Nanoporous and wrinkled electrodes enhance the sensitivity of glucose biosensors

Accepted Manuscript Title: Nanoporous and wrinkled electrodes enhance the sensitivity of glucose biosensors Author: Robert C. Adams-McGavin Yuting Chan Christine M. Gabardo Jie Yang Marta Skreta Barnabas C. Fung Leyla Soleymani PII: DOI: Reference:

S0013-4686(17)30865-4 http://dx.doi.org/doi:10.1016/j.electacta.2017.04.108 EA 29366

To appear in:

Electrochimica Acta

Received date: Revised date: Accepted date:

24-11-2016 20-4-2017 20-4-2017

Please cite this article as: R.C. Adams-McGavin, Y. Chan, C.M. Gabardo, J. Yang, M. Skreta, B.C. Fung, L. Soleymani, Nanoporous and wrinkled electrodes enhance the sensitivity of glucose biosensors, Electrochimica Acta (2017), http://dx.doi.org/10.1016/j.electacta.2017.04.108 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Nanoporous and wrinkled electrodes enhance the sensitivity of glucose biosensors a,†

, Yuting Chan

a,†

, Christine M. Gabardo

Skreta a, Barnabas C. Fung a, and Leyla Soleymani a,b*

McMaster University, Department of Engineering Physics, 1280 Main Street West, Hamilton,

us

L8S 4L7, Canada

McMaster University, School of Biomedical Engineering, 1280 Main Street West, Hamilton,

L8S 4L7, Canada †

These authors contributed equally to this work

Ac ce p

te

d

M

* E-mail: [email protected]

an

b

, Jie Yang b, Marta

cr

a

b,†

ip t

Robert C. Adams-McGavin

Page 1 of 32

Abstract Three-dimensional electrodes improve the performance of biosensors by increasing their surface area to volume ratio, decreasing the analyte diffusion time, and/or improving analyte access or

ip t

capture at the electrode. We demonstrate a rapid and facile method based on electroless deposition and polymer-induced wrinkling for creating three-dimensional multi-lengthscale

cr

electrodes. This all solution-processing method enables the structure of the electrodes to be tuned

us

by inducing continuous or nanoporous wrinkled surfaces. The surface area and analytical sensitivity of the electrodes are tuned by varying the electroless deposition duration, with the

an

nanoporous and wrinkled electrodes demonstrating the highest surface area and analytical sensitivity compared to their wrinkled and planar counterparts. The nanoporous and wrinkled

M

electrodes developed here combine critical lengthscales ranging from the nanoscale to the

d

macroscale by including nanoscale pores, microscale wrinkles and sub-millimetre-scale electrode

te

footprints, and demonstrate a surface area enhancement of more than 5 times compared to the all-solution-processed planar electrodes. These electrodes were applied to glucose sensing, and

Ac ce p

their response was measured using three classes of electrochemical techniques: cyclic voltammetry, chronoamperometry, and pulsed amperometric detection. When using cyclic voltammetry, these electrodes enable enzyme-free glucose sensing with a sensitivity of 591 µA/mM.cm2 in alkaline solutions. This sensitivity is preserved when analysing solutions having a physiologically-relevant concentration of Cl- ions, and is reduced to 38 µA/mM.cm2 when analysing solutions having a neutral pH. Keywords Electrochemical Biosensing, Wrinkling, Glucose Sensing, Nanoporous Electrodes, Rapid Prototyping

Page 2 of 32

Page 3 of 32

d

te

Ac ce p us

an

M

cr

ip t

1.

Introduction

Three-dimensional electrodes with micro/nanoscale features are of tremendous interest in the

ip t

field of electrochemistry, primarily due to their high surface area and the resultant increase in analytical sensitivity. These electrodes have proven to be particularly important for analysing

cr

species, such as glucose that undergo kinetically-controlled reactions.[1] For diffusion-limited

us

electrochemical events, the overall geometric surface area of the electrode is responsible for the magnitude of the measured Faradaic current for most experimental timescales. However, in

an

electrochemical events with slow surface reaction kinetics, the electrode’s micro/nanoscale features contribute to the magnitude of the Faradaic current, which is leveraged towards

M

increasing the sensor’s analytical sensitivity. [2,3]

d

Enzyme-free glucose detection is desirable due to the instabilities associated with enzymes[4],

te

especially for in-situ and in-vivo monitoring applications. The electrocatalytic nature of threedimensional micro/nanostructured electrodes has been used for developing sensitive enzyme-free

Ac ce p

glucose biosensors. Mesoporous platinum fabricated through templated electrodeposition[1], nanoporous Au or Au-Ag created by dealloying Au-Ag alloys[5,6] or anodic polarization[7], Ni microflowers created by electroless deposition, and Co3O4 nanostructured films developed using the hydrothermal method[8] or templated calcination[9] are among the structures previously developed for enzyme-free glucose detection. Polymer-induced wrinkling of conductive thin films has been developed as a method for creating three-dimensional electrodes with micro/nanoscale features[10–13]. In this method, wrinkled films are obtained when compressive stress is applied to a compliant polymer substrate modified with a stiff overlying thin film. Wrinkling has been combined with other rapid prototyping

Page 4 of 32

methods, such as xurography, to develop a facile and benchtop process for creating multilengthscale electrodes. In xurography, a CAD-driven cutting blade is used to pattern a material for masks by removing specific regions of an adhesive film. Xurography with sub-millimetre

ip t

resolution has been used to define the electrode configuration, while wrinkling has been used to create features with critical dimensions in the micron and sub-micron lengthscale[14].

cr

Furthermore, it is possible to wrinkle porous films created by selective dealloying[15],

us

electroless disposition[16], and self-assembly of nanoparticles[17] on shrinkable polymer substrates.

an

Wrinkled electrodes are structurally tunable and their wavelength, porosity and height are tuned by varying the film thickness and continuity, or the ratio of the Young’s moduli of the thin film

M

and the compliant substrate [10,16,18]. This structural tunability can be leveraged towards

d

controlling the electrode’s functional parameters such as sheet resistance, electroactive surface

te

area, and the density and arrangement of self-assembled monolayers.[10,16,19,20] While wrinkled electrodes have been used in electrochemistry experiments[10,19,21], their role in

Ac ce p

kinetically-controlled electrochemical reactions has not been previously investigated. Our vision was to investigate the electrochemical properties of wrinkled, as well as nanoporous and wrinkled electrodes to determine their suitability for enzyme-free glucose detection. For this purpose, we created planar, wrinkled, and nanoporous and wrinkled electrodes by combining xurography with an all-solution-processing method based on nanoparticle self-assembly and electroless deposition on polymer substrates. We measured and compared the sensitivity of the three classes of electrodes created here, and found that the rapidly prototyped nanoporous and wrinkled electrodes developed here demonstrated the highest sensitivity compared to their wrinkled and planar counterparts. [22]

Page 5 of 32

2.

