Probing Photothermal Effects on Optically Trapped Gold Nanorods by

Sep 5, 2017 - (23) However, residual ohmic losses in the particles are unavoidable near resonance, and their advantageous properties therefore come at...
1 downloads 13 Views 1MB Size
Subscriber access provided by Warwick University Library

Article

Probing Photothermal Effects on Optically Trapped Gold Nanorods by Simultaneous Plasmon Spectroscopy and Brownian Dynamics Analysis Daniel Andrén, Lei Shao, Nils Odebo Länk, Srdjan S. A#imovi#, Peter Johansson, and Mikael Käll ACS Nano, Just Accepted Manuscript • DOI: 10.1021/acsnano.7b04302 • Publication Date (Web): 05 Sep 2017 Downloaded from http://pubs.acs.org on September 10, 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.

ACS Nano 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 13

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

ACS Nano

Probing Photothermal Effects on Optically Trapped Gold Nanorods by Simultaneous Plasmon Spectroscopy and Brownian Dynamics Analysis Daniel Andrén†,*, Lei Shao†, Nils Odebo Länk†, Srdjan S. Aćimović†, Peter Johansson†,‡, Mikael Käll†,* †

Department of Physics, Chalmers University of Technology, S-412 96 Gothenburg, Sweden, ‡School of Science and Technology, Örebro University, S-701 82 Örebro, Sweden. *Corresponding authors: [email protected], [email protected] KEYWORDS: photothermal effects, gold nanorod, optical tweezers, nanomotors, thermal reshaping, Brownian dynamics ABSTRACT: Plasmonic gold nanorods are prime candidates for a variety of biomedical, spectroscopy, data storage and sensing applications. It was recently shown that gold nanorods optically trapped by a focused circularly polarized laser beam can function as extremely efficient nanoscopic rotary motors. The system holds promise for applications ranging from nanofluidic flow control and nanorobotics to biomolecular actuation and analysis. However, to fully exploit this potential, one needs to be able to control and understand heating effects associated with laser trapping. We investigated photothermal heating of individual rotating gold nanorods by simultaneously probing their localized surface plasmon resonance spectrum and rotational Brownian dynamics over extended periods of time. The data reveal an extremely slow nanoparticle reshaping process, involving migration of the order of a few hundred atoms per minute, for moderate laser powers and a trapping wavelength close to plasmon resonance. The plasmon spectroscopy and Brownian analysis allows for separate temperature estimates based on the refractive index and the viscosity of the water surrounding a trapped nanorod. We show that both measurements yield similar effective temperatures, which correspond to the actual temperature at a distance of the order 10-15 nm from the particle surface. Our results shed light on photothermal processes on the nanoscale and will be useful in evaluating the applicability and performance of nanorod motors and optically heated nanoparticles for a variety of applications.

I

t is extremely challenging to drive and study nanoscopic processes occurring at and around the surface of a single nanoparticle and the difficulty increases further when the nanoparticle diffuses in solution. The most commonly used approach in this situation is to use optical tweezers, which offer the possibility to trap a single particle using optical gradient forces induced by a tightly focused laser beam.1 The possibility to exert torque on a trapped particle using circularly polarized laser tweezers provides an additional degree of freedom in optomechanical studies.2,3 Laser tweezers can be used to manipulate and analyze dielectric, metallic and biological particles and they have enabled a multitude of intricate studies in, for example, bio and soft matter physics.4,5 Many of these studies are based on analyzing the positional or orientational fluctuations of the trapped object, which allows for measurements of exceedingly weak forces and small distances.6-8 Plasmonic nanostructures, such as gold nanoparticles, interact strongly with optical fields and this has led to a wealth of studies and applications in areas including sensing,9 medicine,10,11 data storage12 and solar harvesting13. In 1994, Svoboda and Block demonstrated that gold colloids could be trapped and manipulated in solution by laser tweezers.14 The combination of optical tweezing and plasmonics has since developed into a new subfield, generating a range of interesting findings.6,15-22 One of the many useful aspects of plasmonic nanoparticles is that they offer several possibilities for spectroscopic analysis of the environment. One example is

