Single Gold Nanorod Charge Modulation in an Ion ... - ACS Publications

Subscriber access provided by Northern Illinois University

Communication

Single Gold Nanorod Charge Modulation in an Ion Gel Device Sean S. E. Collins, Xingzhan Wei, Thomas McKenzie, Alison M. Funston, and Paul Mulvaney Nano Lett., Just Accepted Manuscript • DOI: 10.1021/acs.nanolett.6b02696 • Publication Date (Web): 04 Oct 2016 Downloaded from http://pubs.acs.org on October 5, 2016

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.

Nano Letters 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 17

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

Nano Letters

Resubmission to Nano Letters September 2016

Single Gold Nanorod Charge Modulation in an Ion Gel Device Sean S. E. Collins,1 Xingzhan Wei,1,4 Thomas G. McKenzie,2 Alison Funston,3 Paul Mulvaney1* 1

School of Chemistry and Bio21 Institute, University of Melbourne, Parkville, VIC, 3010, Australia

2

Chemical and Biomolecular Engineering, University of Melbourne, Parkville, VIC, 3010, Australia 3

School of Chemistry, Monash University, Clayton, VIC, 3800, Australia

4

Chongqing Key Laboratory of Multi-scale Manufacturing Technology, Chongqing Institute of Green and Intelligent Technology, Chinese Academy of Sciences, Chongqing, 400714, China

Corresponding Author * E-mail: [email protected]

1 Environment ACS Paragon Plus

Nano Letters

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

Abstract A reliable and reproducible method to rapidly charge single gold nanocrystals in a solid-state device is reported. Gold nanorods (Au NRs) were integrated into an ion gel capacitor, enabling them to be charged in a transparent and highly capacitive device, ideal for optical transmission. Changes in the electron concentration of a single Au NR were observed with dark field imaging spectroscopy via localised surface plasmon resonance (LSPR) shifts in the scattering spectrum. A time-resolved, laser-illuminated, dark field system was developed to enable direct measurement of single particle charging rates with time resolution below one millisecond. The added sensitivity of this new approach has enabled the optical detection of fewer than 110 electrons on a single Au NR. Single wavelength resonance shifts provide a much faster, more sensitive method for all surface plasmon based sensing applications.

Keywords: Gold nanorods, single particle spectroscopy, plasmonics, dark field microscopy, ion gel, solid-state charging


Page 2 of 17

Page 3 of 17

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

Nano Letters

The ability to measure very small and fast localised surface plasmon resonance (LSPR) changes on single metal nanocrystals would be extremely useful for catalysis, sensing and photonics applications. Optical measurement of small numbers of electron transfer events occurring on a single metal nanocrystal could provide new understanding of shape, size and facet dependent catalysis, as well as detailed insights into reaction mechanisms. Resolving the transfer of individual electrons by monitoring changes in LSPR would enable surface plasmon spectroscopy (SPS) measurements to probe the proposed quantum catalysis regime.1 In this regime, all charge transfer is occurring discretely, and thus reactions can be reported as a series of single, time-tagged charge transfer events rather than as sums. Observing discrete chemical reactions occurring on single nanoparticles in real-time is an enormously compelling concept and one that could perhaps reveal rate limiting processes in complex catalytic reactions. For example, in a two-step surface reaction, the electron transfer to or from a primary redox species must occur before the secondary reaction step can proceed. An example of this is CO oxidation on metal nanocrystals, where oxygen adsorption is required before CO can be converted to CO2. A highly sensitive system to measure LSPR as a function of time on a single nanocrystal also could potentially lead to the observation of an optical ‘Coulomb-blockade’ or ‘quantum-staircase’ device. These discrete electron transfer devices have been built and measured electronically,2-5 but are yet to be demonstrated optically. A variety of SPS based measurements of single particle redox processes have been carried out on Au NCs including ascorbic acid oxidation,6 electrochemical charging and electroadsorption,1, 7-10 hydrogen sensing and spillover,11-13 and Ag shell electrochemistry.14,

