1. What Do Scientists Count to Define a Second?
In the past, the length of a second was based on the movement of the Earth. In 1967, the 13th General Conference on Weights and Measures (CGPM) stopped using the motion of the Earth. Instead, it defined one second as the duration of 9,192,631,770 cycles of radiation from a cesium-133 atom. [1] Today, we say this by fixing the frequency of cesium-133 vibrations at 9,192,631,770 Hz. Each type of atom vibrates at its own fixed frequency. An atomic clock counts these vibrations.
The first cesium atomic clock was made in 1955 by Essen and Parry at the National Physical Laboratory in the UK. It was accurate to about one second every 300 years. [2] This way of defining the second has been used for over 50 years. The talk of replacing it comes from the invention of even more accurate clocks.
2. Why Can Light Clocks Measure Time More Precisely Than Cesium Clocks?
Cesium atomic clocks use microwaves with a frequency of about 9.2 GHz. Light clocks use light with a frequency tens of thousands of times higher. [3] The US National Institute of Standards and Technology (NIST) compares this to a ruler with finer markings. If the markings are closer together, you can read smaller differences in length. The International Bureau of Weights and Measures (BIPM) explains that some optical frequency standards have an uncertainty about two orders of magnitude smaller than the best current cesium fountain clocks. [4]
There is a possibility of reaching accuracy 1,000 times better than cesium. [3] However, this is a future outlook, not a value already achieved. Note that "tens of thousands of times" refers to the frequency multiplier, not the accuracy multiplier. Higher frequency does not automatically mean accuracy is tens of thousands of times better.
3. Does Time Pass Faster at Higher Altitudes, as Optical Lattice Clocks Showed?
A research team from RIKEN, the University of Tokyo, and Shimadzu Corporation placed optical lattice clocks at the ground floor and the 450 m observation deck of the Tokyo Skytree in 2020. They compared how time passed. The comparison showed that time passed slightly faster at the observation deck. [8]
According to general relativity, time passes faster in places with weaker gravity, such as higher altitudes. With the precision of these clocks, which have 18 digits of accuracy, they can detect time differences caused by height differences of just a few centimeters. [8] This statement describes the clock's capability; it is separate from the specific 450 m experiment. Light clocks are not just for making seconds; they are also tools for measuring gravity differences. Many types of such clocks exist, and we must choose which one to use for the new definition.
This matters for the definition of the second, too. As clock precision increases, we cannot ignore the effects of gravity and height. When comparing clocks, setting the conditions correctly is very important.
4. Should the New Second Use One Kind of Atom or Several Combined?
There are several candidate atoms. These include strontium, ytterbium, and lutetium atoms, as well as ions of aluminum, strontium, and ytterbium. [3] BIPM news lists ytterbium-171 and strontium-87 as the main candidates reported regularly. BIPM divides the options into three main groups. [4]
First, we could replace cesium with a single optical transition. This is simple but requires global agreement on which atom to use. Second, we could define the second using a weighted average of multiple optical transitions. The weights can be updated to keep up with technological progress. This is not a simple average, but a weighted one. Third, we could use a method based on fundamental physical constants. However, the uncertainty is currently too high for this to be practical. NIST also notes that aluminum ion clocks are promising because they are less affected by temperature. [3]
5. What Conditions Must the New Definition of the Second Meet?
BIPM lists conditions for the replacement: the new definition must immediately improve accuracy by 10 to 100 times, maintain continuity with the cesium definition, and ensure light clocks contribute to the calibration of International Atomic Time (TAI) with appropriate uncertainty. [4] [7] NIST adds a condition that results must agree across three or more laboratories with an uncertainty of less than one second in about 16 billion years. [3] Sixteen billion years is a level where the clock would not drift by even one second since the Big Bang (about 13.8 billion years ago). [3] This is a condition listed by NIST; the official international standard follows the Consultative Committee for Time and Frequency (CCTF) roadmap.
Where are we now? Resolution 5 of the 2022 CGPM urged that a preferred atom or combination be proposed at the 28th General Conference in 2026, aiming to adopt a new definition at the 29th General Conference in 2030. [5] However, the 2025 CCTF recommendation states that there is no agreement yet on whether to use a single atom or multiple, or which atom to choose. [6] Some criteria are still far off. For example, the condition that the uncertainty when comparing the same type of clock in different labs is less than 5 × 10⁻¹⁸ has not been met. Still, it is written that international cooperation could achieve this by 2030. [6] Adoption depends on the conditions and may happen in 2030 or later. [7]
6. How Does Japan Help Compare the Candidates for the New Second?
The optical lattice clock was invented by Hidetoshi Katori at the University of Tokyo. In 2022, data showed that optical lattice clocks from Japan’s NICT and NMIJ contributed 54% to International Atomic Time. [10] Strontium was adopted as a "secondary representation of the second" in 2006, and ytterbium in 2012.
In 2025, NICT and the Paris Observatory announced a new method to compare redefinition proposals. [9] They created graphs of frequency value spaces to intuitively compare single-atom proposals (like strontium or ytterbium) with multi-atom weighted proposals. Multi-atom proposals were hard for experts to understand. [9] How this method will be used in the actual selection process is still to be seen. However, it may help discussions about reaching an agreement by allowing people to look at the same diagrams.
7. How Well Can You Count a Second Yourself?
Look at the second hand on a clock or use a stopwatch. Count "one, two, three..." in your head up to 10 at an even pace, then compare it with the real 10 seconds. You will likely be slightly off.
BIPM explains that even if the definition changes, it will have almost no effect on daily time. Cesium clocks will continue to be used as "secondary representations of the second" after the redefinition. [4] Cesium fountain clocks are also expected to be needed during the transition period [6]. The choice of the next atom will move forward in discussions for the 28th CGPM in 2026 and towards 2030. Try counting 10 seconds three times. Check if your error is always in the same direction or if it varies each time.