Potential-Dependent Stochastic Amperometry ofMultiferrocenylthiophenes in an Electrochemical NanogapTransducerKlaus Mathwig,*,†,‡ Hamid R. Zafarani,§ J. Matthaus Speck,∥ Sahana Sarkar,⊥ Heinrich Lang,∥

Serge G. Lemay,⊥ Liza Rassaei,§ and Oliver G. Schmidt†

†Institute for Integrative Nanosciences, IFW Dresden, Helmholtzstr. 20, 01069 Dresden, Germany‡Pharmaceutical Analysis, Groningen Research Institute of Pharmacy, University of Groningen, P.O. Box 196, 9700 AD Groningen,The Netherlands§Department of Chemical Engineering, Delft University of Technology, Van der Maasweg 9, 2629 HZ Delft, The Netherlands∥Technische Universitat Chemnitz, Fakultat fur Naturwissenschaften, Institut fur Chemie, Anorganische Chemie, D-09107 Chemnitz,Germany⊥MESA+ Institute for Nanotechnology, University of Twente, P.O. Box 217, 7500 AE, Enschede, The Netherlands

*S Supporting Information

ABSTRACT: In nanofluidic electrochemical sensors based onredox cycling, zeptomole quantities of analyte molecules canbe detected as redox-active molecules travel diffusivelybetween two electrodes separated by a nanoscale gap. Thesesensors are employed to study the properties of multi-ferrocenylic compounds in nonpolar media, 2,3,4-triferroce-nylthiophene and 2,5-diferrocenylthiophene, which displaywell-resolved electrochemically reversible one-electron transferprocesses. Using stochastic analysis, we are able to determine,as a function of the oxidation states of a specific redox couple,the effective diffusion coefficient as well as the faradaic current generated per molecule, all in a straightforward experimentrequiring only a mesoscopic amount of molecules in a femtoliter compartment. It was found that diffusive transport is reducedfor higher oxidation states and that analytes yield very high currents per molecule of 15 fA.

1. INTRODUCTION

There is great interest in miniaturized electrochemical sensorsfor their ease of integration in lab-chip applications, highlysensitive detection, the requirement of only the smallest samplevolumes, direct electrical signal transduction, and cost-effectiveness when standard microfabrication is employed.1−8

We previously reported on electrochemical nanogap sensorswhich allow sensitive analysis in femtoliter detection volumes.9

In these systems, electrochemically active analyte moleculesshuttle by diffusion between two electrodes embedded in theroof and ceiling of a nanochannel (Figure 1). Thereby, themolecules undergo redox cycling, i.e., they are repeatedlyoxidized and reduced at the opposing electrodes at kilohertzfrequencies. In this way, each analyte contributes a current ofseveral thousand electrons per second, resulting in a highintrinsic signal amplification (as compared to a single-electrodeconfiguration in which a molecule reacts only once).Nanogap sensors are currently rare in their ability to perform

stochastic amperometric sensing.10 Only a small number ofmolecules are present in a very limited detection volume, evenat high analyte concentrations. Therefore, the system can bedescribed as a mesoscopic regime instead of a continuum: All

molecules undergo a random Brownian walk and diffuse in andout of the nanofluidic channel and a coupled reservoir. Thisresults in pronounced fluctuations of the molecular numberdensity in the detection volume.11 These fluctuations arereflected in the recorded electrical current; they can directly beprobed amperometrically. Molecular-level information is thengleaned from fluctuations in current−time traces. In particular,diffusion coefficients as well as specific dynamic adsorptivitieswere determined by analysis of the amplitude12 and frequencyspectrum13,14 of current fluctuations.To date, nanogap sensors have exclusively been employed for

the detection of archetypical organometallic or metal−organiccompounds such as ferrocenes15,16 or catechol derivatives;7,17,18

experiments were limited to conditions of a polar solvent(water or acetonitrile), a high concentration of fully dissociatedbackground electrolyte, and analytes with typically only onepossible electron-transfer process.

Received: July 21, 2016Revised: September 19, 2016Published: September 20, 2016

Article

pubs.acs.org/JPCC

© 2016 American Chemical Society 23262 DOI: 10.1021/acs.jpcc.6b07320J. Phys. Chem. C 2016, 120, 23262−23267

In the present work, we extend the application of nanogapsensors and stochastic amperometry to more complex analytesunder less polar conditions. Within this context, ferrocenyl-substituted thiophenes represent suitable molecules for suchinvestigations, because the thiophene building block allows aneasy functionalization with redox-active transition-metal-con-taining entities, and the resulting compounds are able to showseveral well-resolved one-electron transfer processes.19−24

Furthermore, these heterocyclic systems are of interestregarding their potential application in molecular electronicsor thiophene-based intrinsic conducting polymers (ICP).25−36

The nanogap geometry enables the determination ofdiffusion coefficients of these compounds as a function ofspecific oxidation states of redox couples. We were able toidentify a trend of decreased diffusive transport for higheroxidations states. Moreover, we found that the analytes yieldthe highest currents per molecule for a three-electron transferprocess.

