AZoOptics speaks to Yishai Klein, Research Physicist from Bar-Ilan University, about his x-ray fluorescence technique based on computational ghost imaging. The new method overcomes current challenges such as limited resolution by removing focusing requirements and significantly reducing measurement times. This could potentially enhance the resolution of chemical element maps and extend the applicability of x-ray fluorescence inspection to new fields.
How did you begin your research into ghost imaging and X-ray fluorescence chemical mapping?
Our group led by Prof. Shwartz focuses on the demonstration and application of quantum and quantum-inspired effects with x-rays. Since the performances of x-ray lenses are very poor, lens-free imaging techniques are very attractive in the x-ray region.
My first research project was to demonstrate the lens-free technique of computational ghost imaging (CGI) with x-rays. This important demonstration showed the basic feasibility to use CGI with x-rays. The next step was to think about what the most suitable applications for CGI could be and where the strengths of CGI will be significant.
Hence, the connection between CGI and x-ray fluorescence was natural. Since XRF is performed by raster scanning of the sample inefficiently, it was clear to us that chemical mapping based on x-ray fluorescence with CGI is more efficient than with the traditional raster scan. Furthermore, the resolution can be improved because it does not require any focusing.
What is X-ray fluorescence and how is it used in several applications?
X-ray fluorescence (XRF) is a powerful method for the identification and mapping of the chemical compositions of samples with intriguing applications that are exploited in a broad range of fields from fundamental science to industry and cultural heritage.
Examples of scientific disciplines where XRF plays a prominent role include materials science, electrochemistry, biology, paleontology, and archaeology. Industrial application examples include analyzers for small parts that are produced by the automotive and aerospace industries.
In cultural heritage, XRF is very useful in providing information on hidden layers of famous paintings.
The basic principle of XRF is simple and is based on the x-ray fluorescence process in which x-ray radiation is used to excite core electrons in the sample. When the core electrons are excited or ejected from the inner shells of the atoms, holes are formed in those shells. The electrons can return to their ground state, or outer electrons can fill the holes leading to the emission of x-ray radiation at photon energies that correspond to the characteristic atomic lines.
The spectrum of the emitted radiation (the fluorescence spectrum) is detected and analyzed, and since each chemical element has unique emission lines, the fluorescence spectrum is used for the characterization of the elemental composition of the sample. The detection can be carried out with simple-to-use energy-resolving detectors, and available components with sufficient energy resolution.
While in its simplest form, XRF provides no spatial information since the detector collects the radiation from large areas. In recent decades, spatially resolved XRF techniques have been developed with their advent opening up appealing opportunities in many fields.
In most cases, two-dimensional chemical maps are reconstructed by focusing the impinging beam and raster-scanning the sample. With this procedure, the spatial information is retrieved since at each measurement point only a small portion of the sample is irradiated and the resolution is determined by the spot size of the input beam. Extensions to three dimensions are also possible by either computed tomography or confocal x-ray microscopy.
What are the current limitations of X-ray fluorescence chemical mapping?
Despite being very successful and widely used, XRF faces two major challenges that hamper its performances and the extension of its applicability to further disciplines:
- Focusing of x-ray radiation is difficult, especially at high photon energies, thus the ability to use small spot sizes in a broad photon energy range is unique to very few synchrotron beamlines and x-ray free electron lasers.
- In almost all practical implementations of micro-XRF, spatial information is obtained by raster scanning. This is a very slow process since the scan is done over every point of the sample. For large samples and for three-dimensional imaging, the measurement time is several days.
Can you explain the techniques used in computational ghost imaging?
In ghost imaging, two sets of data are required to reconstruct an image.
The first, which is called 'reference', contains a set of intensity fluctuation patterns (non-uniform beams), which are introduced into the incoming beam. This set of data does not contain any information about the object and once it is acquired it can be used for the reconstruction of an unlimited number of objects.
In our work, the intensity fluctuations were introduced by a standard sandpaper mounted on motorized stages and were recorded with a slow 2D camera. At each position of the stages, the beam hits a different area of the sandpaper and, since it contains random landscapes with different transmission, it introduces random intensity fluctuations.
After this step, we had the exact knowledge of the intensity fluctuations for each position of the stages. Later, the 2D camera was removed, and the object and a detector that identifies the energy of the radiation were inserted into the beam.
We then scanned the stages to irradiate the object with the intensity fluctuation patterns introduced at the various positions of the sandpaper and measured the emitted radiation from the object by using the energy detector.
By defining carefully the range of energy related to the iron and cobalt emission lines, we separately measured the total emission from each element. This second set of data, which we denote as the 'test', contains the information on the object, but this is a single number per element for each measurement, thus does not provide sufficient information for the 2D image reconstruction.
