Two-Color Laser Printing of Individual Gold Nanorods - Nano Letters

Aug 8, 2013 - The nonresonant laser is used to trap and align individual gold nanorods in two dimensions close to a glass coverslip surface. .... Thus...
0 downloads 22 Views 2MB Size
Letter pubs.acs.org/NanoLett

Two-Color Laser Printing of Individual Gold Nanorods Jaekwon Do, Michael Fedoruk, Frank Jac̈ kel,* and Jochen Feldmann* Photonics and Optoelectronics Group, Department of Physics and Center for NanoScience (CeNS), Ludwig-Maximilians-Universität München, Amalienstr. 54, 80799 Munich, Germany S Supporting Information *

ABSTRACT: We report on the deposition of individual gold nanorods from an optical trap using two different laser wavelengths. Laser light, not being resonant to the plasmon resonances of the nanorods, is used for stable trapping and in situ alignment of individual nanorods. Laser light, being resonant to the transversal mode of the nanorods, is used for depositing nanorods at desired locations. The power and polarization dependence of the process is investigated and discussed in terms of force balances between gradient and scattering forces, plasmonic heating, and rotational diffusion of the nanorods. This two-color approach enables faster printing than its one-color equivalent and provides control over the angular orientation (±16°) and location of the deposited nanorods at the single-nanorod level. KEYWORDS: Gold nanoparticles, optical manipulation, optical trapping, optical tweezers, localized surface plasmon resonance, nanofabrication

A

controlling the nanoparticle orientation14,15,20 but is slow since it requires additional steps (i.e., mechanical vertical motion and chemical modification of substrates) for fixing the particles. Here, we demonstrate two-color laser printing of individual gold nanorods from an optical trap using a nonresonant and a resonant laser simultaneously. The nonresonant laser is used to trap and align individual gold nanorods in two dimensions close to a glass coverslip surface. A second laser, resonant to the transversal plasmon mode of the trapped nanorod, is used for rapid printing by simply opening a shutter once the nanorod is trapped and aligned. Our experiment is illustrated in Figure 1 (full experimental details are given in the Supporting Information). We employ polyethylene glycol (PEG) capped gold nanorods with an average length of 114 nm and an average diameter of 41 nm (aspect ratio ∼2.8) as depicted in Figure 1a. [Comparable results were obtained with nanorods of other dimensions, i.e. average diameter/length (in nm): 40/97, 50/199, 36/162, and longitudinal plasmon resonances at 670 nm/800 nm/850 nm, respectively (data not shown).] The PEG functionalized nanorod exhibit a small negative ζ-potential of −7 mV. The nanorods exhibit a transversal plasmon mode at ∼518 nm and a longitudinal plasmon mode at 724 nm (Figure 1b). Individual gold nanorods are trapped in two dimensions over a glass coverslip using a tightly focused linearly polarized 1064 nm laser as illustrated in Figure 1c. The nanorods align with their long axis parallel to the polarization direction of the laser light

dvanced colloidal chemistry allows preparing macroscopic numbers of both semiconductor and metal nanoparticles with a controlled size, shape, and composition.1−3 This control enables the tuning of the nanoparticles’ properties including but not limited to electronic states, optical resonances, magnetic properties, and catalytic activities.4,5 In particular, complexshaped nanoparticles such as nanowires, -rods, -stars, or cubes are attractive for applications such as optoelectronic devices and sensors due to their anisotropic responses.6−9 Many of these applications require the precise deposition of the nanoparticles onto substrates with a certain orientation in order to make full use of these anisotropies. Optical tweezers, or optical traps, on the other hand, have emerged as versatile tools for studying and manipulating nanoparticles in solution using optical forces.10,11 In particular, complex-shaped nanoparticles such as nanowires or nanorods can be aligned or rotated using well-defined polarizations or structured optical fields for the trap.12−16 The alignment of nanowires and nanorods parallel to the trapping laser polarization is due to their anisotropic, quasi-onedimensional polarizability. Recently, optical forces have been used to print metallic nanoparticles including nanospheres and nanorods onto surfaces.17−21These optical printing methods, in contrast to many other patterning methods, enable the deposition of individual nanoparticles and do not require (chemical) patterning of the surface.22−24 However, if only one color (laser wavelength) is used for the printing, it can be either resonant or nonresonant with the localized surface plasmon resonances of the nanoparticles. The resonant case allows for rapid printing but has limited to no control over the nanoparticle orientation, while the nonresonant printing allows © XXXX American Chemical Society

