Observing light information during atomic time processes to accurately reveal the physical, chemical, and biological phenomena and their evolution during atomic motion has always been a dream of scientists. Atomic time imaging with the sound intrinsic spatial resolution is applicable to studying ultrafast transient events in ultrafast physics, ultrafast chemistry, and ultrafast biology. The events include ultrafast optophysical processes in semiconductor and quantum well microstructures, excitation of excitons and carriers by light, high-order harmonic effects, ultrafast dynamic processes in condensed matter, ultrahigh-intensity laser wake field acceleration, formation and breaking of chemical bonds, transfer of protons and electrons, the influence of molecular vibrations and rotations on chemical reactions, energy transfer processes in photosynthesis, photoisomerization processes in visual systems, and charge and proton transfer in DNA. Currently, only all-optical imaging can be employed to achieve atomic time imaging, where light itself records the modulated light field of transient events or dynamic processes. This is based on the duality of light as both an information carrier and a research resource. From the perspective of the light field, the amplitude, phase, wavelength, wave vector, and polarization of the light field are included. From the perspective of photons, photon energy, momentum, spin angular momentum, orbital angular momentum, and nonlinear and quantum properties of photons are contained.

Ultrafast atomic time imaging has seen significant developments in recent years. In femtosecond holographic imaging, Dr. Martin Centurion was the first to achieve multi-frame encoded femtosecond time-domain holography (Fig. 5), while Dr. N. H. Matlisi first utilized chirped pulses and spectral interferometry to record the frequency-domain holography of laser-induced plasma wake fields (Fig. 6). In scanning streak tube compressed ultrafast photography (CUP), Gao et al. proposed the CUP technique to achieve two-dimensional imaging of non-repetitive ultrafast luminescent phenomena (Fig. 8). This caused a sensation in the ultrafast imaging field, which opened up a new avenue for atomic time imaging via adopting compressed sensing algorithms, thus leading to the emergence of T-CUP, CUSP, and other related derivative technologies. In spectrum-plane encoded atomic time imaging, Ehn et al. proposed the frequency recognition algorithm for multiple exposures (FRAME), which enabled ultrafast multi-frame imaging with high spatial and temporal resolution (Fig. 9). Zhu et al. proposed the frequency domain integration sequential imaging (FISI) technique, achieving the highest space-bandwidth product in ultrafast imaging to date (Fig. 10). In spectral encoding femtosecond imaging, Nakagawa et al. proposed the sequential timed all-optical mapping photography (STAMP), with a maximum framing frequency of 4.4×10¹² frame/s, which was once considered the fastest photography in the world (Fig. 11), and this led to the development of technologies such as SF-STAMP (Fig. 12). The grating-sampling atomic time imaging technique (OPR) combines the grating sampling theory with spectral-time encoding technology by a grating plate, achieving all-optical high spatial and temporal resolution imaging with a grid principle of 2 trillion frames per second (Fig. 13). Multi-stage non-collinear optical parametric amplification (MOPA) idler imaging has parameters such as framing time, exposure time, spatial resolution, and frame size that are independent and unrelated, thus becoming an ideal imaging method. It yields sound effective framing frequency and high spatial resolution in single-shot atomic time-scale imaging (Fig. 14).

Further development of the information theory of atomic time imaging is needed to evaluate and develop atomic imaging technology. We preliminarily improve Schardin’s space-time information theory, explore the optimal atomic time imaging system that is not limited by the Heisenberg uncertainty principle, and always pursue shorter exposure time, finer intrinsic spatial resolution, and greater spatial bandwidth products. Atomic time imaging faces new challenges, including advancing studies on high-speed imaging and computational femtosecond imaging information theory, promoting the combination of femtosecond imaging and picometer spatial resolution technology, and exploring new principles of femtosecond imaging, new optimal imaging systems, and reliable, reasonable enhancement of existing imaging technology performance. Additionally, it is necessary to promote the application of atomic time imaging technology in photon materials, plasma physics, live cells, and neural activity, and to push the timescale from femtoseconds to attoseconds. The development of attosecond imaging already shows its initial signs. Currently, the imaging of the electron wave packet motion in neon atoms and electron motion capturing in nitrogen molecules have been achieved, and the temporal resolution of transmission electron microscopy has been pushed to the attosecond scale. By directly measuring the relationship between the electromagnetic functions of natural and artificial materials with space and time, attosecond electron microscopy provides indispensable information for a deep understanding of the fundamental mechanisms of light-matter interactions, and is expected to promote development in fields such as near-field optics, passive and active deformable materials, photonic integrated circuits, photoperiodic photochemistry, and free-electron cavity optics.