Experimental

2.1 Chemicals Potassium

chloride

(KCl,

99.0%),

≥

hydrogen

tetrachloroaurate(III)

trihydrate

ip t

(HAuCl4·3H2O, > 99.9%), phosphate buffer solution (PBS, 1.0M, pH 7.4), hydrochloric acid (HCl ACS reagent, 37%) were purchased from Sigma-Aldrich (St. Louis, Missouri).

cr

Sulfuric acid (H2SO4, 98%), 2- propanol (99.5%), methanol (≥99.8%), dextrose

us

(CH2OH(CHOH)4CHO, >99%), ascorbic acid (C6H8O6) and sodium chloride (NaCl, ≥99.0%) were purchased from Caledon (Georgetown, Ontario). Ethanol was purchased Commercial

Alcohols

(Brampton,

ON).

Tris(hydroxymethyl)aminomethane

an

from

((HOCH₂)₂CNH₂, ≥99.9%) was purchased from BioShop Canada (Burlington, ON).

M

Sodium hydroxide (NaOH, 1.0M) was purchased from LabChem (Zelienople, PA). All

d

reagents were of analytical grade and were used without further purification. Milli-Q

te

grade water (18.2 M Ω) was used to prepare all solutions. 2.2 Polystyrene Substrate Preparation

Ac ce p

Pre-stressed polystyrene (PS) substrates (Graphix Shrink Film, Graphix, Maple Heights, Ohio) were cut into the designed shape using the Robo Pro CE5000- 40-CRP cutter (Graphtec America Inc., Irvine, CA). The PS substrates were cleaned with ethanol, DI water, and then dried with air. Following the solvent cleaning step, the substrates were placed in an Expanded Plasma Cleaner (Harrick Plasma) and were treated on HIGH RF power setting for 60 seconds. Following the plasma treatment, the substrates were immersed in an aminosilane bath (10% APTES) in an Incubating Mini Shaker (VWR International) for 16 hours at 120 rpm and at room temperature. Following silanization, the substrates were sonicated in DI H2O for 10 minutes, rinsed and dried. The desired

Page 6 of 32

electrode pattern was designed in Adobe Illustrator and cut into the vinyl mask (FDC 4304, FDC graphic films, South Bend, Indiana) using the Robo Pro CE5000- 40-CRP vinyl cutter. The masks were applied to the silanized PS substrates.

ip t

2.3 Gold nanoparticle synthesis and deposition

Gold nanoparticles (Au NPs) were synthesized using previously described methods

cr

(Preparation 1)[23] and were kept at 4OC until used. The PS substrates covered with a

covered with a solution of Au NPs for 16 hours.

an

2.4 Electroless Deposition

us

vinyl mask were fixed in petri dishes using double sided tape. These substrates were then

The PS substrates covered by an AuNP seed layer were placed in an electroless

M

deposition bath containing 5 mL of 0.1% HAuCl4. The bath was placed on the incubation

d

mini shaker at 250 RPM and at room temperature, and 250 µL of 30% H2O2 was added to

te

the solution to initiate the Au deposition. Bubbles forming on the edges of the vinyl mask were eliminated using a pipette tip. After removing the vinyl masks, the devices were

Ac ce p

placed in an oven (ED56, Binder, Tuttlingen, Germany) at 150˚C for 3 minutes for shrinking the pre-stressed polystyrene. 2.5 Electrochemistry

The electrochemistry experiments were performed using Gamry reference 600 potentiostat (Gamry Instruments, Warminister, PA, USA) in a standard three-electrode cell. The electrochemical system consisted of an Ag/AgCl reference electrode, a platinum wire counter electrode, and the all-solution-processed Au electrode as the working electrode. All fabricated electrodes were electrochemically polished prior to use for surface area measurements,

Page 7 of 32

enzyme-free glucose sensing, and DNA hybridization experiments by performing cyclic voltammetry (CV) between 0 and 1.5 V for 80 cycles at 0.1 V/s in 0.05 M H2SO4.

ip t

2.6 Surface Area Measurements Electrochemical surface area measurements were performed by running another 3 cycles of

cr

cyclic voltammetry (in addition to the CVs performed for cleaning) using the same parameters used for electrode cleaning but at 0.05 V/s. The peaks for the reduction portion of the resulting

us

cyclic voltammograms were integrated to obtain the electric charge involved in the redox process

µC/cm2)[2] to acquire the values of surface area.

M

2.7 Enzyme-free Glucose Sensing

an

and divided by the surface charge density involved in forming a monolayer of oxide on Au (386

Enzyme-free glucose sensing was carried out by performing CV in 0, 2.5, 5, 7.5, 10, 15, and 20

d

mM solutions of dextrose dissolved in 0.1 M NaOH, 0.1 M NaOH and 0.1 M KCl, or 0.1 M

te

PBS. CV was performed with the voltage cycled between -0.8 and 0.8 V or -0.2 and 0.7 V at a

Ac ce p

scan rate of 0.1 V/s. To determine the signal generated by oxidation of glucose, the maximum point of the oxidation segment of the scan, which occurred at voltages between 0.2 and 0.3 V, was extracted. In the absence of glucose, the current was sampled in the voltage range for glucose oxidation. Selectivity of the glucose sensor was measured using chronoamperometry. Glucose concentrations from 1 to 10 mM were measured in a solution containing 0.1 M NaOH and 0.1 M KCl. The solution was placed in a mini-shaker operated at 200 rpm and the electrode’s response was measured using chronoamperometry at an applied potential of 0 V or 0.27 V. 0.1 mM ascorbic acid was added to measure the electrode’s response in the presence of interfering agents.