15

Instrumental progress

towards detection of large, single molecules has enabled step-wise detection of individual proteins binding onto the surface of single Au NRs. A white-light laser illuminated dark field microscopy (DFM) system was used by Sönnichsen and co-workers to detect single 450 kDa proteins16 while Orrit and co-workers used a photothermal absorption spectroscopy system to detect molecules as small as 53 kDa.17 A scanning confocal epi-illuminated DFM system was employed by Herrmann and Baumberg to improve the temporal resolution of imaging spectroscopy for measuring single Au


Nano Letters

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

NR growth kinetics18, first reported by Novo et al.6 This enhanced DFM method was effective in determining the kinetics of single Au NR growth with 8 ms temporal resolution and 0.1 nm spectral resolution. Recently the advantage of using narrow-band illumination to simultaneously detect stochastic protein interactions on multiple Au NRs using a superluminescent diode-illuminated hyperspectral DFM was demonstrated.19,

20

It was shown that the measurement integration time

using this method could be reduced to 6 ms whilst retaining the ability to resolve single proteins encountering single Au NRs in a flow cell.20 In the work presented here, narrow-band illumination surface plasmon spectroscopy is developed further to improve temporal resolution of LSPR changes on single Au NRs. The principal goal is to increase the electron sensitivity at the single particle level. Acousto-optical tunable filters (AOTFs) are used to select a monochromatic illumination band (~2-4 nm bandwidth) from a supercontinuum laser source to rapidly monitor the scattered photon flux within a narrow wavelength band. This facilitates faster signal acquisition compared to scanning monochromator or spectral imaging-based techniques, and is ultimately limited only by the signal-to-noise ratio and sampling rate of the photon counting electronics. Additionally, the setup has a much lower photon flux requirement because the full range of resonant energies is not needed during the measurements. A lower photon flux is important in plasmonic studies because the resultant heating at high fluxes can influence charge transfer kinetics,21 alter the local refractive index22 and, for anisotropic nanoparticles, even lead to shape changes.23 Two crucial steps in realising electrically switchable plasmonic devices are high-speed charging rates and sensitive detection methods. To achieve these goals, chemically synthesised Au NRs are employed as individual plasmonic elements in an ion gel capacitor. Due to their very high electrical double-layer capacitance, iongel based condensers can reach higher charge densities than conventional solid-state devices.24 A further advantage of integrating Au NRs into a quasi-solidstate device is that the handling difficulties of fluid cells, such as leaking and evaporation, previously used to electrochemically study single Au NCs are obviated. The resultant ion gel device 4 Environment ACS Paragon Plus

Page 4 of 17

Page 5 of 17

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

Nano Letters

is highly transparent, enabling the scattered light of individual Au NRs to be monitored using the laser DFM system in transmission. The LSPR of a single Au NR can be rapidly switched in the device by periodic modulation of the electron density on the nanorod surface, as we demonstrate.

Figure 1. (a) Schematic representation of the transparent ion gel capacitor device structure. (b) Plot of the reactance, Z'', vs Frequency of the device, which was used to determine the specific capacitance.

The device consisted of a 145 µm thick ion gel layer sandwiched between two pieces of 140 nm ITO on quartz glass. The gel layer was formed by spin casting triblock copolymer poly(styrene)-bpoly(methyl methacrylate)-b-poly(styrene) ([SMS]) codissolved with 1-ethyl-3-methylimidazolium bis(trifluoromethylsulfonyl)imide ([EMI][TFSI]) in ethyl acetate onto a 25 mm × 25 mm photolithographically patterned piece of ITO. This iongel formulation was selected because Lee and co-workers demonstrated that it was a highly capacitive and highly stable solution-processable solid electrolyte.24 The pre-patterning of the ITO was carried out to create six separate cathodes, each with a 2 mm × 10 mm conductive area, to reduce the area of the cathode and increase the number of capacitors per substrate. The patterned substrate was functionalised prior to assembly with a monolayer of (3-mercaptopropyl)trimethoxysilane (MPTMS), followed by a low density submonolayer of Au NRs with average dimensions of 30 nm × 94 nm. The final device configuration is illustrated in Figure 1a.


