# Scandium-based nuclear clocks promise punctuality for next 300 billion years

Set your watches! Scientists have set the clock ticking for the development of a new generation of timepieces with accuracy of up to 1 second in 300 billion years or about 22 times the age of the universe. Researchers working at European XFEL X-ray have examined the potential of scandium as the basis for nuclear clocks, long …

1. #### Roughly accurate

> accuracy of up to 1 second in 300 billion years

Given the claim to accuracy, a time limit of 300 billion years sounds a little approximate. You'd kinda hope that the people who made that forecast would be able to nail down the limit of that 1 second accuracy a bit closer. Say to 316,952,044,811.3503559017 years

1. #### Re: Roughly accurate

316,952,044,811.3503559017

That's only like 73 bit precision?

2. #### The end of time?

Given the estimated age of the universe and the oscillating universe theory, I wonder if this would take accuracy to the end of time?

1. #### Re: The end of time?

Unfortunately it doesn't look like the universe will osscilate. All the math is tending towards an infinite universe, but one stretched so thin that atoms can no-longer hold themselves together.

The lower bounds for the Big Rip are only 100 billion years or so, so these clocks will still be good for that. The upper bounds are out in the 10^12 years sort of number, by which point these clocks will be wrong by around 30 seconds.

3. #### Accuracy v. going rate?

The ability of a chronometer to maintain accuracy (not to drift with respect to its initial setting) has traditionally been referred to as 'going rate'. This is not the same thing as accuracy, which refers to discrepancy between actual time and represented time at a given moment. This timepiece apparently has a going rate of 1 second in 300 billion years, but its absolute accuracy will depend on how it's synchronised with absolute time at the moment it's set running. And by the way, what is absolute time anyway - once we 're talking in terms of seconds per billion years, what's the reference?

1. #### Re: Accuracy v. going rate?

This timepiece apparently has a going rate of 1 second in 300 billion years

Is that 1 second spread uniformly across 300 billion years or say there is going to be a 1 second jump at any moment within 300 billion time period?

Imagine if you always wake up at 6:30 and that one day your alarm clock goes off at 6:29:59. That's easily the whole day ruined.

1. #### Re: Accuracy v. going rate?

But if it went off at 6:30:01 instead you'd have a glorious day, what with the extra second of sleep leaving you rested and ready to go!

1. #### Re: Accuracy v. going rate?

But you'd be fired for being late to work!

2. #### Re: Accuracy v. going rate?

Indeed, what is absolute time? Einstein taught us that observers who are moving, or accelerating with respect to each other, will measure different times.

1. #### Re: Accuracy v. going rate?

Indeed. If you're running to the toilet, time flies - but once you're comfortably in there, time almost stops.

1. #### Re: Accuracy v. going rate?

Well, if you are trying to undo your flies time elapsed approaches infinity.

4. >extreme metrology

I just wanted to say that I love those two words together.

1. Also known as a typical June day in Ireland

2. Is this something metrosexuals are into?

5. #### Clock speed variations

How sensitive would this be to altitude? The closer you are to the ground, the slower time passes. Also, the rotation speed of the Earth varies slightly with the global weather. Would that be something that could measureably affect the rate at which time passes for such a clock?

1. #### Re: Clock speed variations

You can only measure time in your frame of reference. If you had 2 of these clocks, one on the ground and one atop Mt Everest they would differ but as far as the time in their frames of reference are concerned they would both be correct.

It's not just altitude but velocity too, a clock on the equator will differ from an identical one at the pole and don't start on clocks in moving vehicles such as rockets or satellites.

Plainly a standard reference is required from which all other clocks are set, historically this was a clock in Greenwich but now various standards bodies like NIST, NPL, ISO etc cooperate.

1. #### Re: Clock speed variations

Balls - The Greenwich Time Ball is good enough for me :-)

1. #### Re: Clock speed variations

I set my watch when I'm down the pub. Every evening the landlord calls out "time gentlemen, please" and he gets the response "11 o'clock". Good enough for me. Oo, is there time for another?

1. #### Re: Clock speed variations

I'll 'ave 'alf.

1. #### Re: Clock speed variations

Is that you Jacko? I thought you'd be pushing up the daisies by now :-)

Another half, obviously ------------------->

2. #### Re: Clock speed variations

Exactly this question has been investigated, with atomic clocks on a plane and on the ground synchronised, then the plane flies off on a round trip - the flying clock ran slower.

1. #### Re: No plane needed

There's a speculation in the Nature article that the scandium clock could do this with millimetre scale height differences! So I suppose you could use it as a sort of very expensive altimeter.

3. #### Re: Clock speed variations

The closer you are to the ground, the slower time passes.

I thought this is other way around - the higher you get the more you feel like being stuck in eternity.

6. > In other words, if your watch loses a second a year, it will be 9,512 years slow by the time a nuclear clock based on scandium is a second out.

I'm fairly certain I would have noticed and adjusted it before it got to that point.