Page 8 of 32

The pulsed amperometric detection was performed by recording the current after each potential step at 0.27V (detection step). The electrode was subjected to oxidative cleaning and cathodic surface regeneration by applying 20 seconds of 0.8 V and 20 seconds of -0.6 V respectively. The

ip t

analyte was added and mixed at 200 rpm on the shaker before the surface regeneration time period. The shaker was stopped during each detection step. The current used in sensitivity

cr

calculations was a sampled and average current collected during the first 5 seconds of the

us

detection step. 2.8 Sensitivity Calculation

an

The analytical sensitivity of each electrode structure (µA/mM.cm2) is calculated by dividing the slope of the current versus concentration curve by the geometric surface area of the electrode. Results and Discussion

M

3.

d

3.1 Fabrication and Characterization of Nanoporous Wrinkled Electrodes

te

We sought to create hierarchical electrodes that combined features from the millimetre to the nanometre lengthscales inexpensively on the laboratory benchtop. All-solution-processing is

Ac ce p

ideally suited for creating low volume prototypes on the laboratory benchtop; however, it can also be combined with printing-based techniques for industry-scale large volume manufacturing.[24] Here we use an all-solution-processing method for creating threedimensional micro/nanostructured electrodes[16]. A sheet of pre-strained polystyrene (PS) is modified with an aminosiliane molecular linker. This linker contains a silane functional group at one terminus, and an amine functional group at the other terminus, which is used to link a gold nanoparticle (AuNP) seed layer to the substrate (Figure 1-i). In order to create an electrode layout on the substrate, we use xurography to create a vinyl mask defining the electrode pattern, as described previously [10]. The aminosilane-treated substrate is modified with a self-adhesive

Page 9 of 32

vinyl mask, and inserted into a AuNP bath to create a patterned seed layer (Figure 1-ii). The patterned seed layer is then used as a template for autocatalytic electroless deposition of gold (Figure 1-iii). Finally, the all-solution-processed gold electrode is wrinkled by heat shrinking the

ip t

PS substrate (Figure 1-iv). Previous work has demonstrated that shrinking the PS substrate, modified with a stiff thin film, induces wrinkles in the thin film due to the mismatch in the

cr

mechanical properties of the two materials[18].

us

Several of the previous results in polymer-induced conductive thin film wrinkling were obtained using thin films deposited using vapour-based techniques (sputtering, evaporation), which is

an

identified as an obstacle in using these methods in low-cost manufacturing[25]. In this study, we developed wrinkled electrodes by all-solution-processing. Scanning electron microscopy (SEM)

M

images obtained from structures developed at varying deposition times (Figure 1 (b)), reveal that

d

porous wrinkles are obtained at short deposition times (1 min), while non-porous continuous

te

wrinkles having larger features sizes are obtained at longer deposition times (20 min). The increase in wrinkle size is primarily due to the increase in film thickness with increasing

Ac ce p

electroless deposition time. The films deposited for 1 min and 20 min are approximately 17 nm and 56 nm thick, respectively. Increase in the film stiffness at longer deposition times could also contribute to the larger wrinkle sizes observed at 20 min. This is similar to the trend that was previously observed with wrinkled films that were created using vapour-based techniques [11]. It was previously reported that wrinkling can be used as a way of increasing the electroactive surface area per geometric surface area of electrodes [10]. Here, we measured the electroactive surface area of all-solution-processed electrodes by performing cyclic voltammetry (CV) in dilute sulfuric acid (Figure 1(c)). In the forward scan, the CV exhibits an oxidation current with an onset of ~1 V, and in the reverse scan, it displays a reduction current with an onset of ~1 V.

Page 10 of 32

The oxidation current is due to the formation of Au oxides, and the reduction current is the result of the reduction of Au oxides formed in the positive scan [26]. The electroactive surface area of the electrode is calculated using a widely used method for metals such as Au, which exhibit well-

ip t

developed redox features for oxide monolayer formation and reduction, and assumes that oxygen is chemisorbed in a monoatomic layer on the metal surface[27]. We calculate the electroactive

cr

surface area by extracting the charge transferred during the reduction step, and assume that the

us

reduction of a monolayer of Au oxides results in a charge density of 386 µC/cm2[4,27]. The 1 min nanoporous and wrinkled electrodes demonstrate the largest electroactive area with an

an

enhancement in surface area of 5.29 times over planar electrodes, followed by an enhancment of 4.19 times for 20 min wrinkled devices. The 1 min nanoporous wrinkled devices displayed the

M

largest enhancement in the electroactive surface, which can be attributed to the porosity within

d

the wrinkled structures. In summary, it is observed that wrinkling all-solution-processed

te

electrodes increases the electroactive surface area by packing three-dimensional wrinkles into the same footprint taken up by planar electrodes, while incorporating nanopores further enhances the

Ac ce p

electroactive surface area of wrinkled electrodes.

Page 11 of 32

ip t cr us an M d

te

Figure 1 Fabrication and characterization of all-solution-processed electrodes. (a) Fabrication

Ac ce p

of all-solution-processed wrinkled electrodes on pre-strained polystyrene (PS) involves plasma treatment and modification with an amino-silane molecular linker (i), deposition of a gold nanoparticle seed layer (ii), electroless deposition of gold at varying durations (iii), heat treating the substrate at 150OC for 3 minutes (iv) The transmission electron microscopy image in part (iv) demonstrates the cross-sectional view of the fabricated electrode. (b) SEM images of all solution-processed electrodes deposited for 20 min and planar, 1 min and wrinkled, and 20 min and wrinkled. (c) cyclic voltammogram of the fabricated electrodes in 0.05 M H2SO4 at a scan rate of 0.05 V/s performed using a Ag/AgCl reference electrode. d) electroactive surface area of electrodes in (b) measured by cyclic voltammetry in H2SO4. All error bars represent 1 standard deviation.

Page 12 of 32

3.2 Enzyme-free Glucose Sensing Enzyme-free glucose monitoring devices, in contrast to the commonly available systems based on enzyme electrodes, are critical for developing sensing systems that offer

ip t

sufficient stability for in-vivo applications. Given that enzyme-free glucose oxidation is a kinetically-controlled process[28], it is very sensitive to the nano/micro structure of the

cr

sensing electrodes. As a result, several nanostructured materials – nanowires[29],

developed

for

catalytic

oxidation

of

glucose

us

nanoporous structures[5,7,28,30,31], nanoparticles[32], and nanotubes[33] – have been without

the

use

of

enzymatic

an

mediators[34,35]. We hypothesized the nanoporous and high surface area gold electrodes fabricated here to be an effective platform for catalytic oxidation of glucose.