Nano Letters

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

To determine the specific capacitance of the device, impedance spectroscopy analysis was carried out with and without a -1 V DC bias using a potentiostat (AutoLab PGSTAT302N). In Figure 1b, the results of -Z'' against frequency are displayed, where Z'' is the imaginary part of the impedance in the form of impedance Z = Z' + iZ''. A specific capacitance of 5.5 µF/cm2 at 1 Hz and -1 V bias was calculated using the equation from Lee et al.24 In Figure 2a, the position of the LSPR maximum, λmax (red circles and line), of a single Au NR is presented as a function of time, while a square wave potential is applied (blue dotted line), between 0 V and -2 V using a two electrode configuration with the potentiostat. The scattering signal was measured in transmission using halogen lamp illuminated dark field imaging spectroscopy. In this technique the scattered light from individual nanocrystals was reflected onto a spectrograph grating and the resultant diffracted light was then focussed onto an array detector to produce single particle scattering spectra. To isolate point scatterers on a dark background, the sample plane was illuminated at an oblique angle which enabled individual nanocrystal scattering signals to be analysed. A 10 nm blue-shift was observed for the Au NR in Figure 2 when the potential was switched between externally applied voltages of 0 V and -2 V. The significant blue-shift observed here occurred with a sub-second shift rate and, furthermore, the SPR peak position was stable to within a few nanometres. The blue-shifting behaviour was observed for all Au NRs measured, although the magnitude and charging rate both varied somewhat from particle-to-particle on the same cathode. Supplementary graphs illustrating the variations of these responses can be found in the SI (Fig. S3 and S4). The blue-shift magnitude of a single Au NR on switching from 0 V to -2 V was within the range of 5 nm to 10 nm for 75% of the Au NRs measured. During both the charging and discharging stages, the change in λmax for many of the Au NRs occurred at a rate faster than could be reliably resolved using a scientific grade CCD. Full scattering spectra for the same single Au NR in Figure 2a are shown in Figure 2b at 0 V and -2 V. We attribute the blue-shift response following application of a negative bias to an increase in electron concentration on Au NRs via electron transfer from the ITO. Charging via this mechanism has been reported previously by a


Page 6 of 17

Page 7 of 17

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

Nano Letters

number of authors,1, 8, 9, 25 and the optical effects observed here are consistent with these reports as well as the effects observed by chemical charging.6, 26 Another explanation for the optical shifts is that there is a change in the local refractive index occurring due to a tightly bound layer of ions in the Helmholtz layer. However, we do not believe this is the major contributor to the optical shifts, as it is predicted this would increase the refractive index near the gold rod surface,27, 28 leading to a red-shift in the LSPR. Furthermore, an increase in the overall scattering efficiency of the Au NR was observed upon application of -2 V; at this potential the peak intensity increased approximately 20%. Increased scattering has been correlated with increasing electron density on Au NRs with these dimensions and explained in detail in previous work.13 The exposure time for these spectral measurements was 600 ms, with a total acquisition time of 1,000 ms. The noise level of the spectrum in Figure 2b is significant across the band. Using an exposure time any lower than this resulted in a signal-to-noise ratio that was too low for time-resolved spectral analysis.

Figure 2. Single particle spectroscopy data acquired with a standard dark field imaging spectroscopy system for a single Au NR in the ion gel device. The CCD exposure time was 600 ms, with a total acquisition time of 1,000 ms for each spectrum. (a) The wavelength position of the longitudinal LSPR maximum of a single 30 nm × 94 nm AuNR (red trace) against time, during three cycles between 0 V and -2 V displayed on the y-axis on the right hand side (blue trace). (b) Full scattering spectra displayed for the same single rod with an applied voltage of 0 V (red) and -2 V (blue) at 20 s and 50 s into the sequence, respectively. The lines on top of the data points are Lorentzian fits, from which the peak positions were obtained. To resolve the optical time constants of the Au NRs, the conventional DFM was modified to facilitate faster data acquisition. As outlined in the schematic in Figure 3a, we utilised an inverted 7 Environment ACS Paragon Plus

