The shortest time interval ever directly measured is about 247 zeptoseconds, where one zeptosecond is a trillionth of a billionth of a second. Researchers in Germany achieved this in 2020 by tracking how long a single photon took to cross a hydrogen molecule. That number is almost absurdly small, but it sits on a ladder of progressively tinier time slices that physicists have been climbing for decades, and the theoretical floor may be smaller still.
How 247 Zeptoseconds Was Measured
The record-setting experiment, published in the journal Science, worked by firing X-ray photons at a hydrogen molecule and watching what happened to the electrons inside it. A photon traveling at the speed of light still needs a tiny but nonzero amount of time to move from one hydrogen atom to the other. The researchers used an interferometric technique: they measured the wave patterns created by two electrons ejected from the molecule’s two atoms and worked backward to calculate the time delay between those ejections. Because they knew the orientation and bond length of the molecule, the interference pattern gave them a precise time of about 247 zeptoseconds for the photon to make the trip across the average bond length of molecular hydrogen.
The key insight was that even within a single molecule, an electron is not launched from both atomic centers at the same instant. The birth time of each ejected electron depends on how long the photon takes to travel across the molecule. That difference, vanishingly small as it is, showed up clearly in the data.
1Science. Zeptosecond birth time delay in molecular photoionizationThe Ladder From Femtoseconds to Zeptoseconds
To appreciate how physicists reached the zeptosecond regime, it helps to see the milestones that came before it. Each jump down the time scale required fundamentally new tools and ideas, not just faster versions of old ones.
Femtosecond science, dealing in quadrillionths of a second, took off in the late 1980s and 1990s. Ahmed Zewail’s group pioneered a technique called femtosecond transition-state spectroscopy, which used two ultrafast laser pulses: one to start a chemical reaction and a second, fired a controlled fraction of a second later, to catch the molecules in the act. This allowed the first real-time observations of chemical bonds breaking and forming, processes that happen on the scale of about ten to a few hundred femtoseconds.2CaltechAUTHORS. Femtosecond transition-state spectroscopy of chemical reactions The work earned Zewail the Nobel Prize in Chemistry in 1999 and gave birth to the field of femtochemistry. Early experiments captured things like the breaking apart of iodine cyanide molecules and the dissociation of bismuth molecules, watching atoms separate in real time.3The Journal of Chemical Physics. Femtosecond transition-state absorption spectroscopy of Bi atoms produced by photodissociation of gaseous Bi2 molecules
Attosecond science pushed the frontier another thousand-fold shorter. One attosecond is a billionth of a billionth of a second, or a thousandth of a femtosecond. To generate light pulses that brief, researchers used a process called high harmonic generation, where an intense laser pulse interacts with a gas to produce bursts of extreme ultraviolet light lasting just tens to hundreds of attoseconds. This work earned Pierre Agostini, Ferenc Krausz, and Anne L’Huillier the 2023 Nobel Prize in Physics. Attosecond pulses let scientists watch the motion of electrons inside atoms and solids, processes invisible to any slower technique.
And then came zeptoseconds. The 2020 hydrogen-molecule experiment didn’t rely on generating a zeptosecond-long pulse of light. Instead, it used the known speed of light and the tiny distance across a molecule as a kind of built-in stopwatch, extracting the time from the spatial pattern of ejected electrons. The approach was clever precisely because no one can yet build a laser that fires zeptosecond-scale pulses.
What Attosecond and Zeptosecond Measurements Reveal
These ultrashort timescales are not just records for the sake of records. They open windows into processes that were previously too fast to observe directly. At the attosecond level, researchers have begun probing the internal dynamics of solid materials. When light hits a metal surface and ejects an electron, that electron doesn’t pop out instantly. It takes a tiny but measurable amount of time, and that delay varies depending on the material and which electron band the particle came from.
Experiments on copper and nickel surfaces, for instance, have found that the time delay between photoelectrons emitted from different valence bands can reach roughly 220 attoseconds at certain angles and drop to about 30 attoseconds at others. These measurements represent a direct look at electron-electron interactions happening on timescales that were purely theoretical a generation ago. The photoemission time for electrons from copper surfaces was generally found to be about 100 attoseconds longer than for nickel, a difference that reflects the distinct electronic structures of the two metals.4IOP Publishing. Attosecond light science and its application for probing quantum materials
This kind of measurement matters for understanding and eventually designing electronics and quantum devices at their most fundamental level. If you want to control how electrons move through a material, you need to know how fast they move in the first place, and “how fast” in this context means attoseconds.
