Super-resolution imaging using nano-bells
Our group has demonstrated a new scheme for optical super-resolution, inspired, in-part, by PALM and STORM. Here each object in the field of view is tagged with a signal that allows them to be detected separately. By doing this we can identify and locate each object separately with significantly higher resolution than the diffraction limit. We demonstrated this by imaging nanoparticles significantly smaller than the optical resolution limit. In this case the "tag" we have used is the frequency of vibration of nanoscale "bells" made of metallic nanoparticles whose acoustic vibrational frequency is in the multi-GHz range. Since the vibration of the particles can be easily excited and detected and the frequency is directly related to the particle size, we can separate the signals from many particles of sufficiently different sizes even though they are smaller than, and separated by less than, the optical resolution limit. Using this scheme we have been able to localise the nanoparticle position with a precision of ~3nm. This has many potential advantages - such nanoparticles are easily inserted into cells and well tolerated, the particles do not bleach and can be produced easily with very dispersed sizes. We estimate that 50 or more different particles (or frequency channels) can be accessed in each optical point spread function using the vibrational frequencies of gold nanospheres. However, many more channels may be accessed using more complex structures (such as nanorods) and detection techniques (for instance using polarization or wavelength selective detection) opening up this technique as a generalized method of achieving super-optical resolution imaging.
How nano-bells work
In “nano-bells” (see Fig. 1), the tags are metallic nanoparticles where their optical PSF overlap in time and space and hence can not be resolved. However, it is possible to separate the PSF of each particle by measuring their vibrational signatures provided the particles are of different size and hence vibrate at a different frequency (see Fig. 1b). Determining the centroid of the point-spread function for each frequency, super-localisation can be achieved (see Fig. 1c). In “nano-bells”, we propose to use the acoustic characteristics of nanoparticles as a route for super-resolution reconstruction imaging in a similar way to STORM
Experimental measurement of nanoparticle vibrations
Te experimental system consisted of an inverted microscope with additional elements mounted above the optical deck. Te sample was mounted on a positioning stage with 100 nm step size for raster scanning. Laser illumination was delivered from below using a x20 objective lens and the scattered probe light was collected above using an additional x20 objective
Figure 2 shows a typical signal taken from the system: (a) raw signal taken centred on the particle, (b) signal afer processing to extract the change due to scattering cross section modulation as the particle vibrates (inset frequency content), (c) SEM of the particle and (d) optical image of same particle. We can use the frequency of vibration to determine the size of the particle, in this case the size is estimated as 168±2 nm from a 18.9GHz breathing mode.
Super-localisation of multiple nano-bells
Figure 3 shows a sequence of images taken using our GHz/nanoscale phononic microscope of three particles deposited on a gridded coverslip, (a) is the optical image, (b) the SEM image of the same area, (c) the amplitude of the vibrational signal, (d) the acoustic frequency measured from the dominant peak and (e) the reconstruction of the image using (c) to locate the particles and (d) to determine their size.
In Fig. 3a the picture shows two objects well separated and because of this, one can assume their localisation is trivial by optical means. However, the SEM image of the same area (see Fig. 3b) reveals a third object not localised by optical imaging. In this case two of the three particles are close together, separated by less than the width of the optical PSF so the reconstruction becomes more complex.
Here we rely on the vibrational frequency of nano-bells to attribute, like in STORM, a unique point-spread function to each particle. Te acoustic scan provides the amplitude map of the particles which still shows only two objects (see Fig. 3c). However plotting the amplitude of each frequency present in the detected signal, lends to the reconstruction of the PSF of each particle (see Fig. 3d). Once the PSF of each particle is detected using its frequency signature, its location is determined by centroiding its PSF in a similar way to STORM (see Fig. 3e). Note that the PSF of different particles overlap although is not visible in Fig. 3d.
It can be seen that the ultrasonic reconstruction has successfully localised the nanoparticles in the sample at far higher precision than the resolution of the optical microscope is capable of. Te precision in which the size and position of the particles are localised, determined from the signal and noise analysis, can be measured to ~3nm. It is important to note that (with the exception of the SEM images) all the images presented in this paper were made through the same optics and all are constrained by the same conventional optical resolution.
This process breaks down in the event of multiple particles with the same frequency of vibration being imaged in the same optical point spread function similar to when two fluorophores turn on simultaneously within one PSF in STORM. However, if the particles in each optical point spread function are sufficiently size dispersed then they can be easily and reliably imaged. Figure 4 shows a large scale image identifying many particles in a 50×50μm field of view and super-resolution images for three different positions on our sample demonstrating reliable and robust reconstruction.
Towards more complex structures
Related Publications and Talks
Super-resolution imaging using nano-bells. Rafael Fuentes-Dominguez et al, Scientific Reports, 8(1):16373, 2018.
Super-resolution imaging using nano-bells. Rafael Fuentes-Dominguez. LU Nottingham UK (2018). Oral presentation.