M

We created all-solution-processed sensors that consisted of a wrinkled gold electrode

d

connected to a contact pad through a wire covered by an insulating epoxy (Figure 2 (a)).

te

We performed cyclic voltammetry (CV) in the presence and absence of glucose in a basic electrolyte (0.1 NaOH) using the wrinkled and nanoporous electrodes (1 min). The CV

Ac ce p

scans presented in Figure 2 (b) demonstrate that in the absence of glucose, gold oxidation starts at ~0.25 V during the forward scan, and the created gold oxide starts to reduce at ~0.15 V during the revers scan. In addition, the redox feature observed at -0.16 V is attributed to the chemisorption of OH- on gold in alkaline solutions.[36] This CV behaviour is in agreement with previously reported Au CV scans in NaOH solutions.[36,37] In the presence of glucose, the forward scan contains a peak at ~-0.5 V (peak A), and a broad peak with an onset at ~-0.2 V that diminishes at ~0.4 V (peak B). In the reverse scan, the gold oxide reduction feature observed in the absence of glucose is present, and is followed by an oxidation peak with an onset at ~0.1 V (peak A’). Previous

Page 13 of 32

studies have determined that the most negative oxidation peak (peak A) is due to the adsorption and further oxidation of the aldehyde group at C1 of glucose catalyzed by AuOH.[38] The broad peak B can be deconvoluted into three peaks as observed

ip t

previously[36], and can be attributed to the oxidation of the primary alcohols at C6 of glucose, as well as the oxidation of enediol intermediates and the cleavage of the bond

cr

between carbon atoms C1 and C2 in glucose. An OH- layer formed on the electrode is

us

critical for the adsorption of glucose and its intermediates at the electrode surface [39]. In the reverse scan, once the gold oxide layer is reduced, a layer of AuOH is expected to

an

form, which is responsible for the reappearance of the glucose oxidation peak. A rich AuOH layer is important for catalyzing the various glucose oxidation pathways discussed

Ac ce p

te

d

M

here [38,40].

Figure 2-Glucose oxidation on all-solution processed electrodes. (a) Schematic of the fabricated gold electrodes (b) Cyclic voltammogram of 1 min nanoporous and wrinkled electrode in 0.1 M NaOH with (solid) and without (dotted) glucose. The arrows indicate the scan direction. In order to compare the sensitivity of the three structures, nanoporous and wrinkled, wrinkled, and planar, developed here for enzyme-free glucose detection, we analysed solutions containing a clinically-relevant concentration of glucose (3-8 mM)[28,32], as well as higher and lower concentrations outside this range. The CV scans obtained for

Page 14 of 32

varying glucose concentrations using the three structures are presented in Figure 3(a)-(c) for a potential window containing the most significant peaks related to glucose oxidation (peaks B and A’). It is evident that the magnitude of both peaks increase with increasing

ip t

glucose concentration; however, peak B shows a monotonic increase with a larger slope. We plotted the magnitude of the largest peak in the forward scan (in the 0.2-0.3 V range)

cr

as a function of the glucose concentration (Figure 3 (d)), and we summarized the glucose

us

sensitivity and surface area for each of the three structures (Table 1). Through this analysis, we observed that the sensitivity increases with surface area; however, the

an

observed sensitivity enhancement with respect to the planar electrode is larger than the surface area enhancement for the nanoporous structure and smaller than the surface area

M

enhancement for the wrinkled structure. This indicates that while increasing the surface

d

area is critical for increasing sensitivity, the type of structure (for eg. degree of porosity,

te

exposed crystalline structure, etc.)[4] is also important for enhancing the sensitivity. This is likely due to the differences in OH- adsorption on various structures, which influences

Ac ce p

their catalytic activity. It should also be noted that the planar structures developed here demonstrate a sensitivity (89 µA/mM.cm2) that is much larger than the values previously reported using gold disk electrodes (0.72 µA/mM.cm2)[7] [41]. We hypothesize this to be due to the roughness of the all-solution-processed planar electrodes fabricated here (Figure 2(b)), which demonstrate a roughness factor of 1.63 by dividing their electroactive surface area by their geometric surface area.

Page 15 of 32

ip t cr us an M d te Ac ce p

Figure 3 Enzyme-free glucose sensing using all-solution-processed electrodes. Cyclic voltammetry of 1 min and wrinkled (a), 20 min and wrinkled (b), and 20 min and planar (c) devices at varying glucose concentrations (0, 2.5, 5, 7.5, 10, 15, 20 mM). The upward arrows represent an increase in the glucose concentration. (d) The sensitivity of 1 min and wrinkled (red), 20 min and wrinkled (blue), and 20 min planar (black) devices measured from the largest peak observed in cyclic voltammograms in the 0.2-0.3 V range. All of the electrodes investigated here have the same geometric surface area. The inset represents the data displayed in (d)

Page 16 of 32

normalized to the geometric surface area. All voltammograms are plotted with respect to Ag/AgCl. All error bars represent 1 standard deviation. Table 1- Summary of the surface area measured by electrochemical methods and the sensitivity

d

M

an

us

cr

ip t

of the three classes of electrodes measured using cyclic voltammetry as displayed in Figure 3.

te

In order to determine the applicability of the highly sensitive nanoporous and wrinkled

Ac ce p

electrodes to glucose sensing under physiological conditions, we performed additional sensitivity studies under neutral pH and in the presence of interfering species. The presence of chloride ions have previously shown a decrease in the sensitivity of enzymefree glucose biosensors [5]. As a result, we measured the sensitivity of our nanoporous and wrinkled electrodes in the presence of a clinically-relevant concentration (100 mM) of KCl in alkaline solutions (Figure 4 (a)). This shows that the device sensitivity changes insignificantly in the presence of chloride ions. In addition, we performed the CV analysis at varying concentrations of glucose by replacing the alkaline NaOH electrolyte with a neutral phosphate buffer solution (PBS), which is a closer representation of physiological conditions. In the forward cycle, the CV demonstrates a broad oxidation peak with one

Page 17 of 32

peak/plateau in the 0.0 to 0.1 V range and one peak in the 0.25-0.30 V range, followed by a current decrease at potentials larger than 0.45 V due to the formation of gold oxides. In the reverse cycle of the CV, a broad peak with a magnitude larger than the peak seen in

ip t

the forward cycle (> 5mM) is observed with one peak in the 0.0-0.1 V range and one peak in the 0.25-0.30 V range. This behaviour is similar to the previously reported results with

cr

nanostructured gold electrodes in PBS electrolytes[5]. During the forward cycle, the more

us

negative peak is attributed to the formation of the Au(OH)ads layer, electrosorption of glucose, and generation of intermediates, while the more positive peak is described by the

an

oxidation of intermediates and the direct oxidation of glucose. The peaks observed in the reverse cycle can be attributed to the same glucose oxidation pathways as the forward