Received: May 15, 2013 Revised: July 2, 2013

A

dx.doi.org/10.1021/nl401788w | Nano Lett. XXXX, XXX, XXX−XXX

Nano Letters

Letter

used to rapidly print the nanorod onto the surface maintaining the orientation given by the trapping laser polarization (Figure 1d). The polarization of the two laser beams is independently controlled by two λ/2 plates. Simultaneously, the setup allows monitoring the gold nanorods in real time via their white light Rayleigh scattering under dark-field illumination using a consumer digital camera. Scattering and absorption cross sections, particle temperatures induced by plasmonic heating, and optical forces were determined using commercial software (Lumerical Solutions, COMSOL, Mathematica; for full details see Supporting Information). Figure 2 displays our key result: Following the procedure outlined above an OX pattern consisting of perpendicularly oriented gold nanorods is printed on a glass coverslip demonstrating the control over the orientation and position using the rapid two-color laser printing. The printing process, after the nanorod has diffused into the trapping region, takes less than 1 s, which is faster than the reported several seconds in nonresonant one-color printing techniques, where additional surface modification was necessary.20 The laser polarizations were oriented perpendicularly, with the trapping laser determining the orientation of the printed nanorod. Powers were 33 mW and 4.2 mW for the trapping and printing laser, respectively. Figure 2a and b display zoom-in SEM micrographs of the resulting pattern. Clearly, individual nanorods can be recognized with the orientation of the nanorods in the “O” being perpendicular to the nanorods in the “X”. Figure 2c−e displays dark-field white light Rayleigh scattering images of the same structure with and without polarizer in front of the camera. Without polarizer both structures appear in the same color (Figure 2c). If the polarizer direction is aligned along the long axis of the nanorods in one of the two letters, this letter appears red, while the other one appears green (Figure 2d and e). This is due to the dominant scattering contributions of the longitudinal and transversal plasmon mode in the two polarization directions, respectively. It should be noted that several gold nanorods can be printed subsequently to a substrate reproducibly to form larger ordered arrays (see Supporting Information for more details). Using laser radiation that is resonant to the plasmon resonance for printing of gold nanoparticles has previously been shown in a one-color approach to offer control of the particle position beyond the diffraction limit, i.e. sub 100 nm precision.17 A similar precision of ∼50 nm for the particle

Figure 1. (a) Transmission electron microscopy (TEM) image of PEGylated gold nanorods. Inset: TEM image of CTAB coated gold nanorods. Scale bars are 100 nm. (b) Normalized extinction spectra of PEGylated gold nanorods. (c) Schematic illustration of two-color laser printing setup. (d) Cartoon representation of the two-color printing process. First, a nanorod diffusing into the focus of the trapping laser is trapped. Second, the trapped nanorod is aligned along the direction of the trapping laser polarization. Finally, the nanorod is printed by using the scattering force exerted by the collinear printing laser.

due to the larger polarizability of the nanorods at that wavelength compared to the perpendicular orientation.25 After aligning the nanorod in the trap and positioning it at a desired location, a second laser (532 nm) is switched on and

Figure 2. (a,b) Zoom-in SEM images of an “OX” pattern of printed gold nanorods. The nominal orientation of the nanorods in the two letters is perpendicular. (c−e) Dark-field white light Rayleigh scattering images of the same pattern as in (a,b). The direction of the linear polarizer in front of the detector is indicated. Clearly, the two letters exhibit an opposite anisotropic response. Scale bar is 1 and 2 um for (a,b) and (c,d,e), respectively. B

dx.doi.org/10.1021/nl401788w | Nano Lett. XXXX, XXX, XXX−XXX

Nano Letters

Letter

Figure 3. (a−h) Frequency distribution of the angular deviation of the printed nanorods from their nominal orientation. The relative orientation of the trapping and printing laser polarization direction and their powers at the sample are indicated. (i) Cartoon representation of the enhanced nanorod substrate interaction due to a collapsed ligand shell.

