Nanoscopic morphological effect on the optical properties of polymer-grafted gold polyhedra

Plasmonic nanoparticles show highly sensitive optical properties upon local dielectric environment changes. Hybridisation of plasmonic nanoparticles with active polymeric materials can allow stimuli-responsive and multiplex sensing over conventional monotonic sensing capacity. Such heterogeneous adlayers around the plasmonic core component, however, are likely to perturb the local refractive index in the nanometre regime and lead to uncertainty in its intrinsic sensitivity. Herein we prepare a series of polystyrene-grafted polyhedral gold nanoparticles, cubic and concave cubic cores, with different edge lengths and polymer thicknesses with precise synthesis control. Their localised surface plasmon resonance (LSPR) spectral changes are monitored to understand the effect of core morphological details in the interplay of nanoscale polymeric shells. Quantitative image analysis of changes in the core and shell shape contours and finite-difference time-domain simulations of the corresponding LSPR spectra and electric field distributions reveal that the magnitude of the LSPR spectral shift is closely dependent on the core morphology, polymer shell thickness and electric field intensity. We also demonstrate that the polystyrene-grafted gold concave cube displays higher sensitivity for nanoscale refractive index change in the polymer shell than the polystyrene-grafted gold cube at different temperatures. Our systematic investigation will help design polymer-composited plasmonic nanosensors for desirable applications.


Nanorods
Gold nanorods were synthesized by a seed-mediated growth method. 1,2 An aqueous solution of 10 mM HAuCl 4 (125 µL) was added to 5 mL of 100 mM cetyltrimethylammonium bromide (CTAB) in a 20 mL scintillation vial.
While the mixture was stirred at 1150 rpm, an aqueous solution of NaBH 4 (10 mM, 300 µL) cooled in an ice bath was quickly injected to the solution. The gold seed solution was stirred for 1 min and left undisturbed at 30 °C for 20 min. Then, aqueous solutions of HAuCl 4 (10 mM, 10 mL), AgNO 3 (10 mM, 1.8 mL) and ascorbic acid (100 mM, 1.14 mL) were added in sequence into 200 mL of 100 mM CTAB in a 250 mL Erlenmeyer flask at 30 °C.
The prepared seed solution (240 µL) was quickly injected to the mixture while stirred at 500 rpm, and left undisturbed at 30 °C for 2 h. Gold nanorods were collected after twice centrifugations (8000 rpm, 15 min) and dispersed in 5 mL of 50 mM CTAB (λ max = 756 nm).

Nanospheres
Uniform gold nanospheres were prepared according to literature method with modification. 1,2 Test etching reactions were conducted to determine an appropriate volume of 10 mM HAuCl 4 that would be required for S3 spherical shape. Different volumes of 10 mM HAuCl 4 (3 µL, 3.5 µL, 4 µL, 4.5 µL, 5 µL, 5.5 µL and 6 µL) were added to 500 µL of the gold nanorod solution (extinction = 2 at 756 nm) in a 2 mL microtube, respectively, and placed in a thermomixer at 40 ℃ under 300 rpm for 4 h. After etching, their UV-visible spectra were measured, and an optimum volume of 10 mM HAuCl 4 (4.1 µL) could be determined. A scale-up etching reaction was conducted as follows. The gold nanorod solution (4.6 mL) was mixed with 50 mM CTAB (83.57 mL) in a 250 mL Erlenmeyer flask at 40 ℃. An aqueous solution of 10 mM HAuCl 4 (718 µL) was added to the mixture, and it was stirred at 200 rpm for 4 h. Gold nanospheres were collected after twice centrifugations (11000 rpm, 45 min) and dispersed in 100 mM cetylpyridinium chloride monohydrate (CPC).
In order to improve shape uniformity, regrowth and etching processes were applied to the prepared gold nanospheres were collected through twice centrifugations (11000 rpm, 45 min) and dispersed in 100 mM CPC. In order to remove relatively large gold nanospheres, the product solution was centrifugated at 4000 rpm for 4 min and its supernatant was collected. The centrifugation was repeated four times to achieve better shape uniformity for the gold nanosphere.

Section 3. Polymer grafting
Polymer grafting on the polyhedral gold nanoparticles was performed according to a modified literature method. 2,3 First, the CTAB concentration in the gold nanoparticle solution prepared from Section 2.4 and 2.5 was reduced through centrifugation and redispersion in deionized water. Then, the gold nanoparticle solution was concentrated by removing supernatant as much as possible, leaving ~10 µL of the nanoparticle solution. The concentrated particle solution was mixed well with 40 µL of deionized water, which was quickly injected into 1.07 mL of the polystyrene (PS) solution (0.1 mg/mL in N,N-dimethylformamide (DMF)) in a 5 mL scintillation vial while sonicated. The mixture was sonicated for 1 min and left undisturbed for 30 min. The PS-grafted polyhedral gold nanoparticles were collected after six-time centrifugations and dispersed in DMF.

Section 4. Characterisation
A FEI Tecnai 12 transmission electron microscope with a LaB 6 emitter at 120 kV was used for the nanoparticle core and polymer shell characterisation. Specimen was prepared by dropping 10 µL of CTAB-deficient particle solution on a TEM grid, which was fully dried at room temperature for 3 h before the imaging. For PS-grafted nanoparticles, the particle solution in DMF (10 µL) was dropped on a TEM grid, and the droplet was gently wiped after 30 min. UV-visible spectra were obtained using a Genesys 10S UV-Vis spectrophotometer with a quartz cuvette (path length = 1 cm). Temperature-dependent UV-visible spectra were measured using a Thermo Scientific Evolution 600. At each temperature, sample solution was stabilized for 15 min before spectral acquisition. Dynamic light scattering was measured by a Malvern Zetasizer Nano ZS.

S5
Curvature analysis was conducted by ImageJ and our customized Matlab code. 2 A TEM image of a PS-grafted gold nanoparticle was treated via ImageJ. First, the image type was changed to 8 bit, to which Gaussian blur was applied by sigma value of 2.0 pixels. The boundary of the core particle was cut clearly through polygon selections.
Then, a core mask image was obtained using a threshold function. The mask image was further processed via our customized Matlab code, giving rise to curvature distribution and colour map.
Local thickness analysis was conducted by ImageJ. 2 A shell mask could be acquired by the same procedure used for the core mask. The boundaries of the core and shell mask were matched and overlapped each other, followed by applying threshold function and flood fill. Then, local thickness function was executed at the coreshell mask, giving rise to local thickness colour map and distribution.

Section 6. Finite-difference time-domain (FDTD) simulations
Three-dimensional (3D) FDTD simulations were preformed using the commercially available software package FDTD solutions (Lumerical 2020a R7). All 3D models were constructed and exported by 3D MAX software. The simulation boundary was set to perfectly matched layer. The dielectric constant for gold was taken from experimental bulk results previously reported. 4 The refractive index of the surrounding medium was set as 1.3330 for water, 1.4305 for DMF and 1.5916 for PS according to literature. 5 The total-field scattered-field function was used to obtain absorption and scattering spectra. The wavelength of source was adjusted from 300 nm to 1000 nm, and it was directed toward XY plane. The optimum override mesh (1 nm  1 nm  1 nm), mesh accuracy (level 6) and simulation size (3.5 µm) were set after convergence test for reliable data acquisition.     The spectra of gold polyhedra before the PS grafting were simulated in water, and those of PS-grafted gold polyhedra (PS shell thickness = 20 nm) were simulated in DMF.