2. EXPERIMENTAL SECTION2.1. Chemical Reagents and Nanofluidic Device

Fabrication. 2,5-diferrocenylthiophene (2,5-Fc2-cC4H2S) and

2,3,4-triferrocenylthiophene (2,3,4-Fc3-cC4HS) were synthe-

sized in a Pd-promoted Negishi C,C cross-coupling reactionof 2,5-dibromothiophene or 2,3,4-tribromo-thiophene withferrocenyl zinc chloride, as reported previously.21,37−39 Aselectrolyte, an 0.1 M solution of tetrabutylammonium tetrakis-(pentafluorophenyl)borate was used, whereby [NnBu4][B-(C6F5)4] was prepared by metathesis of lithium tetrakis(penta-fluorophenyl)borate etherate (Boulder Scientific) with tetra-butylammonium bromide according to a published procedureand enhanced by a filtration step of the crude product througha pad of silica using dichloromethane as solvent.40

Electrochemical nanogap devices were fabricated by clean-room microfabrication on an oxidized silicon wafer substrate asdescribed previously.41 They consist of a 5 μm wide, 10−20 μmlong, and 70 nm high nanochannel in silicon oxide/siliconnitride; the roof and ceiling of the channel are Pt electrodes.The length and width of the nanochannel are defined byphotolithography; the nanoscale height is defined byevaporation of a sacrificial Cr layer sandwiched between twoPt layers, which were also defined and fabricated byphotolithography and electron-beam deposition, respectively.After these three metal deposition steps, the structure wasburied in a Si3N4/SiO2 passivation layer by plasma-enhancedchemical vapor deposition. Access holes were subsequentlyplasma-etched into this layer to connect the nanodevice to afluidic reservoir. Just before an experiment, the nanochannelwas released by a selective wet-chemical etch of the sacrificialCr layer (Cr etchant Selectipur, BASF). A micrograph andschematic view of a nanofluidic electrochemical device areshown in Figure 1.

2.2. Electrochemical Instrumentation and Experimen-tation. Using a home-built potentiostat setup, two sensitiveoperational amplifiers (Femto DDCPA-300) were separatelyconnected to the top and bottom electrodes of a nanofluidicdevice. The amplifiers were LabView-controlled and used ascurrent meters and voltage sources (Figure 1b). A micro-manipulator was used to position a reservoir in polydimethylsi-loxane (PDMS) on top of the access holes. After the wet-etchof the chromium layer, the etchant was replaced by sulfuric acid(Sigma-Aldrich), prepared as a 0.5 M solution in Milli-Q water,and both Pt electrode surfaces were cleaned by repeatedlysweeping their potential between −0.15 and 1.2 V versus a Ag/AgCl reference electrode (BASi inc.) placed in the reservoir.The PDMS reservoir was then replaced by a reservoir

machined in polytetrafluoro-ethylene and connected to thenanochannel via an O-ring (fluorinated propylene monomer,0.5 mm inner diameter). The reservoir and nanochannel werefilled with the analyte-containing dichloromethane solution,and a double-junction Ag/Ag+ reference electrode (BASi inc.)was immersed into the reservoir. This reference electrode wasconstructed from a silver wire inserted into a Luggin capillarywith a Vycor tip containing a solution of 0.01 M [AgNO3] and0.1 M of the supporting electrolyte in acetonitrile. This Luggincapillary was inserted into a second Luggin capillary with Vycortip filled with a 0.1 M supporting electrolyte solution indichloromethane. The whole setup was shielded frominterference in a Faraday cage.

3. RESULTS AND DISCUSSION

3.1. Cyclic Voltammetry. Cyclic voltammograms (CVs) of2,5-diferrocenylthiophene and 2,3,4-triferrocenylthiophene (1mM in 0.1 M [NnBu4][B(C6F5)4] in dichloromethane)recorded at a Pt ultramicrodisk electrode (UME; BASi inc.)with a diameter of 10 μm are shown in Figure 2. In a stepwiseoxidation, the analytes exhibit two (or three, respectively) well-resolved electrochemically reversible one-electron transferprocesses with equal heights of the current steps. This is ingood agreement with previous results.37

By determining the magnitude of the steady-state currents,we estimate the diffusion coefficients, D, of the fully reducedspecies to be 1.08 × 10−9 m2/s for 2,5-Fc2-

cC4H2S and 0.96 ×10−9 m2/s for 2,3,4-Fc3-

cC4HS. [Diffusion coefficients (D) wereestimated by i = 4nFcDr, where i is the height of a current step,

Figure 1. (a) Top-view optical micrograph of a nanofluidicelectrochemical sensor. (b) Schematic cross section of the deviceand principle of operation of a nanogap sensor. Analyte moleculesundergo redox cycling between the two electrodes. For differentelectrode potentials, redox cycling occurs between different oxidationstates, and different numbers of electrons (1−3) are exchanged at eachtransfer process.

