Ge

Na

b

c

d

a

AA

KNFGECM

1

icrvdabNtaaaa

nT

0h

Sensors and Actuators B 187 (2013) 455– 460

Contents lists available at SciVerse ScienceDirect

Sensors and Actuators B: Chemical

journa l h om epage: www.elsev ier .com/ locat e/snb

as-assisted focused ion beam fabrication of gold nanoelectrode arrays inlectron-beam evaporated alumina films for microfluidic electrochemical sensors

apat Triroja,c,∗, Papot Jaroenapibalb,d, Roderic Beresfordc

Department of Electrical Engineering, Faculty of Engineering, Khon Kaen University, Khon Kaen 40002, ThailandDepartment of Industrial Engineering, Faculty of Engineering, Khon Kaen University, Khon Kaen 40002, ThailandDivision of Engineering, Brown University, 182 Hope Street, Providence, RI 02912, USADepartment of Materials Science and Engineering, University of Pennsylvania, 3231 Walnut Street, Philadelphia, PA 19104, USA

r t i c l e i n f o

rticle history:vailable online 30 January 2013

eywords:anoelectrode arraysocused ion beamas-assisted millinglectron-beam evaporated aluminayclic voltammetryicrofluidic electrochemical sensors

a b s t r a c t

This work reports the fabrication details used to prepare gold nanoelectrode arrays as working electrodesin a microfluidic electrochemical sensor system. We emphasize on the design criteria and the methodsused to construct the nanoelectrode arrays for the purpose of enhancing mass transport of the redoxspecies to the electrodes. The fabrication techniques include electron-beam evaporation of Ti and Auto create the electrode platform on a glass substrate, followed by electron-beam evaporation of thickalumina films as a passivation layer. I2-assisted focused ion beam (FIB) milling is employed to producehigh-aspect-ratio pores in the alumina films. The final nanopores with diameters around 60–120 nm areachieved by ion beam sculpting after the initial milling process. The nanopores are filled with Au viaelectrodeposition to obtain a nanoelectrode structure. The current responses of Ru(NH3)6

3+/Ru(NH3)62+

redox species measured at three-electrode cells in polydimethylsiloxane (PDMS) microfluidic chips areinvestigated by cyclic voltammetry (CV). The current density obtained at the nanoelectrode arrays isincreased by two orders of magnitude compared to that obtained from the microelectrode system. Inaddition, due to the effect of enhanced mass transport, the voltammetric response curve acquired fromthe nanoelectrode array exhibits large deviations from the classical diffusion-limited current transporttheory.

. Introduction

The potential to realize microfluidic electrochemical sensorss recognized as a promising solution for miniaturized analyti-al systems to provide manipulation and detection of chemicaleactions in nano/micro-liter fluid samples. Because of the smallolume, microfluidic devices are ideal for handling costly andifficult-to-obtain samples and reagents. A microchannel platformlso offers fast analysis time due to reduced diffusion distanceetween the sensing units and the target species to be detected.ot only do microfluidic devices make complex analyses simpler

o be performed, but they also have other distinct advantages suchs low power, decreased cost and waste production, portability,

nd disposability. Integrated electrochemical microfluidic systemsre considered to offer important benefits to an extensive range ofpplications, including biological research, real-time detection and

∗ Corresponding author at: Department of Electrical Engineering, Faculty of Engi-eering, Khon Kaen University, Khon Kaen 40002, Thailand.el.: +66 43 202 353x115; fax: +66 43 202 836.

E-mail address: napat@elec.kku.ac.th (N. Triroj).

925-4005/$ – see front matter © 2013 Elsevier B.V. All rights reserved.ttp://dx.doi.org/10.1016/j.snb.2013.01.049

© 2013 Elsevier B.V. All rights reserved.

monitoring of biological threats, and disease diagnosis in point-of-care medical tests.

