We report on a novel fabrication scheme obtaining metal nanoparticles grown in a crystal high index ambient, i.e. SrTiO3 (STO). Starting from a deposited Au layer ellipsoidal and orientated Au nanoantennas can be prepared and controlled via laser ablation. We discuss the deposition conditions and demonstrate the geometrical properties of the Au nanoantennas embedded in the SrTiO3 by Transmission Electron Micrography (TEM) and X-ray Diffraction (XRD) investigations. The plasmonic activity of the resulting highly orientated crystalline particles could be observed in transmission spectroscopy experiments showing an influence on the resulting size, shape and number density of nanoparticles that can be adjusted by precisely controlling the deposition parameters. All performed variations have simultaneously been observed in terms of the spectral shift of the localized plasmon resonance upon illumination. Therefore the resonance shift in the measured transmission spectra can be associated to geometrical changes in the nanoantennas’ shape that can be explained using a quasi-static description of the nanoparticles.
©2011 Optical Society of America
Metal nanoparticles are well known as plasmonic active material due to the excitation of the localized surface plasmon resonances (LSPR ). These resonances depend on material (usually Au and Ag), shape and size of the particles, as well as on the dielectric properties of the ambient medium and on inter-particle distances. There are different possibilities to produce particles - the wet chemical syntheses described in [2,3], the self-organized preparation in different matrices by thin film coatings [4–8] or the use of lithographic processes. One aspect of our work is the investigation of the influence of nanoparticles on the crystalline STO growth. The other aspect is the use of these particles as plasmonic active antennas as sensors in life sciences, where anisotropic particles (which are more difficult to synthesize) promise increased sensitivity . Moreover, increased crystallinity (which is observed in chemical synthesis but not in lithography) is another desired property. Therefore it is favorable to prepare mono-crystalline Au nanoparticles with different aspect ratios, depending on the desired application. The preparation of anisotropic particles without the need for electron beam lithographic processes can be an advantage. If it is possible to control the growth conditions in order to produce nanoantennas at the surface of the matrix, analyt-molecules could be adsorbed specifically at the smallest radii of the nanoantennas experiencing the enhanced electromagnetic fields upon excitation.
Laser-assisted deposition techniques have the capability to achieve epitaxial thin films of high crystalline quality with the advantages of controlling stoichiometry and structural properties. Therefore the method of Pulsed Laser Deposition (PLD) was used to prepare the samples, which exhibit the layer sequence STO-Au-STO.
The following experiments were carried out in a vacuum chamber with a base pressure of p0<5x10−4 Pa. First a gold layer with a thickness of approximately 0.5 nm and 1 nm was deposited onto polished 5x10 mm2 STO (001) single crystals, that were used as substrates for epitaxial film growth. A KrF excimer laser, operating at a wavelength of λ = 248 nm and a pulse width of τ = 25 ns, was set to a repetition rate of 10 Hz and a laser fluence of 1.3 J/cm2 at the target surface. Under these conditions a homogeneous layer growth rate of about 4 nm/min of Au was achieved.
Prior to the second deposition step the sample was heated to 800°C and oxygen was introduced into the chamber up to a pressure of pO2 = 100 Pa. These two parameters were found in an optimization process to represent the ideal conditions for the growth of STO layers onto STO substrates whilst ensuring both smooth surfaces and high crystalline quality.
Using again the KrF excimer laser the top STO layer was then deposited onto the Au covered substrate. At a repetition rate of 5 Hz and a laser fluence of 1.5 J/cm2 a growth rate of 11 nm/min was achieved. After the deposition process the sample was cooled down to room temperature at a rate of 50 K/min in an oxygen rich atmosphere. By varying the deposition time, samples were obtained with thicknesses of the STO top layer ranging from 35 nm to 210 nm.
In order to determine the effect of the gold on the growth and properties of the STO-STO system, the process was modified such that during the Au deposition half the substrate was covered with a shadow mask. In doing so, regions with and without Au exposed to the same conditions, were created. The specific sample layout allows comparative measurements that were used to ascertain the influence of the Au on the STO-STO layer system.
Detailed investigations were then carried out to examine and determine the structural properties. Surface morphology was studied by atomic force microscopy (AFM) and scanning electron microscopy (SEM). Film crystal structure was measured by X-ray diffraction (XRD), whereas transmission electron microscopic (TEM) analyses revealed size, position and distribution of the Au within the layer system. Transmission measurements were done by an UV/VIS/IR spectrometer combined with an integrating sphere (Lambda 900, Perkin Elmer).