The Journal of Physical Chemistry C Article

DOI: 10.1021/acs.jpcc.6b07320J. Phys. Chem. C 2016, 120, 23262−23267

23263

n the number of exchanged electrons, F Faraday constant, c theanalyte concentration, and r the electrode radius.42]3.2. Cyclic Voltammetry in a Nanogap Device. Figure 3

shows voltammograms of both thiophene compounds recorded

in a nanochannel device. The steps are not entirely aspronounced as for the measurements using the UME, whichis most likely caused by the finite rate for heterogeneouselectron transfer, as has been reported earlier for othercouples.43 Nonetheless, all electron-transfer processes are

clearly resolved. In these measurements, one electrode waskept constant at 0 V, while the other electrode was swept withrespect to the reference electrode. The curves depict theoxidation current at one electrode and the reduction current atthe other working electrode. This means that the moleculesundergo redox cycling with one, two, or three electrons permolecule being exchanged during each process. While in thedevice, each analyte molecule undergoes about 60 000 transferprocesses per second. This leads to a considerable amplificationof the current detected per molecule.A difference in the measurements compared to the cyclic

voltammetry at the UME is that the height of the current stepsis reduced for an increasing magnitude of applied potentials oroxidation states, respectively, as indicated by the gray bars inFigure 3. This effect cannot be explained by a reduced diffusioncoefficient of oxidized molecules or an effectively reduceddiffusion due to dynamic adsorption; as such, a reduction isequilibrated by exchange (increase) in local concentration withthe bulk reservoir.44

We believe that the shape of the cyclic voltammogramsrecorded in the nanochannel is dictated by the finite rate forheterogeneous electron transfer, as has been reported earlier forother couples.43 We estimated the electrochemical rateconstants k0. As shown in the dashed lines in Figure 3, theforward sweeps of the CVs were fitted to the Butler−Volmerexpression for thin layer cells,43 using a superposition of j one-electron reactions:

∑α

=

+ − − + − − −=′ ′

i Ei

f E E f E E

( )

1 exp[ ( )] exp[ (1 ) ( )]j

j

jD

zk j1

2 or 3lim,

0 0

j

eff0

(1)

Here ilim,j are limiting currents for the individual waves, Ej0′

formal potentials, α the transfer coefficient, z = 70 nm thenanochannel height, and f = F/RT (F, Faraday constant; R, gasconstant; T, temperature). Deff is the effective diffusioncoefficient; diffusion is effectively slowed by dynamic analyteadsorption due to the high surface-to-volume ratio in thenanochannel of 3 × 107 m−1, typically to Deff = 0.2−0.5D inprevious experiments.10,15 The fitting parameters were ilim andthe dimensionless rate constant Deff/zk

0. The transfercoefficient was approximated by α = 0.5.43 Formal potentials

Figure 2. Cyclic voltammograms of 2,5-diferrocenylthiophene (left)and 2,3,4-triferrocenylthiophene (right) detected in dichloromethanesolutions (1 mM) at a Pt ultramicroelectrode with a diameter of 10 μm(supporting electrolyte 0.1 M [NnBu4][B(C6F5)4]).

Figure 3. Cyclic voltammetry in the nanofluidic device. (a) 50 μM 2,5-Fc2-

cC4H2S in a 20 μm long device. The top electrode is swept (red),while the bottom electrode is kept at 0 V (blue). The dashed line is afit to the Butler−Volmer formalism, eq 1. Gray bars indicate limitingcurrents determined by the same fit. (b) 1 mM 2,3,4-Fc3-

cC4HS in a 10μm long device. Here the bottom electrode is swept (red), while thetop electrode is kept constant at 0 V (blue).

