Article pubs.acs.org/Langmuir
Picosecond-to-Nanosecond Dynamics of Plasmonic Nanobubbles from Pump−Probe Spectral Measurements of Aqueous Colloidal Gold Nanoparticles Tetsuro Katayama,‡,§ Kenji Setoura,† Daniel Werner,† Hiroshi Miyasaka,*,‡ and Shuichi Hashimoto*,† †
Department of Optical Science and Technology, The University of Tokushima, Tokushima 770-8506, Japan Division of Frontier Materials Science and Institute for NanoScience Design, Osaka University, Toyonaka, Osaka 560-8531, Japan § PRESTO, Japan Science and Technology Agency, Kawaguchi 332-0012, Japan ‡
S Supporting Information *
ABSTRACT: The photothermal generation of nanoscale vapor bubbles around noble metal nanoparticles is of significant interest, not only in understanding the underlying mechanisms responsible for photothermal effects, but also to optimize photothermal effects in applications such as photothermal cancer therapies. Here, we describe the dynamics in the 400−900 nm regime of the formation and evolution of nanobubbles around colloidal gold nanoparticles using picosecond pump−probe optical measurements. From excitations of 20−150 nm colloidal gold nanoparticles with a 355 nm, 15 ps laser, time-dependent optical extinction signals corresponding to nanobubble formation were recorded. The extinction spectra associated with nanobubbles of different diameters were simulated by considering a concentric spherical core−shell model within the Mie theory framework. In the simulations, we assumed an increase in particle temperature. From temporal changes in the experimental data of transient extinctions, we estimated the temporal evolution of the nanobubble diameter. Corrections to bubble-free temperature effects on the transient extinction decays were applied in these experiments by suppressing bubble formation using pressures as high as 60 MPa. The results of this study suggest that the nanobubbles generated around a 60 nm-diameter gold nanoparticle using a fluence of 5.2 mJ cm−2 had a maximum diameter of 260 ± 40 nm, and a lifetime of approximately 10 ns. The combination of fast transient extinction spectral measurements and spectral simulations provides insights into plasmonic nanobubble dynamics.
■
INTRODUCTION Gold nanoparticles (Au NPs) exhibit various fascinating phenomena via their interactions with visible light, including the plasmonic enhancement of incident electromagnetic fields.1,2 This is because visible light excites the localized surface plasmon resonance (LSPR) band, leading to extremely efficient extinctions, that is, the absorption and scattering of incident light. The absorption of light also produces increased particle temperatures as a result of the LSPR decay, in which excited electrons couple with phonons within these NPs. Their photothermal response has attracted increasing interest because of their resulting phase changes and nanoscale energy depositions to the surroundings. For example, when pulsed lasers were used to illuminate Au NPs, particle expansion, melting, evaporation, and fragmentation were observed.3−5 As for the effects on surrounding medium, local heating, acoustic emission, and vapor-bubble generation have been observed.3−5 In particular, UV and visible laser-induced heating of Au NPs in aqueous solutions causes superheating of the water medium at the particle−medium interface, resulting in explosive vaporbubble generation around the particles.6,7 The transient expansion and collapse of these bubbles is relevant in photothermal therapy of malignant tissues by killing nearby cells.8,9 © 2014 American Chemical Society
Neumann and Brinkmann performed the initial studies on particle heating-induced bubbles, produced using pulsed-laser excitation.10−12 They investigated the nucleation and initial expansion of a bubble around a single micrometer-sized absorber following microsecond and nanosecond laser irradiation. Transient microbubbles on the particle surface were directly imaged and simultaneously detected by measuring a transient decrease in the transmitted probe laser intensity. Using dark-field microscopy,13−15 Lapotko and coworkers performed imaging and optical detection of nanosecond pulsed laser-induced nanobubbles generated from single and clustered Au NPs.16−18 For example, they measured the optical extinction signal of a bubble that was generated by a single 90 nm diameter Au NP after exciting the LSPR band at a monitoring wavelength of 633 nm over a nanosecond time scale.16 A threshold fluence of 360 mJ cm−2 was observed for the excitation produced by 0.5 ns, 532 nm laser light. Using femtosecond laser excitation (400 nm, 100 fs) and timeresolved X-ray scattering detection, the Plech group observed in an ensemble study that 9 nm-diameter Au NPs produce vapor Received: October 24, 2013 Revised: June 25, 2014 Published: July 23, 2014 9504
dx.doi.org/10.1021/la500663x | Langmuir 2014, 30, 9504−9513
Langmuir
■