The image can be reconstructed by correlating the two sets of data for each position of the object. We produced the chemical mapping by reconstructing each element separately and merging them together.
Can you tell the readers more about your team’s focus-free technique that uses ghost imaging to speed up X-ray fluorescence chemical mapping?
We have extended the capabilities of the x-ray ghost imaging technique to measure chemical mapping and addressed the challenge of detecting two-dimension chemical mapping.
Schematics of the ghost fluorescence setup - The x-ray beam irradiates a mask with an inhomogeneous transmission that induces intensity fluctuations in the beam. When each different illuminating structure hits the object the x-ray fluorescence is measured by a photon-energy-resolving detector. Next, by correlating the data of the known structures with the fluorescence measurement and by using a compressed sensing algorithm the chemical element mapping is reconstructed. Image Credit: Bar-Ilan University
The process of x-ray imaging in general requires a system with capabilities to resolve the spatial distribution of the radiation that is transmitted by the object and collected by the detector. This is usually carried out using a pixelated detector (a camera) where each pixel provides the intensity level at the specific position. The image is reconstructed from the intensity distribution as measured by the detector. However, in contrast to x-ray-transmitted radiation, which propagates relatively parallel, the emitted radiation scatters in all directions so the image measured by the camera will be completely blurred.
Hence, measuring each point separately is required, which slows the process significantly and limits the resolution and focusing capabilities.
In this work, we have shown how to overcome this challenge and demonstrated that the spatial resolution is determined by the feature size of the intensity fluctuations of the irradiating beam by using the ghost imaging approach. Moreover, the scanning points when using the GI approach could be much smaller than the number of pixels in the chemical maps and thus the measurement is faster.
What key benefits does your new technique offer and how could it be applied in the field?
We have developed a new method to detect chemical mapping. So far, x-ray chemical mapping measurements were limited to applications where the duration of the measurement is not crucial. The technique we have demonstrated is fast, low cost, simple and robust. Therefore, it opens the possibility for measurements of chemical mapping for new field-like medical imaging.
Did you experience any key challenges when developing your technology and how were they overcome?
One of the most challenging parts of my work was to deal with noises on the system. One of the elements we imaged was iron, which is very common in almost every component in the system, including screws and holders.
The detected fluorescence of iron includes the signal from the object and the fluorescence from other components. In addition, some overlaps exist between the different photon energies emitted from the object; among themselves and with the photon energy of the input beam. Finally, we filter out the noises by measuring the noise and normalizing the result. We carefully chose the spectral areas where the noise is less dominant.
How might medical imaging in particular benefit from your research?
Our technique can form the foundation for medical x-ray imaging systems with high contrast and high resolution. In x-ray medical imaging, the low contrast between various tissues is the main challenge. Using our technique opens the possibility of obtaining colored x-ray medical images so that the difference between tissues is not only the different absorption but in the elements that make them up. This will be a new diagnostic tool for physicians.
How do you expect X-ray fluorescence to develop in the future?
I expect that the CGI technique will be embedded in many applications of XRF. Since the addition to the existing XRF system is just a mask, the application of the technique is simple, and the benefits are significant. The parameters of the mask and the algorithm could be optimized to achieve even more impressive resolution and performance.
What are the next steps for the project?
We have some directions for the future. Firstly, we want to show the effect at much higher photon energies where the focusing is even more challenging. We also want to extend our procedure to three-dimension reconstruction, where we believe the improvement in measurement time will be very significant.
Do you have any other research that you would like to discuss?
More information on my first demonstration of CGI with x-rays can be found here:
https://opg.optica.org/oe/fulltext.cfm?uri=oe-27-3-3284&id=404491
Where can readers find more information?
All the information about our work is included in the paper: https://opg.optica.org/optica/fulltext.cfm?uri=optica-9-1-63&id=468456
About Yishai Klein
Yishai (35) lives in a small village in Israel with his wife and five children. He graduated from Open University and is currently working as a research physicist in the field of x-ray optics.
Yishai is reaching the end of his direct Ph.D. program at Bar-Ilan University, supervised by Prof. Sharon Shwartz. He was granted a prestigious scholarship from the Israeli Ministry of Science & Technologies to develop a new method of x-ray imaging.
During his Ph.D., Yishai has performed successful experiments demonstrating novel approaches for x-ray imaging and spectroscopy in the lab, synchrotrons, and x-ray free-electron lasers around the world.
Biography and banner portrait image credit: Ben Rapaport
Disclaimer: The views expressed here are those of the interviewee and do not necessarily represent the views of AZoM.com Limited (T/A) AZoNetwork, the owner and operator of this website. This disclaimer forms part of the Terms and Conditions of use of this website.