensitively monitor conductivity, steric, dielectric or charging capacity characteristics at an electrode/electrolyte interface and become powerful sensory tools when these surfaces are suitably receptor modified. On the immersion of such films into electrolyte, an electrical double layer will be established, the analysis of which is commonly within a Randles equivalent circuit framework. Though undoubtedly powerful, this approach is limited by both the assumption of suitable equivalent circuit (and “fudge factors” that are often implemented in attaining a “good fit”) and the need to acquire data across a full range of frequencies. Ultimately, this time consuming process also reports on one analysing electrochemical parameter only for faradaic studies (Rct), and normally one (capacitance) non optimized parameter for non-faradaic analyses. In this paper, we introduce the concept of Immittance Functions (), which can be directly computed from raw experimental data (without any presumptions). We use these to dramatically improve assaying characteristics and demonstrate that the component function parameters respond differently 11

ACS Paragon Plus Environment

Analytical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 12 of 27

to the same receptor-target interaction. We also show that they do so at a characteristic sensitivity optimized analytical frequency. The key aspect of this approach was to use EIS as transfer function to monitor interfacial events without focus on, or pre-modelling of, an associated physical picture. Once derived ImFs are known, subsequent analyses can not only be multi-parameter in nature but also frequency optimized, more sensitive and associated with assay times between 5 and 10 fold faster than traditional analyses. Exemplified here with a clinically important target, it is clear that faradaic and non-faradaic  approaches have their own advantages in terms of sensitivity, range and practicality. This study has ultimately sought to provide the electroanalytical chemist with a new analytical toolbox.

Acknowledgements The authors acknowledge São Paulo state research funding agency (FAPESP) and CNPq.

Supporting Information Supplementary data and figures explain the equivalent circuit used for calculating resistance to charge transfer. This material is available free of charge via the Internet at http://pubs.acs.org

12

ACS Paragon Plus Environment

Page 13 of 27

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

References (1) Yan, S.; He, N.; Song, Y.; Zhang, Z.; Qian, J.; Xiao, Z. Journal of Electroanalytical Chemistry 2010, 641, 136-140. (2) Pournaghi-Azar, M. H.; Ahour, F.; Hejazi, M. S. Electroanalysis 2009, 21, 1822-1828. (3) Li, L.; Zhao, H.; Chen, Z.; Mu, X.; Guo, L. Biosensors and Bioelectronics 2011, 30, 261-266. (4) Telsnig, D.; Kassarnig, V.; Zapf, C.; Leitinger, G.; Kalcher, K.; Ortner, A. Int. J. Electrochem. Sci. 2012, 7, 1047610486 (5) Lin, J.; Wei, Z.; Zhang, H.; Shao, M. Biosens. Bioelectron. 2013, 41, 342-347. (6) Sahoo, P.; Suresh, S.; Dhara, S.; Saini, G.; Rangarajan, S.; Tyagi, A. K. Biosens. Bioelectron. 2013, 44, 164-170. (7) Venkatanarayanan, A.; Keyes, T. E.; Forster, R. J. Anal. Chem. 2013, 85, 2216-2222. (8) Ohno, R.; Ohnuki, H.; Wang, H.; Yokoyama, T.; Endo, H.; Tsuya, D.; Izumi, M. Biosens. Bioelectron. 2013, 40, 422426. (9) Daniels, J. S.; Pourmand, N. Electroanalysis 2007, 19, 1239-1257. (10) Yang, L.; Li, Y.; Erf, G. F. Anal. Chem. 2004, 76, 1107-1113. (11) Bryan, T.; Luo, X.; Bueno, P. R.; Davis, J. J. Biosensors and Bioelectronics 2013, 39, 94-98. (12) García-Jareño, J. J.; Giménez-Romero, D.; Keddam, M.; Vicente, F. J. Phys. Chem. B 2005, 109, 4584-4592. (13) Nandakumar, V.; La Belle, J. T.; Reed, J.; Shah, M.; Cochran, D.; Joshi, L.; Alford, T. L. Biosensors and Bioelectronics 2008, 24, 1039-1042. (14) Quershi, A.; Gurbuz, Y.; Kang, W. P.; Davidson, J. L. Biosensors and Bioelectronics 2009, 25, 877-882. (15) Özcan, B.; Demirbakan, B.; Yeşiller, G.; Sezgintürk, M. K. Talanta 2014, 125, 7-13. (16) Rickert, J.; Göpel, W.; Beck, W.; Jung, G.; Heiduschka, P. Biosens. Bioelectron. 1996, 11, 757-768. (17) Luo, X.; Xu, Q.; James, T.; Davis, J. J. Anal. Chem. 2014, 86, 5553-5558. (18) Lehr, J.; Fernandes, F. C. B.; Bueno, P. R.; Davis, J. J. Anal. Chem. 2014, 86, 2559-2564. (19) Xu, M.; Luo, X.; Davis, J. J. Biosensors and Bioelectronics 2013, 39, 21-25. (20) Johnson, A.; Song, Q.; Ko Ferrigno, P.; Bueno, P. R.; Davis, J. J. Analytical Chemistry 2012, 84, 6553-6560. (21) Bryan, T.; Luo, X.; Forsgren, L.; Morozova-Roche, L. A.; Davis, J. J. Chemical Science 2012, 3, 3468-3473. (22) Ionescu, R. E.; Gondran, C.; Bouffier, L.; Jaffrezic-Renault, N.; Martelet, C.; Cosnier, S. Electrochimica Acta 2010, 55, 6228-6232. (23) Ramón-Azcón, J.; Valera, E.; Rodríguez, Á.; Barranco, A.; Alfaro, B.; Sanchez-Baeza, F.; Marco, M. P. Biosensors and Bioelectronics 2008, 23, 1367-1373.