Nano Letters

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

optical microscope with a supercontinuum light source and a photon multiplier tube (PMT). The increased illumination power of the laser in conjunction with a PMT in single-photon counting mode were implemented to achieve a significant enhancement of the sensitivity compared to the standard dark field imaging spectroscopy technique (typically a 100 W halogen lamp and a scientific grade CCD array). Imaging DFM spectroscopy sensitivity benchmarks have been set much higher18 than shown in Figure 2. However, the purpose of initially using a broadband spectrometer was not to show improvements in sensitivity but rather to verify our observations at selected wavelengths when measuring single Au NR charging with a narrow-band spectroscopy system.

Figure 3. (a) Schematic diagram of the laser spectroscopy setup. AOTF = acousto-optical tunable filter, MM = multimode, M = mirror, DFC = dry dark field condenser (0.8-0.95 NA), TL = tube lens, PH = 100 µm pin hole, PCL = planoconvex lens, FM = flip mirror, PMT = photon multiplier tube, CMOS = complementary metal oxide semiconductor, CPU = computer. (b) AOTF scangenerated scattering spectra of a single Au NR under zero external bias using the laser DFM system.

A pseudo-white light source output from the laser (Fianium SC450-6 with dual-AOTF band selection) in the setup shown in Figure 3a was used to initially locate single Au NRs in the device. The AOTF filtered laser light was launched into a multimode fibre and then collimated upon exiting 8 Environment ACS Paragon Plus

Page 8 of 17

Page 9 of 17

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

Nano Letters

with an achromatic lens assembly. The resultant laser light then entered the dark field condenser off-centre, yet parallel to the optical axis, ensuring the light was aligned outside the internal beam block (see Figure S3). The beam was incident at the oblique angle required for dark field illumination. The beam spot size diameter was approximately 30 µm when focussed on the top surface of the device, but was stretched elliptically when focussed onto the bottom electrode surface, where the Au NRs were deposited due to the beam passing through the gel layer at an oblique angle. Once a single Au NR of interest was identified in the microscope, it was centred on the CMOS with a mechanically-driven stage. To isolate the target particle from nearby particles and background scattering, a 50 µm pinhole was inserted into the light path at the first image plane after the microscope tube lens. Following the insertion of the pinhole, a flip-mirror was then retracted to direct the light path to the PMT. The scattering spectrum of an Au NR was generated by scanning the laser wavelength range from 600 nm to 850 nm in 1 nm steps at a rate of 0.545 nm/s (maximum power of 12 mW at 700 nm). At each step, a photon count from the PMT was collected to produce a full scattering spectrum of the longitudinal band of a single particle. The wavelength dependency of the laser, quantum efficiency of the detector, and background signal, were accounted for by repeating each scan in an area free of Au NRs. This background signal was subtracted from the particle signal, and then divided by it to remove the background and to normalise the spectrum. An example of a single Au NR scattering spectrum is shown in Figure 3b. The raw data was fitted by a Lorentzian function and the λmax and half width at half maximum (hwhm) were extracted; these were 761 nm and 34 nm for this specific particle, respectively. This scanning technique enabled the subsequent time-dependent measurements to be made without the need to switch between a broadband diffraction-based spectroscopy setup. Although a longer time was required to collect a full scan compared to that required for conventional dark-field scattering, this circumvented the need to locate the same particle twice. After determining the spectral parameters of an Au NR (λmax and hwhm) fast charging measurements could be conducted. The laser was set to 727 nm where the largest change in