Can Anything Be Shorter Than a Zeptosecond?
Yes, at least in theory and in some specialized predictions. Physicists have proposed that even shorter time intervals exist in extreme environments. One striking example comes from the quark-gluon plasma, the ultra-hot soup of fundamental particles produced in heavy-ion collisions at facilities like the Relativistic Heavy Ion Collider. Theoretical work has shown that high-energy photon pulses produced during the initial expansion of such a plasma could have durations on the yoctosecond scale, where one yoctosecond is a thousandth of a zeptosecond.5PubMed. Yoctosecond photon pulses from quark-gluon plasmas Nobody has directly measured a yoctosecond-scale event the way the German team measured 247 zeptoseconds, but the physics strongly suggests that processes at these scales are real.
Below yoctoseconds, the next named unit is the Planck time, roughly 5.4 × 10⁻⁴⁴ seconds. This is not just another step down the ladder. It represents a theoretical boundary below which our current understanding of physics breaks down entirely. The Planck time is defined by combining three fundamental constants: the speed of light, the gravitational constant, and the quantum of action. At this scale, quantum effects and gravity become inseparable, and the usual rules of spacetime may no longer hold.6arXiv. Physics of Limit Values at Planck scale No experiment has come remotely close to probing times this short, and it remains an open question whether “shorter than the Planck time” is even a meaningful concept.
The Tunneling Time Debate
One of the most fascinating uses of ultrashort time measurement has been the attempt to answer a surprisingly contentious question: when a quantum particle tunnels through a barrier, how long does it spend inside?
Quantum tunneling is the phenomenon where a particle passes through an energy barrier it shouldn’t classically be able to cross. It happens constantly in nature and is the basis for technologies like tunnel diodes and scanning tunneling microscopes. But whether the particle spends any measurable time inside the barrier, or whether it somehow traverses the barrier instantaneously, has been debated for decades.
The experimental evidence is genuinely conflicting, which is unusual for a question that seems so straightforward. A 2017 experiment using the “attoclock” technique on helium atoms found clear evidence for a nonzero tunneling time delay, suggesting the electron does spend a small but real amount of time inside the barrier.7PubMed. Experimental Evidence for Quantum Tunneling Time But a 2019 experiment applying the same attoclock approach to atomic hydrogen, a simpler system, reached the opposite conclusion. That team put an upper limit of 1.8 attoseconds on any tunneling delay and argued that the angles previously attributed to finite tunneling times were actually caused entirely by the Coulomb potential, the electric field of the atomic nucleus pulling on the departing electron.8Nature. Attosecond angular streaking and tunnelling time in atomic hydrogen
A 2022 study attempted to resolve the question using a more comprehensive streaking method and concluded that the tunneling time is very close to zero, in agreement with the hydrogen result. The researchers found that the most probable electron emission angle almost exactly coincided with the peak of the driving laser field, which is what you would expect if the electron exits the barrier essentially at the same instant the field is strongest.9Light: Science & Applications. Full experimental determination of tunneling time with attosecond-scale streaking method
The upshot is that most recent evidence tilts toward tunneling being practically instantaneous, or at least taking less time than current tools can resolve. But the 2017 helium result has not been formally overturned, and the theoretical community continues to debate what “tunneling time” even means, since different valid definitions of the concept give different answers. This is one of those areas where a question that sounds simple turns out to probe deep and unresolved issues in quantum mechanics.