Figure 4. (a) Effective diffusion coefficient of 2,3,4-Fc3-cC4HS determined by the magnitude of fluctuations istd in current−time traces. The “mean

charge in e” corresponds to the applied potentials 0.5e ≙ 0−0.25 V; 1e ≙ 0−0.5 V; 1.5e (“0..3”) ≙ 0−0.9 V; 1.5 e (“1..2”) ≙ 0.25−0.5 V; 2 e ≙ 0.25−0.9 V; 2.5 e ≙ 0.5−0.9 V. (b) Corresponding currents per molecule for 2,3,4-Fc3-

cC4HS. The gray lines indicate one-electron (lower line) and two-electron (upper line) transfer processes. (c) Effective diffusivities and currents per molecule for 2,5-Fc2-

cC4H2S. Here the mean charge in ecorresponds to the potentials 0.5e ≙ 0−0.25 V; 1e ≙ 0−0.55 V; 1.5 e ≙ 0.25−0.55 V.

The Journal of Physical Chemistry C Article

DOI: 10.1021/acs.jpcc.6b07320J. Phys. Chem. C 2016, 120, 23262−23267

23264

Ej0′ were determined as the inflection points of the CVs. Using

the effective diffusion coefficients determined as shown below(see Figure 4), rate constants were extracted. They decreaseconsiderably with a higher electron number from 1.4 to 0.50cm/s for 2,5-Fc2-

cC4H2S and from 0.40 to 0.18 to 0.07 cm/s forthe 2,3,4-Fc3-

cC4HS compound for the first and second (andthird) wave, respectively.Moreover, electrical migration might have an adverse effect

on transfer kinetics and analyte transport perpendicular to theelectrodes; this effect was previously studied in nanogapgeometries.45−47 [NnBu4][B(C6F5)4] has a large ionic dissoci-ation constant KA > 1000.48 Therefore, a thick electrical doublelayer of λD > 10 nm (at actual electrolyte ion concentrations<0.1 mM) could lead to repulsion of charged analytes fromcharge electrodes and thereby to reduced limiting currents forhigher oxidation states.3.3. Stochastic Amperometry. We biased the electrodes a

specific fixed potential, recorded current−time traces for bothelectrodes, and analyzed the mesoscopic number densityfluctuations mirrored in the current fluctuations. In thisdetection scheme, the nanogap geometry offers the uniqueadvantage that an arbitrary couple of oxidation states can bechosen for redox cycling between the two working electrodes.This means that also cycling between two positive oxidationstates is possible, independent of the fully reduced bulk state.Properties were analyzed by determining the magnitude of

the fluctuations,12 i.e., the standard deviation istd and comparingit to the magnitude of the Faradaic limiting current ⟨ilim⟩ forspecific combination of electrode potentials. Here, istd dependsonly on the current generated per molecule, ip, and the meannumber of analyte molecules ⟨N⟩ present in the detectionvolume:

= ⟨ ⟩i i Nstd p1/2

(2)

The current per molecule, ip, depends only on thenanochannel height, z; the number of electrons transferred,n; and the diffusivity, Deff (and on the elementary charge, e):

= =i neD z n/ ( 1, 2, 3)p eff2

(3)

By using eqs 2 and 3 as well as the limiting current ⟨ilim⟩ =ip⟨N⟩, Deff and ip can directly be determined, i.e., ip = i2std/⟨ilim⟩.The results of this fluctuation analysis are shown in Figure

4a. Here, the effective diffusion coefficient, Deff, is plotted as afunction of applied potential represented by the nominal “meancharge in e”. This means that for a value of 0.5, the analyte2,3,4-Fc3-

cC4HS cycles between the fully reduced state and astate with only one oxidized ferrocenyl group, potentials of 0 Vand 250 mV are applied, and the molecules in the device havean average charge of about 0.5e. Our measurements showunequivocally that the effective diffusivity decreases withincreasing mean charge or with the mean of the potentialapplied at the top and bottom electrodes. Figure 4b shows thecurrent per molecule, ip, as determined by eq 3. This currentincreases with the number of electrons exchanged but decreasesfor higher oxidation numbers. In Figure 4c, the same trends ofsmaller currents and diffusion coefficients for an increasedmean charge are depicted for the 2,5-Fc2-

cC4H2S analyte.Deff can also be extracted from a current−time trace by

determining its power spectral density and determining thecharacteristic crossover frequency, f 0.