M

scan, which is enabled by the reduction of gold oxide on the electrode surface. Previous

d

reports have indicated that the oxygen released during the reduction of the gold oxide

te

could act as an intermediate for glucose oxidation. This could explain the enhancement of the oxidation peaks observed in the reverse cycle. [42] In order to determine the

Ac ce p

sensitivity of the nanoporous and wrinkled structures in detecting glucose in PBS solutions, we plotted the magnitude of the largest peak in the reverse cycle of the CV scan versus the glucose concentration (Figure 4 (c)). Through this analysis, we obtain a curve with a dual slope having a sensitivity of 38 µA/mM.cm2 until 7.5 mM and a sensitivity of 23 µA/ mM.cm2 from 7.5 mM to 20 mM. This dual-slope behaviour has been previously reported, and is likely due to the decreased interactions between the catalytic Au(OH)ads layer and glucose molecules at increased glucose concentrations [39]. For practical sensing applications, the first regime lies within the reported normal physiological range of glucose. The second regime could also be used to quantify glucose at higher concentrations with less

Page 18 of 32

sensitivity if a dual slope calibration is performed. It was previously determined that glucose oxidation is kinetically-controlled at concentrations larger than 2 mM [39]. As a result, we hypothesize the reduced slope observed here (with an onset at 7.5 mM) to be related to the

ip t

changes occurring at the electrode surface. Overall, the sensitivity of the nanoporous and wrinkled electrodes significantly diminishes in PBS compared to NaOH. This has been

cr

previously reported [4]and is related to the role of NaOH in increasing the rate of

us

dehydrogenation processes involved in glucose oxidation, the formation of a more easily oxidized glucose intermediates, and the increased coverage of the catalytic Au(OH)ads layer in

Ac ce p

te

d

M

an

alkaline solutions compared to neutral solutions [39].

Page 19 of 32

ip t cr us an M d te Ac ce p

Figure 4 Interference study of the nanoporous and wrinkled enzyme-free glucose biosensor. (a) Sensitivity analysis of 1 min devices at varying concentrations of glucose in the presence (black) and absence of 0.1 M KCl (red). The peak current was extracted from the CV peak at ~0.27 V. (b) Cyclic voltammogram of 1 min devices in 1X PBS solution at 0, 2.5, 5, 7.5, 10, 15, 20 mM glucose concentrations. The arrow demonstrates an increase in the glucose concentration. (c) The sensitivity of the 1 min devices in 1x PBS extracted from the largest reverse peak of the CV curves in (b). The chronoamperometry response of a 1 min device in a 0.1 M KCl and 0.1 M NaOH electrolyte under an applied potential of 0.27 V (d) and 0 V (e). 0.1 mM ascorbic acid

Page 20 of 32

was added at 60 s and glucose was added every 30 s after that to obtain an increase in the glucose concentration of 1 mM, until a final (10 mM) glucose concentration was achieved. (f) Current versus concentration curves extracted from the data presented in (e) and (f). Red

ip t

squares indicate measurements at 0.27 V and black circles indicate measurement at 0 V. All error bars represent 1 standard deviation.

cr

In order to benchmark the sensitivity of the nanoporous and wrinkled electrodes against

us

other enzyme-free glucose sensors and to evaluate their response in the presence of interfering species, we performed chronoamperometry (CA) scans in the presence of KCl

an

and ascorbic acid and at varying glucose concentrations (Figure 4 (d)-(f)). We initially monitored the current at an applied potential of 0.27 V, which is centered around the peak

M

potentials observed in CV scans obtained in alkaline electrolytes (NaOH and KCl)

d

(Figure 4 (d), (f)). It was observed that adding ascorbic acid (0.1 mM), changes the

te

current magnitude by an amount that is considerably lower (seven times) than the change observed with adding glucose. In spite of this, 0.27 V (w.r.t. Ag/AgCl) is considered a

Ac ce p

relatively high oxidation potential, which might lead to the generation of large background currents when analyzing complex chemical or biological samples. This might require the integration of the electrodes developed here with separation techniques or membranes.

In addition, we observe that the slope of the current versus concentration curve (Figure 4 (f)) decreases beyond the glucose concentration of 3 mM. This decrease in sensitivity at increased glucose concentration has been previously observed and explained as follows. [39] The sensitivity of the electrode to glucose is concentration and potential dependent. At the anodic peak ~0.27 V, glucose oxidation is dominated by the pairing of the enediol

Page 21 of 32

conformation of glucose with the Au(OH)ads layer. Through this process, a single glucose molecule interacts with multiple Au(OH) complexes, resulting in multiple oxidations per glucose molecule. As the glucose concentration increases, there will be a smaller ratio of

ip t

glucose to Au(OH) complexes reducing the rate of oxidation. In addition, the intermediate products generated in previous oxidative steps could be accumulating and reducing the

cr

access to the Au(OH)ads layer. The observation that this reduction in slope was not

us

observed in sensitivity measurements using CV scanning also supports this explanation. We hypothesized that operating the electrode at a lower potential would alleviate the

an

diminishing glucose sensitivity at higher glucose concentrations (> 3mM) due to the decreased oxidation rates at lower potentials and the subsequent reduction of the

M

accumulation of the reaction intermediates or products. A linear trend was observed upon

d

measuring the CA response of the electrodes at 0 V in a 0-10 mM glucose solution

te

(Figure 4 (e)), and a sensitivity of 222 µA/mM.cm2 was measured (Figure 4 (f)). This sensitivity is among the highest ever reported for enzyme-free glucose biosensing using

Ac ce p

gold electrodes[4]; however, it is lower than the sensitivity measured here using CV scanning. We expect the cycling performed during CV to be responsible for regenerating the Au(OH)ads layer after each scan and the resultant enhanced performance of the glucose biosensor.

