Subscriber access provided by Purdue University Libraries
Article
Electrochemical imaging of dense molecular nanoarrays Khalil Chennit, Jorge Trasobares, Agnès Anne, Edmond Cambril, Arnaud Chovin, Nicolas Clement, and Christophe Demaille Anal. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.analchem.7b03111 • Publication Date (Web): 26 Sep 2017 Downloaded from http://pubs.acs.org on September 27, 2017
Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Analytical Chemistry is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.
Page 1 of 9
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Analytical Chemistry
Electrochemical imaging of dense molecular nanoarrays Khalil Chennit,† Jorge Trasobares,‡ Agnès Anne,*,† Edmond Cambril,§ Arnaud Chovin,† Nicolas Clément,*,‡ and Christophe Demaille*,† Laboratoire d'Electrochimie Moléculaire, UMR 7591 CNRS, Université Paris Diderot, Sorbonne Paris Cité, 15 rue JeanAntoine de Baïf, F-75205 Paris Cedex 13, France. ‡ Institute of Electronics, Microelectronics and Nanotechnology, CNRS, University of Lille, Avenue Poincaré, BP60069, 59652, Villeneuve d’Ascq, France. § Centre de Nanosciences et de Nanotechnologies, CNRS, Univ. Paris-Sud, Université Paris-Saclay, C2N-Marcoussis, 91460 Marcoussis, France. ABSTRACT: The aim of the present work is to explore the combination of atomic force electrochemical microscopy, operated in molecule touching mode (Mt/AFM-SECM), and of dense nanodot arrays, for designing an electrochemically addressable molecular nanoarray platform. A high density nanoarray of single grained gold nanodots (~15 nm-diameter nanoparticles, 100 nm pitch) is decorated by a model molecular system, consisting in ferrocene (Fc) labeled PEG disulfide chains. We show that the high resolution of Mt/AFM-SECM enables the electrochemical interrogation of several hundreds of individual nanodots in a single image acquisition. As a result the statistical dispersion of the nanodot molecular occupancy by Fc-PEG chains can be reliably quantified, evidencing that as little as a few tens of copies of redox-labeled macromolecules immobilized on individual nanodots can be detected. The electrochemical reactivity of individual nanodots can also be reliably sampled over a large population of nanodots. We evidence that the heterogeneous rate constant characterizing the electron transfer between the nanodots and the Fc heads displays some quantifiable variability, but that the electron transfer remains in any case in the quasi-reversible regime. Overall, we demonstrate that Mt/AFM-SECM enables high throughput reading of dense nanoarrays, with a sensitivity and a read-out speed considerably higher than previously reported for SECM imaging of molecular microarrays.
Molecular arrays, in which sensing (bio)molecules are displayed on a surface in an array format, have become universal tools for a broad range of studies and applications ranging from genomics to proteomics.1 Their working principle is that minute amounts of probe biomolecules, which may be DNA strands, peptides, antibodies, or antigenic molecules, are deposited as micrometer sized spots on a surface and their binding to target molecules in solution is monitored. The array format enables the multiplexed simultaneous reading of a large number of spots for high throughput sensing, provided a suitably fast and spatially resolved read-out technique is used. Fluorescence imaging fulfills these criteria and has become the gold standard as a microarray read-out method. However other techniques, and in particular scanning electrochemical microscopy (SECM), have also been proposed to read, i.e. to image, molecular arrays.2 In SECM a microelectrode (tip), used as a local probe, is rastered over the sample to be interrogated and, schematically, collects electroactive species generated locally at a sample surface.3 SECM was notably used to characterize antibody microarrays in enzyme linked immunosorbant assay (ELISA)-like configurations,4 where capture primary antibodies were deposited in micrometric spots on a glass surface. Binding of the sought antigens was then revealed by making use of secondary antibodies labeled with an enzyme whose product was detected at the tip. The possibility of using SECM to image DNA microarrays was demonstrated as well.5,6,7,8,9 However, in spite of these pioneering works it is now largely admitted that the applicability of SECM for imaging sensing microarrays is seriously limited by its intrinsically slow