bubbles that grow and decay within a period of less than a nanosecond.6 They used fluences in the 30−75 mJ cm−2 range because the bubble generation threshold was found to be 29 mJ cm−2. A maximum bubble diameter of 20 nm was estimated after 300 ps of excitation. Single-particle studies performed using microscopy have the potential to capture the geometry of microbubbles directly. However, imaging nanosized objects using optical microscopy is difficult, because of the optical diffraction limit. Despite this, the remarkable light-scattering abilities of Au NPs enabled the optical detection of single plasmonic nanobubbles.16−18 Potentially, this also enables the characterization of timeresolved particle heating and related phenomena to be performed without the inaccuracy produced by the ensemble averaging of particle-to-particle variations that is inherent in the measurement of colloidal solutions with a distribution of particle size. Nevertheless, studies of a single particle−single laser shot revealed that the detection of single nanobubbles suffers limited detection sensitivity; fluences 1 order of magnitude larger would have been required for the determination of the bubble generation thresholds.16−18 Another disadvantage is that differences between the individual particles affect the results from observations of any type of single-particle event, and many particles must therefore be examined to obtain a statistically meaningful conclusion. Ensemble studies using colloidal solutions demonstrated superb detection sensitivity and high-quality nanobubble signals, revealing averaged bubble dynamics, although only indirect evidence of nanobubbles was obtained.6,7,19,20 Although time-resolved X-ray scattering is excellent in revealing nanoscale structural dynamics, large-scale instrumentation and complicated data analysis are necessary. In contrast, optical spectroscopy uses instrumentation that is readily available, and the data analysis required is simple and straightforward. The Plech group performed a transientextinction study using four monitoring wavelengths (405, 488, 635, and 660 nm), and nanosecond laser excitations with 80 nm-diameter colloidal Au NPs (at fluences of 30 and 45 mJ cm−2).19 Previous experiments using optical detection, both single-particle studies and ensemble measurements alike, produced results with limited time resolution. Because of this, the early events in the nanobubble dynamics are obscured. Additionally, spectral information on these nanobubbles is still lacking because of the limited range of available detection wavelengths. Optical detection using extinction signals could be more useful if the time resolution could be further increased to several tens of picoseconds, and could also provide spectral information. We performed an optical extinction study to reveal nanobubble dynamics, using a picosecond pump−probe method. Although previous nanosecond experiments identified bubble growth and collapse from the rise and decay of the optical scattering signal, improved time resolution could provide a more detailed temporal profile. Such a profile could fill the gap between previous femtosecond-laser-pump picosecond-X-ray-probe studies and optical studies with microsecond time resolution. We used an ensemble method rather than performing a single-particle study, because of the superior detection sensitivity of the former. However, it would be desirable to extend this method for single-particle measurements, to obtain single-particle accuracy, rather than ensembleaveraged figures.
Article
MATERIALS AND METHODS
Samples. Aqueous solutions of Au NPs with nominal diameters of 20 nm (cat. no. EMGC20, 19 ± 3 nm measured using transmission electron microscopy (TEM)), 60 nm (cat. no. EMGC60, 58 ± 5 nm measured using TEM), 100 nm (cat. no. EMGC100, 100 ± 8 nm measured using TEM), and 150 nm (cat. no. EMGC150, 155 ± 18 nm measured using TEM) were purchased from BBI Solutions (Cardiff, UK). These particles were synthesized using a variation of the Frens citrate reduction method, and were stabilized with citrate.21 TEM images and corresponding histograms of the size distributions are given in the Supporting Information, Figure S1. Five milliliters of the sample solutions (particle concentrations of 7.0 × 1011 mL−1 (20 nm), 2.6 × 1010 mL−1 (60 nm), 5.6 × 109 mL−1 (100 nm), and 1.7 × 109 mL−1 (150 nm)) were contained in a quartz cuvette (1 cm × 1 cm × 5 cm) with an optical path length of 1 cm. The extinction spectra of these particles are given in the Supporting Information, Figure S2. The sample solution was stirred magnetically during the laser experiments and was replaced every 500 shots. Instrumentation. A picosecond pump−probe system with a custom-built mode-locked Nd3+:YAG laser was used to measure the transient spectra at various time delays in the picosecond-tonanosecond time regime.22 Details of the experimental setup are given in Supporting Information, S3. Briefly, third-harmonic light (355 nm) with an fwhm of 15 ps was used as a pump pulse, and was focused using an f = 200 mm lens into a spot of diameter 2.0 mm. Under these experimental conditions, the laser light formed an approximately parallel beam. The irradiation with 355 nm wavelength light (power fluctuation 600 nm increased being correlated with decreases in extinction (or increases in bleaching) at the LSPR peak position. These results indicated that the transient species that induced the scattering were selectively inhibited under the high pressure, and from the observed effects of the pressure, these scattering species could be ascribed to the bubbles produced around the Au NPs. Note that bubble formation is distinct from acoustic phonon vibrations that occur on time scales of less than 300 ps and at a much lower fluence than that required for bubble generation.27 One additional point we noted is that at a fluence of 5.2 mJ cm−2 the bleaching signal at the LSPR peak position persisted over the optical delay range of our pump−probe instrument. To explain this phenomenon, we measured the transient extinction signal at 532 nm as a function of time, using a continuous-wave laser beam as a probe light, and a photodiode (rise time