precision of ±16° was achieved with perpendicularly oriented printing and trapping lasers with powers of 8.3 mW and 27 mW at the sample, respectively. The above observations demonstrate that there is a subtle interplay of trapping and printing laser powers with respect to the angular printing precision. In particular, it appears that simply increasing both laser powers independently does not result in better precision. In the following we discuss the printing process in detail. A focused laser beam exerts two types of forces on a plasmonic nanoparticle: a gradient force and a scattering force. If the laser is not resonant to any (plasmon) resonances the resulting force is dominated by the gradient force which pulls the particle into the region of highest intensities and thus allows for stable trapping of the particle. Under resonant conditions, however, the resulting scattering force points mainly along the direction of beam propagation and propels the particle forward. In the present case, the maximum trapping force (total gradient force of both trapping and printing laser) is ∼1000 fN for trapping and printing laser power of 15 mW and 4.2 mW at the sample, respectively (see Supporting Information for details). To print the particle with our two-color approach the scattering force needs to overcome the trapping force. With the same laser powers as above, the maximum scattering force of both printing and trapping laser combined is∼2650 fN. This is clearly larger than the maximum total trapping force and will allow for printing the particle. It should be noted, however, that this force comparison is only semiquantitative since, first, it does not include spatial or rotational diffusion of the nanorod nor its

position is achieved here (see Supporting Information for details). Interestingly, trying to print two nanorods subsequently at the same position results in the second nanorod being printed about 350 nm away from the first (data not shown). In the two-color printing technique, however, in addition to the location of the nanorod, the precision with which the nanorod’s orientation can be controlled is of interest. In order to gain more insight into the printing process and its precision, we varied both the intensities and the relative polarization direction of the trapping and printing lasers. Figure 3 displays histograms of the angular deviation of the printed nanorod’s orientation from the nominal orientation. From Figure 3a−c it can be seen that the lowest deviation is achieved if trapping and printing laser are polarized perpendicularly to each other. The deviation for parallel orientation of the two laser polarizations is comparable to one-color resonant printing of the nanorods. Figure 3d−h provide evidence for the importance of the relative laser intensities employed. It appears that for given printing laser power there is an intermediate trapping laser intensity that results in highest precision (Figure 3d−f). Similarly, from the comparison of Figure 3e, g, and h it can be concluded that lower printing laser powers lead to higher precision for a given trapping power. However, the printing laser power needs to overcome a certain (trapping laser power-dependent) threshold to successfully print the nanorods. For instance, at a trapping power of 15 mW a printing power of at least 4.2 mW was required for successful printing. In the investigated range of parameters the highest C

dx.doi.org/10.1021/nl401788w | Nano Lett. XXXX, XXX, XXX−XXX

Nano Letters

Letter

finite size, and, second, the maximum values for the combined trapping and scattering forces do not occur at the same spatial positions in the trap (see Supporting Information for details). The requirement of perpendicular orientation of the trapping and printing laser polarization directions can be understood considering the polarizability of the nanorod. Due to its larger polarizability along its long axis the near-infrared trapping laser aligns the nanorod parallel to its polarization direction.25 However, the printing laser being resonant to the transverse plasmon mode will have a stronger polarizing effect on the nanorod perpendicular to its long axis. Thus, upon parallel orientation of the two laser polarizations the linearly polarized printing laser can induce either a left or right-handed torque on the prealigned nanorod during the printing. This will reduce the final angular printing precision. Consequently, optimal angular printing precision will be achieved with perpendicular orientation of the two laser polarizations. We now turn to the discussion of the observed dependencies of the angular printing precision on trapping and printing laser powers, respectively. The angular printing precision will be determined by Brownian motion due to rotational diffusion of the nanorod both in the trap and during printing. Strong rotational diffusion will lead to a low precision, while more restricted rotational diffusion will result in more precise printing. As outlined above, there is a subtle interplay of trapping and printing power with respect to angular printing precision. We therefore discuss the dependence of the printing precision on each laser power separately with keeping the other one fixed. We first consider the dependence on trapping power for a given (fixed) printing power. At low trapping powers the trap will exhibit the lowest stiffness which results in more pronounced rotational diffusion of the nanorod. With increasing trapping power the trap becomes stiffer, and the rotational diffusion will be more limited. Thus increasing trapping power will initially result in more precise printing. However, when the trapping forces become comparable to the printing (i.e., scattering) forces, the acceleration of the nanorod toward the surface during the printing step becomes smaller. Thus there will be more time for rotational diffusion of the nanorod reducing the angular printing precision. Consequently, there will be an optimal trapping force balancing the counteracting effects of rotational diffusion in the trap and during the printing. The latter argument is also supported by the calculated combined (i.e., trapping and printing laser) maximum scattering and gradient forces of the two beams. For instance, for a fixed printing power of 4.2 mW, the ratio of maximum combined scattering force to maximum combined trapping (i.e., gradient) force decreases from 2.6 via 1.8 to 1.4 when the trapping power is increased from 15 mW via 33 mW to 52 mW. To understand the printing power dependence of the angular printing precision for a given trapping power, plasmonic heating, Brownian motion due to rotational diffusion, and printing times need to be considered. At typical laser powers of 33 mW and 4.2 mW for the trapping and printing laser the surface temperature of the particle will rise by 57 K and 147 K, respectively. When both lasers are present the rise in temperature amounts to 204 K. This has major impact on the dynamics of the printing process. However, increasing the printing laser power will also reduce the printing time. Assuming a constant printing force and a linear dependence of the scattering force on laser power P, the printing time will scale as P−1/2. The rotational diffusion coefficient on the other

