Validation of openFrame LED Autofocus using simultaneous astigmatic axial readout -Sara Habte (Biophotonics and Imaging)
Previously the group has presented a novel optical autofocus module as a part of the openFrame using two orthogonal cylindrical lenses to collimate the autofocus laser beam emerging from the single mode fibre. Another approach presented is to simultaneously provide extended range of operation (>100 µm) and high precision (<600 nm) using machine learning in a 2-step approach or providing closed loop “single-shot” operation over up to ±37 μm with <50 nm accuracy. These levels of performance are realised by focusing an infrared laser beam onto the microscope coverslip with the back reflection being imaged on a dedicated autofocus camera. We derive a metric that quantifies defocus from the size of the light distribution at the autofocus camera that can be independent of laser power and insensitive to drift in the optical alignment. The operating range and precision depend on the confocal parameter of the autofocus laser beam after being focused by the objective lens. This can be adjusted by changing the diameter of the autofocus laser beam incident at the objective lens. By contriving a different beam diameter in two orthogonal planes using either a rectangular aperture in the collimated autofocus laser beam, or using different orthogonal cylindrical lenses to collimate the autofocus laser beam emerging from the single mode fibre that delivers it to the autofocus module, we can maximise precision (with maximum beam diameter) and extend operating range (reducing orthogonal beam diameter), making both measurements simultaneously by resolving the autofocus camera image along orthogonal directions.
While these two approaches can provide months of stable operation, they each have their drawbacks. The machine learning approach with the rectangular apertured beam [1] requires a convolutional neural network to be trained to determine magnitude and sign of defocus from the autofocus camera image and we found it necessary to train it over ~10 days to make it independent of any system variations impacting the autofocus camera image. For the second approach, we slightly offset the collimation of the cylindrical lenses such that the measured defocus is different for the two planes defined by the orthogonal cylindrical lenses, and this enables the magnitude and sign of the defocus to be calculated from a single autofocus camera image following calibration of the system. However, while the system reported in utilised a low-cost single-mode fibre (SMF)-coupled laser diode, we used a super luminescent diode (SLD) in the system reported in since its performance was impacted by interference between the autofocus laser beam reflected from the coverslip and unwanted beam(s) reflected from other surfaces in the optical system. Using the SLD removed this interference. Unfortunately, SLDs are significantly more expensive than laser diodes, and availability can be intermittent. Accordingly, we are redesigning the optical system and analysis method to enable the closed-loop approach of to be used with a simple multimode fibre-coupled LED for implementation in slide scanning and automated multi-well plate microscopy.
Furthermore, we propose the use of a sub resolution beads imaged using added astigmatism as an independent verification of the performance of autofocus systems. We modify the openFrame through the addition of a cylindrical lens in front of the imaging sensor. We determine the offset of the imaged plane from the focal plane via an autocorrelation of the bead images to make use of signals from across the imaged field with minimal processing effort. A z-stack of reference images can be taken, and the autocorrelation of these images can be used to determine the z offset. By providing an independent means of determining the focal offset, we can determine the performance of an autofocus system without simply relying on the feedback it provides, which may not be correct if the system is performing poorly.
Bayesian optimization of resonant dispersive wave generation in hollow capillary fibres – Tim Klee (Attosecond Optical Science)
Resonant dispersive wave (RDW) generation in hollow capillary fibres (HCF) is a powerful technique for producing ultrashort light pulses in the deep ultraviolet range, which are important for ultrafast spectroscopy and material processing. However, the complex nonlinear dynamics governing this process and the large associated parameter space make it challenging to achieve optimal RDW pulses with the highest peak power. In this study, Bayesian optimisation (BO) is coupled with the open source \texttt{Luna.jl} simulation framework to optimise the HCF and pump pulse paramters for less than 5 femtosecond (fs) RDW generation at a target wavelength of 200 nm. Temporally non-structured RDW were consistently identified with peak powers of up to 14 GW, exceeding experimentally published values by up to 70 \%. Furthermore, a subset of the RDW optima exhibited an energy stability that is better than that of the pump pulse. Given that this approach can be generalised to other RDW wavelengths, our findings suggest that BO is a valuable tool in developing HCF systems that support RDW generation tailored to a particular experimental need.
Locally Chiral Evanescent Waves for Efficient Enantio-Discrimination – Peilin Yang (Attosecond Optical Science)
Distinguishing between the left- and right-handed versions of a chiral molecule (enantiomers) is important in a wide range of disciplines. However, it can also be a difficult task, as oppo-site molecular enantiomers behave identically unless they interact with another chiral object.
Traditional chiral spectroscopy is not efficient because it relies on the handedness of circularly polarized light. The enantio-sensitive response of the molecules to such light arises beyond the electric-dipole approximation, and it is only a small fraction of the total intensity of the optical response, usually below 0.1%. This creates important limitations for ultrafast chiral spectroscopy.
To overcome these limitations, one can tailor the sub-cycle oscillations of the electric-field vector of a laser field in 3D, creating synthetic chiral light, which is locally chiral: the tip of the electric-field vector draws a (3D) chiral Lissajous figure in time, at every point in space. The enantio-sensitive response of the chiral molecules to such light arises within the electric-dipole approximation, and it is stronger by orders of magnitude.
The recipe to create synthetic chiral light has two ingredients. First, two phase-locked frequencies. Phase locking is crucial, as the field’s handedness is sensitive to phase delays. Second, a longitudinal field, i.e. a frequency component polarized in the propagation direction, which al-lows the Lissajous figure of the field to be 3D and chiral. This component can be created using a non-collinear optical setup, but it also arises naturally when light is confined in space, as it in tightly focused laser configurations, as well as in optical nanofibers and other nano-photonic structures.
We are investigating how we can apply modern nanophotonics to tailor the polarization of light at will, in order to create synthetic chiral light. We take advantage of the fact that, when light propagates through an optical nanofiber, with diameter of only a few hundreds of nanometers, the strong (sub-wavelength) confinement gives rise to strong longitudinal field com-ponents, but also to strong evanescent waves around the fiber.
Here we will show how we can combine several optical nanofibers for efficient enantio-discrimination. The evanescent waves arising in the vicinity of the nanofibers can create locally chiral fields in the surrounding medium. Importantly, by controlling the properties of the fibers, we can design such locally chiral evanescent waves in a way that they maintain their handedness in space.
Our numerical simulations show that the low-order nonlinear optical response of the chiral molecule carvone to such a locally chiral evanescent wave is strongly enantio-sensitive. This work creates exciting opportunities for imaging molecular chirality in the liquid phase, the natural environment of biomolecules, with high enantio-sensitivity and on ultrafast timescales.