13 Both methods arecompared in the Supporting Information, showing higher

diffusivities for the latter methods but the same trend ofreduced Deff for higher potentials and charges.The slightly faster diffusion of 2,5-Fc2-

cC4H2S compared to2,3,4-Fc3-

cC4HS is expected because of its smaller size. Theobserved reduced diffusivities at higher oxidation states are alsonot unexpected. For example, the diffusion coefficient offerrocenedimethanol in aqueous solution is reduced by 20% inits oxidized state44 because of a changed solvation shell. In thecase of the ferrocenylthiophenes, a similarly strong effect is notexpected because of the very large size of the electrolyte ions ina nonpolar solvent. Several origins or a combination of effectscan lead to the observed change in Deff, namely, a change inbulk diffusivity, a change in dynamic adsorptivity, a slightlyreduced limiting current due to slow kinetics, as well as acontribution by electrical migration. The increased Deff for thespectroscopic determination compared to the current magni-tude analysis (Figure S2 in the Supporting Information) hintsat an anisotropic contribution by migration which is captured inistd but not in f 0 because only longitudinal diffusion contributesto the crossover frequency (shuttling in and out of thenanochannel), while fluctuations on all time scales (includingdiffusion across the channel) are captured in istd.A change in electrostatic adsorption can be ruled out as a

cause for the decreased effective diffusivity because morepositively charged analytes would be repelled from morepositively charged electrode surfaces, but a decrease in Deff isobserved for these conditions.Ultimately, while clear trends are observed for Deff and ip, so

far it is not possible to distinguish the contribution ofadsorption and changed diffusivities to these observations.While transport is hindered in the nanochannel, very largecurrents per molecule of up to 15 fA at room temperature weremeasured for high overpotentials (Figure 4b). This shows thatcompounds exhibiting multielectron transfer processes can havea potential application in single-molecule electrochemistry.

4. CONCLUSIONS

We introduced a new electrochemical method to determine thediffusive properties as a function of the oxidation state ofcomplex electrochemically active molecules in a stochasticamperometric measurement. The experiments show thatelectrochemical nanogap sensors are suitable to work underconditions of a very volatile solvent, enabling a more complexanalysis of novel compounds. For the investigated oligoferro-cenyl thiophenes, our measurements show a decrease ineffective diffusivity for an increase in applied potentials andoxidation state of the compounds. Moreover, highest currentsper molecule of 15 fA per molecule were determined, whichcompares well to previous single-molecule experiments.16 Forfuture experiments, we plan to investigate the origins ofchanges in effective diffusivity by determining the influence ofelectrical migration as well as to extend the method fromidentifying the properties of a redox couple to determining thediffusion of a specific oxidation state.

■ ASSOCIATED CONTENT

*S Supporting InformationThe Supporting Information is available free of charge on theACS Publications website at DOI: 10.1021/acs.jpcc.6b07320.

Power spectral density analysis of current−time traces(PDF)

The Journal of Physical Chemistry C Article

DOI: 10.1021/acs.jpcc.6b07320J. Phys. Chem. C 2016, 120, 23262−23267

23265

■ AUTHOR INFORMATIONCorresponding Author*E-mail: k.h.mathwig@rug.nl. Phone: +31 50 363 3022.NotesThe authors declare no competing financial interest.

■ ACKNOWLEDGMENTSS.S. and S.G.L. acknowledge financial support from theEuropean Research Council (ERC) under Project 278801.