While miniaturization of fully integrated microfluidic electro-chemical sensors is the ultimate goal for decentralized analysis,there is a tradeoff between the sample size and the detection limit.When the volume of the solution is reduced, in a fixed concentra-tion, the total number of molecules available to be detected in aminute volume is also reduced. In some cases, the measured cur-rent can be as low as sub-picoampere. Transducing such extremelylow-level signal requires a highly sensitive detector. This limi-tation is circumvented by employing highly efficient electrodesof nanometer sizes in microfluidic systems. These electrodespossess unique properties which create the possibility of acquiringsuch low-level electrochemical signals. Theoretical studies haveshown that the electrochemical behavior of nanoscale electrodesdeviates significantly from the predictions of traditional theories.The departure is found to be dependent on the electrode size as itapproaches values of molecular dimensions. Those deviations are,

for example, the high electric field generated near the electrodesurface that causes an enhancement or inhibition of the flux of theelectroactive species to the electrode surface [1], the departureof the steady-state current response from that under the classical

4 Actua

ddb

iioaa

wcrtdcn

i

werhTpnntTecta

�

wεoewaosotwomnsdte

nasnseag

insulation may affect the current response when the films areused for the passivation layer of nanoelectrodes. Electron-beamevaporated alumina (Al2O3) is developed as an insulating layer inthis work because the fabrication method is compatible with the

Fig. 1. The fabrication steps used to prepare a nanoelectrode array in a three-

56 N. Triroj et al. / Sensors and

iffusion control theory [2], and the shift of the half-wave potentialue to the increase in the current density [3]. Such deviations cane observed from the shape of the voltammetric response curves.

As the electrode size decreases, radial diffusion becomes dom-nant. When a net current flows across the electrode/solutionnterface, a depletion layer close to the electrode surface is devel-ped. The thickness of the depletion layer around the electrode witht least one dimension smaller than 10 �m becomes hemisphericalnd is given by Eq. (1) [4]:

1ı(t)

= 1√�Dt

+ 1r

(1)

here ı is the thickness of the depletion layer, D is the diffusionoefficient of the species, t is the time scale of the experiment, and

is the radius of the electrode. As the electrode decreases in size,he depletion layer thickness approaches the size of the electrodeimension. Besides, the steady-state diffusion-controlled limitingurrent is proportional to the inverse of the depletion layer thick-ess as shown in Eq. (2) [4]:

(t → ∞) = nFDAC∗

ı(t → ∞)(2)

here n is electron stoichiometry, F is faraday constant, A is thelectrode surface area, and C* is the bulk concentration of theedox species. According to Eq. (2), smaller electrode providesigher measured current density due to thinner depletion layer.his phenomenon is called enhanced mass transport, which is therimary reason to employ nanoscale electrodes in low-level sig-al detection. Furthermore, the electric field and charge within theear-electrode layers become significant influences on the massransport and electron transfer kinetics at the nanoscale electrodes.he region close to the electrode surface where a high electric fieldxists is described as the electrical double layer whose thickness isharacterized by the Debye length, �−1. The Debye length is propor-ional to the inverse square root of the ionic concentration, C−1/2,s predicted in the Debye–Huckel theory [5] as shown in Eq. (3):

−1 = 1zF

√ε0εrRT

2C(3)

here z is the charge magnitude of each ion in a (z, z) electrolyte,0 is the permittivity of vacuum, εr is the static dielectric constantf the medium, R is the gas constant, and T is the temperature. Thelectric field established at equilibrium is found to be containedithin a distance of at most 3�−1 [5]. For electrodes having radius

bout a few tens of nanometers, the electrical double layer begins toccupy an appreciable fraction of the depletion layer. The electro-tatic force within the double layer will accelerate or retard the fluxf the charged species to the electrode surface. Therefore, migra-ion resulting from the high electric field within the double layerill be added to the mass transport equation. The small feature size

f the nanoelectrodes also makes their use desirable because theeasured currents at the nanoelectrodes exhibit low background

oise owing to minimal charging current that is proportional to theurface area of the electrode. Therefore, nanoscale electrodes areesirable in low-level signal measurements to improve the signal-o-noise ratio and to enhance the sensitivity of the microfluidiclectrochemical sensors.

Taking into account that the preceding unique properties ofanoelectrodes give rise to the measured current density, researchctivities in the development of innovative electrochemical sen-ors have been focusing on the fabrication and characterization ofanoscale electrodes. However, the field of nanoelectrochemistry

till remains in its infancy considering the difficulty to producespecially well-aligned and low-packing-density nanoelectroderrays that yield multiple responses of that obtained from a sin-le nanoelectrode. Approaches used to produce nanopore array

tors B 187 (2013) 455– 460

electrodes include electron beam lithography [6], nanoimprintlithography [6], and focused ion beam technique [7]. To maxi-mize the signal-to-noise ratio, the concerns for the nanoelectrodearray fabrication are rather with the inter-electrode distance, whichshould be sufficiently large compared to the electrode radius inorder to prevent overlapping of the depletion zones [8–10]. Inaddition, the choice of materials used for the nanoelectrodes’ pas-sivation layer is also crucial such that the films should have highdensity to prevent any leakage current paths, as well as they shouldbe thick enough to minimize parasitic coupling. In this work, wereport detailed fabrication steps used to construct the nanoelec-trode arrays employed as the working electrodes in the microfluidicelectrochemical sensor system. We also show an experimentalevidence of the enhanced mass transport effect occurred at thenanoelectrode array.