13

ACS Paragon Plus Environment

Analytical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 14 of 27

(24) Bart, M.; Stigter, E. C. A.; Stapert, H. R.; de Jong, G. J.; van Bennekom, W. P. Biosensors and Bioelectronics 2005, 21, 49-59. (25) Qureshi, A.; Gurbuz, Y.; Kallempudi, S.; Niazi, J. H. Physical Chemistry Chemical Physics 2010, 12, 9176-9182. (26) Chuang, Y.-H.; Chang, Y.-T.; Liu, K.-L.; Chang, H.-Y.; Yew, T.-R. Biosensors and Bioelectronics 2011, 28, 368372. (27) Lin, K.-C.; Kunduru, V.; Bothara, M.; Rege, K.; Prasad, S.; Ramakrishna, B. L. Biosensors and Bioelectronics 2010, 25, 2336-2342. (28) Bisquert, J.; Garcia-Belmonte, G.; Fabregat-Santiago, F.; Bueno, P. R. J. Electroanal. Chem. 1999, 475, 152-163. (29) Bisquert, J.; Compte, A. J. Electroanal. Chem. 2001, 499, 112-120. (30) Lenzi, E. K.; de Paula, J. L.; Silva, F. R. G. B.; Evangelista, L. R. J. Phys. Chem. C 2013, 117, 23685-23690. (31) Macdonald, J. R. J. Phys.: Condens. Matter 2010, 22, 495101. (32) Luo, X.; Xu, M.; Freeman, C.; James, T.; Davis, J. J. Analytical Chemistry 2013, 85, 4129-4134. (33) Benilova, I. V.; Soldatkin, A. P.; Martelet, C.; Jaffrezic-Renault, N. Electroanalysis 2006, 18, 1950-1956. (34) Gautier, C.; Esnault, C.; Cougnon, C.; Pilard, J.-F.; Casse, N.; Chénais, B. Journal of Electroanalytical Chemistry 2007, 610, 227-233. (35) Fang, X.; Tan, O. K.; Tse, M. S.; Ooi, E. E. Biosensors and Bioelectronics 2010, 25, 1137-1142. (36) Gabay, C.; Kushner, I. New England Journal of Medicine 1999, 340, 448-454. (37) Sattar, N.; Gaw, A.; Scherbakova, O.; Ford, I.; O’Reilly, D. S. J.; Haffner, S. M.; Isles, C.; Macfarlane, P. W.; Packard, C. J.; Cobbe, S. M.; Shepherd, J. Circulation 2003, 108, 414-419. (38) Ridker, P. M.; Buring, J. E.; Cook, N. R.; Rifai, N. Circulation 2003, 107, 391-397. (39) Güven, E.; Duus, K.; Lydolph, M. C.; Jørgensen, C. S.; Laursen, I.; Houen, G. J. Immunol. Methods 2014, 403, 2636. (40) Kim, C.-H.; Ahn, J.-H.; Kim, J.-Y.; Choi, J.-M.; Lim, K.-C.; Jung Park, T.; Su Heo, N.; Gu Lee, H.; Kim, J.-W.; Choi, Y.-K. Biosens. Bioelectron. 2013, 41, 322-327. (41) Mishra, S.; Sharma, V.; Kumar, D.; Rajesh. Appl Biochem Biotechnol 2014, 1-14. (42) Chen, X.; Wang, Y.; Zhou, J.; Yan, W.; Li, X.; Zhu, J.-J. Anal. Chem. 2008, 80, 2133-2140.

14

ACS Paragon Plus Environment

Page 15 of 27

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

Figure Annotations Figure 1. Batch microfabricated array with six individually addressable gold working electrodes (200 microns in diameter) and a shared counter electrode, utilised herein.