localized surface plasmon resonance (LSPR) sensing, which utilizes the extremely high sensitivity of the surface plasmon resonance to changes in the refractive index of the medium close to the metal surface. We recently showed that single crystalline gold nanorods, one of the most important classes of plasmonic nanostructures, could be trapped and rotated at record rotation frequencies in water by circularly polarized laser tweezers.23 The possibility to utilize such optically driven nanomotors for molecular analysis was demonstrated as a proof of principle and several other biomedical and nanofluidic applications were suggested. The high rotation frequencies observed, reaching several tens of kHz, are possible because of the combined effect of a small particle size, which decreases friction, and an amplification of the scattering component of the optical torque for laser wavelengths near the LSPR.23 However, residual ohmic losses in the particles are unavoidable near resonance and their advantageous properties therefore come at the price of significant heat generation. This can in turn affect molecular adlayers, result in reshaping of the trapped particle and even induce bubble formation around the particle surface. Photothermal effects can be problematic in some applications, for example in delicate biomolecular experiments, but they can also be put to an advantage, such as in photothermal imaging,24 plasmonic photothermal therapy25 and nanosurgery.26 In this work, we examine photothermal processes affecting single gold nanorods that are trapped and actively rotated

1

ACS Paragon Plus Environment

ACS Nano

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

using optical tweezers. We first briefly describe the sample characteristics and the experimental methodology, which is based on simultaneous plasmon spectroscopy and rotational Brownian motion analysis. Next, we show how these techniques can be used to reveal and quantify miniscule laser induced changes in nanorod shape as a function of time. We then utilize the refractive index sensitivity of the LSPR and the viscosity sensitivity of the Brownian fluctuations, respectively, to establish two separate particle temperature estimates, which we record as functions of laser power. Finally, the meaning of these two temperatures is discussed in relation to the calculated spatial temperature profile around photothermally heated gold nanorods. We conclude by summarizing the work and by pointing out some possible directions for further research.

RESULTS AND DISCUSSION Probing photothermal processes via two separate channels. Figure 1 illustrates the sample characteristics and the experimental methods used. We focus on two types of single crystalline gold nanorods, hereafter referred to as “small” and “large”, both prepared via seed mediated growth.23 The particle dimensions were chosen such that their plasmon spectra overlapped the trapping laser wavelength at 660 nm in order to induce strong photothermal interactions and optical torques. The morphological and spectroscopic differences between the two batches are illustrated by the SEM images and ensemble averaged extinction spectra shown in Figure 1a and 1b, respectively. The lengths and diameters of the “small” (“large”) nanorods were 108±7 (155±9) nm and 65±5 (88±5) nm, corresponding to aspect ratios (ARs) of ~1.66 (~1.76). The spectroscopic and optomechanical properties of a trapped nanorod can be used to access a considerable amount of information about several particle-specific photothermal processes. In order to probe these on the single particle level, an experiment was constructed around two-dimensional (2D) optical tweezers interfaced to a grating spectrometer and a photon correlation system (Figure 1c and Figure S1). The setup and analysis routines allow for continuous and parallel measurements of dark-field (DF) scattering spectra and autocorrelation functions () versus laser power (see Methods for details). A gold nanorod supports two electric dipolar LSPRs, a longitudinal mode at long wavelengths and a two-fold degenerate transverse mode at shorter wavelengths. The mode splitting monotonically increases as a function of increasing aspect ratio. The resonance wavelengths can be accessed through a bi-Lorentzian curve fit of DF scattering spectra (Figure 1d). The LSPRs provide information on the dimensions of the nanorods and of the (temperature dependent) refractive index (RI, ) of the surrounding medium. The laser light backscattered from a trapped and rotating nanorod can be used to determine its average rotation frequency  and rotational decay time  due to Brownian fluctuations (Figure 1e). These parameters can be extracted by polarized photon correlation spectroscopy and curve fitting to a theoretical autocorrelation function (see Methods). The rotation parameters are determined by the optical torque  ,

Page 2 of 13

the rotational friction coefficient  , and the temperature  according to:  ()=

 ()=



 ()

 ()  

,

(1)

For a nanorod with length , we have  () = !"()# $ , where "() is the temperature dependent dynamical viscosity of the surrounding water and # is a geometrical factor depending on the nanorod eccentricity.23,27 (2).

