SMART researchers create world’s smallest LED and holographic microscopes that can transform existing cell phone cameras into high-definition microscopes
Disruptive and Sustainable Technologies for Agricultural Precision (DiSTAP) and Critical Analysis for Personalized Medicine Manufacturing (CAMP) of the Singapore MIT Research and Technology Alliance (SMART), an MIT research enterprise in Singapore. Research Group (IRG) researchers have developed the world’s smallest LED (Light Emitting Diode) that can transform an existing cell phone camera into a high-definition microscope.
A new LED, smaller than the wavelength of light, will be used to build the world’s smallest holographic microscope, which can be used to replace existing cameras found in everyday devices such as mobile phones, simply by changing silicon chips and software. paved the way for converting the to a microscope. This technology also represents a significant advance in diagnostic miniaturization for indoor farms and sustainable agriculture.
This breakthrough is complemented by the researchers’ development of innovative neural networking algorithms that can reconstruct objects measured by holographic microscopy, allowing cells and bacteria to be visualized without the need for bulky conventional microscopes or additional optics. Allows for enhanced inspection of microscopic objects such as This research also paves the way for major advances in photonics. This is the construction of powerful on-chip emitters smaller than a micrometer, which has long been a challenge in the field.
Light in most photonic chips originates from off-chip sources, reducing overall energy efficiency and fundamentally limiting the scalability of these chips. To address this issue, researchers have developed on-chip emitters using various materials such as rare-earth-doped glasses, Ge-on-Si, and heterogeneously integrated III-V materials. Emitters based on these materials have shown promising device performance, but integration of their fabrication processes into standard complementary metal oxide semiconductor (CMOS) platforms remains challenging. Silicon (Si) has shown potential as a candidate material for nanoscale, individually controllable emitters, but Si emitters have low quantum efficiency due to the indirect bandgap, limiting the availability of materials and fabrication tools. This fundamental drawback, coupled with the limitations set by Realization of small native Si emitters in CMOS.
In a recently published Nature Communications paper, “Sub-wavelength Si LEDs integrated into a CMOS platform,” SMART researchers describe the development of the smallest reported Si emitters with light intensities comparable to state-of-the-art Si. explained. An emitter with a much larger emitting area. In a related breakthrough, SMART researchers, in a paper titled “Simultaneous Spectral Recovery and CMOS Micro-LED Holography Using Untrained Deep Neural Networks,” describe a new training technique that can reconstruct images from holographic microscopy. We have also revealed the construction of a deep neural network architecture that has not been developed. Recently published in Optica magazine.
A new LED developed by SMART researchers exhibits high spatial intensity (102 ± 48 mW/cm2) at room temperature and has the smallest emitting area (0.09 ± 0.04 μm2) of all known Si emitters in CMOS An integrated sub-wavelength scale LED. scientific literature. To demonstrate its practical potential, the researchers will integrate the LED into an in-line, centimeter-scale, all-silicon holographic microscope that does not require lenses or pinholes and is essential in a field known as lensless holography. Integrated.
A common obstacle faced in lensless holography is the computational reconstruction of the imaged object. Conventional reconstruction methods require detailed knowledge of the experimental settings for accurate reconstruction and are sensitive to variables that are difficult to control, such as optical aberrations, presence of noise, and twin image issues.
The research team also developed a deep neural network architecture that improves the quality of image reconstruction. This novel, untrained deep neural network incorporates total variation regularization to enhance contrast and takes into account the wide spectral bandwidth of the source. Unlike traditional methods of computational reconstruction that require training data, this neural network eliminates the need for training by embedding a physical model within the algorithm. In addition to holographic image reconstruction, the neutral network also provides blind source spectral recovery from a single diffraction intensity pattern.
The untrained neural network demonstrated in this work allows researchers to manufacture via fully commercial, unmodified bulk CMOS microelectronics without any prior knowledge of the source spectrum or beam profile. , can use novel light sources such as the novel and smallest known Si LEDs above.
The researchers believe that this synergistic combination of CMOS microLEDs and neural networks can be used in other computational imaging applications, such as miniature microscopes for live-cell tracking and spectroscopic imaging of biological tissues such as living plants. is assumed. This work also demonstrates the feasibility of next-generation on-chip imaging systems. In-line holographic microscopy has already been adopted for a variety of applications including particle tracking, environmental monitoring, biological sample imaging and metrology. Further applications include arranging these LEDs in his CMOS to generate programmable coherent illumination for future more complex systems.
Iksung Kang, lead author of the Optica paper and research assistant at MIT at the time of the study, said: For example, these LEDs can be combined in arrays to achieve the high levels of illumination required for large-scale applications. Moreover, the low cost and scalability of microelectronics CMOS processes allows this to be done without increasing system complexity, cost, or form factor. This makes it relatively easy to convert your phone’s camera into this type of holographic microscope. Additionally, by leveraging the electronics available in the process, control electronics and imagers can be integrated on the same chip, creating an ‘all-in-one’ micro-LED that could potentially revolutionize the field. . “
“Besides the immense potential in lensless holography, our new LEDs have many other possible applications. With an emission area of 200 psi, our LEDs are ideal for bioimaging and biosensing applications, including near-field microscopy and embedded CMOS devices.” Principal Investigator at SMART CAMP and DiSTAP and Professor of Electrical Engineering at MIT and Rajeev Ram, co-author of both papers. “It is also possible to integrate this LED with an on-chip photodetector, and find further applications in his testing of on-chip communications, NIR proximity sensing, and photonics on-wafer.”
This research was conducted by SMART and supported under the National Research Foundation (NRF) Singapore’s Research Excellence and Technology Enterprises (CREATE) programme.
Subwavelength Si LEDs integrated into a CMOS platform, Nature Communications (Open Access)