3. Results and Discussion
3.1 Formation and Characterization of Au Nanoparticles in SrTiO3
Several samples were prepared for TEM cross section imaging using a focused ion beam. TEM micrographs show that during the deposition of the STO film the gold layer self-assembles to small particles with sizes in the order of a few tens of nanometers (cf. Fig. 1a ). These clusters are embedded into the STO matrix. This result is somewhat unexpected compared to former experiments of the formation of Au particles in oxide thin films. In  it was shown that Au particles preferentially grow at the film surface or at the interface between the deposited film and the substrate. Only a small number was observed within the film. In contrast, the present experiments display a unique appearance: Homogeneously arranged and isolated small particles are located at a characteristic depth within the deposited STO layer. Their ellipsoidal shape is consistent amongst all the samples. The gold clusters are elongated into the growth direction of the surrounding STO layer. They exhibit a general aspect ratio c/a of 2:1 with typical size dimensions of about 8-11 nm in the a-axis and 17-23 nm in the c-axis of the ellipsoid.
Whereas XRD measurements have proven the high overall crystalline quality of the STO film, high resolution TEM images allow crystal structure analyses at selected points of interest, (Fig. 1b). For the sample without gold homoepitaxial growth of the deposited STO onto the substrate can be confirmed. The actual lattice plane spacing of the STO matrix matches almost perfectly the value of the lattice plane spacing for the substrate. The lack of an apparent interface between film and substrate is additional evidence of the accurate lattice accommodation. For the Au modified STO-STO system the nanoparticles do not seem to have any negative effect on the layer growth of the STO. Given that the matrix lattice plane spacing is about the same as for the system described above, homoepitaxial growth in the presence of small particles can still be assumed. It is remarkable that even in the vicinity of the Au particles the STO layer shows no indications of defects or dislocations. The crystal structure appears undisturbed. The lack of an interface structure between STO substrate and STO layer could be the reason that no nanoparticles remain at the substrate surface after initiating the STO growth.
The strong interactions between STO and Au during layer growth lead to a preferential orientation of the nanoparticles according to the characteristics of the surrounding matrix. The high crystalline order - indeed monocrystallinity - of the particles can be revealed. The reconstructed lattice plane spacing of 2.32 Å indicates a <111> orientation of the gold crystals (theoretical value of 2.35 Å) that correlates well with the STO structure.
The results of samples with varying STO layer thickness but identical amount of Au of about 1 nm are summarized in Fig. 2 . The strict alignment of the particles within the matrix is consistent. Surprisingly the characteristic depth, at which the particles seemed to be lined up, correlates with the thickness of the STO layer. The more STO is deposited and the longer the deposition process lasts the further away the clusters are located from their original position (Fig. 2a). Additionally, an increase of the STO film thickness causes a variation of the particle dimensions. Changes of the c-axis as well as the a-axis can be observed (Fig. 2b). A tendency to a smaller aspect ratio and a more spherical shape is apparent.
The exact particle formation mechanism needs further investigations but we assume that the driving forces of layer growth lead to a detachment of the gold from the substrate and their transport within the forming STO layer. However, particles are not observed at the sample surface when the STO layer thickness exceeds the dimension of the main axis of the elliptic particles. Why this process comes to a halt at a certain depth within the matrix is not completely understood yet.
3.2 Theoretical Description of Plasmonic Resonances of Au Nanoparticles in Environmental Material
In the following the plasmonic properties of the fabricated nanoantennas upon optical excitation and their spectral effect will be investigated in terms of a quasi-static analysis. For this purpose the nanoparticles’ shape, see Fig. 1, will be approximated by ellipsoids. The quasi-static description is valid for small nanoparticle dimensions, i.e. particles that are small compared to the impinging wavelength [10,11]. This is the case for the prepared nanoparticles, see e.g. Fig. 1, where the main axis of the ellipsoidal nanoantennas are about 15-30 nm whereas their fundamental dipole mode is expected at visible wavelengths. For such nanoparticles the scattering, absorption and extinction cross sections, i.e. the ratio between the scattered power and the incident intensity, can be calculated analytically by Eqs. (1) k(ω) is the wave vector and α(ω) corresponds to the nanoparticles polarizability that can be calculated with Eq. (2) it can be concluded that the resonance occurs as soon as the denominator becomes zero. This condition, i.e. the localized plasmon-polariton resonance frequency of the particle, is dependent on the permittivities of the ellipsoid εi(ω) and the ambient dielectric εa(ω). Finally, the resonance frequency is affected by the geometry of the ellipsoid that enters Eq. (2) with the depolarization factor Ll, which is purely related to the three semi-axis of the ellipsoid (index l)Eq. (3) the cross sections, i.e. Eqs. (1), can be evaluated including the dispersive properties of the involved materials as well as the particles dimensions that can be depicted from the TEM analysis in Fig. 1. For the dielectric function of Au as well as for STO the materials data have been taken from literature [12,13]. In Fig. 3(a) the effect of the particle shape on the extinction cross section is shown. For simplicity all presented cross sections have been normalized by the ellipsoidal particles cross section normal to the direction of incidence. As expected, the plasmon resonance peak shifts to the red as soon as the particle shape changes from spherical towards elliptical. In a second step the ambient index has been varied in the calculations. As it is shown in Fig. 3(b), a change in the ambient material from vacuum (n = 1) to STO provides a resonance red-shift as well. Moreover, it can be seen, that for the spectral interval of interest a constant approximation of the exact STO dispersion can be done since differences between the extinction cross sections obtained with the exact dispersion data and the constant approximation are negligible.