imaging (scanning) rate.2 This limitation notably arises from the fact that the tip scanning rate has to be kept slow enough to avoid disturbing diffusion of the electroactive species to be detected. In recent years, progresses in tip miniaturization,10 and the development of new SECM imaging modes11,12,13 and instrumentation,14,15 have considerably improved the performances of SECM, notably bringing its resolution down to the nanometer range.16,17,18 Yet, to this date the possibilities offered by such nano-SECM techniques as read-out tools for array-based sensing is still unexplored. This is highly regrettable since SECM, once endowed with nanometer resolution, potentially enables the interrogation of the new generations of very high density molecular arrays, namely molecular nanoarrays,19 where both the spot size and spacing tend toward a few tens of nanometers or less, and for which new highly resolved read-out techniques are still in need. Indeed, even though “regular” fluorescence imaging displays sufficient sensitivity to interrogate nanometer-sized spots containing a few or even a single fluorescent (or fluorescently labeled) biomolecule,20,21 its lateral resolution, limited by light diffraction, is insufficient to read high density nanoarrays where the nanodot separation falls below a few hundred nanometers. The aim of the present work is to demonstrate that atomic force electrochemical microscopy (AFM-SECM), a highresolution version of SECM combining atomic force microscopy and electrochemical imaging, operated in molecule touching mode (Mt),22 can be used as a high throughput serial read-out technique for imaging high density molecular nanoarrays. An high-speed electron beam lithography-based process
ACS Paragon Plus Environment
Analytical Chemistry
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
is used to pattern conducting silicon surfaces with dense and large area gold dot nanoarrays.23,24 Thermal treatment gives the nanodots the unique geometry of facetted single grain gold nanocrystals (nanoparticles) ideally suited for interrogation by AFM-based techniques.25–28 A model nanometer-sized redox labeled macromolecule, consisting in a flexible linear polyethylene glycol (PEG) - disulfide chain bearing a ferrocene (Fc) head is selectively grafted onto the gold nanodots by thioadsorption. Mt/AFM-SECM imaging of the nanoarray then allows topography and current images of the resulting molecular nanoarray to be simultaneously acquired. The topography image enables to identify individual nanodots, while the current image, reflecting the electrochemical detection of the FcPEG chains directly contacted by the AFM-SECM microelectrode tip, allows both the molecular occupancy, and the electrochemical reactivity of nanodots to be characterized at the single nanodot level.
EXPERIMENTAL SECTION Chemicals and solutions. The linear Fc-PEG3400-disulfide molecules (~79 (CH2-CH2O) monomer units per PEG chain) were custom-synthesized as described elsewhere.22 All other chemicals were analytical grade products and used as received. All solutions were prepared with double-deionized water (Milli-Q Millipore 18.2 MΩ cm-1 resistivity). All samples were protected from light during preparation and investigation. Nanoarray fabrication. Gold nanocrystal arrays were fabricated as described previously on a highly-doped n++ silicon substrate.23,24 In short, 10-nm-diameter holes are formed in a 45 nm-thick polymethyl methacrylate layer by e-beam exposure at 100 keV and development. A short dip in HF (1%, 30 seconds) removes the native silicon oxide in the holes, after which 8 nm of gold are evaporated at 3Å /s. A thermal annealing at 260°C leads to the formation of facetted gold nanocrystals. The bottom of the gold nanocrystals are buried in the silicon substrate and a silicon dioxide layer covers the dots due to a diffusion of silicon atoms through the gold dot and subsequent oxidation in air. This oxide layer is removed before use. Nanodots Fc-PEGylation. The nanoarray surface was first etched by HF treatment (1%, 2 min) to remove the thin SiO2 layer covering the gold nanodots. The treated surface was then rinsed in ethanol, and water, dried under N2, exposed to 5 min UV-ozone cleaning, then mounted in a O-ring fluid cell, and immediately exposed to a ~100 µM aqueous solution of FcPEG-disulfide for 15 h (overnight). The Fc-PEGylated surface was then thoroughly rinsed with ethanol in order to efficiently remove unwanted loosely adsorbed Fc-PEG chains from thedot-interspaces. After rinsing with H2O, the surface was finally covered with aqueous 1 M NaClO4 electrolyte solution for immediate in-situ characterizations. CV and AFM-SECM experiments. The O-ring fluid electrochemical cell was equipped with a Pt counter electrode, a Pt /polypyrrole quasi-reference electrode (~+ 100 mV/SCE),22 using the “nanoarray-substrate” as the working electrode. All potentials are given vs. SCE. The low sensitivity channel of the home-made AFM-SECM bipotentiostat was used to record substrate CVs. The Mt/AFM-SECM experiments were carried out with a Molecular Imaging PICOSPM I AFM microscope (Agilent), which was modified as described previously.22 The AFM-SECM tips were hand-fabricated according to a procedure adapted from literature,29 and detailed elsewhere.30 The
Page 2 of 9
topography images were first ordered flattened and the current images are raw data.