Continuous reshaping at constant trapping power. The nanomotors were first operated at constant power for extended periods of time. Figure 2 illustrates typical behaviors observed for one exemplary nanorod of each type. During 30 minutes of continuous trapping, the “small” particle exhibits a pronounced decrease in rotation frequency, from ~16.5 to ~14.5 kHz, whereas the “large” particle rotated with a constant, or even slightly increasing, frequency of ~14.5 kHz (Figure 2a). This indicates that the focused laser field induces structural changes in the “small” but not in the “large” nanorod. However, the recorded plasmon spectra reveal clear changes due to photothermal reshaping for both nanorod types (Figure 2b). Specifically, the long-wavelength longitudinal LSPR peaks continuously blue-shifted by a few nm during the time course of the experiment, indicating a slow decrease in aspect ratio towards a more spherical shape, for both the “small” and the “large” rod (Figure 2c,d). The transverse LSPR of the “small” nanorod exhibits a concomitant red-shift as is expected for a decreasing aspect ratio (Figure S2a). We performed electrodynamic simulations of changes in spectral and rotational properties to explain this behavior (see Methods for details). As seen in Figure 2e, the driving optical torque  at the trapping wavelength decreases significantly if the aspect ratio of the “small” nanorod is slightly decreased (by ~3%, keeping the volume constant). This will result in a decreased  according to eq 1. In contrast, the torque driving the “large” nanorod is essentially constant for a similar change in particle shape, in spite of an even larger spectral shift towards shorter wavelengths (Figure 2f, see Figure S3 for calculated scattering spectra). Figure 2e and 2f shows the corresponding changes in absorption spectra caused by the decrease in aspect ratio. The absorption cross-section %&' at 660 nm obviously decreases for the “small” particle due to the reshaping. This will result in a lower surface temperature, which will in turn increase the viscosity and rotational friction coefficient around the particle, thus further reducing the rotation frequency. The “large” particle, in contrast, shows a minor increase in absorption at 660 nm, which will thus tend to increase its rotation frequency over time. The simulations hence qualitatively explain the contrasting experimental behavior observed for the two types of nanorods and they illustrate the rather complex relation between particle shape, the optical torque spectrum and the photothermal properties of gold nanorods.13 Quantification of photothermal reshaping. Photothermal reshaping of gold nanorods can be understood as a diffusion process where gold atoms primarily migrate from high to low curvature areas, driving the particle towards a