2. Experimental

2.1. Fabrication of the electrode platforms and the aluminainsulation layer on a glass substrate

The fabrication process consists of preparation of dense alu-mina films as the insulating layer via electron-beam evaporation,fabrication of a high-aspect-ratio nanoscale pore structure in theinsulating layer through I2 gas-assisted focused ion beam (FIB)milling, followed by ion beam sculpting, and electrodepositionof Au to fill the nanopores. Fig. 1 illustrates the fabrication stepsused to prepare a nanoelectrode array in a three-electrode cell,consisting of a working electrode (WE), a reference electrode (RE),and a counter electrode (CE), embedded in a polydimethylsiloxanePDMS microchannel.

The electrode platforms composed of 20-nm Ti as adhesionlayer and 300-nm thick Au are prepared on a cleaned glass slideusing UV lithography, electron-beam evaporation, and lift-off.To prepare the electrodes utilized in aqueous solutions, densefilms for the insulating sheath are preferred because they arechemically resistant. Besides, the leakage current of the porous

electrode cell embedded in a PDMS microchannel. (a) Patterning of the electrodesand the alumina insulating layer using photolithography, e-beam evaporation, anda lift-off method. (b) I2-assisted FIB milling and unfocused ion beam sculpting ofthe pore openings. (c) Electrodeposition of Au into the FIB milling nanopores. (d)Ag/AgCl electrode fabrication and assembly with a PDMS microchannel network.

Actuators B 187 (2013) 455– 460 457

bhgS(da

qlrsIatii(stpdrwotHsesoidas5id7taDatiromtpfwesst

2s

pcpms

N. Triroj et al. / Sensors and

atch microfabrication process, i.e. the fabrication does not requireigh temperature processing such as those employed in thermallyrown SiO2 and low-pressure chemical vapor deposited (LPCVD)i3N4. In addition, the Al2O3 films are found to have higher density3.1 g/cm3 [11]) than that of plasma-enhanced chemical vaporeposited (PECVD) Si3N4 (2.5 g/cm3 [12]), which is normally useds the passivation layer for the microelectrodes [13].

We want to emphasize that it is important to fabricate highuality alumina films for the passivation layer in order to prevent

eakage current. Obtaining dense alumina films of a given thicknessequires close monitoring of several parameters such as substrateurface cleanliness, substrate temperature, and deposition rate.f the substrate surface is unclean, deposited alumina films willdhere poorly and peel away from the substrate. The preparation ofhe alumina films for the insulating layer is described in the follow-ng fabrication steps. The electrode pattern on the glass substrates processed with UV lithography using 4.5-�m thick AZ® 9245Clariant Corporation, Somerville, NJ) photoresist pattern for theubsequent lift-off of Al2O3 at reference electrodes, counter elec-rodes, and probing contact pads. After the photoresist developingrocess, the substrate is cleaned thoroughly by rinsing in DI water,rying with N2, and baking on a hot plate at 110 ◦C for 10 min toemove any water molecules on the surface. The incorporation ofater into the growing films may give rise to a reduction of density

f the films [14]. Preceding the deposition of alumina, the substrateemperature is ramped slowly to 180–200 ◦C and is then stabilized.eating the substrate to at least 200 ◦C during evaporation is neces-

ary to ensure good adhesion and improved step coverage [15]. Thelectron beam is then swept to heat and fuse the top surface of theingle-crystal sapphire source for 5–10 min with the power levelf 490 W (7 kV, 70 mA). If the beam power is too high, the result-ng alumina films will appear darkly colored, indicating oxygeneficiency of the films and induced traps. Such films are undesir-ble because they exhibit a relatively high conductivity and chargetorage [16]. The pressure in the deposition chamber is started at