Quantum Speed Limits
Separate from tunneling, there is a broader theoretical question about how fast any quantum process can happen. These are called quantum speed limits, and they put a floor on how quickly a quantum system can evolve from one state to another. Two foundational results set the terms of the debate. The Mandelstam-Tamm bound ties the minimum evolution time to the energy uncertainty of the system: the more uncertain the energy, the faster the system can change. The Margolus-Levitin bound ties it instead to the system’s average energy: more energy means faster possible change.10Journal of Physics A: Mathematical and Theoretical. Quantum speed limits: from Heisenberg’s uncertainty principle to optimal quantum control
These limits are not just abstract curiosities. They set hard boundaries on how fast a quantum computer can perform operations, how quickly energy can be transferred in quantum thermodynamic processes, and how rapidly information can be communicated in a quantum network. In a sense, quantum speed limits define the ultimate clock speed of nature itself, below which no physical process can operate regardless of how clever the engineering is.
Is Time Itself Made of Smallest Pieces?
All of the measurements and limits described so far treat time as a continuous quantity that can in principle be divided endlessly. But some theoretical frameworks challenge this assumption. The “chronon” model, for instance, proposes that at a microscopic level, time advances in discrete steps rather than flowing smoothly. In this picture, macroscopic time is still continuous for all practical purposes, but at some fundamental scale, time ticks forward in indivisible units, much like how matter appears smooth at large scales but is made of atoms at small scales.11Symmetry. Complementary Continuous-Discrete Time, Chronon Layering and Temporal Folding
Whether time is truly discrete or continuous remains an open question. Our best-tested theories, quantum mechanics and general relativity, both treat time as continuous. But these theories are known to be incomplete, since they don’t agree with each other at extreme scales. A future theory of quantum gravity might reveal that time has a granular structure at the Planck scale or something near it. For now, the question is theoretical, because no experiment can probe times anywhere close to the Planck time. The 247-zeptosecond measurement, record-breaking as it is, sits roughly 20 orders of magnitude above the Planck time. That gap is larger than the one separating human-perceptible time from the zeptosecond record.
Precision Clocks and the Art of Counting Time
Measuring extremely short durations is a different challenge from keeping extremely accurate time, though the two are related. The most precise clocks in the world are optical lattice clocks, which trap atoms in grids made of crisscrossing laser beams and count the oscillations of light as atoms transition between energy states. A strontium optical lattice clock, for example, measures a specific atomic transition at a frequency of about 429 trillion oscillations per second, pinning down the tick with a fractional uncertainty comparable to the best optical clocks ever built.12PubMed. Accurate optical lattice clock with 87Sr atoms
These clocks don’t measure zeptoseconds, but they define the second with such precision that they can detect the gravitational time dilation caused by lifting the clock just a centimeter higher in Earth’s gravity. The technology behind them, laser control of individual atoms, is closely related to the ultrafast laser techniques used in attosecond and femtosecond science. Advances in one field feed directly into the other.
How Human Perception Compares
It is worth stepping back to appreciate how far removed all of this is from human experience. Your visual system does not operate on anything close to these timescales. When you see something happen, the perception is established in areas of your brain’s visual cortex roughly 150 to 200 milliseconds before you consciously react to it.13PubMed Central. Estimation of the timing of human visual perception from magnetoencephalography That is not a single number, either. Your visual system has multiple temporal limits set by different underlying mechanisms, and these split into at least two groups: a fast group that handles things like motion direction and depth edges, and a slower group for tasks like recognizing words or pairing color with motion.14Trends in Cognitive Sciences. Multiple temporal limits in vision
When it comes to actively judging durations rather than just reacting, people can discriminate time intervals in the range of 150 to 1,500 milliseconds with a precision of about 5 percent, meaning you can tell two half-second intervals apart if one is roughly 25 milliseconds longer than the other. That threshold shifts depending on whether the signal arrives through your eyes, ears, or skin.15PubMed. Discrimination of short time intervals by the human observer Your ears are generally better at time discrimination than your eyes, which is one reason why musicians rely on auditory rather than visual cues for rhythm.
The gulf between these human-scale time limits and the zeptosecond record is staggering. A single blink of your eye lasts about 300 milliseconds. The shortest time interval you could hope to consciously distinguish is tens of milliseconds at best. The zeptosecond measurement is roughly 10¹⁸ times shorter than that, the same ratio as one second compared to several times the age of the universe. We can talk about and measure zeptoseconds, but we will never directly experience them. They belong to a domain of reality that only instruments can access, where photons cross molecules and electrons are born from atomic orbitals on timescales that make the blink of an eye look geological.