RESULTS AND DISCUSSION The single-grain gold nanodots arrays used in the present work were prepared using an original process which results in faceted gold nanocrystals half buried in highly doped silicon, having the shape of a truncated octahedron and displaying an atomically smooth flat top (scheme in Figure 1)23,27 Previous studies showed that these nanodots provide a nice ohmic contact with the underlying conducting silicon surface, meaning that they can be collectively electrically biased by applying a chosen voltage at the silicon material.25,27,28 In the present case, the nanodot fabrication process was tuned in order to form nanocrystals characterized by an average diameter, δ ~ 15 nm, a height, h ~2-3 nm, and arranged in a square pattern with 100 nm pitch (see Figure 1). Scanning electron microscopy (SEM) revealed the perfectly ordered nature of the pattern, confirmed the expected pitch, but also evidenced a 20 % relative standard deviation for the nanodot diameter in agreement with previous work.23 As a model macromolecule to be immobilized on the nanodots, we chose a home-synthesized Fc-PEG3400-disulfide molecule, featuring a polyethylene glycol chain and a redox ferrocene head.22 This nanometer sized, linear, flexible polymer molecule is a valid conformational model for Fc-DNA and Fcpeptide chains that one would ultimately want to immobilize on the nanodots for sensing applications. We previously showed that thanks to its disulfide group this molecule spontaneously forms end-thio-attached monolayers on gold surfaces22 and gold nanoparticles.31 Importantly, we have also previously shown that Fc-PEG chains are ideal markers for rendering redox-inactive globular biomolecules, such as antibodies, detectable by Mt/AFM-SECM.32,33 Ensemble scale CV characterization of the Fc-PEGylated nanoarray. A typical cyclic voltammogram (CV) recorded at a Fc-PEGylated nanodot array is presented in Figure 1a (solid line). It displays a pair of symmetrical peaks, which can be even better visualized after background subtraction, as evidenced in Figure 1b (green solid line). These peaks share a common peak potential value of ~ 0.13 V/SCE, which is close to the standard potential E° = 0.15 V/SCE of the Fc/Fc+ heads.34 We also verified that the intensity of the peaks increased linearly with scan rate. All of these characteristics demonstrate the presence of Fc-PEG chains anchored to the nanodots. Moreover, a very good fit of the experimental peaks was obtained using Laviron’s expression for the reversible CV expected for a surface-confined species undergoing simple fast (Nernstian) redox reaction (red dotted line in Figure 1b).35 This later result indicates that the Fc-heads form a single, homogeneous population of non-interacting redox species located outside of the electrochemical double layer developing around each nanodot. This is in contrast with what was previously observed for short-chained Fc-alkyl(C11)-thiol immobilized onto similar nanodot arrays, which gave rise to CVs featuring very narrow double peaks indicative of Fc-Fc interactions and double layer effects.28 Hence, the uncomplicated CV behavior observed here is to be attributed to the relatively long PEG tethers which maintain their Fc heads as isolated entities in a solution-like environment.