× 10−6 torr and rises to 1 × 10−5 torr during the process, indicat-ng the liberation of oxygen from the sapphire source. Alumina iseposited at the rate of 3 A/s until the resulting film thickness of20–800 nm is obtained. After cooling down the substrate to roomemperature, the patterned photoresist is removed by immersing inn acetone-assisted ultrasonication, rinsing with isopropanol andI water, and drying with N2. The substrate is then annealed in airt 300 ◦C for 4 h in order to reduce the density of interfacial traps,hus increase the packing density of the alumina sheath for betternsulating properties [12,17]. This post-annealing step also helpseducing both the charge storage capability and the conductivityf the films [16]. With a careful preparation, the evaporated alu-ina films appear smooth and clear, and have good adhesion to

he substrate. Fig. 2 shows the fabricated three-electrode cell afteratterning the alumina passivation layer over the electrode plat-orms on a glass substrate. The three-electrode cell consists of aorking electrode (WE), a reference electrode (RE), and a counter

lectrode (CE). The electrode platforms are then subjected to theubsequent FIB fabrication to create nanopore arrays. The insethows an intended nanopore array pattern at the working elec-rode.

.2. Gas-assisted focused ion beam milling and ion beamculpting techniques

Nanopore arrays can be prepared by creating a high-aspect-ratioore structure in the insulating layer such that the pores make

ontact with the underlying Au layer. Etching of a deep nanoporeresents a unique challenge – to reach the desired depth and toaintain the size of the pore top to be within the critical dimen-

ion for the nanoscale effects to be observed. We employ the FIB

Fig. 2. A SEM image showing the three-electrode cell after patterning the aluminapassivation layer over the electrodes. The inset shows the pattern of a nanoporearray at the working electrode.

(FEI DB235) to mill nanopores through the alumina insulating layerat the working electrode sensing area. Since the thickness of thealumina layer is on the order of 720–800 nm, the conventionalFIB milling technique does not work well. The reason is that asthe ion beam mills deeper into the sample, sputtered materialsare not completely ejected from the crater but remain within it.Re-deposition limits the depth and the aspect ratio of the milledpattern. This phenomenon can be thought of as digging a deep holewhile throwing dirt over the shoulder. To overcome this problem,we employ a gas chemistry to assist in the milling of high-aspect-ratio nanopores. I2 gas is directed toward the surface of the samplewhile alumina is being milled with the ion beam. Sputtered materialchemically reacts with the assisting gas to form a volatile com-pound that can be pumped away for faster etching. The Ga+ ionfocused beam current is 10 pA and ion beam energy is 30 keV. Thedwelling time and the overlap between pixels are set to 0.5 �s and50%, respectively. A 5 × 5 array pattern of 50 nm × 50 nm squarewindows with 1-�m pore-to-pore spacing is defined at each work-ing electrode sensing area as illustrated in Fig. 2. The milling timehas been optimized using test samples to obtain the minimum poreopenings to expose the buried Au layer. The resulting diametersof the nanopores range between 120–200 nm as shown in Fig. 3a.Due to a charging effect, the FIB milling process on an insulat-ing substrate needs to be performed at lower magnification. Thisrestriction can affect the resolution of the resulting pattern. It isobserved that the actual pore is much larger than the intended pat-tern size. To reduce the size of the pore tops, we use the ion beamsculpting method [18]. The underlying phenomena can be summa-rized as follows. Unfocused ion beam bombardment increases themobility and diffusivity of the atoms at the surface. As a result, masstransport of the exposed material along the surface promotes par-tial closing of the pore tops. We use a high speed scanning and lowion beam current to avoid heavy ion etching of the materials. Fol-lowing the initial FIB milling of nanpores, rastering of the ion beamover the whole array using ion beam energy of 30 keV, ion beam cur-rent of 10 pA, and scanning time of 0.724 s per scan for 15–20 minreduces the initial pore opening diameters to 60–120 nm. Fig. 3bshows a SEM image of the nanopore array after the ion beam sculpt-ing process.

2.3. Electrodeposition technique, fabrication of Ag/AgCl referenceelectrodes, and fabrication of the microfluidic chips

To create hemispherical nanoelectrodes protruding from alu-mina substrate, the pores are then filled by dc electrodepositionof Au at −0.8 V for approximately 2 h in Orotemp 24 gold platingsolution (KAu(CN)2 containing 1 troy oz/gallon from Technic Inc.,

458 N. Triroj et al. / Sensors and Actuators B 187 (2013) 455– 460

F elects

Ctonfilfitouost

tplNpdpE1imto

p

Fe

ig. 3. SEM images of a nanopore array in alumina insulating layer on the workingculpting process.

ranston, RI), using a platinum wire as an anode. The underpo-ential deposition technique is recommended for the fabricationf nanowires of uniform density over the whole depth of theanopores. This is because focusing effect due to the high electriceld concentrated at the pore top can be neglected as a result of the

ow deposition rate. At high applied voltage (high current density),ast and inhomogeneous deposition takes place. Besides, the result-ng nanowires are mechanically unstable with highly porous struc-ures [19–21]. Fig. 4a shows a SEM image of the nanoelectrode arrayn the working electrode after electrodeposition of Au. The non-niformity of the filling of Au is likely due to the surface roughnessf the substrate, which leads to an uneven milled depth. It can beeen that Au is deposited only inside the nanopore area, indicatinghat there is no leakage path anywhere else on the alumina surface.

