Timekeeping: From the Sun to the Atomic Second

well hello and welcome to npl my name's ann curtis i'm a senior research scientist in the time and frequency group here and npl is first and foremost a metrology institute which means we deal with measurement science and that really at its most basic level deals with the si base units and these measure things such as temperature mass and length but i'm specifically going to talk to you today about the si second which has quite a convoluted definition and i'll read it for you the second is the duration of 9 billion 192 million 631 770 periods of the radiation corresponding to the transition between two hyperfine levels of the ground state of the cesium-133 atom the definition and what i'd like to get across to you guys during this talk is how did we get from the point where the sun was the thing that told time i think a lot of us still feel very intrinsically that the sun is what tells us what time it is how did we get from that to this very specific definition of a second that has to do with atoms um so my talk will basically start with an a historical overview of timekeeping but really in terms of the drivers of timekeeping what drives our needs for clocks and what's the technology of measurement science over the years and then at the end what's our current state of the art how do you make an atomic clock and what do we actually need these types of clocks for today before i get deep into this i'd like to go through a couple definitions and i'll start with just showing you the difference between terms such as accuracy and precision one way to describe this is in terms of you have an archer who is trying to hit the center of a bullseye so here's our first archer they're not very accurate and they're certainly not very precise and trying to hit the center of this target as we increase in accuracy the archer becomes more accurate these arrow points are now centered a lot closer to the point they were trying to hit but they're spread out they're not very precise we can also increase in precision and we can see the archer now has hit basically all the arrows in the same place they're quite precise but it's not very accurate you see there's an offset now between where they were trying to hit and where they actually did hit and of course the perfect archer who's both accurate and precise and this is related to measurement science in the sense that you can imagine in this system here where you're accurate but imprecise two things could happen depending on what kind of shot this person is over time you could get more and more arrows kind of centered around this bit here and you can imagine that statistically you could start to believe yeah i really am i really know where the center of that target is even though my bits are my arrows are a bit spread out but over time instead of being centered around here the whole kind of average could drift off to the side and then you wouldn't have a very stable system so that's the idea of stability over time then if you look down at this kind of inaccurate precise here i mean clearly this archer knows what they're doing they're always hitting the same spot but there's some kind of offset some systematic reason that they're not hitting the center every time and if we could understand that offset and correct for it or even just subtract it off or in the case of this archer just move the target over a couple meters then you would be in a position where you're both accurate and precise again so these are the kind of terminology i'll be using during the talk and when you go and tour the time and frequency section of npl after the lecture you'll hear these terms come up over and over again so to summarize accuracy is what helps you get to work on time stability or precision it doesn't matter what time it is but it matters that your clock is ticking at a regular stable rate you know for things such as measuring heartbeat for example and then there's a third thing that we like to talk about which is called reproducibility and that really has to do with how much do you trust your clock if you're trying to compare your clock to other clocks how does it hold up um so then i should really start with the basic basic question of what is a clock so a clock is made up of an oscillator an oscillator is something that's periodic it means it repeats so in terms of something you all are very familiar with a clock with a pendulum this periodic swing of the pendulum it also includes electromagnetic radiation which also oscillates in time on the second part of a clock is a counter because you've got this radiation but you don't know anything about it or you have this pendulum moving but there's no way to look at how it's moving over time the counter then measures the oscillations and in terms of the clock that's the mechanism that takes each tick and talk and turns it into seconds minutes and hours and keeps track of time as it's going um the third thing that can be very useful for a clock is a reference and the reference is really how you check that your clock is on time and we'll talk more about that as we go on so in terms of measuring time if we're looking at the historical perspective you really want to ask two questions why do people need to know what time it is this need is what drives clock development and how do those needs change over time so you start with your kind of in in historical times local solar time which is simply based on the motion of the sun across the sky um you have shadow clocks even portable shadow clocks that can measure this but the systematics the things that cause your measurements not to be as accurate or precise as you'd like is that the way the sun moves across the sky changes throughout the year of course there are overcast days and worse yet there are 12 hours a day where you have no access to your clock what evolved out of this were water clocks and this is an innovation in time keeping where for the first time you didn't have some motion of a body in space that gave you your time and the way these work is you're either measuring how long it takes for one cis turn to empty