ACS Paragon Plus Environment
Page 3 of 9
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Analytical Chemistry
Figure 1. (Top) Left – Schematic representation of Fc-PEGylated gold nanodots interrogated by CV. Right –SEM image of a 100 nm pitch, 15 nm nominal dot size nanoarray. (Bottom) (a) Raw CV recorded at a Fc-PEGylated nanodot array. The dotted trace is a background signal recorded at a bare nanoarray. (b) Background subtracted CV (faradaic signal) (green trace) of the Fc-PEGylated nanoarray superimposed to the ideal Nernstian surface CV (red dotted trace). Scan rate 2 V/s in 1 M NaClO4 electrolyte. (c) Histogram of the nanodot diameter as measured from the SEM image.
Integration of the faradaic CV signal yields, N, the total number of Fc heads, hence of Fc-PEG chains, present on all of the nanodots borne by the sample surface (see Supporting Information). Thus, dividing N by the number of nanodots present on the sample, which is known by design, yields an accurate value of n, the average number of Fc-PEG chains per nanodot, also referred to as the average molecular occupancy of the nanodots. In the case of Figure 1, we found n ~ 75, whereas n derived from separate experiments falls in the range of 90 ± 20, independently of the nanodot pitch (varying from 50 – 200 nm). This later result confirms that the Fc-heads interrogated by CV are selectively immobilized on the gold nanodots. Comparing the respective sizes of the Fc-PEG chain (Flory radius, RF ~5 nm) and of the nanodot (~15 nm), the n value found here may seem high. However, this simply evidences that PEG chains terminally anchored to nanometer sized particles form high-density spherical brushes.36,37,38 Estimating the characteristic averaged nanodot surface area (Sdot = 200-250 nm2, see Supporting Information), allows the typical n value to be converted into a nanodot coverage, γ = 0.3-0.6 PEG chain /nm2. This γ value is very similar to the grafting density of ~0.4-0.8 PEG / nm2 we previously determined for surface-adsorbed ~20 nm diameter gold nanoparticles functionalized by Fc-PEG3400 disulfide.31 Such a result indicates that the nanodots display the same accessibility and reactivity for thio-anchoring of macromolecules as do gold nanoparticles pre-synthesized in solution. Mt/AFM-SECM interrogation of individual nanodots. Mt/AFM-SECM imaging was carried-out in situ in 1 M NaClO4, in tapping mode and using a home-made combined AFM-SECM probe (tip) oscillated at its fundamental flexure frequency. The probe was biased to Etip = + 0.3 V/SCE and the surface to Esub = 0.0 V/SCE. These potentials were chosen so as to be respectively very anodic and cathodic with respect to
the standard potential of the Fc heads. Simultaneously acquired topography and tip current images of the nanoarray are shown in Figure 2a and Figure 2b. Individual nanodots forming an almost perfect square array, can be unambiguously identified and precisely located from the topography image. The current image shows a similar array, but consisting in current spots, ascribable to the detection of the Fc-PEG chains borne by individual nanodots. The tip current is generated by the redox cycling of the Fc heads, which are alternatively oxidized at the tip and re-reduced at the nanodot, as depicted in the scheme presented in Figure 2. Comparing the topography and current images, but also their cross sections, evidences the very good correlation between the locations of the nanodots and the current spots. This result confirms the selective immobilization of the Fc-PEG chains on the nanodots. As evidenced in Figure 2b, a well-defined current spot can be unambiguously assigned to each of the nanodots identified in the topography image showing that all of the nanodots bear Fc-PEG chains. However, the current image also shows that there exists some variability in the spot intensity, suggesting some degree of dispersion of the nanodot molecular occupancy. Quantitative data can be obtained by examining the correlated cross sections taken along the Mt/AFM-SECM topography and current images. As illustrated in Figure 2a, cross section of the topography image enables to measure accurately the dimensions of all of the nanodots. The high resolution of Mt/AFM-SECM, and the high density of the nanodot array offer in particular the possibility of measuring the height of a large population of ~110 dots in a single 1 µm x 1 µm image (see Figure 2a). This allows for a reliable statistical analysis of the nanodot height, by constructing histograms such as the one presented in Figure 2c. It can be seen that the nanodot height distribution is characterized by an