In order to alleviate the loss of performance observed when using chronoamperometry (reduced slope and linear range in case of 0.27 V), we used pulsed amperometric detection (PAD) with electrode regeneration steps to measure the analytical sensitivity of our biosensors. It was previously demonstrated that a multi-step potential waveform can be used to regenerate the electrode surface between subsequent measurements to improve the

Page 22 of 32

electrode’s sensitivity in multiple measurements. [43,44] For this purpose, we applied a potential waveform (Figure 2(a)-(b)) with a positive step (0.8 V), a negative step (-0.6 V), and a detection step (0.27 V) for oxidative cleaning, cathodic regeneration, and analyte detection

ip t

respectively.[44] This potential waveform was periodically repeated and the analyte of interest was added right before the cathodic regeneration step. It was observed that adding interference

cr

agents (KCl and AA) did not cause a significant signal increase; however increasing the

us

solution’s glucose concentration by 1 mM increments caused an increase in the measured signal during the 0.27 potential step (Figure 5 (a)). The value of the measured current was sampled and

an

averaged during the first 5 seconds of the detection step, and was plotted as a function of the solution concentration (Figure 5 (c)). The porous and wrinkled electrodes evaluated using PAD

M

demonstrate a sensitivity of 794 µA/ (mM.cm2). This sensitivity did not diminish throughout the

d

interrogated concentration range, and was higher than the sensitivity measured using CV and CA

te

(Table 2). We believe this enhanced sensitivity to be related to the added oxidative cleaning and cathodic regeneration steps of the PAD method. It is hypothesized that these steps remove the

layer.

Ac ce p

reaction intermediates from the surface and enable the analytes to access the catalytic Au(OH)ads

Page 23 of 32

ip t cr us an M

d

Figure 5-Sensitivity measurement using Pulsed Amperometric Detection (PAD). (a) The PAD

te

signal of a nanoporous and wrinkled electrode. The arrows indicate the first (1 mM KCl), second

Ac ce p

(0.01 mM AA), and third (1 mM glucose) addition of analyte, while the subsequent additions are not marked. The letters on the graph represent the potential steps demonstrated in (b). (b) Demonstration of one period of the multi-step waveform applied in (a). (c) The plot of the current versus concentration curve. The data is extracted and averaged from the first 5 s of the measured PAD current during the detection step. 4.

Conclusions

In this work, we demonstrated a facile and benchtop method for creating nanoporous and wrinkled biosensing electrodes. It was shown that wrinkling increases the electrode surface area, while nanoporosity results in additional enhancements in surface area. It was further demonstrated that the nanoporous and wrinkled electrodes are applicable to enzyme-free glucose

Page 24 of 32

detection, and they offer the highest sensitivity amongst their planar and wrinkled counterparts when investigated using CV in alkaline solutions. The sensitivity of these electrodes changed insignificantly when analyzed in alkaline electrolytes that contained 0.1 M KCl; however it

ip t

reduced significantly when analysed in neutral phosphate buffer solution electrolytes. In addition to CV, the sensitivity of the nanoporous and wrinkled electrodes were measured using CA and

cr

the effect of a physiologically-relevant ascorbic acid concentration was evaluated. The electrode

us

sensitivity was reduced when using CA instead of CV; however, the change induced by the ascorbic acid addition was insignificant. In addition to CA and CV, the electrode sensitivity was

an

measured using PAD, which demonstrated the highest sensitivity amongst the three measurement methods. The sensitivity of the nanoporous and wrinkled electrodes significantly varies

M

depending on the type of measurement used, with methods that offer electrode regeneration

d

displaying the highest sensitivity. Overall, the nanoporous and wrinkled electrodes developed

te

here present an excellent sensitivity compared to the previously developed enzyme-free glucose biosensors as summarized in Table 2.

Ac ce p

Table 2. Summary of the sensitivity of enzyme-free glucose biosensors made from gold. Structure Solution Measurement Sensitivity Measured Reference 2 type method (µA/mM.cm ) Linear Range (mM) Au-wrinkled Alkaline Voltammetric 591 2.5-20 This work and nanoporous Au-wrinkled Alkaline Amperometric 222 1-10 This work and nanoporous Au-wrinkled Alkaline Pulsed 794 1-10 This work and Amperometric nanoporous Au-wrinkled Neutral Amperometric 38/23 2.5-20 This work and nanoporous AuAlkaline Amperometric 10.8 0.01-28 [45] nanoporous

Page 25 of 32

Amperometric

20.1

1-20

[5]

Neutral Alkaline

Amperometric Voltammetric

22.6 87.5

0.05-30 0.1-25

[4] [46]

Neutral Neutral Neutral

Amperometric Amperometric Amperometric

179 191 232

0-8 0.1-25 0.01-11

[47] [48] [7]

ip t

Neutral

cr

Aunanoporous Au-nanocoral Au-NP-glassy carbon Au-NP-Au Au-dendrites Aunanoporous

us

Acknowledgements

The financial support for this work was provided by the Natural Sciences and Engineering

an

Council of Canada (NSERC), Ontario Ministry of Research and Innovation, and the Ontario Graduate Scholarship. L.S is the Canada Research Chair in Miniaturized Biomedical Devices.

M

The electron microscopy was carried out at the Canadian Centre for Electron Microscopy (CCEM), a national facility supported by the NSERC and McMaster University.

d

We acknowledge the support of Calvin Wong for setting up the chronoamperometry experiments

[1]

Ac ce p

References

te

and Travis Casagrande for SEM imaging.

S. Park, T.D. Chung, H.C. Kim, Nonenzymatic glucose detection using mesoporous

platinum, Anal. Chem. 75 (2003) 3046–3049. doi:10.1021/ac0263465. [2]

A.J. Bard, L.R. Faulkner, Electrochemical methods: fundamentals and applications, 2nd

ed., John Wiley & Son, Inc., New York, 2001. [3]

L. Soleymani, F. Li, The mechanistic challenges and advantages of biosensor

miniaturization into the nanoscale, (2017). doi:10.1021/acssensors.7b00069. [4]

T.M. Cheng, T.K. Huang, H.K. Lin, S.P. Tung, Y.L. Chen, C.Y. Lee, H.T. Chiu, (110)-

Exposed gold nanocoral electrode as low onset potential selective glucose sensor, ACS Appl. Mater. Interfaces. 2 (2010) 2773–2780. doi:10.1021/am100432a.