Nano Letters

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

scattering cross section, ∆Qsca, in response to blue-shifting was expected to occur.1 This wavelength was the half maximum on the blue side of the peak for the Au NR shown in Figure 3b, where the gradient of the Lorentzian function is steepest. Figure 4a shows the scattering signal at 727 nm of the Au NR in Figure 3b as a function of time upon oscillation of the applied potential between 0.0 V and -2.0 V at 0.1 Hz. The initial signal (at 0.0 V) was stable (blue trace) with a value of approximately 2,500 photons/ms. Photons were collected with 1 ms bin time. When the voltage was switched to -2.0 V, the signal level increased by approximately 22% to 3,100 counts/ms. When the voltage was returned to 0.0 V, the signal returned to a similar count as the initial value and varied by just 8% over the four cycles shown here, once it had stabilised. From Figure 4a, it can be seen that the signal level for the particle at -2.0 V is slightly more consistent and stable than at 0.0 V. To confirm that the change in signal intensity was due to the blue-shifting of the LSPR, the red flank of the peak was monitored at the half maximum (795 nm) during a subsequent repeat of the sequence (red trace). The red trace displays a lower initial signal of ~720 counts/ms, resulting from the lower initial laser power at this relatively long wavelength and the drop in sensitivity of the PMT in the NIR spectra regime. Monitoring the shifts on the longer wavelength side of the resonance leads to shifts upon application of a negative bias. The red trace exhibits an 18% decrease in scattering upon switching to -2.0 V, confirming that the changes in signal were indeed due to a blue-shift. This result is consistent with that obtained by the standard DFM technique in Figure 2a. The asymmetry in the two sets of traces is due to the fact the laser is not tuned precisely to the most sensitive wavelengths for monitoring SP shifts. This issue is addressed below.


Page 10 of 17

Page 11 of 17

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

Nano Letters

Figure 4. Time-resolved kinetic charging data for a single 30 nm × 94 nm gold nanorod, acquired using the laser DFM setup. (a) Photon count per millisecond plotted against time of the blue flank at half maximum (blue trace) and red flank at half maximum (red trace), during 4 cycles of 0.0 V (white background) to -2.0 V (grey background). (b) Optical response of charging (dots) with single exponential fits (lines) for the blue and red flanks.

In Figure 4b, the second cycle for both the blue flank and the red flank are shown together with fitted exponential functions. The exponential fitted data allowed the optical time constant, τOPT, to be calculated for each cycle. For the specific Au NR in Figure 4b, sequential measurements of τOPT for charging were found to be: 50 ms ± 1 ms, 82 ms ± 2 ms, 119 ms ± 5 ms and 123 ms ± 5 ms, with an average value of 90 ms ± 30 ms. The apparent trend of increasing τOPT was not always observed for other Au NRs as shown in the supplementary data in the SI (Figure S6, Table S1). Similar τOPT values were obtained on the red flank of the resonance (264 ms ± 7 ms, 151 ms ± 5 ms, 116 ms ± 3 ms, 111 ms ± 2 ms). Assuming that the Au NRs are charging at the same rate as the entire ITO layer, i.e. their Fermi levels remain in equilibrium at all times, then the electrical RC time constant of the total ion gel capacitor, τELEC, should be equal to the optically measured τOPT of the Au NR. This theory was tested by monitoring the sequence with the potentiostat and fitting exponential functions to the electrical data (current vs. time data) in the same way. The uniformity of the chronoamperograms recorded simultaneously with the optical measurements during charging and discharging is demonstrated in the plot of Figure S7 in the Supporting Information. During charging, the τELEC for each cycle were found to be 109 ms ± 6 ms, 89 ms ± 5 ms, 87 ms ± 5 ms 11 Environment ACS Paragon Plus

Nano Letters

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

Page 12 of 17

and 88 ms ± 5 ms, with an average value of 90 ms ± 10 ms. These values closely match the average τOPT of the Au NR under charging. Conversely, during the discharging steps the agreement between τELEC and τOPT was weaker, 230 ms ± 50 ms for the optical value of the Au NR taken from the blue flank and 71 ms ± 2 ms for the τELEC of the device. A longer τOPT for discharging was observed for all particles, suggesting that the Au NRs do not electrically discharge at the same rate as the ITO plates. It is evident from measurements on other particles (Figure S6), as well as those measured with the lamp-illuminated DFM (Figure S4 and S5), that not all particles are charging at the same rate

as

the

whole

cathode.

