Mid-infrared single-photon computational temporal ghost imaging
Wen Zhang, Kun Huang, Xu Wang, Ben Sun, Jianan Fang, Yijing Li, Heping Zeng · 2026
Read Mid-infrared single-photon computational temporal ghost imaging by Wen Zhang, Kun Huang, Xu Wang, Ben Sun, Jianan Fang, Yijing Li, Heping Zeng (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2605-25336.txt.
Introduction
Wen Zhang and colleagues build a mid-infrared ghost imaging system for fast waveforms at single-photon sensitivity. Infrared detectors are too slow and too deaf for such signals. The team converts the light upward in frequency and encodes it with patterns from a near-infrared modulator. Matching intensities against those patterns rebuilds waveforms at 80 picosecond resolution, beyond the detectors' speed and jitter. The system recovers waveforms below 0.1 photons per bit and, with compressive sampling, cuts capture time by over 90 percent. The approach could extend to terahertz sensing.
Catalog record
| CREATOR | Wen Zhang, Kun Huang, Xu Wang, Ben Sun, Jianan Fang, Yijing Li, Heping Zeng |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 39868 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2605.25336
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Mid-infrared single-photon computational temporal ghost imaging Wen Zhang, Kun Huang, Xu Wang, Ben Sun, Jianan Fang, Yijing Li, Heping Zeng Abstract The capture of transient optical waveforms is critical to reveal dynamical phenomena in various fields. However, fast and sensitive mid-infrared (MIR) measurements are typically limited by processing bandwidth and detection sensitivity of conventional infrared detectors. Here, we propose and implement a computational temporal ghost imaging system, which favors high-speed and high-sensitivity characterization of MIR temporal objects. The core process relies on high-fidelity nonlinear optical transduction for facilitating both the programmable structured illumination and frequency upconversion detection based on the high-performance near-infrared light modulator and detector, respectively. Consequently, the correlation between the recorded integral upconversion intensity and the designated encoding patterns allows one to reconstruct the MIR profiles with a temporal resolution of 80 ps, well beyond the intrinsic bandwidth or timing jitter of the involved detectors. Moreover, a record-high detection sensitivity is manifested by recovering single-photon MIR waveforms with an incident flux below 0.1 photon/bit. Additionally, faithful reconstructions at sub-Nyquist sampling rates are demonstrated using the compressive sensing algorithm, which can reduce the data acquisition time by over 90%. The presented paradigm features high timing precision, single-photon sensitivity, and efficient data sampling, which could be extended into far-infrared or terahertz regions to address pressing demands in fast and sensitive sensing. I Introduction High-speed detection of optical signals is essential across a wide range of applications that require the rapid and accurate processing of light-based information, such as fiber-optic communication Assefa2010Nature , photonic computing Shastri2021NP , ultrafast spectroscopy Mahjoubfar2017NP , and time-of-flight imaging Kim2021NN . In these applications, a high temporal resolution of the light detection is usually required to realize low time latency, fast data rate, and enhanced depth resolution. Nowadays, the detection bandwidth for optical detectors at telecom wavelengths has reached impressive values over 100 GHz Shi2024NP ; Wey1995JLT , with an aim to fulfill the expanding demand for data communication capacity in big data, cloud computing and artificial intelligence Shastri2021NP . However, the detection performances in other spectral regions are relatively lagging behind, especially in the mid-infrared (MIR) range Razeghi2014RPP . Specifically, the speed response of conventional MIR detectors based on mercury-cadmium-telluride (MCT) alloys is severely limited by the long carrier lifetime Razeghi2014RPP , which typically results in a bandwidth below GHz for commercially available products. Recently, MIR uni-traveling carrier (UTC) photodetectors based on InAs/InAsSb type-II superlattice Huang2022EDL and quantum-well infrared photodetectors (QWIPs) based on group III-V materials Palaferri2018Nature have been implemented to offer an operation bandwidth beyond 10 GHz, albeit with a smaller detection responsivity due to the reduced thickness of the constituent absorption layer. Additionally, narrow-bandgap MIR detectors have long been plagued by dark current and thermal perturbation, which severely limits the detection sensitivity at the room-temperature condition Wang2019Small . Consequently, enhanced sensitivity usually resorts to the cryogenic operation, like superconducting nanowire detectors Taylor2023Optica ; Pan2022OE ; Chen2021SB . Notably, emerging low-dimensional materials and novel nanophotonic structures have been investigated to realize