The fabrication of Cl2-plasma treated Ag/AgCl reference elec-rodes was reported previously [13]. The process begins withatterning of 4.5-�m thick AZ® 9245 photoresist for the succeeding

ift-off of the Ag. We use AZ® Developer (Clariant Corp., Somerville,J; aqueous solution of trisodium phosphate with silicic acid) in thehotoresist development step. Note that the standard photoresisteveloper AZ® 400 K (Clariant) containing buffered KOH is not com-atible with this process because it can etch the alumina films [22].lectron beam evaporation is then used to deposit 20 nm Ti and

�m Ag on the reference electrode area. The Ag layer is convertednto an Ag/AgCl stack by Cl2 plasma treatment. The photoresist

ask is then removed after plasma chlorination of Ag. Fig. 4b shows

he optical microscope image of the complete three-electrode celln the glass substrate.

For the fabrication of the PDMS microchannels, the negativehotoresist (SU-8 50, MicroChem Corp., Newton, MA) is spin coated

ig. 4. Fabrication results. (a) A nanoelectrode array after the Au electrodeposition procelectrochemical sensor chip after the bonding process.

rode (a) after the I2 gas assisted FIB milling and (b) after the subsequent ion beam

onto a cleaned silicon wafer to pattern the resist with a thicknessof 30 �m, which later becomes the depth of the microchannels.This photoresist structure creates a master mold for casting PDMSchannels. The vapor of tridecafluoro-1,1,2,2-tetrahydrooctyl-1-1-trichlorosilane (United Chemical Technologies, Bristol, PA) appliedunder vacuum for 2 h is used to coat the master mold to preventirreversible bonding between silicon and PDMS. A 10:1 mixtureof PDMS prepolymer and curing agent (Sylgard 184, Dow Corning,Midland, MI) is degassed under vacuum, poured over the master,and cured at 75 ◦C for 1 h. The negative relief of the structure ispeeled off from the master after curing. Channel inlets and outletsare made by puncturing the PDMS slab at the patterned reservoirsusing a pipette tip. After the cleaning process, the complete three-electrode cell array on the glass slide and a PDMS microchannelslab are treated with air plasma, quickly aligned and bonded. Anirreversible bond is formed after 24 h of light pressure applied tothe bonded substrates. Microtubes are then installed into the inletsand outlets of the microchannels. The PDMS prepolymer is pouredover the top of the PDMS microchannel slab to seal the connections.The prepolymer PDMS layer is cured by heating the chip to 65 ◦Cfor 30 min. The final microfluidic electrochemical sensor chip inte-grated with the nanoelectrode arrays is shown in Fig. 4c. For thechannel width of 1 mm and the length of 6 mm, the volume of eachmicrochannel is 180 nL.

2.4. Cyclic voltammetry measurements

The instrument for cyclic voltammetry measurements consistsof a custom-built potentiostat with a computer-controlled userinterface, a Sub-femtoamp Remote SourceMeter (Keithley model

ss. (b) The complete three-electrode cell on the glass substrate. (c) The microfluidic

Actuators B 187 (2013) 455– 460 459

6pCTStoacIntsi

eR7s[toc(MamcotdttaAm

3

3t

anfaemERtHurl

aeiimaa

N. Triroj et al. / Sensors and

430), and a shielded probe station. Tungsten tips attached torobe tip holders of the S-725 micropositioners (Signatone, Gilroy,A) forms good contact with the gold electrode pads on the chip.he LMC662 operational amplifier (National Semiconductor Corp.,anta Clara, CA) is implemented for the potentiostat circuit owingo the low input bias current (2 fA). Keithley 6430 SourceMeter isperated by a LabVIEW program (National Instrument, Austin, TX)nd general purpose interface bus (GPIB) communication. A triggerable is use to synchronize the Keithley 6430 with the Nationalnstrument PCI-6024E data acquisition (DAQ). The DAQ is con-ected to the potentiostat to output the potential scan and to collecthe resulting current response of the system from the data pointstored in the buffer layer of Keithley 6430. The recorded data aremported into Excel or MATLAB for analysis.