or how long it takes for one cistern to fill um these were invented essentially to measure time at night you had night watches and getting the timings right for that you had religious ceremonies you could use them indoors to measure how long you were debating in court you had portable versions for measuring pulse rate and things like that but the systematics are you're not going to have a constant flow rate either into or out of one of these cisterns and there are things that affect the flow rate as well though you can calibrate these kind of things with a sundial and that's what they did in terms of innovations where you started putting markers to measure time so you could see what time it was by the level of your water or even more ingenious devices whereas the water fills up it pushes a lever which turns the gear which then indicates the time on the face of a clock and these are starting to look like real clocks now as we know them and then of course you could have a really elaborate example which played music and created motion and could even show you the movement of celestial bodies over time through the mechanical motion of this water clock but what we really needed in terms of increasing the accuracy and stability of these clocks was something that would oscillate that would have a periodicity that was much more constant than this flow rate of these water clocks and that came about with the invention of pendulum clocks and why this is really so powerful has to do with the intrinsic systematics of the system if you look at what defines the period of a pendulum that is the time it takes for a pendulum from one point to swing and come back to that same point is simply a square root of the length of that pendulum and that is the length from the point of which it's swinging to its center of mass and the gravi acceleration due to gravity where that clock is those are the only two things as long as your oscillation amplitude is small so small little oscillations like in a clock those are the only things that affect the period the timing of this period what is not in this equation is mass it doesn't matter how heavy your pendulum is and the other thing that's not in there is the amplitude of the swing so if that changes a little it actually doesn't affect the oscillation period of your clock so the systematics are now much more subtle it has to do with small changes in the length of the pendulum with temperature or friction due to the air that's involved near your pendulum swinging and people found ways around these to compensate for these changes and then you could have pendulum clocks that had errors of less than 10 seconds in a day which means that the deviation from the steady calculable oscillation frequency would only be 10 seconds lost or gained in one day what was driving the need for these more and more accurate clocks why couldn't we have just stayed with water clocks well one thing that was really driving things both economically and otherwise was navigation in the 17th century and around that time you had all these observatories that were opened in these different maritime nations because they wanted to know where they were when they were traveling in the ocean and this all has to do with the longitude problem which i'm sure many of you are quite aware of but i'll explain it in some detail here so your latitude which is your measurement of where you are north and south is is actually fairly simple to measure um it has to do with the sun's angle at noon so you can figure out where you are north to south fairly easily and in fact you can measure what time it is where you are by the sun quite easily as well but what you can't do easily is find your horizontal or longitudinal position on the globe unless you have a no unless you know the time at some reference point and so what you can think is if you can draw a circle around the earth here you can split that circle up into 360 degrees and if we have a 24-hour day that means that as you travel 15 degrees that's equivalent of a time change of one hour now let's say you started in greenwich and you synchronized the clock in your hand to some greenwich clock and then you traveled some distance from greenwich and you wanted to know where you are because you're in the middle of the ocean now well all you need to do is lo measure your local time compare it to the time on your clock and let's say you saw exactly a three hour time difference you would know you had traveled precisely 45 degrees from greenwich and you would know your position very accurately clocks at this time these pendulum clocks could make these measurements at the level necessary to know where you were but unfortunately pendulum clocks were not really good on a boat and this led to this longitude act of 1714 where the government put together a 20 000 prize if you could measure the longitudinal position to half a degree and the error on that that kind of error is about 35 miles at the equator and equivalent to a clock that can keep time to three seconds per day and to give you an idea of how important this was both economically and militarily to the government at the time twenty thousand pounds prize money then would be the equivalent of about 1.5 million pounds prize money now so they were taking this quite seriously and took john harrison a long time to get the money but he did demonstrate a clock and you can go if you go to greenwich they have the whole series of clocks he worked on towards his work and a whole history of of this longitudinal prize laid out quite well i highly recommend it but he created a sea going clock that is actually not that large that was accurate to about .2 seconds in one day but what that meant in reality was over a seven week trip he took to prove that his clock was good the air in their position was only about 30 kilometers and so this meant for the first time you could safely sail across the oceans and and the kind of navigational needs were being met by the clocks of the time um but of course as you try to get better and better timing and know your position better and better you run into the subtleties of the