room-temperature MIR detectors Wu2021NR ; Xue2024LSA ; Mu2024NP , yet the attainable sensitivity is currently far away from the single-photon level. Therefore, it remains a long-sought-after goal for direct MIR detectors to achieve high sensitivity and high speed, as well as being compatible with room-temperature operation. Alternatively, the temporal ghost imaging (TGI) technique provides an appealing solution to mitigate the stringent bandwidth requirement for high-speed signal detection Chen2013OE ; Ryczkowski2016NP . In this scenario, a reference beam without ‘seeing’ the temporal object is dynamically modulated with random or structured patterns, while a test beam after passing through the target is detected by a non-temporally-resolved ‘bucket’ detector Ryczkowski2016NP . Consequently, the fast temporal object can be reconstructed with a slow integrating photodiode based on the intensity correlation between the two beams, although neither of them independently carries the temporal characteristics of the object Tian2020PRAp ; Wu2023AO . Fundamentally, the TGI is analogous to the spatial counterpart from the perspective of the space-time duality Patera2018PRA , which favors a strong immunity to the waveform distortion due to noises and perturbations along the path between the object and detector Erkmen2010AOP ; Moreau2018LPR . To date, TGI has been demonstrated by using a variety of light sources, including spontaneous parametric downconversion (SPDC) photon pairs Denis2017JO ; Dong2016SR , chaotic laser sources Ryczkowski2016NP ; Wu2020OE , and thermal light sources Devaux2017JO ; Yao2018OL . Meanwhile, enhanced TGI variants based on temporal magnification Ryczkowski2017APLP and differential measurement Oka2017APL have been investigated to promote the temporal resolution and signal-to-noise ratio. In particular, computational TGI uses a pre-programmed light intensity modulator to generate a sequence of known illumination probe patterns, which removes the need for detecting the reference beam Devaux2016Optica ; Xu2018OE ; Chen2021OL . These developments in TGI have benefited a number of applications, such as long-distance communication Chen2021OL ; Yao2018OL , information cryptography Jiang2017SR , quantum device evaluation Wu2019OL , and high-speed spectroscopy Janassek2018PRAp . However, the aforementioned instantiations of TGI have mostly been restricted in the visible or near-infrared (NIR) regime due to the relatively easier accessibility of high-speed detectors or modulators. To leverage the full potential of TGI, it is desirable to extend the operation wavelength into the MIR region, as being fueled by the demand for fast infrared sensing in various fields such as high-throughput free-space communication Zou2022NC , time-resolved photoluminescence spectroscopy Julsgaard2020PR , and high-resolution depth tomography Fang2023LSA . In this context, the nonlinear frequency conversion strategy has been exploited in the TGI architecture, where the technical challenge in MIR detection can be mitigated by translating the infrared radiation into visible or NIR spectral region Wu2019Optica . In this approach, a two-color scheme is implemented by converting the probe temporal intensity pattern into the second-harmonic replica via a nonlinear crystal Wu2019Optica . The involved wavelength conversion allows the shift of the probe laser into a spectral region where ultrafast detectors are available, which facilitates the imaging of a temporal object at 2 \mu m. However, this technique requires an infrared chaotic source with random intensity fluctuations, and relies on a large number of measurement realizations to obtain a high-contrast profile of the temporal object. Additionally, the conversion efficiency of the harmonic generation depends on the laser intensity, which would impose difficulty in preparing high-power sources in the MIR spectral region. To this end, MIR computation TGI has been proposed and implemented based on the pre-programmed structured illumination Wu2024LSA , where the required MIR probe patterns are generated by the difference frequency generation (DFG) process from the fast-modulated NIR light. The presented method could in principle generate probe patterns at arbitrary wavelengths, and significantly improve the reconstruction speed by using the invertible Hadamard matrix Edgar2019NP . However, in these reported TGI modalities, the MIR measurement for the test beam through the temporal object has still been conducted by using conventional narrow-bandgap infrared detectors, thus precluding the realization of ultra-sensitive operation at the single-photon level. So far, MIR single-photon computational TGI has not yet been demonstrated, which urgently calls for effective solutions to realize high-speed intensity modulation and high-sensitivity photon counting in the MIR region Kutas2022AQT . In this work, we devise and implement an ultra-sensitive MIR computational TGI system, which uniquely integrates both the nonlinear downconversion