■ REFERENCES(1) Rothberg, J. M.; Hinz, W.; Rearick, T. M.; Schultz, J.; Mileski, W.;Davey, M.; Leamon, J. H.; Johnson, K.; Milgrew, M. J.; Edwards, M.;et al. An Integrated Semiconductor Device Enabling Non-OpticalGenome Sequencing. Nature 2011, 475, 348−352.(2) Sage, A. T.; Besant, J. D.; Lam, B.; Sargent, E. H.; Kelley, S. O.Ultrasensitive Electrochemical Biomolecular Detection Using Nano-structured Microelectrodes. Acc. Chem. Res. 2014, 47, 2417−2425.(3) Laborde, C.; Pittino, F.; Verhoeven, H. A.; Lemay, S. G.; Selmi,L.; Jongsma, M. A.; Widdershoven, F. P. Real-Time Imaging ofMicroparticles and Living Cells with CMOS Nanocapacitor Arrays.Nat. Nanotechnol. 2015, 10, 791−795.(4) Rassaei, L.; Singh, P. S.; Lemay, S. G. Lithography-BasedNanoelectrochemistry. Anal. Chem. 2011, 83, 3974−3980.(5) Singh, P. S. From Sensors to Systems: CMOS-IntegratedElectrochemical Biosensors. IEEE Access 2015, 3, 249−259.(6) Mathwig, K.; Albrecht, T.; Goluch, E. D.; Rassaei, L. Challengesof Biomolecular Detection at the Nanoscale: Nanopores andMicroelectrodes. Anal. Chem. 2015, 87, 5470−5475.(7) Rassaei, L.; Mathwig, K.; Kang, S.; Heering, H. A.; Lemay, S. G.Integrated Biodetection in a Nanofluidic Device. ACS Nano 2014, 8,8278−8284.(8) Dawson, K.; O’Riordan, A. Electroanalysis at the Nanoscale.Annu. Rev. Anal. Chem. 2014, 7, 163−181.(9) Zevenbergen, M. A.; Krapf, D.; Zuiddam, M. R.; Lemay, S. G.Mesoscopic Concentration Fluctuations in a Fluidic NanocavityDetected by Redox Cycling. Nano Lett. 2007, 7, 384−388.(10) Katelhon, E.; Krause, K. J.; Mathwig, K.; Lemay, S. G.; Wolfrum,B. Noise Phenomena Caused by Reversible Adsorption in NanoscaleElectrochemical Devices. ACS Nano 2014, 8, 4924−4930.(11) Krause, K. J.; Mathwig, K.; Wolfrum, B.; Lemay, S. G. BrownianMotion in Electrochemical Nanodevices. Eur. Phys. J.: Spec. Top. 2014,223, 3165−3178.(12) Singh, P. S.; Chan, H.-S. M.; Kang, S.; Lemay, S. G. StochasticAmperometric Fluctuations as a Probe for Dynamic Adsorption inNanofluidic Electrochemical Systems. J. Am. Chem. Soc. 2011, 133,18289−18295.(13) Zevenbergen, M. A. G.; Singh, P. S.; Goluch, E. D.; Wolfrum, B.L.; Lemay, S. G. Electrochemical Correlation Spectroscopy inNanofluidic Cavities. Anal. Chem. 2009, 81, 8203−8212.(14) Mathwig, K.; Mampallil, D.; Kang, S.; Lemay, S. G. ElectricalCross-Correlation Spectroscopy: Measuring Picoliter-per-MinuteFlows in Nanochannels. Phys. Rev. Lett. 2012, 109, 118302.(15) Mampallil, D.; Mathwig, K.; Kang, S.; Lemay, S. G. ReversibleAdsorption of Outer-Sphere Redox Molecules at Pt Electrodes. J. Phys.Chem. Lett. 2014, 5, 636−640.(16) Kang, S.; Nieuwenhuis, A. F.; Mathwig, K.; Mampallil, D.;Lemay, S. G. Electrochemical Single-Molecule Detection in AqueousSolution Using Self-Aligned Nanogap Transducers. ACS Nano 2013, 7,10931−10937.(17) Wolfrum, B.; Zevenbergen, M.; Lemay, S. Nanofluidic RedoxCycling Amplification for the Selective Detection of Catechol. Anal.Chem. 2008, 80, 972−977.(18) Kanno, Y.; Ino, K.; Shiku, H.; Matsue, T. A Local RedoxCycling-Based Electrochemical Chip Device with Nanocavities forMulti-Electrochemical Evaluation of Embryoid Bodies. Lab Chip 2015,15, 4404−4414.