1. You'd think so. But don't we all have that one clock that's slowly drifted out and you never got round to setting it, and now you're reluctant to change it because you're so used to taking into account that it's 9512 years slow?

1. Off topic, but ICBA figuring out how to change the oven & microwave clocks every 6 months (stupid, fiddly button combos); So one is set to GMT, the other to BST. Good enough for me!

1. We have power cuts often enough for me to remember how to set the clock on the oven.

2. #### That oh \$*** moment

When you realize you looked at the wrong clock and now you're 9512 years late for the bus.

1. #### Re: That oh \$*** moment

Here it is now coming round the corner, just behind the one that is only 9511 years late.

1. #### Re: That oh \$*** moment

I'm sure many UK readers will sympathise! Depending where you live, you've very likely just been through a bus strike, are going through a bus strike, or about to go through a bus strike. Let's not mention trains, shall we?

3. #### Re:- one clock

Our central heating controller clock is always two minutes late even after being reset.

“It’s five minutes to opening time. It’s always five minutes to opening time in Purgatory.”

1. #### Re: one clock

The clock in my car is fairly accurate unless I tick the "synchronise to GPS" option, in which case it runs six minutes fast!?

1. #### Re: one clock

You're driving much too fast.

1. #### Re: one clock

88mph should just about do it

2. My mechanical watch runs about 3 secs per day fast. I set it 30 seconds slow on the first of the month and re-set it at the end of the month when it's running 1 minute fast. That's over a month between settings and if I need to know the time more precisely it's easy to calculate to within a few seconds based on the date. I can't think of the last time I needed to know the time with such accuracy - nothing else happens in my life that starts or finishes on time to this precision. If I were in charge of this scandiuum clock it would run for almost 900,000 years before it needed adjusting.

3. I'm fairly certain I would have noticed and adjusted it before it got to that point.

I'm fairly certain I wouldn't be that worried about it after 9,512 years.

7. #### Shouldn't Tim Worstall pop up here...

...to say something about the scandium market and to explain to us why Brexit is such a great idea?

1. #### Re: Shouldn't Tim Worstall pop up here...

I've just sent him a link to the article. Let's see if he appears when invoked.

1. #### Re: Shouldn't Tim Worstall pop up here...

And I appear. Umm, yes, OK. Ata guess you need one atom to make such a clock? Which would make the lb of Sc I've got floating around in the souvenir box enough to make one for ever El Reg reader?

There's the Xmas offer sorted then.

BTW, Brexit was great, Yah?

1. #### Re: Shouldn't Tim Worstall pop up here...

Oooh a Register commemorative Scandium based nuclear clock? That's SO much better than a blue checkmark, sign me up!

8. #### Price of scandium.

Currently 3789 ISK in Dodixie.

9. #### Of Course

It's easy enough to keep time by microwaves, since they can be put into an electrical circuit that connects to a clock.

And it's easy enough to measure distances with the wavelength of light, just by using an interferometer.

That's why, in the past, while the second was defined as 9,192,631,770 oscillations of a hyperfine transition in Cesium-133, the metre had been defined as 1,650,763.73 wavelengths of the spectral line associated with the 2p10 to 5d5 transition of Krypton-86.

But then they decided to keep only the second, and define the speed of light as exactly 2.99792458 * 10^8 metres per second, requiring distance to be derived from microwaves, which are more suitable to measuring time than distance. I felt this was a mistake, but maybe I'm wrong, and they have found ways to manage to get an improvement in the accuracy of the metre out of this.

Now, we would have to switch to X-Rays. Which can't be seen in an interferometer, making it hard to measure distance with them, and whose oscillations we have no way to turn into electrical pulses for counting either. So we could have a supremely accurate definition of the metre and the second, but one that is impossible to actually put to use to calibrate anything.

Right now, the particular resonance involved is known to have an energy of 12.38959 keV. This is only seven digits, and we will need to know the value of the resonance to nineteen digits to even use it as a theoretical definition to the possible accurace of one part in 10^19.

So, while it is an interesting first step, who knows how long it's going to take to fill in the missing pieces?

Oh, wait a moment. It is possible to perform X-Ray diffraction experiments. So you diffract these X-Rays from scandium with a crystal, and once you measure the angle by which they've diffracted... you have now measured the spacing between atoms in the crystal by one part in 10^19, and you can use that to calibrate something because that might actually be measurable directly. And no doubt there are other ways that scientists can figure out.

1. #### Re: Of Course

Our microwave has a mechanical timer that goes Ping when my cook-chill curry is done. Does that count?

10. #### Doesn't matter.

I'll still be late on Monday mornings.

:)

11. How do they know the scandium clock is showing the 'right' time....

Sure it doesn't drift, but for 300b years it shows a slightly wrong time, all the time.

Which of course, as we know, is an illusion - lunch time doubly so.

12. How do they measure that accuracy of a Scandium clock to deterine the error ?

13. Looks like Scandium was appropriately named then, as Scandinavians are sticklers for punctuality.

## POST COMMENT House rules

Not a member of The Register? Create a new account here.