We are at an inflection point in our control of light, beyond 2D transverse intensity patterns and towards tailored light in space and time, for complete 4D control. When new degrees of freedom are added to the mix, the potential is enormous. It is novel spatiotemporal optical wavepackets that are lighting the way to this exciting future.

Controlling light can be traced back thousands of years, with stories of directing sunlight from mirrors to burn attacking ships, an early form of incoherent light shaping^[1]. In this example, when light is added to light, the outcome is proportionally more light. This paradigm is broken when the light can be treated as coherent waves: light added to light can result in darkness. Thomas Young did exactly this to create spatial intensity structure in the form of “fringes”. Moving beyond just two displaced splits, his notion of fringes can be generalized to any geometry and any degree of freedom^[2]. His experiment revealed just how easy it is to control the spatial structure of light by simply adding plane waves, initially in the transverse plane for 2D structured light in intensity, but now in more abstract degrees of freedom of light^[3].

The power and ease with which light can be structured in space is best illustrated by perhaps the simplest of beam-shaping elements, the lens. In a single element, positions can be mapped to angles and angles to positions, and equivalently, space to spatial frequency and vice versa, courtesy of a Fourier transforming action. With a little care to distances before and after the lens, planes can be made spatially correlated, essential for imaging. The ability to spatially structure light in any amplitude and phase exploded in the early 1990s with the advent of diffractive optical elements, ushering in control of light by propagation phase using depth-structured optics, but this was expensive and time-consuming. The field of structured light became accessible to all with the availability of commercial spatial light modulators and digital micromirror technologies, opening a path to the digital on-demand control of structured light^[4,5]. In an interesting reversal of direction, metasurfaces have since become highly topical^[6], bringing back prefabricated optics. Such nano-scale optics has allowed new control of light by geometric phase as well as polarization-dependent propagation phase, new and essential tools for the structured light toolbox. This has facilitated the creation of new forms of light, now well documented in many review articles^[7–11].

The tailoring of light in time had a slower start. The first obstacle is fundamental: optical frequencies are very fast, too fast for any direct modulation of the wave in time. Controlling the frequency spectrum instead suffers from practical obstacles: lenses do not Fourier transform the time component, so dispersive elements such as gratings are needed. However, most coherent sources of light are narrow bandwidth, making such control difficult and severely limiting. Wide bandwidth coherent sources were introduced into the mainstream when the late Wilson Sibbett demonstrated Kerr mode locking in Ti:Sapphire lasers^[12,13], quoting that he knew this was going to be a success when he saw the commercial members of the audience race out after his talk to get started. This launch point coincided with the explosion in spatial shaping tools, so that at last, time could catch up. The digital spatial technology of spatial light modulators could now be used with dispersive elements such as gratings to independently control each frequency component before recombining them in what was called a “4f shaper” at the time, largely pioneered by the late Andrew Weiner^[14,15]. This led to the emergence of the ultrafast community, in particular femtosecond chemistry^[16], bringing temporal control to pump-probe spectroscopy for studying and controlling complex systems, leading to the Nobel prize for chemistry.

At the time, it was well known that optical elements such as lenses could lead to coupling of the space and time degrees of freedom, but this was considered more of an unwanted hindrance than a resource to be harnessed. In recent times, the two fields have merged, making it possible to structure light in both space and time. Mirroring the coherent control approaches of decades past, spatially tailored vectorial light, can imprint its structure onto matter using temporal control as an enabling mechanism^[17,18]. Similarly, light’s degrees of freedom can be imprinted into the time domain, for instance, for light with time-varying orbital angular momentum^[19], used recently to create accelerating temporal trajectories of light^[20]. The coupling of space and time has been explored as a new resource for non-separable states of light in space and frequency^[21], a classical analogue to quantum entangled states. The technology of digital control not only can benefit the temporal control of light but also can lessons from spatially structured light, for instance, enhancing the control in the frequency domain^[22] by leveraging tools from Bessel beams. Space-time coupling has been fully embraced, with the notion illustrated in Fig. 1.

These advances have given rise to exotic forms of light in 4D, which collectively can be grouped under spatiotemporal optical wavepackets, the subject of a very comprehensive review by Liu et al. in Photonics Insights^[23]. The review begins with an overview of structured light, spatial and temporal, before moving on to the advances made in recent years, from tools for creation, detection, and control, to fundamental and applied demonstrations. The emphasis is on the coupling of space and time, where the whole is more than just the sum of the parts. The authors neatly capture the myriad tools now available, using illustrative examples to highlight the importance of each. A major challenge in the field is the analysis of such complex forms of light, a topic that has matured vastly in a very short time, yet still requires more effort to make it as ubiquitous as its spatial counterpart.

To illustrate a key message of the review, consider the perspective of creating new forms of light beyond those of the textbook. What has held the field back for so long? One answer is that it is difficult to induce the necessary higher-order multipoles (beyond dipoles) and toroidal excitations in matter^[24]. Here, it can be parallel to the spatial domain: just as rare quadrupole transitions held back the ability to create light with orbital angular momentum (OAM)^[25], exotic higher-order and toroidal multipoles also held back new radiative states of light. The new spatiotemporal toolbox has been essential in moving the community forward, for example, borrowing notions from controlling OAM, such as conformal mapping and spiral phases, to produce these exotic fields without the need for exotic materials.

The scope of the review also makes clear that there are gaps yet to be filled. So far, spatiotemporal light has been demonstrated as classical wavepackets, and here mostly from external control and not directly at the source. In the quantum realm, spatial control has led to high-dimensional entangled states, while temporal control has seen high-bit-rate quantum communication. A promising avenue is to explore the interface of these exciting directions. For the most part, spatiotemporal control has been used to demonstrate new forms of light, probing the underlying physics and opening new research directions. True applications have been slower to materialize, with promising laboratory-based experiments starting to show potential. These examples highlight that, rather than a closed book, spatiotemporal light has many new chapters to write.