Figure 2. Relative Response (RR) of the sixteen parameters analysed across a scanned frequency range of (0.4-10 kHz) at varied CRP concentrations (10 pM (square), 50 pM (circle), 100 pM (up triangle), 300 pM (down triangle) and 500 pM (diamond) on PEG-thiol modified electrode. As C*, M* & Y* are phasorially derived from Z* it is evident that several functions show a similar response (Z’’ and M’, M’’ and Z, for example).

Figure 3. A typical example showing the change in Relative response (RR) of Z’’ of a faradaic CRP EIS assay on an antibody modified PEG thiol gold electrode as concentration is varied from 10 to 500 pM. Response is maximal around 12 Hz, the thus indicated optimum analytical frequency.

Figure 4. Comparative function responses to CRP at PEG-Thiol modified gold, within faradaic assays; |Z| (at an optimum frequency, f, of 12.6 Hz) has sensitivity (S) of 264, superior to Rct (sensitivity 108). As a comparison, Y’’ even at its maximal optimum frequency of 0.72 Hz shows a sensitivity (34) that is lower than Rct while Z''/ Z' has equivalent sensitivity

Figure 5. Optimal analytical curve of CRP on 16-MHDA modified gold electrode, obtained using faradaic assay, showing that |Z| and Z’’/Z’ ratio (both with optimal frequency = 0.1 Hz) show larger (12 times and 7 times respectively) sensitivity (S) when compared with Rct.

Figure 6. Within the highlighted frequency zone, spanning 4 to 20 Hz, the averaged associated sensitivity for Z’’ has a value of 215.88 ± 29 which is, for analysis purposes, equivalent to the maximal sensitivity of 245 seen at optimal frequency of 12.6 Hz.

Figure 7. Comparative ImF parameter sensitivities (each obtained from analytical curves constructed within the calculated optimal frequency range of 0.1-10Hz and from three repetitions on three different PEG thiol modified electrodes) showing that most sensitive parameters like |Z|,|M| and M’ are confined with a frequency range of 2 to 14 Hz

15

ACS Paragon Plus Environment

Analytical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 16 of 27

Figure 8. a) Optimal analytical curve of non-faradaic analysis of CRP on 16-MHDA modified gold electrode showing that Z’’ and Mratio (M/M’’) showing highest relative sensitivity (for Z’’: sensitivity =3.4, R2=0.99) while M’/M’’ shows a lower value (sensitivity =1.49, R2=0.97). The value and errors associated for each point were acquired from analyses in the 0.05 to 1.0 Hz range. b) A summary of relative ImF parameter sensitivity in a non-faradaic CRP detection assay (0.5-100 nM). Data here was obtained by averaging across an optimal frequency range of 0.1-1 Hz and from analytical curves constructed through three repetitions on each of three different electrodes.

16

ACS Paragon Plus Environment

Page 17 of 27

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

Figure 1

17

ACS Paragon Plus Environment

Analytical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 18 of 27

Figure 2

18

ACS Paragon Plus Environment

Page 19 of 27

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

Figure 3

19

ACS Paragon Plus Environment

Analytical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 20 of 27

Figure 4

20

ACS Paragon Plus Environment

Page 21 of 27

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

Figure 5

21

ACS Paragon Plus Environment

Analytical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 22 of 27

Figure 6

22

ACS Paragon Plus Environment

Page 23 of 27

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

Figure 7

23

ACS Paragon Plus Environment

Analytical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 24 of 27

Figure 8

24

ACS Paragon Plus Environment

Page 25 of 27

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

Table annotation Table 1. Optimal ImF parameter responses for CRP detection on antibody modified PEG-thiolated gold. The representative data here is based upon one electrode with 3 repetitions per concentration. Sensitivity and LoD were calculated by linear fits to averaged relative response (RR) across the repetitions.

25

ACS Paragon Plus Environment

Analytical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 26 of 27

Table 1 ImF

′ ′′  ,, /′ || ′ ′′  ,, /′ || ′ ′′  ,, /′ || ′ ′′  ,, /′ ||

Optimized Frequency (Hz)

R2

3.0 12.6 85.3 12.6 4.8 0.7 85.3 0.45 0.7 4.8 85.3 6210 12.6 3.0 85.3 12.6

0.99 0.98 0.99 0.98 0.97 0.95 0.99 0.95 0.95 0.97 0.95 0.95 0.98 0.99 0.95 0.98

Sensitivity

LoD (pM)

(% conc-1) 102 245 84 264 -36 -34 84 47 34 -36 30 5.2 245 102 -30 262

6 8 6 8 3 1 6 2 868627 3 12681899 5.9E-39 8 6 3 8

26

ACS Paragon Plus Environment

Page 27 of 27

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

For TOC Only

27

ACS Paragon Plus Environment