1. An anti-eavesdropping quantum network in development empowers quantum key exchange—in this case, an exchange of symmetric keys and individual photons between people in a network to encrypt messages so outside parties cannot view or access them. This work involves a time bin entanglement protocol.
The basic idea features a pump photon sent through an imbalanced interferometer, basically splitting it into a two-part wave function. The photon is then sent through a parametric downconversion crystal, where it is converted into two photons. In the study, these were then divided and sent to two parties.
Those two parties could each identify their photon in one of two detectors in the interferometer. When the phases of several interferometers are very precisely controlled, interferences can be detected and observed. This strongly protects all communication within the network.
2. Diagnosing diseases and monitoring treatment outcomes could get a lot easier and more precise, thanks to a new technology that dramatically improves the signal in a fluorescent sensor.
Called wavelength-induced frequency filtering, MIT engineers based the new technique on a laser setup that fluctuates excitation wavelengths at a specific frequency. In response, an implant containing nanosensors emits light that doubles that frequency. With the change, the fluorescent signal can be separated from background autofluorescence. This improves the signal, reduces noise, and increases the penetration depth from which the implant can be detected.
Fluorescent sensor signals are often too weak when implanted in tissue that’s not close to the body’s surface. The new technique produces a strong signal in tissue, as deep as 5.5 cm, in areas not typically accessible without invasive measures such as wired implants or surgery.
3. The tedious, time-consuming, and overall very expensive task of staring into a microscope to search for 2D monolayers in a sample is set to become much easier.
A new automated scanning device can actually detect monolayers with 99.9% accuracy. Jesús Sánchez Juárez, a Ph.D. student at the University of Rochester, developed the device using an existing AI neural network that is already programmed to recognize objects. From there, he and his team made the device replicable with a standard five-time magnification objective lens and a standard OEM camera.
The device can invert image colors in a sample’s monolayers to make them more visible, and basically easier for a computer system to separate them from substrates. Unlike the existing laborious manual-search task, this new system is based on inexpensive components and is so efficient it can already accurately process 100 images over one-square centimeter samples in under 10 minutes.