hand scales linearly with temperature and thus will linearly increase with laser power due to plasmonic heating. Thus the angular diffusion of the nanorod is expected to increase during the printing process with printing laser power despite shorter printing times at higher powers. Thus a reduced angular precision is expected at higher powers. [The temperature/ power dependency of the viscosity has been neglected in this argument since it enters linearly both into the printing force and diffusion coefficient via the expressions for the Stokes and frictional rotational drag, respectively. Including viscosity would thus only amplify the described power dependency.] It should be noted that the reduced angular precision could in part, and in addition to the above argument, be attributed to changes in fluid flows due to the plasmonic heating induced temperature gradients. However, such contribution cannot be quantified directly from our experiment. It should be noted, that the increased particle temperature due to plasmonic heating also facilitates the adhesion of the nanorods to the glass surface at the end of the printing process. The above estimates for the nanorod surface temperature exceed the temperatures for which collapse of the PEG ligand shell has been previously reported in similar systems.26 Indeed, partial removal of ligands at these temperatures appears feasible.27 A collapsed ligand shell will allow for a stronger interaction between the gold nanorod and the glass surface promoting a stronger adhesion as illustrated in Figure 3i. Despite this significant temperature increase no evidence for tip melting was seen in SEM images of printed particles, which may be due to the fact that the nanorods were only shortly exposed to both lasers. Finally, it should be noted that our twocolor laser printing technique can be combined with the recently reported use of spatial light modulators for parallel printing of many nanoparticles.18 Thus a further acceleration of the printing process with positional and orientational control seems feasible. In conclusion, we demonstrated two-color printing of individual gold nanorods onto a glass surface, which enables rapid printing of individual gold nanorods with control of location and orientation. A nonresonant laser is used for stable trapping, alignment, and positioning of individual gold nanorods, while a second laser resonant to the transverse plasmon mode of the gold nanorod is used to print the nanoparticle maintaining the orientation given by the trapping laser. The polarization and power dependence of the two-color printing method was investigated in detail and explained by a complex interplay of the force balance between gradient and scattering force, plasmonic heating of the nanorod, and its surrounding and rotational diffusion of the nanorod both in the trap and during the printing process.



ASSOCIATED CONTENT

S Supporting Information *

Full experimental details and additional experimental and computational results. This material is available free of charge via the Internet at http://pubs.acs.org.



AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected] and [email protected]. D

dx.doi.org/10.1021/nl401788w | Nano Lett. XXXX, XXX, XXX−XXX

Nano Letters

Letter

Present Address

(27) Ni, W.; Ba, H.; Lutich, A. A.; Jäckel, F.; Feldmann, J. Nano Lett. 2012, 12, 4647−4650.

F.J.: Department of Physics and Stephenson Institute for Renewable Energy, University of Liverpool, Oliver Lodge Building, Oxford Street, L69 7ZE Liverpool, United Kingdom. Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS Financial support from the European Research Council (ERC) through the Advanced Investigator Grant HYMEM, the European Union Commission through the Marie Curie Research Training Network ICARUS, and the Deutsche Forschungsgemeinschaft (German Research Foundation, DFG) through the Nanosystems Initiative Munich (NIM) is gratefully acknowledged.