ACS Paragon Plus Environment

2

Page 3 of 13

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

ACS Nano

thermodynamically stable spherical shape.28 Previous studies of this process have mainly concerned immobilized particles subject to ultrafast laser pulses,12,28-33 and there are few studies based on CW laser illumination.31,34,35 Thermal reshaping of optically trapped gold nanorods have only been mentioned in passing (e.g. in SI of ref. 6). It is therefore interesting to quantify the reshaping in terms of actual number of atoms involved. To do this, we first need to connect the measured LSPR shifts to actual changes in aspect ratio. We used anisotropic etching36 to shrink the length of ensembles of particles while monitoring the longitudinal LSPR and aspect ratio changes through extinction spectroscopy and SEM analysis, respectively. The LSPR shifts observed in this experiment were similar to those seen in the single nanoparticle trapping experiments. We obtained values for d )*+,- ⁄dAR of 148 nm and 221 nm for the “small” and the “large” particle types, respectively, in excellent agreement with FDTD simulations of particles with constant volume and decreasing aspect ratio (Figure S4). By approximating the nanorod shape with a hemispherically capped cylinder (capsule), we then estimated the absolute change in nanorod dimensions during the time course of an experiment (Supporting Information Discussion S1, Figure S5). This analysis shows that the nanorods shrink in length by only a few Å during the 30 min experiments shown in Figure 2. The corresponding migration rate amounts to a few hundred atoms per minute (Supporting Information Table S1). We find it quite remarkable that such modest structural changes can be resolved for single optically trapped nanoparticles. Reversible photothermal heating effects. We next tried to separate the slow irreversible reshaping process from reversible heating effects caused by an increase in temperature of the water surrounding the trapped particle. We applied a saw-tooth laser power pattern with linearly increasing amplitude while tracking the DF spectra and rotational dynamics of the trapped particles. Figure 3 summarizes data for one exemplary particle of each kind, “small” and “large”. It is obvious that the saw-tooth laser pattern is essentially reproduced in the three extracted parameters, the longitudinal LSPR peak position ()*+,- ), the rotation frequency () and the autocorrelation decay time ( ), for both particle types (see shaded area in each panel). Because of the high thermal conductivity of Au, the particle temperature equilibrate within nanoseconds of changing the laser power.37 An increase in laser intensity will thus, almost instantaneously, increase the temperature of the water surrounding the particle surface and decrease its refractive index n. Since d )*+,- ⁄d > 0, this qualitatively explains the observed negative correlation between )*+,- and laser power seen in Figure 3 (see Figure S2b for data on the “small” nanorod’s transverse LSPR mode). Further, according to eqs 1 and 2, an increased laser power will increase the optical driving torque, hence the positive correlation between  and P, and the concomitant temperature rise will decrease the viscosity of the interfacial water, resulting in the observed decrease in decay time  . Nevertheless, as in the constant power case, there is a longterm trend caused by a continuous reshaping of the nanorods. Reshaping clearly occurs with a more or less constant rate throughout the experiment in the “small” particle

case (Figure 3a) and up until the last power ramp for the “large” particle (Figure 3b). Estimates of the corresponding shape changes and atomic migration rates can be found in the Supporting Information (Discussion S1 and Table S1). However, the final power ramp for the “large” nanorod induces an irreversible blue-shift that is even more pronounced (~4nm). It is also possible to discern a rather abrupt change in slope of the )*+,- (3) trace at a particular threshold power 45 ≈ 22 mW. The same effect was observed for several “large” nanorods at similar laser powers. We tentatively interpret this point as the onset of vapor formation around the trapped particle, since this would abruptly decrease the RI sensed by the particle, leading to a more pronounced LSPR blue-shift. Vapor formation would at the same time reduce the heat conduction away from the nanorod, which could explain the accelerated reshaping and the considerable irreversible peak shift. We return to the question of possible vapor formation at the end of the article. Two separate temperature estimations. We now try to estimate the absolute temperature near the nanorods surface based on the Brownian dynamics and spectroscopic measurements in Figure 3. As described in ref. 23 and 38, the effective rotational Brownian temperature  is most easily obtained from the decay time  because this parameter, in contrast to  , does not explicitly depend on the optical torque. We thus first set the nanorods lengths L and shape parameters g equal to average values obtained from SEM analysis of “small” and “large” particles.  can then be obtained from eq 2 by inserting an analytical expression for the viscosity of water "().38,39 A more in-depth analysis of temperature estimates from rotational dynamics, including stochastic simulations and a discussion of the role of translational diffusion can be found in ref. 40. The spectroscopic data can be translated to a corresponding temperature, which we call -6 , in a similar fashion. We first estimate the reversible LSPR shifts 7)*+,- induced by the RI change, by fitting and subtracting the linear trend from the spectral shifts in Figure 3 measured at the lowest laser power (Figure S6a). This procedure hence effectively removes the contribution from reshaping (Figure S6b). We here assume that no adsorption or desorption of stabilizing ligand molecules occur during the experiment, which is reasonable considering the highly diluted colloid solution used in the experiments. To proceed, we then utilize the chain rule 89 89 8> ( :;