average value of 2.3 nm and a 26 % relative standard deviation. Interestingly, a similar height value is also found for bare gold nanodots (see Figure S3), showing that, as evidenced in previous works,22,31 flexible end-attached PEG chains are not sensed in topography by our combined AFM-SECM tips. The width of the nanodots visible in Figure 2a was also measured, and we found an average value of ~ 50 nm. This value is larger than the actual nanodot width, due to tip convolution effects, but it can be used to derive an estimate for the tip radius, Rtip, yielding Rtip ~ 90 nm (see Supporting Information). This radius is typical of a moderately sharp tip as produced in-house. Turning now to the current image, one can see that, in the cross section in Figure 2b, the current spots appear as well resolved bell-shaped curves, which can be nicely fitted by series of Gaussian curves displaying minimal lateral overlap. This fitting procedure enables the accurate measurement of the maximum intensity of each current spot (referred to below as the “spot current”, ispot) with no contribution from the current of neighboring spots. Please note that vertical rather than horizontal cross sections of the current image were selectively considered since, as is visible in Figure 2b, current dots tend to show some tailing in the image scan direction, an artefact ascribable to the slow response time of the high sensitivity current measurer. It is also worth mentioning that, to the best of our
ACS Paragon Plus Environment
Analytical Chemistry
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 4 of 9
Figure 2. Mt/AFM-SECM in situ imaging of a Fc-PEGylated nanodot array. Simultaneously acquired topography (a) and tip current (b) images. Cross sections of the images taken along the column of nanodots indicated by a vertical white dotted line are also shown. (c) and (d) are histograms of the nanodot height and spot current values. (e) Cross correlation plot showing the spot current of the nanodots as a function of their respective height. The yellow line is calculated using Equation S3 with the best fit value α = 60°.
knowledge, and as a benefit of the nanoarray format, this is the first time that such a large population of gold nanoparticles are electrochemically individually probed in a single image by a SECM-based microscopy techniques. Overall, the statistical distribution of the spot current measured from Figure 2b is captured by the histogram shown in Figure 2d, and is characterized by an average spot current of 80 fA and a relative standard deviation of 18 %. Quantifying the molecular occupancy of the nanodots. As demonstrated previously,39 the Mt/AFM-SECM tip current depends sharply on the tip-to-substrate distance, the motional dynamics of the PEG-attached Fc heads, and the number of Fc-PEG chains contacted by the tip. In the present case, the large tip radius, as compared to the nanodot size, and the low height of the nanodots, as compared to distance over which Fc-PEG chains can convey electrons by elastic bounded diffusion (~10 nm),39 allow us to consider that, when the tip is located exactly above a nanodot, it efficiently contacts all of the Fc-PEG chains it bears. As a result, the intensity of the spot currents can be considered as being simply proportional to the number of Fc-PEG chains borne by the nanodots. Hence, the spot current distribution presented in Figure 2d simply reflects the variability in the molecular occupancy of the nanodots. However, the proportionality constant between the current generated by the oscillating Mt/AFM-SECM tip and the actual number of chains contacted is not known a priori. Nevertheless, its value can be determined experimentally by noting that the average spot current derived from statistical analysis of the current image (80 fA in Figure 2) necessarily corresponds to the average number of chains per spot that was determined by CV, n = 75 for this surface. Hence, one obtains a current per chain parameter, i*, of ~ 1.1 fA/ chain. This value is purely phenomenological, as it is for example expected to depend on the tip oscillation frequency and imaging parameters, but is valid for analyzing a given experiment. Throughout the experiments i* values ranging from 1 to 1.5 fA/chain were consistently obtained. Knowing the i* value enables to convert the spot current histogram into a molecular occupancy histogram. In the case of Figure 2d one can then