Page 26 of 32

[5]

L.Y. Chen, X.Y. Lang, T. Fujita, M.W. Chen, Nanoporous gold for enzyme-free

electrochemical glucose sensors, Scr. Mater. 65 (2011) 17–20. doi:10.1016/j.scriptamat.2011.03.025. Z. Liu, L. Huang, L. Zhang, H. Ma, Y. Ding, Electrocatalytic oxidation of d-glucose at

ip t

[6]

nanoporous Au and Au-Ag alloy electrodes in alkaline aqueous solutions, Electrochim. Acta. 54

Y. Xia, W. Huang, J. Zheng, Z. Niu, Z. Li, Nonenzymatic amperometric response of

us

[7]

cr

(2009) 7286–7293. doi:10.1016/j.electacta.2009.07.049.

glucose on a nanoporous gold film electrode fabricated by a rapid and simple electrochemical

[8]

an

method, Biosens. Bioelectron. 26 (2011) 3555–3561. doi:10.1016/j.bios.2011.01.044. M.H. Yang, J.M. Jeong, K.G. Lee, D.H. Kim, S.J. Lee, B.G. Choi, Hierarchical porous

M

microspheres of the Co3O4@graphene with enhanced electrocatalytic performance for

S. Fan, M. Zhao, L. Ding, J. Liang, J. Chen, Y. Li, S. Chen, Synthesis of 3D hierarchical

te

[9]

d

electrochemical biosensors, Biosens. Bioelectron. (2015) 1–8. doi:10.1016/j.bios.2016.01.075.

porous Co3O4 film by eggshell membrane for non-enzymatic glucose detection, J. Electroanal.

[10]

Ac ce p

Chem. 775 (2016) 52–57. doi:10.1016/j.jelechem.2016.05.035. C.M. Gabardo, Y. Zhu, L. Soleymani, J.M. Moran-Mirabal, Bench-Top Fabrication of

Hierarchically Structured High-Surface-Area Electrodes, Adv. Funct. Mater. 23 (2013) 3030– 3039. doi:10.1002/adfm.201203220. [11]

C.C. Fu, A. Grimes, M. Long, C.G.L. Ferri, B.D. Rich, S. Ghosh, S. Ghosh, L.P. Lee, A.

Gopinathan, M. Khine, Tunable nanowrinkles on shape memory polymer sheets, Adv. Mater. 21 (2009) 4472–4476. doi:10.1002/adma.200902294. [12]

J.D. Pegan, A.Y. Ho, M. Bachman, M. Khine, Flexible shrink-induced high surface area

electrodes for electrochemiluminescent sensing., Lab Chip. 13 (2013) 4205–9.

Page 27 of 32

doi:10.1039/c3lc50588j. [13]

S. Sonney, N. Shek, J.M. Moran-Mirabal, Rapid bench-top fabrication of

electrodes, Biomicrofluidics. (2015) 26501. doi:10.1063/1.4918596. [14]

ip t

poly(dimethylsiloxane)/polystyrene microfluidic devices incorporating high-surface-area sensing

C.M. Gabardo, A.M. Kwong, L. Soleymani, Rapidly prototyped multi-scale electrodes to

cr

minimize the voltage requirements for bacterial cell lysis, Analyst. 140 (2015) 1599–1608.

[15]

us

doi:10.1039/C4AN02150A.

L. Zhang, X. Lang, A. Hirata, M. Chen, Wrinkled nanoporous gold films with ultrahigh

an

surface-enhanced raman scattering enhancement, ACS Nano. 5 (2011) 4407–4413. doi:10.1021/nn201443p.

C.M. Gabardo, R.C. Adams-McGavin, B.C. Fung, E.J. Mahoney, Q. Fang, L. Soleymani,

M

[16]

d

Rapid prototyping of all-solution-processed multi-lengthscale electrodes using polymer-induced

[17]

te

thin film wrinkling, Sci. Rep. 7 (2017) 42543. doi:10.1038/srep42543. C.M. Gabardo, J. Yang, N.J. Smith, R.C. Adams-McGavin, L. Soleymani, Programmable

Ac ce p

Wrinkling of Self-Assembled Nanoparticle Films on Shape Memory Polymers, ACS Nano. 10 (2016) 8829–8836. doi:10.1021/acsnano.6b04584. [18]

J. Genzer, J. Groenewold, Soft matter with hard skin: From skin wrinkles to templating

and material characterization, Soft Matter. 2 (2006) 310. doi:10.1039/b516741h. [19]

C.M. Gabardo, R.C. Adams-McGavin, O.M. Vanderfleet, L. Soleymani, Rapid

prototyping of microfluidic devices with integrated wrinkled gold micro-/nano textured electrodes for electrochemical analysis, Analyst. 140 (2015) 5781–5788. doi:10.1039/C5AN00774G. [20]

S.M. Woo, C.M. Gabardo, L. Soleymani, Prototyping of wrinkled nano-/microstructured

Page 28 of 32

electrodes for electrochemical DNA detection., Anal. Chem. 86 (2014) 12341–7. doi:10.1021/ac5036296. [21]

S. Lin, E.K. Lee, N. Nguyen, M. Khine, Thermally-induced miniaturization for micro-

ip t

and nanofabrication: progress and updates., Lab Chip. 14 (2014) 3475–88. doi:10.1039/c4lc00528g.

M. Tominaga, T. Shimazoe, M. Nagashima, I. Taniguchi, Electrocatalytic oxidation of

cr

[22]

us

glucose at gold nanoparticle-modified carbon electrodes in alkaline and neutral solutions, Electrochem. Commun. 7 (2005) 189–193. doi:10.1016/j.elecom.2004.12.006. K.C. Grabar, R.G. Freeman, M.B. Hommer, M.J. Natan, Preparation and Characterization

an

[23]

of Au Colloid Monolayers, Anal. Chem. 67 (1995) 735–743. doi:10.1021/ac00100a008. C.M. Gabardo, L. Soleymani, Deposition, patterning, and utility of conductive materials

M

[24]

[25]

te

doi:10.1039/C6AN00210B.

d

for the rapid prototyping of chemical and bioanalytical devices, Analyst. (2016) 3511–3525.