system you're using noon i'm saying it's quite easy to know what time it is where you are but noon is not the same as the sun being directly overhead it turns out and this varies by about plus or minus 15 minutes over the course of a year so the international community decided to use instead a mean annual average of this non-uniform motion of the real sun and it was at that time that the second was then defined as some fractional um some fraction of the mean solar day and this greenwich mean time was established as a global standard in 1884. the problem with this although it was good enough at the time is any definition of time that's based on the rotation of the earth is going to run into trouble because the rotation of the earth is not constant so why does the earth's rotation have some variation well it really has to do with conservation of angular momentum at its heart but the idea is that the mass of the earth it doesn't all just stay in one place and you also have the moon and earth mass system as well and as you change the mass distribution around it changes how fast the earth rotates and these can be kind of large changes or small changes fast or slow people have known that there's a difference between earth rotational time and this means solar day time and have been measuring it since kind of the 1650s and you can see that the the differences can be on the order of tens of seconds over these centuries it's been measured but it's not just a long-term deviation on a daily basis or a yearly basis these are year indicators here basically you can see that the the speed of rotation differs from this earth's rotational time on the order of milliseconds now so if milliseconds are kind of the timing that you need you have a problem if you're defining time in terms of this rotation so what this is really telling us is you know the sun moving around the earth is not a good enough reference the earth is not a good enough reference we really need to reference our clocks to something more stable and that's where the introduction of atomic time comes from this idea that we can reference our time keeping to atoms and the first season of atomic clocks were produced at npl in the 50s and they had an accurate they were accurate to apart in 10 to the 10th now to compare that to the clocks we've been talking about that's as if you ran your cesium clock for 300 years and it was only changed in time had an error of about one second and that's comparing this 0.2 seconds per day of this kind of best navigational clock before that other labs quickly reproduced this and showed that yes this clock does keep this kind of time and because of this reproducibility the season clock was adopted as the basis for this international definition of time which i read to you at the beginning this gmt then was replaced by an international time reference utc which is still used today which is maintained by an ensemble of atomic clocks to which our cesium fountains give data every month and just as a note leap seconds were introduced at this time because although we're now keeping time with atoms which don't care how the earth is rotating people still live on the earth and it rotates and we still want the sun to be approximately at noon when it's noon and not drift away from that over time and so that's why the leap seconds are added in it just keeps utc and earth's rotational time constant within a second um so in terms of microwave standards today because this 9 billion oscillations per second of the cesium clock definition is a microwave transition these kind of clocks which you can see if you go visit down in our labs they have an accuracy of about a part and 10 to the 15 which in the terms i've been talking about is a clock who would only lose or gain a second every three billion years um so the real question is how does having on a clock reference to atoms take us so give us so much better accuracy than these clocks that are based on pendulums and internal workings and that really has to do with the fact that these atomic clocks the oscillator in the clock the clock itself is electromagnetic radiation that's then referenced to an atomic sample and i'll go through that because this is quite quite important to understanding how atomic clocks work so if you have kind of an electromagnetic spectrum and energy is increasing on this axis you have the kind of low energy radio waves the visible spectrum from red to blue and then kind of uv and x-ray as the energy increases and you can also talk about frequency the frequency of the oscillations increases as well frequency and energy are proportional so you can talk about them in the same way according to quantum mechanics atoms and atomic systems have quantized energy levels and very basically that means that when you have your atom with an arrangement of electrons there's kind of a low energy state where these electrons want to be in and you can shine any kind of energy you want on the system and nothing will happen until you shine just the precisely correct radiation on your atom to excite those electrons into a new configurational state so you can shine whatever you want on this atom but it won't become excited until you get the frequency exactly right and it's the precision at which you have to get this frequency right that makes these clocks so stable so the atomic clock is basically made up of an oscillator that produces the radiation so if it's a microwave clock that's a microwave cavity producing microwaves if it's an optical clock that means it's in the visible and that's laser technology that produces the radiation you have a counter of some sort that measures the radiation and then here's your atomic reference and the reason again this is so um important to the atomic clock is this oscillator is not going to do anything to this atomic reference until you get the radiation frequency exactly right in which case your atomic reference will produce a signal that you can use to tell the oscillator it's gotten the frequency right and this is the basics for the atomic clocks and the reason that atomic clocks are again so powerful on top of that is