(19) Iyoda, M.; Kondo, T.; Okabe, T.; Matsuyama, H.; Sasaki, S.;Kuwatani, Y. A Simple and Efficient Synthesis of Di-, Tri-, andTetraferrocenylarenes. Chem. Lett. 1997, 26, 35−36.(20) Jin, B.; Tao, F.; Liu, P. Rapid-Scan Time-Resolved FT-IRSpectroelectrochemistry − Study on the Electron Transfer ofFerrocene-Substituted Thiophenes. J. Electroanal. Chem. 2008, 624,179−185.(21) Hildebrandt, A.; Schaarschmidt, D.; Claus, R.; Lang, H.Influence of Electron Delocalization in Heterocyclic Core Systemson the Electrochemical Communication in 2,5-Di- and 2,3,4,5-Tetraferrocenyl Thiophenes, Furans, and Pyrroles. Inorg. Chem.2011, 50, 10623−10632.(22) Speck, J. M.; Schaarschmidt, D.; Lang, H. Atropisomeric3,3′,4,4′,5,5′-Hexaferrocenyl-2,2′-Bithiophene: Synthesis, Solid-StateStructure, and Electrochemistry. Organometallics 2012, 31, 1975−1982.(23) Hildebrandt, A.; Ruffer, T.; Erasmus, E.; Swarts, J. C.; Lang, H.A Star-Shaped Supercrowded 2,3,4,5-Tetraferrocenylthiophene: Syn-thesis, Solid-State Structure, and Electrochemistry. Organometallics2010, 29, 4900−4905.(24) Hildebrandt, A.; Lang, H. (Multi) Ferrocenyl Five-MemberedHeterocycles. Organometallics 2013, 32, 5640−5653.(25) Heeger, A. J. Semiconducting and Metallic Polymers: TheFourth Generation of Polymeric Materials (Nobel Lecture). Angew.Chem., Int. Ed. 2001, 40, 2591−2611.(26) MacDiarmid, A. G. “Synthetic Metals”: A Novel Role forOrganic Polymers (Nobel Lecture). Angew. Chem., Int. Ed. 2001, 40,2581−2590.(27) Zhang, D.; Ryu, K.; Liu, X.; Polikarpov, E.; Ly, J.; Tompson, M.E.; Zhou, C. Transparent, Conductive, and Flexible Carbon NanotubeFilms and Their Application in Organic Light-Emitting Diodes. NanoLett. 2006, 6, 1880−1886.(28) Kim, Y. H.; Lee, J.; Hofmann, S.; Gather, M. C.; Muller-Meskamp, L.; Leo, K. Achieving High Efficiency and ImprovedStability in ITO-Free Transparent Organic Light-Emitting Diodes withConductive Polymer Electrodes. Adv. Funct. Mater. 2013, 23, 3763−3769.(29) Roncali, J.; Blanchard, P.; Frere, P. 3,4-Ethylenedioxythiophene(EDOT) as a Versatile Building Block for Advanced Functional π-Conjugated Systems. J. Mater. Chem. 2005, 15, 1589−1610.(30) Paul, F.; Lapinte, C. Organometallic Molecular Wires and OtherNanoscale-Sized Devices. Coord. Chem. Rev. 1998, 178−180, 431−509.(31) Ward, M. D. Metal-Metal Interactions in Binuclear ComplexesExhibiting Mixed Valency; Molecular Wires and Switches. Chem. Soc.Rev. 1995, 24, 121−134.(32) Akita, M.; Koike, T. Chemistry of Polycarbon Species: FromClusters to Molecular Devices. Dalton Trans. 2008, 9226, 3523−3530.(33) Ceccon, A.; Santi, S.; Orian, L.; Bisello, A. ElectronicCommunication in Heterobinuclear Organometallic Complexesthrough Unsaturated Hydrocarbon Bridges. Coord. Chem. Rev. 2004,248, 683−724.(34) Barigelletti, F.; Flamigni, L. Photoactive Molecular Wires Basedon Metal Complexes. Chem. Soc. Rev. 2000, 29, 1−12.(35) Barlow, S.; O’Hare, D. Metal−Metal Interactions in LinkedMetallocenes. Chem. Rev. 1997, 97, 637−670.(36) Hildebrandt, A.; Lang, H. (Multi)ferrocenyl Five-MemberedHeterocycles: Excellent Connecting Units for Electron TransferStudies. Organometallics 2013, 32, 5640−5653.(37) Speck, J. M.; Claus, R.; Hildebrandt, A.; Ruffer, T.; Erasmus, E.;van As, L.; Swarts, J. C.; Lang, H. Electron Transfer Studies onFerrocenylthiophenes: Synthesis, Properties, and Electrochemistry.Organometallics 2012, 31, 6373−6380.(38) Speck, J. M.; Korb, M.; Ruffer, T.; Hildebrandt, A.; Lang, H.Substituent Influence on Charge Transfer Interactions in α,α′-Diferrocenylthiophenes. Organometallics 2014, 33, 4813−4823.(39) Poppitz, E. A.; Hildebrandt, A.; Korb, M.; Schaarschmidt, D.;Lang, H. Synthesis, Properties, and Electron Transfer Studies ofFerrocenyl Thiophenes. Z. Anorg. Allg. Chem. 2014, 640, 2809−2816.

