Keywords: experimental thesis, non-linear optics, microscopy, spectroscopy, femtosecond laser pulses, biological applications, artificial intelligence/deep learning

A major challenge in medicine is the rapid identification of tumors and their precise classification in one of several possible types. The doctor can then choose a targeted care. Today, unfortunately, this is not always possible, since the tumor diagnosis is based on visual (and therefore subjective) inspection of biopsies under the microscope. In the future, the solution could be provided by the Raman scattering. This technique exploits the specific vibrational response of the molecules as a fingerprint for their unique identification: strong ties, such as stiffer springs, produce vibrations at higher frequencies; heavy atoms, on the contrary, will generate vibrations at lower frequencies. It is thus possible to obtain quantitative and reproducible information on cells and tissues for their characterization. Unfortunately, "spontaneous" Raman scattering (based on the use of a continuous laser beam) takes a long time, so in general is used to characterize a single point of the tissue. To obtain two-dimensional images of large areas of tissues in a short time you need to use the "coherent" Raman scattering technique, illuminating the sample with two ultra-shortand synchronized laser pulses. To obtain chemical information as comprehensive as possible, one of the two pulses must be broadband (polychromatic).

 
This work of thesis will take place in an ultra-fast optical laboratory that has received funding by the European Community (see www.vibra.polimi.it and https://www.crimson-project.eu/) to create a microscope that can identify and characterize tumors by precisely this technique of spectroscopy and coherent Raman microscopy. The student will learn to align broadband femtosecond laser beams at high repetition rate in the near infrared. With the help of other researchers, he will build complex experimental setups to perform measurements on biological samples, mixing basic physics (radiation-matter interaction activities, Raman effect ...), non-linear optics (white light generation, parametric amplification, second harmonic generation), microscopy of biological samples and data analysis using "chemometric" approaches (multivariate statistical analysis).
Here is an example of images collected in breast cancer cells with this technique:

Here is another example of image collected in murine spine section:

Please have a look at www.vibra.polimi.it and http://polli.faculty.polimi.it/ for further details.
Here are a few pictures of the recently developed microscope: