back to article Boffins build super-accurate atomic clock

The atomic clocks currently used for regulating international time zones are great and all, but who has the time every few million years to adjust them? Fortunately, physicists in the US have figured out how to control seemingly "forbidden" collisions between neutral strontium atoms to make a clock that neither loses nor gains …


This topic is closed for new posts.
  1. Simon Waddington
    Paris Hilton


    I thought it was a perfect cock,not a perfect clock, that got one laid.

  2. Chris Fleming


    I just tripped out the power cord... Just give me a sec to plug it back in.

    Mines the one with 300 million year battery in the pocket.

  3. Kelley Johnston

    inaccurate explanation

    "Strontium belongs to a class of atoms called fermions" - dear gods. Crossed a knowledge band gap here we did, didn't we?

    Fermions are a class of elementary particles. Atoms are made up of elementary particles. Strontium is an atom composed of elementary particles like protons and electrons, both of which are classed as fermions.

  4. Anonymous Coward
    Thumb Up


    I want one on my wrist. And that ASAP.

  5. Anonymously Deflowered


    Isn't that degree of accuracy over that timespan rendered fairly pointless due to clock drift caused by relativistic effects?

    /waiting to have my wikipedia level physics flamed by someone who knows what he's talking about

  6. Anonymous Coward

    Yes but....

    didn't we have to adjust the clocks last year just before New Years Eve because the world itself can't keep time ? 1 seconds every ten years or so isn't it ?

    So after 300 million years this uber-clock will be out by 30 million seconds. Even FarEastern Rolex ripoffs are more accurate than that over their lifetime.

  7. breakfast Silver badge

    Useful but

    You know, after the first 299 million years I'm pretty sure that most people would have forgotten about the extra second anyway. Probably the sentient goo that we have evolved into by then will have transcended the notion of time altogether...

  8. Toby Murcott
    Paris Hilton

    Fermions and watch winding

    Actually the comment about strontium atoms being fermions is correct. Some atoms can be fermions depending on their constituent sub-atomic particles.

    However, it's the line "resulting in a very impressive feat of not needing significant winding for more than 300 million years." that got my pedantic back up. Winding provides energy to power a clock, not its means of determining time. The clock will need powering for all the time it's in operation, but if you are happy with an accuracy of 1 second in 300 million years it won't need recalibrating within that period.

    Paris, because she has plenty of energy but never turns up on time

  9. Anon


    Even wikipedia knows that "composite particles (such as hadrons, nuclei, and atoms) can be ... fermions ..."

  10. Charles Manning

    re: Wristwatch

    Not that sort of "getting laid".


    btw: A clock this accurate is few orders of magnitude better than is needed for any telecom network. GPS is probably enough for most telecom applications (where a lot of GPS get used). Apart from the fact that it is going to be difficult to hide all the coolants etc in that clandestine cell site.

  11. BlueGreen

    ...accurate to one second for 300 million years ... mmkay... measured against ... what?

    (today is National Ellipses Day. Take one to work. Or eat it. Or paint it or ...something.)

  12. Anonymous Coward

    Put your money where your mouth is!

    ...I'm willing to bet one 100 trillion of your US Dollars (or the US national debt, which ever is greater) that it loses two seconds before the 300 million years is up.

  13. Dr Patrick J R Harkin

    Does it change automatically for BST?

    And does it have a snooze button?

  14. Herby

    But what happens when a leap second happens?

    Oh, we need to speed up this silly planet to eliminate those pesky added seconds. Make sure you account for them, as some operating systems think that there are always 31536000 seconds in a non leap year. Then there are those that keep time by pairs of seconds.

    Oh, well.

    Tick, tock, tick, tock, etc.

  15. Richard Duncan
    Thumb Up

    I think my Casio does this.

    No, honestly.

  16. John Robson Silver badge

    No winding?

    It's being continually wound by that secondary red laser.

    It might not need resetting though...

  17. Anonymous Coward
    Anonymous Coward

    Strontium Atom = Fermions?

    I thought Helium 3 was a fermion - so why not Strontium?

  18. Richie


    "Also, casually mentioning that you made a clock that's accurate to a second for 300 million years is almost guaranteed to get you laid."

    I'm afraid that's a common geek misconception.

  19. This post has been deleted by its author

  20. Chris Bradshaw

    get laid??

    I think another knowledge band has been crossed - this one social. There are some people who can tell a girl that they have designed an atomic clock accurate to one second every 300 million years in such a way as to get laid. There are other people who can actually design such an atomic clock. But the overlap between these two groups of people is extremely small - in fact, I would guess it has roughly the same degree of smallness as the afore-mentioned one second in 300 million years...

  21. Andy Barber

    @ Wristwatch

    How 20th century can you get?

    Mine's the one with a Strontium mobile phone in the pocket!

  22. Darryl

    Yeah, but...

    Does it automatically switch over for Daylight Savings Time (Summer Time)??

  23. Richard Kay


    You don't need one on your wrist. My £20 wristwatch synchronises to atomic clocks by picking up a radio broadcast. It also handles GMT/BST seasonal changes so doesn't need any adjustment until I travel to a different timezone. Interestingly, my computers now also synchronise to atomic time using NTP (Network Time Protocol) so I can read logged adjustments relating to when they have leap seconds, resulting from atomic time being more accurate than earth rotation time, needing a minute with 61 seconds every few years to keep the earth's rotation adjusted to atomic time.

  24. Pascal Monett Silver badge

    "atoms tend to want to mingle"

    The horny little bastards !

  25. Anonymous Coward
    Anonymous Coward

    Really necessary?

    I already get in trouble for being just a few seconds late, why couldn't they just make one that was accurate to the nearest half hour and give us late risers a decent chance

  26. Ken Hagan Gold badge

    Re: inaccurate explanation

    If the wavefunction describing the whole atom is anti-symmetric then it quacks like a fermion. That's good enough for me. Protons (composed of three quarks and a shed-load of virtual rubbish) would be a case in point.

  27. John Smith Gold badge

    But for real cleverness

    Build it into a mems chip.

  28. Anonymous Coward

    Re: inaccurate explanation

    Fermions are just any particle/collection of particles with odd half integer spin i.e. 1/2 3/2 .....

    As you correctly stated electrons, protons and for that matter neutrons are all fermions.

    Now quarks are also fermions having a spin of 1/2, protons are a collection of 3 quarks and as such have a spin of 3/2 making them fermions.

    So to work out if your atom is a fermion or a boson all you have to do is add all the spins of the constituent particles, if you get an integer value its a boson if not its a fermion.

    OK so Strontium... goes of to wiki to check the atomic structure.. ah isotopes.

    Most conman isotope SR 88(80% abundance) 50 neutrons, 38 protons and 38 electrons.

    So that's 75 + 57 + 19.. whole number so a boson... doh.

    OK so they must be using a different isotope and its got to be stable so they must be using SR 87(7% abundance), 49 neutrons, 38 protons and 38 electrons.

    So that's 73.5 + 57 + 19... not a whole number so a fermion.

    Of course if you ionise the atom you are removing an electron and so losing 1/2 spin...

  29. Christoph

    @ Bernie

    No problem - just get a watch that's set by the radio signal, they're pretty cheap these days. As soon as these new clocks are ready they'll be used to control that signal.

  30. Harry

    You can't claim time to be accurate ...

    ... unless you have an even MORE accurate time source to compare it with.

    Their clock has the claimed "accuracy" only because the DEFINITION of time has been arbitrarily defined to coincide with the method they're using to measure it.

    Once upon a time, time used to be defined relative to the rotation of the earth. A definition that probably still makes the most sense to most people, because the primary purpose of time is to ensure we do things at the appropriate part of the day. If the earth were to mysteriously start rotating at half its normal speed, it would make no sense to use a supposedly "more accurate" time source that caused us to sleep only on alternate days.

    Thus, the aim should probably be to find a time source which more accurately reflects the slightly-changing speed of rotation of the earth, so we don't have to artificially "increase" the year with leap-seconds whose very existence in effect proves "we got it very slightly wrong, but just like politicians it must be right because WE say it is".

  31. Luther Blissett

    @Simon Waddington

    > I thought it was a perfect cock,not a perfect clock, that got one laid.

    In which case I hope you haven't been waiting 300 million years for the perfect hen (to evolve).

    You'd be better making sure *talk* about having the perfect cock/clock/whatever reaches the right ears (and the left ears) (and the bits in between).

  32. Murray Pearson

    Even more inaccurate terminology

    Saying "it doesn't need to be wound for 300MY" means it needs no power input, which is clearly wrong. You mean it doesn't need to be SET for that time period, assuming it has power input — meaning it's being CONSTANTLY wound up for that entire time period. Winding ain't the same as setting.


  33. Graham Marsden

    Got the right time, mate...?

    ... err, yes, but by the time I tell you, it'll be wrong...

  34. Anonymous Coward
    Paris Hilton

    I want one on my wrist

    Really? Highly contagious stuff this Strontium isn't it? So you don't want to father children, and you won't need the flourescent jacket when you go out at night either.

    However, what's the point of this? Humankind won't last out this current century, never mind having a Timex which doesn't need winding for 300 million years. Future researchers will dig this stuff up and claim "pah! One second in 300 million years! What p1ss-poor quality!".

    Paris because I fancy keeping a beat going.

  35. pctechxp

    And the point was?

    Surely there are more important things to research as we have pretty darn good atomic clocks now and who is going to care in 300 million years time assuming we haven't been exterminated by a machine uprising, boiled to death by global warming, frozen to death by the after effects of a asteroid/meteroite hit or nuclear war?

    By that time I;m sure all of us reading this will be part of the soil.

  36. Matthew
    Thumb Up

    re: Bernie

    Who wouldn't want a nuclear device on their wrist!

  37. Anonymous Coward

    @Simon Waddington

    Simeon - you want perfect clock OR large ClOCK. Best is to have both.

  38. Andy ORourke

    @ bernie

    You can have one on your wrist, there are some really nice watches which receive regular updates from the current generation of atomic clocks in Greenwich, I am sure that once this new generation of atomic clocks are available the timekeepers there will be upgrading.

  39. Anonymous Coward
    Anonymous Coward


    How do you actually do anything with a clock that precise?

    How do you plug it into something in order to synch it, without disturbing it?

    If you use a long cable, instead of a short one, isn't it going to effect the readings at that level?

    If it's so cold, how can you possibly set the alarm without freezing your finger tips?

    What about if you spill some tea on it, will it work again once its dry?

    And most importantly, where on earth do I find girls that will sleep with me if I tell them I've made a very accurate clock?

    Please, I really want to know. It's... important.

  40. Garret Cotter


    Fermions are particles which obey Fermi-Dirac statistics - that is all. Composite particles such as atoms (and indeed protons and neutrons) can be bosons or fermions. E.g. He3 is a fermion and He4 is a boson.

  41. Eddy Ito

    In other news

    A new super accurate clock was adjusted one hour ahead as boffins, who previously had refused to get out of bed earlier, no longer fear becoming discombobulated with a different prime time schedule.

  42. Anonymous Coward


    Actually, the original explanation in the article was quite accurate as popular science goes.

    A "fermion" is any particle - elementary or otherwise - with a half-integer total angular momentum. (For an elementary particle, its intrinsic angular momentum is "spin"). Turning a fermion around any axis by 360 degrees changes the sign of it's wavefunction. If particle's wavefunction does not change upon a 360-degree rotation, it is called a "boson", and has an integer spin. The total intrinsic angular momentum of a particle has important consequences at sufficiently low temperatures: two (or more) bosons can occupy the same quantum state (or be at the same place if you will). Two fermions cannot.

    Whether a compound particle, such as an atom or a molecule, is a fermion or a boson depends on two things: the spins of it's constituent parts, and the way these spins are added together, or "coupled". The way these rules work is that a compound particle containing an even number of fermions (and an arbitrary number of bosons) will be a boson. A particle with an odd number of fermion constituents (and again any number of bosons) will be a fermion.

    Strontium has 38 electrons (spin-1/2 particles). In its ground state, all its electrons are "paired", so that their spins cancel out, and total electron wavefunction has integer (zero) angular momentum. Most of the isotopes of its nucleus also have an even number of particles of each type (protons and neutrons), and thus integer spin as well. However, Strontium-87, which has natural abundance of about 7% has spin 9/2. As a result, the Sr-87 atoms have half-integer total angular momentum, and are fermions.

  43. Anonymous Coward


    Yeah, right.

    With another atomic clock making scientist, maybe.

    (But what DO you do if you get bored at work? Watch the clock?)

  44. Anonymous Coward
    Anonymous Coward

    @Kelley Johnston

    Thanks for bringing it up... you saved me the bother. OTOH, there's still time for Austin Modine to update the wikipedia page for Fermions ...

  45. Flocke Kroes Silver badge

    @Kelley Johnston

    Protons are not elementary particles. Composite particles that contain an odd number of fermions are fermions. Quarks are fermions and protons are three quarks, so a proton is a composite fermion.

    Atoms are made of electrons, protons and neutrons. Electrons and neutrons are also fermions. Atoms have the same number of protons as electrons, so atoms with an odd number of neutrons are fermions (unless there is a force that makes atoms pair up to form bosons).

  46. jake Silver badge

    @Kelley Johnston

    "Strontium is an atom composed of elementary particles like protons and electrons, both of which are classed as fermions."

    Actually, while electrons are a type of fermion called "leptons", protons are comprised of fermions, and along with neutrons are part of the sub-atomic set we call "baryons". Some call protons and neutrons "composite fermions", but in my mind that's pushing it. They are more correctly referred to as "nucleons". The rest of yours was accurate.

    I knew that particle physics minor would come in handy someday ...

  47. Anonymous Coward
    Anonymous Coward

    Fermions, Bosons and knowledge gaps

    It seems the knowledge gap is with Kelley Johnston.

    Fermions are /not/ a class of elementary particles, although it is true that some elementary particles are fermions. Others are bosons.

    Atoms can be either, depending on whether they contain an even or an odd number of fermions.

    (4)He is a Boson. (3)He is a fermion.

    (87)Strontium is also a fermion, and is presumably what they're using in their clock, since it's the only stable isotope of strontium that's a fermion.

  48. Chris Miller

    Accurate explanation

    Fundamental particles can be divided into fermions and bosons - fermions are those such as quarks and leptons with half-integer spin that obey the Fermi exclusion principle. The same distinction can also be applied to composite particles, such as protons and neutrons or atomic nuclei like carbon-13 and helium-3, you sum the spins of their components, and if there's an odd half left over they behave as fermions.

    So it's quite valid to describe (some isotopes of) strontium as fermions.

  49. Xris Middleton

    Useful but...

    "reducing the strontium atoms' temperature to a few millionths of a degree closer to absolute zero"

    Am I being led to believe that the atoms are now a few millionths of a degree above absolute zero or were they originally at room temperature and they just lowered that by a few millionths of a degree closer to absolute zero.

    I only ask because ıf you were to use this for improving the synchronization of telecom networks or on deep-space communications, what size is the kit necessary to reduce the temperature to near absolute zero?

    I would imagine that it isn't small or cheep enough to be used that extensively.

  50. Anonymous Coward

    What's the point?

    They'll only have to put it back an hour in October...

  51. Schultz

    getting laid...

    within the 300 Ma timeframe, surely.

  52. Peter Depledge
    Paris Hilton

    Oh, I need a title

    Why do I get the impression that these comments will form an episode of the Big Bang Theory?

    Paris, cos there's no Penny icon

  53. Anonymous Coward
    Thumb Down

    re. And the point was?


    The point of this is that timing signals are very important for a number of applications.

    The most obvious one is GPS; a more accurate clock will allow more accurate positioning. The current technology is not accurate enough for applications like mapping as it's accuracy is around 10m, these new clocks could give accuracy to less than 1m.

    One issue, pointed out in a few comments, is that the Earth's gravitational field is strong enough to bring time-dilation into effect, which in the case of these clocks will alter the "speed" at which they tick just by moving one a few metres away from the Earth. Variations in the Earth's gravitational field will also affect them (current atomic clocks are only accurate enough to show these effects when put into aeroplanes or into orbit).

    Make no mistake, this is a valuable technology.

  54. Dave


    So, they _think_ that they have doubled the accuracy of their previous clock. How long was it running for? And what was the reference source that proved the inaccuracy??

    As for:

    "casually mentioning that you made a clock that's accurate to a second for 300 million years is almost guaranteed to get you laid."

    I smell excretia there...

  55. Anonymous Coward


    Does the laser have a frikkin shark attached?

    posting anon due to the shame of joining in the awfully predictable joke submissions above

  56. Cullen Newsom

    @Bernie et al, wristwatches, etc.

    Pshaw! at all of those radio-controlled "atomic wristwatches". Cheap hacks! This is the man you want to see about your wristwatch.

  57. Charles Manning

    @AC re: More accurate GPS

    Naah, sorry mate, you need to understand a bit more about how GPS works.

    Current GPS accuracy is not constrained by the clock. GPS accuracy is constrained by variance in the propagation delays due to ionospheric disturbances etc.

    Current commercial RTK receivers can give you approx 20mm of error. At those accuracies even the earth tide starts to show up.

    Improving the clock is not going to make any measurable difference because the clock is not the current source of any significant error.

  58. YerWhat

    There are some problems with your comment

    Of course, the rotation of the earth will have slowed somewhat over time and other effects of floating about in space.

    Does improved timings not improve GPS heights measurements?

This topic is closed for new posts.