see that the nanodots actually bore from ~40 to ~100 Fc-PEG chains. Similar results were obtained by imaging other nanoarray surfaces. A benefit of the dual imaging principle of Mt/AFM-SECM is that two different informations are acquired for each individual nano-object imaged. In the present case, both the height and the spot current of all of the nanodots imaged were actually measured. This uniquely allows us to investigate the possibility of any existing correlation between the height of the nanodots and their molecular occupancy. Such a correlation can be expected since higher nanodots necessarily expose more gold surface, and potentially bear more Fc-PEG chains (see scheme in Figure 1). To address this issue, cross correlation plots such as the ones presented in Figure 2e were constructed, where the spot current of each nanodot interrogated was plotted as a function of its respective height, h. As seen in Figure 2e, the upward orientation of the thus obtained point cloud is a visual indication that some degree of correlation exists. More quantitatively, ispot, is theoretically related to the nanodot size by: ispot = i* γ Sdot. Giving δ its average value of 15 nm, it is thus possible to use this later relation to fit the spot current vs. h data presented in Figure 2e, using only α, the facet angle defined in Figure 1, and the i*γ product as adjustable parameters. The later parameter simply acts as a proportionality constant whereas α actually sets the curvature of the computed i vs. h variation. It can be seen in Figure 2e that a good fit of the data can be obtained using the above equation and the following best fit values: i*γ = 0.39 fA/nm2 and α = 60°. This value of α is in agreement with the expected value of 30-60°,26,28 while the i*γ value translates into a γ value of ~ 0.35 PEG chain / nm2, which is close to the PEG grafting density value derived above by CV. These results can be regarded as satisfying selfconsistency tests validating our quantitative analysis approach. Importantly, we obtained similar results by analyzing images acquired in several separate experiments using distinct nanoarray surfaces (see Supporting Information).
ACS Paragon Plus Environment
Page 5 of 9
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Analytical Chemistry
Figure 3. Mt/AFM-SECM imaging of a Fc-PEGylated nanodot array as a way of recording CVs at individual nanodots. Simultaneously acquired topography (a), tip current (b), and substrate potential, Esub (c) images. Cross sections of the topography, current and Esub images, taken along the vertical green line shown in the “hold slow” region of the scan, are plotted below the images as red, purple and green traces respectively. (d) Single nanodot CV (NCV) obtained by plotting the spot current ispot as a function of the corresponding Esub, both read from the cross section. (e) Contour plot obtained by superimposing NCVs of 42 different individual Fc-PEGylated nanodots. Individual NCVs were normalized by setting their plateau current to unity. Red trace: average of all of the NCVs. White dotted curve: best fit of Equation (1) to the whole family of NCVs. (f) Histogram of the Λ parameter obtained by fitting Equation (1) to each of the individual NCVs.
Overall, we showed that differences in nanodot height largely contribute to the observed dispersion in spot current, i.e. in nanodot molecular occupancy. Yet, the fact that the spot current values are relatively broadly dispersed around the computed ispot vs. h variation shown in Figure 2e, with a distribution visibly exceeding the ~5% measurement error we estimated for the spot currents, indicates that other randomly scattered parameters, such as the nanodot width or the PEG grafting density, may also sensibly modulate the nanodot molecular occupancy. At this stage, we have thus validated the possibility of using Mt/AFM-SECM to detect a few tens of redox-labeled macromolecules immobilized on nanodot arrays. This of course represents a major improvement in terms of sensitivity as compared to the SECM detection of the billions of molecules immobilized on much larger microdots reported in literature.2 Such a result opens the possibility of monitoring in situany event that would translate into modulating the electrochemical response of individual nanodots, either by altering the molecular occupancy of the dots, or hampering/slowing access of the Fc heads to the nanodot surface, e.g. by binding to some other macromolecule. The former scenario would for example correspond to the detection of protease action onto immobilized Fc-labeled peptides, and the latter to biorecognition events such as Fc-PEGylated antigen / antibody reactions or Fc-DNA hybridization. Probing the electrochemical reactivity of individual nanodots. In order to fully validate the Mt/AFM-SECM electrochemical reading of the nanoarray it is required that the electrochemical reactivity of the nanodots, i.e. the rate at which they can exchange electrons with the Fc heads is quantified. This can be achieved by studying the dependence of the spot current associated with individual nanodots as a function of the potential applied to the substrate, Esub, while keeping the