Exact calculations done by FDTD (Finite-difference time-domain method) regarding the influence of inter-particle distances on the extinction spectra are shown in the following (Fig. 4 ). The frequency behaviour of interacting particles with the same ellipsoidic shape and equal distances is calculated. The distance variation from Λ = 15 nm up to Λ = 50 nm shows in all cases a double peak formation shifting to higher wavelength with decreasing particle distance.
3.3 Measurements of Plasmonic Frequencies of Au Nanoparticles in STO
The prepared samples were investigated by transmission experiments in an UV/VIS/IR spectrometer. After transmittance through the sample the detected light consists of the forward scattered emission and the directly transmitted light. Hence, the residual losses are defined by absorption and backward scattered emission. The resulting extinction curves, calculated as 1-T, are used to interpret the spectral behavior of the samples with the focus on the position of extinction maxima, the shape of the curves and the appearance of more than one resonance peak. All plotted curves were normalized in respect to reference transmitted intensity (Fig. 5 ).
What we found for samples with 1 nm Au in general are extinction spectra (Fig. 6 ) that show a double peak characteristic. As shown in Fig. 4 an interaction between the gold nanoantennas is a reason for the development of more than one resonance peak. Furthermore, increasing the STO film thickness up to 210 nm leads to a distinctive red-shift of the main extinction maximum up to 650 nm. If the distance between the particles is changed according to the change in particle aspect ratio (see Fig. 2b) such shift could be assumed. Superposition of resonances of the particle ensemble cause a broadening of the curves.
A contrary frequency behavior for samples with Au layer thicknesses of 0.5 nm and 1 nm, both suited to two different STO layer thicknesses, 35 nm and 70 nm can be observed when the film thickness is in the range of the main particle dimensions (Fig. 7 ). For curves b)-d) the formation of a double peak system is clearly suggested, but for the graph a) with the lowest film thickness a smooth curve is found. According to the simulations, the lack of such a double peak system suggests particles that exhibit a typical extinction spectrum for separated metal particles without coupling to other particles.
Additionally it is shown that the position of the resonance peak for samples with a 35 nm STO layer is found at higher wavelengths than for samples with a 70 nm STO layer. Spectra of samples featuring a Au thickness of 1 nm exhibit a resonance maximum at 560 nm (35 nm STO) compared to 535 nm (70 nm STO). The same characteristic is observed for samples with 0.5 nm Au: The resonance maximum is at 530 nm (35 nm STO) whereas the peak for samples with 70 nm STO lies at 515 nm. This behavior correlates to the aspect ratio for smaller STO film thicknesses presented in Fig. 2.
The high resonance peak at 460 nm in case of 1 nm Au and 35 nm STO layer thickness cannot be explained with typical plasmonic characteristics of particles within a matrix. We used different methods to obtain detailed information about size and distribution of gold particles on the STO surface of such a sample. In structural investigations using AFM we observed that some of the embedded particles can be found at the STO surface. Based on estimations of the entire gold volume approximately 20% of the nanoparticles can be found on top of the STO surface. The maximum difference in height between gold particles and the STO surface was found to be 15-20 nm. Therefore a partially change of the environmental medium refractive index from n = 2.4 (STO) to n = 1 (air) is considered as to be a system of air and STO that leads to a different extinction behavior compared to other samples.