S. Lin, E.K. Lee, M. Khine, Thermally-induced miniaturization for micro- and

Ac ce p

nanofabrication: progress and updates, Lab Chip. 14 (2014) 3475–3488. doi:10.1039/C4LC00528G. [26]

E. Rouya, S. Cattarin, M.L. Reed, R.G. Kelly, G. Zangari, Electrochemical

Characterization of the Surface Are of Nanoporous Gold Films, J. Electrochem. Soc. 159 (2012) K97–K102. doi:10.1149/2.097204jes. [27]

S. Trasatti, O.A. Petrii, Real Surface Area Measurements In Electrochemistry, Pure Appl.

Chem. 63 (1991) 711–734. doi:10.1351/pac199163050711. [28]

S. Park, T.D. Chung, H.C. Kim, Nonenzymatic glucose detection using mesoporous

platinum, Anal. Chem. 75 (2003) 3046–3049. doi:10.1021/ac0263465.

Page 29 of 32

[29]

Y. Zhang, L. Su, D. Manuzzi, H.V.E. de los Monteros, W. Jia, D. Huo, C. Hou, Y. Lei,

Ultrasensitive and selective non-enzymatic glucose detection using copper nanowires, Biosens. Bioelectron. 31 (2012) 426–432. doi:10.1016/j.bios.2011.11.006. H. Boo, S. Park, B. Ku, Y. Kim, J.H. Park, H.C. Kim, T.D. Chung, Ionic Strength-

ip t

[30]

Controlled Virtual Area of Mesoporous Platinum Electrode, J. Am. Chem. Soc. 126 (2004)

Y. Li, Y.Y. Song, C. Yang, X.H. Xia, Hydrogen bubble dynamic template synthesis of

us

[31]

cr

4524–4525. doi:10.1021/ja0398316.

porous gold for nonenzymatic electrochemical detection of glucose, Electrochem. Commun. 9

[32]

an

(2007) 981–988. doi:10.1016/j.elecom.2006.11.035.

F. Kurniawan, V. Tsakova, V.M. Mirsky, Gold nanoparticles in nonenzymatic

M

electrochemical detection of sugars, Electroanalysis. 18 (2006) 1937–1942.

K.B. Male, S. Hrapovic, Y. Liu, D. Wang, J.H.T. Luong, Electrochemical detection of

te

[33]

d

doi:10.1002/elan.200603607.

carbohydrates using copper nanoparticles and carbon nanotubes, Anal. Chim. Acta. 516 (2004)

[34]

Ac ce p

35–41. doi:10.1016/j.aca.2004.03.075.

S. Park, H. Boo, T.D. Chung, Electrochemical non-enzymatic glucose sensors, Anal.

Chim. Acta. 556 (2006) 46–57. doi:10.1016/j.aca.2005.05.080. [35]

P. Si, Y. Huang, T. Wang, J. Ma, Nanomaterials for electrochemical non-enzymatic

glucose biosensors, RSC Adv. 3 (2013) 3487–3502. doi:10.1039/c2ra22360k. [36]

S.-P. Tung, T.-K. Huang, C.-Y. Lee, H.-T. Chiu, Electrochemical growth of gold

nanostructures on carbon paper for alkaline direct glucose fuel cell, RSC Adv. 2 (2012) 1068– 1073. doi:10.1039/c1ra00611h. [37]

P. Rodriguez, N. Garcia-Araez, M.T.M. Koper, Self-promotion mechanism for CO

Page 30 of 32

electrooxidation on gold., Phys. Chem. Chem. Phys. 12 (2010) 9373–80. doi:10.1039/b926365a. [38]

P. Parpot, N. Nunes, A.P. Bettencourt, Electrocatalytic oxidation of monosaccharides on

gold electrode in alkaline medium: Structure-reactivity relationship, J. Electroanal. Chem. 596

[39]

ip t

(2006) 65–73. doi:10.1016/j.jelechem.2006.07.006.

L. A. Larew, D.C. Johnson, Concentration dependence of the mechanism of glucose

cr

oxidation at gold electrodes in alkaline media, J. Electroanal. Chem. Interfacial Electrochem. 262

[40]

us

(1989) 167–182. doi:10.1016/0022-0728(89)80020-8.

P. Parpot, P.R.B. Santos, A.P. Bettencourt, Electro-oxidation of d-mannose on platinum,

an

gold and nickel electrodes in aqueous medium, J. Electroanal. Chem. 610 (2007) 154–162. doi:10.1016/j.jelechem.2007.07.011.

M. Wooten, J.H. Shim, W. Gorski, Amperometric determination of glucose at

M

[41]

d

conventional vs. nanostructured gold electrodes in neutral solutions, Electroanalysis. 22 (2010)

[42]

te

1275–1277. doi:10.1002/elan.201000006.

E.B. MAKOVOS, C.C. LIU, a Cyclic-Voltammetric Study of Glucose-Oxidation on a

Ac ce p

Gold Electrode, Bioelectrochemistry Bioenerg. 15 (1986) 157–165. doi:10.1016/03024598(86)80023-X. [43]

D.C. Johnson, W.R. LaCourse, Liquid Chromatography with Pulsed Electrochemical

Detection at Gold and Platinum Electrodes, Anal. Chem. 62 (1990) 589A–597A. doi:10.1021/ac00209a715. [44]

R.D. Rocklin, A.P. Clarke, M. Weitzhandler, Improved Long-Term Reproducibility for

Pulsed Amperometric Detection of Carbohydrates via a New Quadruple-Potential Waveform, Anal. Chem. 70 (1998) 1496–1501. doi:10.1021/ac970906w. [45]

L. Han, S. Zhang, L. Han, D.P. Yang, C. Hou, A. Liu, Porous gold cluster film prepared

Page 31 of 32

from Au@BSA microspheres for electrochemical nonenzymatic glucose sensor, Electrochim. Acta. 138 (2014) 109–114. doi:10.1016/j.electacta.2014.06.095. [46]

G. Chang, H. Shu, K. Ji, M. Oyama, X. Liu, Y. He, Gold nanoparticles directly modified

ip t

glassy carbon electrode for non-enzymatic detection of glucose, Appl. Surf. Sci. 288 (2014) 524– 529. doi:10.1016/j.apsusc.2013.10.064.

B.K. Jena, C.R. Raj, Enzyme-Free Amperometric Sensing of Glucose by Using Gold

[48]

us

Nanoparticles, (2006) 2702–2708. doi:10.1002/chem.200501051.

cr

[47]

H. Shu, L. Cao, G. Chang, H. He, Y. Zhang, Y. He, Direct electrodeposition of gold

an

nanostructures onto glassy carbon electrodes for non-enzymatic detection of glucose,

Ac ce p

te

d

M

Electrochim. Acta. 132 (2014) 524–532. doi:10.1016/j.electacta.2014.04.031.

Page 32 of 32