Harnessing Light: Measuring the color of a single photon

Image courtesy of Miriam Kopyto

Photons, little packets of light, are everywhere. By some estimates, the Earth is bombarded with 1035 photons from the sun per second. Considering that there are 1020 grains of sand on earth, the number of photons is estimated to be about a quadrillion earths’ worth of sand. Despite being bathed in an essentially steady stream of light, we have the technology to isolate and measure the presence of individual photons. Such technology is aptly referred to as “single photon counters.” While trying to measure single photons may seem contrived and unimportant, photon counting is in fact central to astronomy, bioluminescence, medicine, genomics, quantum communication, and materials science, among other fields.

While we have been counting photons for decades—on the Hubble Space Telescope, for instance—engineering limitations have made it near impossible to measure the wavelength of a single photon. The wavelength of a photon, or the distance between successive peaks or troughs in the wave, translates into the color of light, which is the photon’s most important physical property. Recently, a research team led by Hong Tang—the Llewellyn West Jones, Jr. professor of electrical engineering at Yale—and Risheng Cheng—a graduate student in the Tang lab—engineered a single-photon spectrometer that can detect both the presence and wavelength of an individual photon.

Frequency Asked Questions

The goal of spectroscopy is to study the interactions between matter and light. There is a plethora of spectroscopic devices, which are commonly used to measure absorption and emission of chemical compounds. Modern spectrometers are highly sensitive and can measure light from stars millions of light years away, allowing researchers to gain insight on the spectrum of electromagnetic rays that radiate from celestial bodies. The next logical step would be to marry spectroscopic capabilities with single photon counting technology, yet this step has proven more challenging than expected. Attempts to simultaneously detect a single photon without ignoring its spectral information have resulted in resolutions too low for accurate measurement and machinery too bulky for practical use. Single photon spectrometers would benefit fields such as astronomy, which rely on very weak sources of light and a limited source of photons. “The single photon is a very weak energy packet. Not all conventional detectors can detect a single photon,” Cheng said. But as with all engineering, the goal is to extract as much information as possible out of a single measurement.

Devices known as superconducting nanowire single photon detectors (SNSPDs) are state-of-the-art in single proton detection. Current technology depends on the use of either large wavelength scanners and signal amplifiers or semiconductor-based single photon counters with a small number of detection channels. These methods present a tradeoff between photon sensitivity and spectral sensitivity. “Semiconductors all amplify single photon energy thousands or millions of times [to an electrically detectable level], but you lose all information on different photon energies and wavelengths,” Cheng said. This equipment can also succumb to a phenomenon known as “dark count,” in which detectors report a photon despite an absence of light. Dark counts are caused by electrodes releasing electrons at high temperatures.

An Electrifying Solution

Tang and his collaborators recently reported a novel solution that could be implemented at the millimeter scale. “There was a lot of design simulation, application, and testing,” Tang said. “There are many stringent requirements and critical design parameters.” Their broadband on-chip single-photon spectrometer does not suffer from any of the issues mentioned prior—a large amount of effort was put into finely tuning its spectroscopic sensitivity. The device is capable of measuring wavelengths ranging from six hundred to two thousand nanometers, encapsulating both infrared and visible light regions. This feat was achieved through the coupling of two distinct circuits: one to disperse and diffract the light, and the other to detect the light. Incredibly, all of this circuitry was contained in a single chip.

The first circuit contains perhaps the most important mechanical part of the device, known as an echelle grating. A type of diffraction grating, echelle grating is comprised of a periodic series of grooves that split photons based on their wavelength, a common feature of spectrometers. Echelle grating was used for its particularly high dispersive power. Following diffraction, photons are scattered outwards—like a garden hose spray with low to high wavelength photons from one end to the other— towards a curved superconducting nanowire. Depending on its wavelength, the diffracted photon can strike various points on the nanowire. Together, the grating and nanowire form a structure known as Rowland mounting. When a diffracted photon hits the nanowire, it generates a signal that is propagated in both directions towards detectors on either end. By measuring the time difference between these two signals and knowing the speed of these signals, the point on the wire where the photon struck, and in turn, the angle of diffraction, can be calculated. With the angle of diffraction, the wavelength of the photon can be determined.

Small Innovations Make All the Difference

To increase the resolution of the spectrometer, the researchers also equipped the superconducting nanowire with a microwave delay line, which is capable of reducing propagation speed along the wire to 0.73 percent of that in a vacuum, the slowest speed reported among superconducting nanowire delay lines. This attenuation in velocity is possible due to the capping of the wire with insulating aluminum oxides. By reducing the velocity of the signal, the time delay is naturally amplified, translating into higher resolution.

To construct the nanowire, the researchers employed a method called atomic layer deposition. A material base is exposed to multiple chemical precursors sequentially. Each of these precursors then reacts and attaches to the material surface in a self-contained manner. This process is highly efficient and can be used to develop pure superconductive nanowire. “If there are any defects in the device, it will not work as we would like. So far, this is the best [nanowire] that can be grown,” Cheng said. Testing various frequencies of light ranging from 600 nanometers to 1970 nanometers within the broadband device, the researchers found that the spectrometer can differentiate between two wavelengths at least seven nanometers apart, which is a remarkable level of sensitivity for an instrument on a chip.


The researchers considered various applications as they completed this device. One potential use is the detection of biological stains, multicolored dyes that typically mark particular aspects of tissues and cells. “In most cases, biological luminescence occurs at the single photon level,” Cheng said. Being able to accurately and reliably measure the strength of a small number of photons could allow for a more detailed investigation into human tissue. Additionally, while there are a variety of medical imaging modalities, many—including X-ray—rely on the detection of high energy particles. Implementation of more efficient photon counting chips can be used to reduce radiation doses while maintaining image quality and improving the signal-to-noise ratio.

Another field that could benefit from this research is quantum communication, a new way of transmitting information. In quantum communication, detectors take advantage of the superposition of quantum bits, which is the ability of to represent a combination of 0 and 1 simultaneously. The spectrum could be divided into several different channels, enabling parallel transmission of the signal in different wavelengths. With this technology, the communication capacity within a single photon detector could be increased one hundred-fold. “If you use conventional transmission with one fiber, you’ll need ten detectors to receive ten wavelengths,” Cheng said. “In our detector, we only need one recipient to differentiate these wavelengths, reducing the cost.” With the advent of this on-chip single-photon spectrometer, it is now possible to extract even more information from individual photons, and its small scale will no doubt facilitate its application to bringing various dark corners of nature to light.

Extra Reading:

Cheng, C., Chang-Ling, Z., Guo, X., Wang, S., Han, X., Tang, H. 2019. Broadband on-chip

single-photon spectrometer. Nature Communications, 10.