Investigation of the electrochemical properties of the nano-lectrode array is carried out using the analyte solution of 5 mMu(NH3)6Cl3 in 25 mM sodium phosphate and 25 mM NaCl (pH.0). Ru(NH3)6

3+/Ru(NH3)62+ redox couples are used as the test

pecies because they do not adsorb onto the electrode surface23]. It is noted that adsorption can severely affect the elec-rode reactions and the current responses especially for the casef nanoelectrodes. Without further cleaning, a freshly fabricatedhannel is loaded with the analyte solution using a syringe pumpHarvard Apparatus model pump 11 plus dual syringe, Holliston,

A) at the speed of 2 �L/min. The syringe pump is turned offfter the solution completely fills up the channel. Cyclic voltam-ograms of the redox reactions from Ru(NH3)6

3+/Ru(NH3)62+

ouples are obtained at various scan rates of 20–200 mV/s. Inrder to investigate the enhanced mass transport due to attrac-ive migration case, the potential is initially scanned in a negativeirection from 0 V to −0.45 V, causing a reduction of Ru(NH3)6

3+

o Ru(NH3)62+, followed by the subsequent anodic scan. Fur-

hermore, the voltammetric data obtained at the nanoelectroderray is compared with the voltammetric result obtained at anu microelectrode (3 �m diameter) in the microfluidic environ-ent.

. Results and discussion

.1. Nanoelectrode array configuration, the depletion layer, andhe electrical double layer

The nanoelectrodes observed from the SEM image (Fig. 4a)ppear to have a hemispherical structure with surface rough-ess. The apparent curvature radius of the electrodes estimated

rom the SEM image ranges from 45 nm to 192 nm, with theverage electrode radius of 87 nm. This contributes to the totallectrode surface area of 1.47 × 10−8 cm2. In respect to the experi-ental time scale, the depletion layer thickness (ı) derived from

q. (1) is approximately the same size as the electrode radius.egarding the inter-electrode distance (d) of 1 �m, this condi-ion can be considered as non-overlap diffusion (d > 2ı) [9,10].ence, each nanoelectrode is expected to behave as an individ-al element such that the current response of the whole arrayesembles that of individual nanoelectrodes working in paral-el.

For the parameters used in the electrochemical measurements,ccording to Eq. (3), the thickness of the double layer (the high-lectric field region) is less than 10 nm. While the depletion layers within the nanoscale range, the size of the double layer region

s comparable to the thickness of the depletion layer. As a result,

igration-related mass transport should also be considered in thenalysis of the current response obtained at the nanoelectroderray.

Fig. 5. CV results showing (a) the response of a nanoelectrode array and (b) theresponse of a microelectrode with the diameter of 3 �m.

3.2. Cyclic voltammetric responses of Ru(NH3)63+/Ru(NH3)6

2+

redox species

When performing the electrochemical measurements of theredox process involving Ru(NH3)6

3+/Ru(NH3)62+ in analyte solu-

tion at various scan rates (20–200 mV/s), the current responsesobtained at the nanoelectrode array appear to be sigmoidallyshaped with not as well-defined steady-state limiting currentplateau. Insignificant discrepancies between the current responsesobtained at different scan rates suggest that the measurementsare performed under steady-state conditions. Steady-state voltam-mograms at the scan rate of 100 mV/s obtained at the nanoelectrodearrays and the microelectrode are compared in Fig. 5. As observedfrom the voltammograms at −0.45 V, the current density obtainedat the nanoelectrode arrays is increased about two orders ofmagnitude (1.36 × 106 A cm mol−1) compared to that obtained atthe microelectrode (1.55 × 104 A cm mol−1). The measured cur-rent density at the nanoelectrode array is also much higher thanthe steady-state diffusion-controlled limiting current. However,it should be emphasized that the surface roughness could in factunderestimate the actual electrode surface area, and thus overes-timate the true current enhancement factor. The increase in thecurrent density may derive from the high mass transport rate dueto the reduced depletion layer thickness and the migration effectwhich is added to the mass transport equation. For the reductionof a positively charged species (Ru(NH3)6

3+), the migration effectserves to accelerate mass transport of the redox species to the elec-trodes, and hence the increase in the measured current is acquired.We also notice that for the case of the nanoelectrode array, thesteady-state limiting current plateau is not as well-defined at−0.45 V. Because of the high mass transport rate, the electrodeprocess is controlled by the heterogeneous electrode kinetics [24].