tip potential sufficiently anodic for the electron transfer at the tip to remain infinitely fast (Etip = 0.3 V/SCE). This challenging measurement is carried-out by imaging the Fc-PEGylated nanoarray, until a well-defined line of nanodots is detected (e.g. scan line marked by the upper dotted white line in the topography and current images in Figure 3). At this stage, the slow scan axis direction is disabled, while the topography and current scan lines continue to be plotted in their respective images. Provided care has been taken to align the scan and nanoarray lattice directions, several particles (up to 9 in Figure 3) are then thus recursively interrogated by the tip in this socalled “hold slow mode”. From this moment-on the nanodots appear consequently as bands in the topography image. Shortly after the hold-slow mode is triggered, the substrate potential, Esub, is swept at 5 mV/s from its initial value of -0.025 V/SCE toward 0.3 V/SCE and back, while image acquisition continues. In order to visualize this potential ramp, the value of Esub is acquired as a third image as shown in Figure 3c. It can be seen, by comparing Figure 3b and Figure 3c, that the Esub sweep appears as a dark band in the current image. Once the Esub potential ramp is completed the slow scan axis is enabled again, at the level of the lower dashed line in Figure 3a and 3b, and regular image acquisition continues. Cross sections of the tip current and Esub images taken at the level of the center of nanodots, and in the region where Esub was swept, allow spot current and Esub paired data corresponding to individual nanoparticles to be obtained (Figure 3b, cross sections). By plotting the spot current as a function of the corresponding Esub value, S-shaped curves such as the one presented in Figure 3d are obtained. This curve actually corresponds to a stationary CV recorded at one well identified nanodot acting as an individual nano-electrode. The plateau current of such nanodot cyclic voltammogram (NCV) is proportional to the number of Fc heads addressed by the tip, i.e. to the nanodot molecular occupancy, while its rising portion
ACS Paragon Plus Environment
Analytical Chemistry
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
contains information regarding the rate of electron exchange between the nanodot and the Fc heads. As a benefit of the nanoarray format, NCVs can be recorded for all of the nanodots visible in each of the current images acquired. As a result, individual voltammograms can easily be obtained for tens of differing nanodots. Figure 3e actually shows the superimposition of 42 of such CVs displayed as a color coded contour plot. Please note that to correct for the dispersion in molecular occupancy, the NCVs were normalized by setting their plateau current to unity. Also note that even though several experimental SECM-based approaches have recently been described to achieve single nanoparticle voltammetry,17,16,40 this is the first time that such a parallelized acquisition of a large number of single particle CVs is reported. Figure 3e shows that all of the NCVs form a relatively narrow family of curve, indicating a moderate degree of dispersion of the electrochemical response of the nanodots. Averaging all of the data points of the spot current vs. Esub dataset thus yields a statistically significant single NCV, plotted as a red trace in Figure 3e, which is representative of the average electrochemical response of the nanodots. In order to analyze this average NCV, we first made use of a general theoretical treatment, that is suitable for analyzing quasi-reversible stationary voltammograms,41 and that is based on comparing the differences between the measured half-wave (E1/2) and quartile potentials (E1/3, E3/4) of the CV to tabulated values. This treatment has the interest of yielding simultaneously all of the parameters characterizing the electron transfer process in the framework of Butler-Volmer kinetics. These parameters are the apparent standard potential of the redox couple, E°’, the charge transfer coefficient, αBV, and a dimensionless parameter, Λ = ks /m, which compares the rate constant of the heterogeneous electron transfer, ks, to the mass transport coefficient, m. The definition of m depends on the