The second has served us well for hundreds of years, but it's not perfect – the atomic clocks that are our official timekeepers lose a second of time every 200 million years or so.

That may not be enough to make you late for choir practice, but in the fields of space travel and physics even small errors can make big differences. With that in mind, efforts are underway to build clocks that are more accurate than ever at keeping time.

The key is developing clocks able to see higher-frequency visible light, which means greater precision in measuring time.

As Edwin Cartlidge reports at Science, metrologists from the National Institute of Standards and Technology (NIST) Boulder Laboratories in Colorado are now working on devices up to 100 times more accurate than existing clocks.

Every clock marks time based on a certain action, whether it's the swing of a pendulum, or – in the case of modern-day atomic clocks – the oscillations of a microwave beam set at the exact wavelength needed to excite the chemical element caesium.

As the caesium electrons jump back and forth between two energy states, the frequency of these jumps can be used to measure time.

The next-generation clocks currently in development at NIST are optical atomic clocks, measuring light waves with a resonance frequency some 100,000 times higher than microwave radiation, allowing for much more precision.

In fact these new clocks could get so precise they would only lose one second every 15 billion years.

yit-clockA ytterbium optical atomic clock. (NIST)

Using two optical clocks deploying lasers to cool and trap ytterbium atoms, researchers at NIST have managed to get them ticking together to an accuracy level 100 times greater than caesium clocks – technically to within 1.4 parts in 1018.

"It would be the first time that two clocks of the same species have been shown to agree at that level," Ludlow told Science, though the experiments have yet to be peer-reviewed and published in a journal.

If verified, the accuracy would meet one of the requirements set down by the International Bureau of Weights and Measures (BIPM) in France, requirements that need to be met before these new optical clocks can replace their older caesium counterparts.

Now further research is required to make sure the accuracies can be maintained and in several different labs simultaneously. Because of their complexity, current optical clock setups tend to only be run in short bursts.

Only when all the cross-checks are completed will we have a new, more accurate second – and that might not happen until the 2026 get-together of the the world's top metrological body, the General Conference on Weights and Measures (GCPM).

The GCPM meets every four years and this year has four new updated definitions in its sights: for Planck's constant, for the Boltzmann constant, for the Avogadro constant, and for the amount of electric charge in a single proton.

More precise measurements for these values will enable scientists to provide more precise reference points for what constitutes a kilogram, a kelvin, and an ampere, among other measurements. After that, the second could be next in line for a new definition.

One of the questions still to be answered is which type of new optical atomic clock to use – some use lattices of neutral atoms, like the clocks at NIST, while others use single trapped ions, and progress is being made across all of them.

"These clocks are getting more reliable, more robust all the time," said the head of NIST's Time and Frequency Division, Chris Oates, in November.