Chemical Identification of Single Ultrafine Particles Using Surface-Enhanced Infrared Absorption

In the past decade, it has been demonstrated that surface-enhanced infrared absorption (SEIRA) is a powerful method to enhance vibrational signals of thin molecular layers. Much less attention has so far been given to the possibility of using SEIRA for the detection and characterization of nanometer-sized particles, such as ultrafine dust particles. Here, we report on SEIRA measurements demonstrating that even one single particle with a deeply subwavelength dimension of less than 100 nm can be detected and chemically characterized with standard infrared microspectroscopy. Our approach is based on plasmonic resonances of bowtie-shaped $\mathrm{Au}$ apertures that are designed to extraordinarily enhance the material-specific phononic excitations of a nanometer-sized silica particle. We show that the bowtie geometry is especially suited for single-particle spectroscopy, as it combines the advantage of an intense electromagnetic hot spot, the size of which can be adjusted to the particle dimension, with easy positioning of ultrafine dust particles inside that hot spot. In agreement with numerical calculations, we show that a detection limit in terms of a particle diameter of less than 20 nm can be achieved, which corresponds to a ratio of the diameter to the vacuum wavelength below 0.002. Our approach offers the possibility of analyzing infrared bands from tiniest particles and thus paves the way toward SEIRA-based devices that can sense ultrafine dust.

Figure 1

The positioning of the silica spheres. The spheres are moved by centrifugal forces through the bowtie structure, where they get stuck in the center of the bowtie due to the tapering part.

Figure 2

The calculated near-field intensity of nanobowtie antennas for the electric-field direction ($\overrightarrow{E}$) as indicated. (a) The normalized near-field intensity distribution in a plane parallel to the substrate at the half height of the bowtie antenna. The base side of the bowtie is varied from 100 nm up to 3000 nm, whereas the length is adjusted to achieve a constant resonance frequency of ${\omega }_{\mathrm{res}}=1100{\mathrm{cm}}^{-1}$. The hot spot changes from an extended hot spot with low intensity to a highly localized hot spot with high intensity. For comparison, the size of an ultrafine dust particle is indicated by blue circles. (b) The near-field intensity read-out at the tapering part as a function of the bowtie base side $b$.

Figure 3

SEM images of bowtie antennas after sphere positioning. Bowties with a base side between 100 nm and $2\mu \mathrm{m}$ are fabricated and 85 nm silica spheres (colored red) are positioned at the center of the structures.

Figure 4

(a) The relative IR reflectance of the bowtie apertures shown in Fig. 3. A broad plasmonic resonance containing the Fano-type absorption peak of the silica sphere is observed. For the smallest ($b=100\mathrm{nm}$), the signature of the silica sphere can barely be seen. However, the signal size increases if the base side of the bowtie structure is increased. The lower panel shows the baseline-corrected signal, where the increase in the signal can be seen clearly. The corresponding baseline is shown in gray in the upper panel. (b) FDTD simulations of the same geometry as shown in (a). The same trend as in the experimental data is found.

Figure 5

(a) The SEIRA signal-size read-out as the difference between the minimum and maximum of the baseline-corrected curve shown in Fig. 4. A strong increase of the SEIRA signal is observed for small values of $b$. For high values of $b$, the signal is decreasing again. The signal decrease can be explained by the size of the hot spot, which is shown in (b). An optimum SEIRA signal is found if the hot-spot size matches the sphere diameter.

Figure 6

The detection limits. (a) SEM images of nanospheres positioned in the bowtie gap ($b=1000\mathrm{nm}$, $L=2100\mathrm{nm}$). The sphere diameter is changing from 34 nm through 55 nm to 85 nm. (b) The IR relative reflectance spectra and the baseline-corrected vibrational signal of the structures shown in (a). The signal of the smallest sphere is only slightly above the noise level, whereas the bigger spheres can easily be detected. The right-hand panel shows the corresponding FDTD simulations. (c) The SEIRA signal-size read-out from (b) plotted against the sphere diameter. The simulated signal is linearly increasing up to $d=50\mathrm{nm}$. The red dashed line indicates the noise level of the experiment, whereas the blue dashed line indicates the detection limit of the simulated data, which we choose to be in accordance with the experimental data.

Figure 7

The detection limit for silica particles. (a) The simulated SEIRA signal size of silica particles with an increasing diameter placed in the center of bowtie nanoapertures with different base sides $b$. The crossing point with the noise-level horizontal is defined as the detection limit and is plotted in (b) as a function of the triangle base side $b$. By increasing the base side, the detection limit can be improved down to a sphere diameter of less than 20 nm for the most powerful bowties ($b=2500\mathrm{nm}$).

Figure 8

The calculated extinction cross section of a fine dust particle with a diameter of 50 nm consisting of the organic material CBP (upper panel). The middle panel shows FDTD simulations of such a CBP sphere in bowtie antennas with $b=2500\mathrm{nm}$ and different lengths. The lower panel shows the baseline-corrected vibrational signal. The strongest vibrational bands of CBP are marked with dashed lines and can be used to identify and characterize the molecular species.