The Journal of Physical Chemistry C Article

DOI: 10.1021/acs.jpcc.6b07320J. Phys. Chem. C 2016, 120, 23262−23267

23266

(40) Speck, J. M.; Korb, M.; Schade, A.; Spange, S.; Lang, H.Ferrocenes Bridged by Ethylenediamino Thiophene: Varying ChargeTransfer Properties in a Series of 3,4-Di-N-Substituted 2,5-Diferrocenyl Thiophenes. Organometallics 2015, 34, 3788−3798.(41) Mathwig, K.; Lemay, S. G. Pushing the Limits of ElectricalDetection of Ultralow Flows in Nanofluidic Channels. Micromachines2013, 4, 138−148.(42) Aoki, K. Theory of Ultramicroelectrodes. Electroanalysis 1993, 5,627−639.(43) Zevenbergen, M. A. G.; Wolfrum, B. L.; Goluch, E. D.; Singh, P.S.; Lemay, S. G. Fast Electron-Transfer Kinetics Probed in NanofluidicChannels. J. Am. Chem. Soc. 2009, 131, 11471−11477.(44) Mampallil, D.; Mathwig, K.; Kang, S.; Lemay, S. G. RedoxCouples with Unequal Diffusion Coefficients: Effect on RedoxCycling. Anal. Chem. 2013, 85, 6053−6058.(45) Hyk, W.; Stojek, Z. Thin and Ultra-Thin Layer Dual ElectrodeElectrochemistry: Theory of Steady-State Voltammetry withoutSupporting Electrolyte. Electrochem. Commun. 2013, 34, 192−195.(46) Xiong, J.; Chen, Q.; Edwards, M. A.; White, H. S. Ion Transportwithin High Electric Fields in Nanogap Electrochemical Cells. ACSNano 2015, 9, 8520−8529.(47) Chen, Q.; McKelvey, K.; Edwards, M. A.; White, H. S. RedoxCycling in Nanogap Electrochemical Cells. The Role of Electrostaticsin Determining the Cell Response. J. Phys. Chem. C 2016, 120, 17251.(48) LeSuer, R. J.; Buttolph, C.; Geiger, W. E. Comparison of theConductivity Properties of the Tetrabutylammonium Salt of Tetrakis-(pentafluorophenyl)borate Anion with Those of Traditional Support-ing Electrolyte Anions in Nonaqueous Solvents. Anal. Chem. 2004, 76,6395−6401.

The Journal of Physical Chemistry C Article

DOI: 10.1021/acs.jpcc.6b07320J. Phys. Chem. C 2016, 120, 23262−23267

23267

S1

Supporting Information for: Potential-Dependent Stochastic Amperometry of Multiferrocenylthiophenes in an Electrochemical Nanogap Transducer Klaus Mathwig,,†,§ Hamid R. Zafarani,# J. Matthäus Speck,‡ Sahana Sarkar,¶ Heinrich Lang,‡

Serge G. Lemay,¶ Liza Rassaei,# and Oliver G. Schmidt†

†Institute for Integrative Nanosciences, IFW Dresden, Helmholtzstr. 20, 01069 Dresden, Germany §Pharmaceutical Analysis, Groningen Research Institute of Pharmacy, University of Groningen, P.O. Box 196, 9700 AD Groningen, The Netherlands #Department of Chemical Engineering, Delft University of Technology, Van der Maasweg 9, 2629 HZ Delft, The Netherlands ‡Technische Universität Chemnitz, Fakultät für Naturwissenschaften, Institut für Chemie, Anorganische Chemie, D-09107 Chemnitz, Germany ¶MESA+ Institute for Nanotechnology, University of Twente, P.O. Box 217, 7500 AE, Enschede, The Netherlands

Electrochemical correlation spectroscopy

Figure S1. Power spectral densities (PSDs). a) PSDs for oxidation potentials of 0.9 V, 0.5 V and 0.25 V for 0.5 mM 2,3,4-Fc3-cC4HS undergoing redox cycling in a 10 µm long device (the reduction potential is kept at 0 V). Gray lines are a fit to the theoretical model. Inset: raw current-time traces for potentials of 0.25 V and 0.9 V and the top and bottom electrode, respectively. b) Power spectra for 10 mM 2,5-Fc2-cC4H2S in a 20 µm long device.

S2

Exemplary power spectra extracted from current-time traces measured in a nanofluidic device are shown in Figure S1. Traces were recorded for a length of 60 s (100 s for 2,5-Fc2-cC4H2S) at a bandwidth of 150 Hz. Low-frequency drift was filtered, and spectra of three current-time traces were averaged for each depicted power spectrum. The PSDs S(f) can be characterized by two characteristic properties: the magnitude of the plateau at lower frequencies, which for a given geometry and diffusion coefficient depends only on the analyte concentration, and the crossover frequency f0, which indicates the transition from the plateau at lower frequencies to a f–3/2 decay at high frequencies [1]. The value of f0 was extracted by a fit to S(f) = S0/(1 + (f/f0)3/2) and Deff evaluated using f0 = (Deff/π)(3/L2

a(La + 6Le)))2/3. Here La is the top electrode length and Le = 2 µm is the distance between the access holes and the top electrode). Figure S2 shows corresponding diffusivities for 2,5-Fc2-cC4H2S in comparison to Deff determined by evaluating the current fluctuation magnitude (as shown also in Fig. 4c).

Figure S2. Comparison of Deff of 2,5-Fc2-cC4H2S calculated by determining istd and the crossover frequency f0 of the power spectral density of a current-time trace. Reference (1) Zevenbergen, M. A. G.; Singh, P. S.; Goluch, E. D.; Wolfrum, B. L.; Lemay, S. G.

Electrochemical Correlation Spectroscopy in Nanofluidic Cavities. Anal. Chem. 2009, 81 (19), 8203–8212.