Such an effect is likely the source of the unusual steady-statevoltammogram. The voltammetric response at the nanoelectrodearray also shows the shift of the half-wave potential as compared

4 Actua

trtsptmt

4

apeoeaantrcolpdtisi

A

a

R

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

60 N. Triroj et al. / Sensors and

o the position of the half-wave potential from the voltammet-ic response at the microelectrode system. The negative shift ofhe half-wave potential may derive from the high current den-ity which alters the mass transport from the diffusion-controlledrocess to the charge-transfer-controlled process [3]. Accordingly,hese significant deviations observed from the shape of the voltam-

etric response curve confirm the evidence of the enhanced massransport effect occurred at the nanoelectrode array.

. Conclusions

We describe the detailed fabrication steps used to construct microfluidic electrochemical sensor with well-aligned and low-acking-density nanoelectrode arrays. The nanoelectrodes aremployed as working electrodes to improve the current read-ut in a nano-liter volume testbed. Alumina films prepared bylectron-beam evaporation contain no leakage current path. High-spect-ratio nanopores with desirable pore-to-pore spacing can bechieved by I2 gas-assisted FIB milling and ion beam sculpting tech-iques. Cyclic voltammetry with Ru(NH3)6

3+/Ru(NH3)62+ species at

he nanoelectrode array exhibits a significant increase in the cur-ent density. The deviations from the classical diffusion-limitedurrent transport theory observed from the voltammetric curvef the nanoelectrode arrays are the absence of the steady-stateimiting current plateau and the cathodic shift of the half-waveotential. The effects of high mass transport rate due to the reducedepletion layer thickness and the high electric field generated nearhe electrode surface are accounted for such deviations. The exper-mental results demonstrate the extraction of electrochemicalignals using an ultrasensitive transducer based on nanoelectrodesn a microfluidic chip platform.

cknowledgment

This work was previously supported by AFOSR under MURIward F49620-03-1-0365.

eferences

[1] J.D. Norton, H.S. White, S.W. Feldberg, Effect of the electrical double layeron voltammetry at microelectrodes, Journal of Physical Chemistry 94 (1990)6772–6780.

[2] R. He, S. Chen, F. Yang, B. Wu, Dynamic diffuse double-layer model for theelectrochemistry of nanometer-sized electrodes, Journal of Physical ChemistryB 110 (2006) 3262–3270.

[3] J. Chen, K. Aoki, Overestimation of heterogeneous rate constants of hex-acyanoferrate at nanometer-sized ultramicroelectrodes, ElectrochemistryCommunications 4 (2002) 24–29.

[4] A.J. Bard, L.R. Faulkner, Electrochemical Methods: Fundamentals and Applica-tion, first ed., John Wiley & Sons, New York, 1980.

[5] W.J. Albery, Electrode Kinetics, Clarendon Press, Oxford, 1975.[6] M.E. Sandison, J.M. Cooper, Nanofabrication of electrode arrays by electron-

beam and nanoimprint lithographies, Lab Chip 6 (2006) 1020–1025.[7] Y.H. Lanyon, G. De Marzi, Y.E. Watson, A.J. Quinn, J.P. Gleeson, G. Redmond,

D.W.M. Arrigan, Fabrication of nanopore array electrodes by focused ion beammilling, Analytical Chemistry 79 (2007) 3048–3055.

[8] M.E. Sandison, N. Anicet, A. Glidle, J.M. Cooper, Optimization of the geometry

and porosity of microelectrode arrays for sensor design, Analytical Chemistry74 (2002) 5717–5725.

[9] W.E. Morf, Theoretical treatment of the amperometric current responseof multiple microelectrode arrays, Analytica Chimica Acta 330 (1996)139–149.

tors B 187 (2013) 455– 460

10] W.E. Morf, N.F. de Rooij, Performance of amperometric sensors based on mul-tiple microelectrode arrays, Sensors and Actuators B 44 (1997) 538–541.

11] W.M. Gosney, R.S. Miller, Aluminum oxide films made from evaporated sap-phire, Thin Solid Films 14 (1972) 255–266.

12] M. Arps, A. Markwitz, Improved current–voltage characteristics of downstreamplasma enhanced chemical vapor deposition SiNx deposited at low temperatureby using He as a dilution gas, Journal of Vacuum Science & Technology A 15(1997) 1864–1873.