atoms of the same element in the same isotope that's the 133 bit of the cesium are identical and that means if you make your clock somewhere else the atoms are going to behave exactly the same way so all of a sudden these clocks are extremely reproducible unlike any kind of pocket watch or a pendulum clock you can imagine so what's driving this modern day clock development to you know just a few seconds in a billion years well navigation is still driving it except now on much larger time scales and much larger length scales you have synchronization where you have large arrays of things that need fine timing control and now this idea of radio telescopes is just fascinating and i highly recommend looking into it um so you need the clock on each of these to know the time and the synchronization between all the other clocks and again just the kind of economic and public needs are still driving this forward so to finish up i wanted to give one concrete example so you really really understand how the accuracy of these clocks directly feeds in to these modern day drivers and i thought i'd do that by explaining how global navigation satellite systems work because although everyone i bet in this room has a phone that can tell you where you are i don't think many people are aware exactly how that works and how vitally important the clock development that we're doing here and at other national metrology labs feeds into this kind of thing so first i'll just explain how does it work and then hopefully by the end of it you'll understand why clocks are so important so i'll do that with a just a basic example of time of arrival ranging and it calls the foghorn example because the idea is the sound of a fog horn is going to allow us to know where we are to find our location so let's imagine we're a boat and we're about to sail off and before we do we synchronize clocks with a whole bunch of other boats so we've all synchronized so at five o'clock it's five o'clock in all the boats the reason this works is we know the speed of sound is about 335 meters per second now velocity is simply how far something travels over a certain amount of time in this case meters per second is how we've defined this but you can change that equation around and you can from knowing the velocity and the transit time of a signal you can know the distance that that signal is traveled and that is basically the whole idea here so if we set off a fog horn at a certain time say five o'clock and i on my boat measure that fog horn coming in a couple seconds later you can then from that information of knowing the velocity and the time it took between when the signal was sent and when you received it you can now know your distance from that signal now of course you could be anywhere along the circumference of the circle that's played out so then you just need a second boat who also has a foghorn it'll subscribe another circumscribe another circle so now there's two points you could be at where those two circles intersect so with a third boat you then precisely know where you're located so what do clocks have to do with this in terms of accuracy and precision well let's imagine there was a slight timing offset after you synchronized your clocks one of the clocks has moved a bit let's say it's moved by about a second so there's a second offset between the two clocks but of course you don't know this because you aren't next to each other to check the speed of sound is 335 meters per second so this small error in your location is going to be equal to the speed of sound times whatever this small offset is and so that leads you to a 335 meter error and of course that gets added to the errors produced by the other timing errors on these other boats and instead of one very precise point that you know you're at there's now this large area where you could be located and so the moral of the story is that timing errors errors in your clocks reduce the accuracy of your position measurement satellite navigation works in precisely the same way you have a receiver on the earth a satellite in the sky they have clocks that are synchronized at some level there's some distance delta x you would like to measure the same equations apply but now instead of the speed of sound you're dealing with the speed of electromagnetic radiation um which is much higher and this of course leads to and if you want say sub meter accuracy you're going to need nanosecond timing in your clocks now the clocks we have aboard these satellites are microwave clocks and they can very easily get this kind of timing um and and get this kind of precision and placement so what's next well we keep making better and better clocks because there seem to be more and more applications for them as we do the optical clocks we're developing here and at other places in the world they are clocks that would not lose one second over the lifetime of the universe and there are actually reasons why that's necessary which i very briefly will show you um and at this level earthquakes clocks start to run into problems which i don't have time to go into but please feel free to ask me afterwards um and so your solution really is to put your optical clocks in space and that gets around all these issues and then you suddenly can access better navigation tests of fundamental physics there's geoscience you can do in terms of motion of tectonic plates in terms of synchronization you can now have even larger arrays of things even in space where you're trying to do say gravity wave detection radio astronomy in space deep space communications where the distances are again now astronomical and what has history shown us it's said that in terms of clock development it's always kind of economic drive that pushes things forward but really the better we make the clock the more things we can do with it so thank you for your time and if you um have questions i'll be in the cafeteria just down the way and i'll be happy to answer them

Channel: National Physical Laboratory

Views: 7,996

Rating: 4.8899083 out of 5

Keywords:

Id: UPVZ7wxCQwg

Length: 25min 12sec (1512 seconds)

Published: Wed May 28 2014