This

may

be

due

to

local

differences

in

the

3-

mercaptopropyl)trimethoxysilane layer thickness used to functionalize the cathode or different orientations of the Au NR on the cathode. Differences in optical and electrical trends across cycles are likely to be due to local variations of charge density on the cathode at certain times. Charge retention on the Au NRs may also explain why the τOPT for charging fluctuates between cycles. Note that the diffuse double layer relaxation time is of order microseconds, and does not limit the charge transfer processes. The actual shape of the surface plasmon resonance is rarely perfectly Lorentzian. Inhomogeneity in shape, the dielectric mismatch at the surface, and adsorbate damping can all cause the SP resonance to be asymmetrical. Furthermore, the decrease in damping by gold, leads to asymmetry in the scattering signal across the resonance, i.e. to dispersion in the response. Hence in order to determine the most sensitive wavelengths at which to carry out time resolved measurements, it is useful to do potential step scattering measurements across the SP resonance. To determine ∆Qsca over the full range of the longitundal LSPR band, the laser was set to a single wavelength and then the scattering signal was recorded in 16 nm intervals separately as the potential was brought to -2.0 V from 0.0 V in 0.5 V steps. After each change in potential, the signal was left for several seconds to stabilise, after which time data was collected over 30 s to determine ∆Qsca. The average values of ∆Qsca over the full potential change are plotted in Figure 5a. Error bars are estimated as the standard deviation of the photon count/s. Qsca increases at shorter wavelengths, below the isosbestic-like point at about 12 Environment ACS Paragon Plus

Page 13 of 17

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

Nano Letters

755 nm, and decreases at longer wavelengths. This behaviour occurs due to a blue-shift and an increase in scattering, resulting from the increase in electron density. The asymmetric plot shape either side of the isosbestic-like point is due to the blue-shift and increase in scattering occurring together. A blue-shift alone would result in equal but opposite amplitude of ∆Qsca either side of the isosbestic-like point.

Figure 5. (a) Percentage change in the scattering cross section, ∆Qsca, for set wavelengths across the longitudinal surface plasmon band of a single Au NR at different applied voltages; (b) Average percentage change in the scattering cross section, ∆Qsca, relative to ∆Qsca at 1 V (at the blue-flank of hwhm) as a function of the potential wave magnitude for three measured Au NRs. Inset shows low magnitude range below -50 mV.

The method used to determine ∆Qsca produces uncertainties through low signal-to-noise ratios, which prevent measurement of values below a 5% change in relative scattering signal. For example, a 200 mV modulation resulted in a ∆Qsca signal-to-noise ratio of approximately 2:1. An averaging technique suitable for digital photon counting applications was applied to overcome this limitation. A square wave function generator (Agilent 33220A Function/Arbitrary Waveform Generator) was used to charge the device by alternating between a high and low voltage at a frequency of 0.5 Hz, a frequency value low enough to allow sufficient time for signal collection. The average scattered photon count at the blue flank of the half-maximum for each particle was calculated after 200 cycles by separately binning the high and low voltage periods, counting only the last 800 ms of each 1,000 ms half-cycle (50% duty cycle). The averages for ∆Qsca relative to ∆Qsca at 1 V as a function of the 13 Environment ACS Paragon Plus

Nano Letters

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

potential wave magnitude are displayed for three measured Au NRs in Figure 5b down to 10 mV. The averaging technique was shown to be effective in measuring ∆Qsca values below 0.1%. Error bars were estimated within a 95% confidence range and the response to the voltage applied was fitted with a linear curve (R2 = 0.62), an expected response for an ideal capacitor. From ∆Qsca, it is possible to calculate the average number of electrons transferred during each wave cycle. A change in ∆Qsca of ~0.05% was obtained in the low potential range for a specific Au NR (see Fig. S8b) which can be converted to a number of electrons by modelling the expected shifts with the finite element method in COMSOL. Scattering spectra of a 30 nm × 94 nm Au NR in an estimated refractive index polymer environment of 1.45 was simulated with charge density changes ∆N between 0% and 1% (Figure 6a). Figure 6a shows that there is very little overall change to the LSPR lineshape or position at these low ∆N values. If the ∆Qsca at the blue flank hwhm position is plotted against ∆N (Figure 6b), the gradient of the linear relationship between these parameters can be obtained. At 0.05% ∆Qsca the value for ∆N is 0.0033%. The total number of atoms in the nanorod was calculated to be 3.5 × 106, using the volume and density of an average gold nanorod from this sample. The number of free electrons is equal to the number of atoms for gold. Using this value, it was calculated that there were 110 ±10 electrons on average transferring on and off a single Au NR within each cycle. A change of 110 electrons is equivalent to a 0.02 nm blue-shift, below the detection limits of the CCD imaging-based single particle spectroscopy systems reported.18, 20 The uncertainty reported has been estimated from the size distribution of the rods. This is the smallest optically measured number of electrons transferred on a single gold nanocrystal reported to date.