13] N. Triroj, M.A. Lapierre-Devlin, S.O. Kelley, R. Beresford, Microfluidic three-electrode cell array for low-current electrochemical detection, IEEE SensorsJournal 6 (2006) 1395–1402.

14] K.K. Shih, D.B. Dove, Deposition of aluminum oxide films with high refractiveindex, Journal of Vacuum Science & Technology A 12 (1994) 321–322.

15] B. Lewis, The deposition of alumina, silica and magnesia films by elec-tron bombardment evaporation, Microelectronics and Reliability 3 (1964)109–120.

16] K.M. Eisele, Charge storage and stoichiometry in electron beam evaporatedalumina, Journal of the Electrochemical Society 122 (1975) 148–152.

17] S. Shuzhen, C. Lei, H. Haihong, Y. Kui, F. Zhengxiu, S. Jianda, Annealing effectson electron-beam evaporated Al2O3 films, Applied Surface Science 242 (2005)437–442.

18] J. Li, D. Stein, C. McMullan, D. Branton, M.J. Aziz, J.A. Golovchenko, Ion-beamsculpting at nanometre length scales, Nature 412 (2001) 166–169.

19] W.-C. Yoo, J.-K. Lee, Field-dependent growth patterns of metals electroplated innanoporous alumina membranes, Advanced Materials 16 (2004) 1097–1101.

20] W. Lee, R. Scholz, K. Nielsch, U. Gösele, A template-based electrochemicalmethod for the synthesis of multisegmented metallic nanotubes, AngewandteChemie International Edition 44 (2005) 6050–6054.

21] S. Karim, M.E. Toimil-Molares, F. Maurer, G. Miehe, W. Ensinger, J. Liu, T.W.Cornelius, R. Neumann, Synthesis of gold nanowires with controlled crystallo-graphic characteristics, Applied Physics A 84 (2006) 403–407.

22] J.-A. Paik, S.-K. Fan, C.-J. Kim, M.C. Wu, B. Dunn, Micromachining of mesoporousoxide films for microelectromechanical system structures, Journal of MaterialsResearch 17 (2002) 2121–2129.

23] R.M. Penner, M.J. Heben, T.L. Longin, N.S. Lewis, Fabrication and useof nanometer-sized electrodes in electrochemistry, Science 250 (1990)1118–1121.

24] C. Amatore, C. Lefrou, Is cyclic voltammetry above a few hundred kilovolts persecond still cyclic voltammetry? Journal of Electroanalytical Chemistry 296(1990) 335–358.

Biographies

Napat Triroj received the B.S. degree from Columbia University in 2001, the M.S.degree from the University of Washington in 2003, and the Ph.D. degree fromBrown University in 2008, all in electrical engineering. She is currently a Lecturerat the Department of Electrical Engineering, Faculty of Engineering, Khon KaenUniversity, Thailand. Her research is focused on synthesis and characterizationsof semiconductor-based thin films and nanostructured materials with the appli-cations to sensing and energy harvesting, and fabrication of microfluidic chips forelectrochemical biosensors.

Papot Jaroenapibal received the B.S. degree from Columbia University in 2002,the M.S. degree from the University of Pennsylvania in 2005, and the Ph.D. degreefrom the University of Pennsylvania in 2007, all in materials science and engi-neering. He is currently a Lecturer at the Department of Industrial Engineering,Faculty of Engineering, Khon Kaen University, Thailand. His current research activ-ities include synthesis and characterizations of one-dimensional nanostructuredmaterials, diamond-related thin films, and electron microscopy.

Roderic Beresford earned degrees in electrical engineering from Yale University(B.S. 1979, M.S. 1981) and from Columbia University (Ph.D. 1990). He is cur-rently Associate Provost and Professor of Engineering at Brown University. Priorpositions held include editor-in-chief of VLSI Design, senior editor of Electron-ics, and engineer at the IBM T.J. Watson Research Center. He has published over70 scientific and technical papers on semiconductor materials and devices. Hiscurrent research emphasizes synthesis of nanostructures and microelectronic /

microfluidic systems. He has worked on molecular beam epitaxial growth of III-Vsemiconductors since 1987 and serves on the Advisory Board of the North Amer-ican MBE Conference. Prof. Beresford is also a senior member of the Institute ofElectrical and Electronics Engineers, member of the American Physical Society, andSigma Xi.