stationary cyclic voltammetry technique actually used. In the present case analysis of the average NCV shown in Figure 3e, yielded: E°’ = 0.2 V/SCE, αBV = 0.36, Λ = 1.14. The value found for E°’ is surprisingly about 70 mV higher than the standard potential of the Fc heads, as derived from substrate CVs. This qualitatively indicates that confinement of the Fc-heads in the oscillating tip-substrate nanogap thermodynamically disfavors Fc oxidation into Fc+. This may result from hindered access of the ClO4- counter ion of the supporting electrolyte to within the nanogap, which is a necessarily step for Fc oxidation. We also note that the value obtained for αBV is somewhat lower than the value of 0.5 expected for a species undergoing outer-sphere electron transfer.42 For this reason, the NCV was also analyzed on the basis of the following equation, valid for any stationary CV,41,43 but imposing this time αBV = 0.5: ߰=
ଵ
ಷ భ బ.ఱಷ ሺா ሺாೞೠ್ିா°ᇲ ሻ൰ ିா°ᇲ ሻ൰ା ୣ୶୮൬ି ೃ ೞೠ್ ౻ ೃ
ଵାୣ୶୮൬ି
(1)
where ψ is the normalized spot current. As seen in Figure 3e (white dash trace), a very good fit is obtained using E°’ and Λ as the sole adjustable parameters, yielding: E°’ = 0.22 V/SCE, Λ = 0.51. Taken together the above two NCV analyses consistently point to a high apparent standard potential of the confined Fc heads, in the order of ~0.2 V/SCE, and to a moderately slow electron transfer rate (Λ = 0.5 -1). Extracting the exact value of ks from the value of Λ requires that the transport coefficient m is known. However, deriving an accurate value for m is
Page 6 of 9
difficult here since the Fc heads actually undergo a complex elastic bounded diffusion motion in a fluctuating tip-substrate gap. Yet, previous theoretical and experimental works aiming at characterizing electron transport by elastic bounded diffusion of the Fc heads of terminally attached Fc-PEG3400 chains have shown that for this system, m should be in the order of ~D/RF, with D the effective diffusion coefficient of the Fc head (D ~ 2 10-8cm2/s).39,44 Hence, considering that Λ values fall in the 0.5 -1 range, one obtains ks values in the ~ (2.0 – 4.0) 10-2 cm/s range. This value is in good agreement with the one of (4 ± 1) 10-2 cm/s that was measured by CV at a gold electrode for Fc-PEG3400 chains freely diffusing in solution.34 This quantitative agreement shows that the gold nanodots do behave as fully active gold nanoelectrodes. Note that ks values consistently found for the PEG-bond Fc heads are very small as compared to those found for soluble Fc species such as ferrocene methanol, which range from a few to a few tens of cm/s.45 This shows that the PEG chain sterically slows the electron transfer at nanodot surfaces. However, the important result here is that, due to an equally relatively slow mass transport, the electron transfer appears as being quasi-reversible (Λ >> 0.1). This means that for any practical purposes it will be sufficient to apply a moderately cathodic potential to the nanodots for the spot current to reflect solely the molecular occupancy of any Fc-PEG labeled molecular species borne by the nanodots, with no interference of the electron transfer rate. Actually, Figure 3e evidences that a current corresponding to 90 % of the plateau current value is reached for a Esub value of no more than +0.05 V/SCE. Being able to minimize the value applied to the gold nanodot is important for practical molecular array applications, notably in order to avoid interfering currents due to the detection of solution species, but also to preserve fragile species immobilized on the nanodots. The above considerations are relative to the average behavior of nanodots but in order to get some insights into the dispersity of their electrochemical reactivity we analyzed the data further. We fitted the NCVs recorded for each of the tens of individual nanodots interrogated with the help of Equation (1) using a common value of E°’ = 0.2 V and Λ as the sole adjustable parameter. The resulting Λ histogram, presented in Figure 3f, captures quantitatively the dispersity in the electrochemical reactivity of the nanodots. An average Λ value of Λav = 0.57 is obtained, which, is, as expected, close to the value obtained above by analysis the averaged NCV. One can also determine that the statistical distribution of Λ, and consequently of ks, is characterized by a relative standard deviation of 48%. This confirms the moderate but actual degree of dispersion of the nanodot electrochemical reactivity. Yet one can also determine that 43% of the nanodots actually display Λ values higher than Λav, whereas only a very small fraction of the nanodots (