Page 14 of 17

Page 15 of 17

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

Nano Letters

Figure 6. Modelling data generated by COMSOL. (a) Calculated scattering spectra of a gold nanorod with different charge densities, ∆N, assuming refractive index = 1.45. Dielectric data for gold is from Johnson and Christy.29 Dotted line represents the blue flank of the hwhm. (b) Plot of ∆Qsca against ∆N taken from the generated spectra on the left at the blue flank of the hwhm.

Reporting optical detection of 100 ms RC charging kinetics on a single particle and the detection of close to 100 electrons surpasses previous reports in this field.1, 6 The method of optically tracking electron transfer introduced in this work would be useful for measuring previously undetected surface processes on gold nanoparticles. Two such examples are the detection of small numbers of adsorbing low molecular weight molecules and low activity redox reactions. Another useful prospect for this method is detecting spectral shifts on metal nanocrystals with highly damped LSPR, such as palladium and platinum NCs. Both materials are industrially important catalysts but their catalytic activities are notoriously difficult to study directly using surface plasmon spectroscopy.30 The super-continuum laser approach here may enable the detection of very small changes in LSPR lineshape at specific wavelengths, providing a sensitive method toward optical catalyst screening. In summary, a spectroscopic laser DFM system has been developed to resolve the optical switching constant of individual Au NRs in a novel ion gel capacitor device. The optical switching time constant τOPT was found to be less than 100 ms and this is easily tuned by reducing the thickness of


Nano Letters

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

the ion gel spacer layer, which determines the DC resistance of the structure. Less than 0.1% changes in scattering cross section at the flank of the half maximum were detected, corresponding to the transfer of around 100 electrons. These findings exceed the sensitivity of charge-induced changes in any previously reported work on single metal NCs.

Supporting Information Details on the gold nanorod synthesis, polymer synthesis, device assembly and characterisation, and dark-field microscopy are supplied in the Supporting Information. This material is available free of charge via the Internet at http://pubs.acs.org. The authors declare no competing financial interest.

Acknowledgements This research was supported by ARC Grant LF100100117 and USAF AOARD Grant FA2386-131-4072. SC and TM acknowledge the financial support of a Melbourne Research Scholarship and an Australian Postgraduate Award respectively.

References 1.

Novo, C.; Funston, A. M.; Gooding, A. K.; Mulvaney, P., J. Am. Chem. Soc. 2009, 131, 14664-14666.