REFERENCES

(1) Li, H.; Kanaras, A. G.; Manna, L. Acc. Chem. Res. 2013, DOI: 10.1021/ar3002409. (2) Hughes, B. K.; Luther, J. M.; Beard, M. C. ACS Nano 2012, 6, 4573−4579. (3) Sau, T. K.; Rogach, A. L. Adv. Mater. 2010, 22, 1781−1804. (4) Choi, C. L.; Alivisatos, A. P. Annu. Rev. Phys. Chem. 2010, 61, 369−389. (5) Sau, T. K.; Rogach, A. L.; Jäckel, F.; Klar, T. A.; Feldmann, J. Adv. Mater. 2010, 22, 1805−1825. (6) Cui, Y.; Wei, Q.; Park, H.; Lieber, C. M. Science 2001, 293, 1289−1292. (7) Cui, Y.; Lieber, C. Science 2001, 291, 851−853. (8) Hrelescu, C.; Sau, T. K.; Rogach, A. L.; Jäckel, F.; Laurent, G.; Douillard, L.; Charra, F. Nano Lett. 2011, 11, 402−407. (9) Murphy, C. J.; Sau, T. K.; Gole, A. M.; Orendorff, C. J.; Gao, J.; Gou, L.; Hunyadi, S. E.; Li, T. J. Phys. Chem. B 2005, 109, 13857− 13870. (10) Ashkin, A.; Dziedzic, J. M.; Bjorkholm, J. E.; Chu, S. Opt. Lett. 1986, 11, 288−290. (11) Dienerowitz, M.; Mazilu, M.; Dholakia, K. J. Nanophotonics 2008, 2, 021875. (12) Bonin, K.; Kourmanov, B.; Walker, T. Opt. Express 2002, 10, 984−989. (13) Pauzauskie, P. J.; Radenovic, A.; Trepagnier, E.; Shroff, H.; Yang, P.; Liphardt, J. Nat. Mater. 2006, 5, 97−101. (14) Pelton, M.; Liu, M.; Kim, H. Y.; Smith, G.; Guyot-Sionnest, P.; Scherer, N. F. Opt. Lett. 2006, 31, 2075−2077. (15) Tong, L.; Miljković, V. D.; Käll, M. Nano Lett. 2010, 10, 268− 273. (16) Yan, Z.; Sweet, J.; Jureller, J. E.; Guffey, M. J.; Pelton, M.; Scherer, N. F. ACS Nano 2012, 6, 8144−8155. (17) Urban, A. S.; Lutich, A. A.; Stefani, F. D.; Feldmann, J. Nano Lett. 2010, 10, 4794−4798. (18) Nedev, S.; Urban, A. S.; Lutich, A. A.; Feldmann, J. Nano Lett. 2011, 11, 5066−5070. (19) Ito, S.; Yoshikawa, H.; Masuhara, H. Appl. Phys. Lett. 2002, 80, 482−484. (20) Ling, L.; Guo, H.-L.; Zhong, X.-L.; Huang, L.; Li, J.-F.; Gan, L.; Li, Z.-Y. Nanotechnology 2012, 23, 215302. (21) Do, J.; Schreiber, R.; Lutich, A. A.; Liedl, T.; RodríguezFernández, J.; Feldmann, J. Nano Lett. 2012, 12, 5008−5013. (22) Nepal, D.; Onses, M. S.; Park, K.; Jespersen, M.; Thode, C. J.; Nealey, P. F.; Vaia, R. A. ACS Nano 2012, 6, 5693−5701. (23) Kraus, T.; Malaquin, L.; Schmid, H.; Riess, W.; Spencer, N. D.; Wolf, H. Nat. Nanotechnol. 2007, 2, 570−576. (24) Kuemin, C.; Nowack, L.; Bozano, L.; Spencer, N. D.; Wolf, H. Adv. Funct. Mater. 2012, 22, 702−708. (25) Selhuber-Unkel, C.; Zins, I.; Schubert, O.; Sönnichsen, C.; Oddershede, L. B. Nano Lett. 2008, 8, 2998−3003. (26) Boca, S. C.; Astilean, S. Nanotechnology 2010, 21, 235601. E

dx.doi.org/10.1021/nl401788w | Nano Lett. XXXX, XXX, XXX−XXX