Although gold crystals occur at the surface of samples with the thinnest STO top layer, this cannot be observed in thicker STO layers. To overcome this restriction, a wet chemical etching method appears most promising. Preliminary experiments towards a selective etching of the STO top layer while leaving the gold particles unaffected were successful.
All discussed effects most likely interfere with each other so that an exact interpretation is rather difficult. However, a basic qualitative description of the behavior of nanoparticles within a matrix could be given, but further especially structural investigations must be done.
Homoepitaxial STO thin films with thicknesses in the range of 35 up to 210 nm modified by embedded crystalline gold nanoparticles were prepared. The growth conditions during STO deposition determine the layer system morphology leading to anisotropic, equal-sized Au particles incorporated into the matrix. Slight changes in microstructure, primarily the variation of aspect ratio, cause distinctive changes in the optical properties. A red-shift of the plasmonic resonances with increasing STO layer thickness is detected. Films with the STO layer in the dimensions of an average Au particle exhibit contrary behavior. In conclusion, the STO-Au-STO system provides the possibility to tune the plasmonic properties of the films through directed microstructure engineering. Additionally, it allows the formation of crystalline and orientated plasmonic nanostructures without the requirement for lithographic approaches.
The authors would like to thank C. Voigt for her technical assistance and I. Uschmann for helpful discussions.
References and links
1. K. L. Kelly, E. Coronado, L. L. Zhao, and G. C. Schatz, “The optical properties of meta nanoparticles: the influence of size, shape, and dielectric environment,” J. Phys. Chem. B 107(3), 668–677 (2003). [CrossRef]
2. A. Csaki, T. Schneider, J. Wirth, N. Jahr, A. Steinbrück, O. Stranik, F. Garwe, R. Müller, and W. Fritzsche, “Molecular plasmonics: light meets molecules at the nanoscale,” Philos. Trans. R. Soc. London, Ser. A (in press).
3. A. Csaki, S. Berg, N. Jahr, C. Leiterer, T. Schneider, A. Steinbrück, D. Zopf, and W. Fritzsche, in Gold Nanoparticles: Properties, Characterization, Application, P. E. Chow, ed. (Nova Science, 2010).
4. W. Jacak, J. Krasnyj, J. Jacak, R. Gonczarek, A. Chepok, L. Jacak, D. Z. Hu, and D. Schaadt, “Radius dependent shift in surface plasmon frequency in large metallic nanospheres: theory and experiment,” J. Appl. Phys. 107(12), 124317 (2010). [CrossRef]
5. M. Torrell, L. Cunha, M. R. Kabir, A. Cavaleiro, M. I. Vasilevskiy, and F. Vaz, “Nanoscale color control of TiO2 films with embedded Au nanoparticles,” Mater. Lett. 64(23), 2624–2626 (2010). [CrossRef]
6. M. M. Kjeldsen, J. L. Hansen, T. G. Pedersen, P. Gaiduk, and A. N. Larsen, “Tuning the plasmon resonance of metallic tin nanocrystals in Si-based materials,” Appl. Phys., A Mater. Sci. Process. 100(1), 31–37 (2010). [CrossRef]
7. S. Cho, S. Lee, S. Oh, S. J. Park, W. M. Kim, B. Cheong, M. Chung, K. B. Song, T. S. Lee, and S. G. Kim, “Optical properties of Au nanocluster embedded dielectric films,” Thin Solid Films 377–378(1-2), 97–102 (2000). [CrossRef]
8. B. Karthikeyan, “Fluorescent glass embedded silver nanoclusters: an optical study,” J. Appl. Phys. 103(11), 114313 (2008). [CrossRef]
9. V. Grosse, S. Engmann, F. Schmidl, A. Undisz, M. Rettenmayr, and P. Seidel, “Formation of gold nano-particles during pulsed laser deposition of YBa2Cu3O7-δ thin films,” Phys. Status Solidi (RRL) 4(5–6), 97–99 (2010). [CrossRef]
10. C. F. Bohren and D. R. Huffman, Absorption and Scattering of Light by Small Particles (Wiley, 1983).
11. U. Kreibig and M. Vollmer, Optical Properties of Metal Clusters (Springer, 1995).
12. P. B. Johnson and R. W. Christy, “Optical constants of the noble metals,” Phys. Rev. B 6(12), 4370–4379 (1972). [CrossRef]
13. M. Bass, C. DeCusatis, G. Li, V. N. Mahajan, and E. Van Stryland, Handbook of Optics: Optical Properties of Materials, Nonlinear Optics, Quantum Optics (McGraw-Hill, 2009).