Page 16 of 17

Page 17 of 17

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

Nano Letters

2. Thelander, C.; Magnusson, M. H.; Deppert, K.; Samuelson, L.; Poulsen, P. R.; Nygård, J.; Borggreen, J., Appl. Phys. Lett. 2001, 79, 2106. 3. Bolotin, K. I.; Kuemmeth, F.; Pasupathy, A. N.; Ralph, D. C., Appl. Phys. Lett. 2004, 84, 3154. 4. Pradhan, S.; Sun, J.; Deng, F.; Chen, S., Adv. Mater. 2006, 18, 3279-3283. 5. Guttman, A.; Mahalu, D.; Sperling, J.; Cohen-Hoshen, E.; Bar-Joseph, I., Appl. Phys. Lett. 2011, 99, 063113. 6. Novo, C.; Funston, A. M.; Mulvaney, P., Nat. Nanotechnol. 2008, 3, 598-602. 7. Hoener, B. S.; Byers, C. P.; Heiderscheit, T. S.; De Silva Indrasekara, A. S.; Hoggard, A.; Chang, W.-S.; Link, S.; Landes, C. F., J. Phys. Chem. C 2016, 120, 20604-20612. 8. Byers, C. P.; Hoener, B. S.; Chang, W. S.; Yorulmaz, M.; Link, S.; Landes, C. F., J. Phys. Chem. B 2014, 118, 14047-14055. 9. Dondapati, S. K.; Ludemann, M.; Müller, R.; Schwieger, S.; Schwemer, A.; Händel, B.; Kwiatkowski, D.; Djiango, M.; Runge, E.; Klar, T. A., Nano Lett. 2012, 12, 1247-1252. 10. Byers, C. P.; Hoener, B. S.; Chang, W. S.; Link, S.; Landes, C. F., Nano Lett. 2016, 16, 2314-2321. 11. Seo, D.; Yoo, C. I.; Chung, I. S.; Park, S. M.; Ryu, S.; Song, H., J. of Phys. Chem. C 2008, 112, 24692475. 12. Sil, D.; Gilroy, K. D.; Niaux, A.; Boulesbaa, A.; Neretina, S.; Borguet, E., ACS Nano 2014, 8, 77557762. 13. Collins, S. S. E.; Cittadini, M.; Pecharromán, C.; Martucci, A.; Mulvaney, P., ACS Nano 2015, 9, 78467856. 14. Chirea, M.; Collins, S. S. E.; Wei, X.; Mulvaney, P., J. Phys. Chem. Lett. 2014, 5, 4331-4335. 15. Byers, C. P.; Zhang, H.; Swearer, D. F.; Yorulmaz, M.; Hoener, B. S.; Huang, D.; Hoggard, A.; Chang, W. S.; Mulvaney, P.; Ringe, E.; Halas, N. J.; Nordlander, P.; Link, S.; Landes, C. F., Sci. Adv. 2015, 1, e1500988. 16. Ament, I.; Prasad, J.; Henkel, A.; Schmachtel, S.; Sönnichsen, C., Nano Lett. 2012, 12, 1092-1095. 17. Zijlstra, P.; Paulo, P. M.; Orrit, M., Nat. Nanotechnol. 2012, 7, 379-382. 18. Herrmann, L. O.; Baumberg, J. J., Small 2013, 9, 3743-3747. 19. Becker, J.; Trügler, A.; Jakab, A.; Hohenester, U.; Sönnichsen, C., Plasmonics 2010, 5, 161-167. 20. Beuwer, M. A.; Prins, M. W. J.; Zijlstra, P., Nano Lett. 2015, 15, 3507-3511. 21. Hung, W. H.; Aykol, M.; Valley, D.; Hou, W.; Cronin, S. B., Nano Lett. 2010, 10, 1314-1318. 22. Link, S.; El-Sayed, M. A., J. Phys. Chem. B 1999, 103, 4212-4217. 23. Link, S.; Burda, C.; Nikoobakht, B.; El-Sayed, M. A., The Journal of Physical Chemistry B 2000, 104, 6152-6163. 24. Lee, K. H.; Zhang, S.; Lodge, T. P.; Frisbie, C. D., J. Phys. Chem. B 2011, 115, 3315-3321. 25. Chapman, R.; Mulvaney, P., Chem. Phys. Lett. 2001, 349, 358-362. 26. Novo, C.; Mulvaney, P., Nano Lett. 2007, 7, 520-524. 27. Brown, A. M.; Sheldon, M. T.; Atwater, H. A., ACS Photonics 2015, 2, 459-464. 28. Templeton, A. C.; Pietron, J. J.; Murray, R. W.; Mulvaney, P., J. Phys. Chem. B 2000, 104, 564-570. 29. Johnson, P. B.; Christy, R. W., Phys. Rev. B 1972, 6, 4370-4379. 30. Tittl, A.; Giessen, H.; Liu, N., Nanophotonics 2014, 3, 157–180.

TOC image:


Single Gold Nanorod Charge Modulation in an Ion ... - ACS Publications

Recommend Documents