Keeping Time at NIST | Hacker News Reader

Keeping Time at NIST

nist.gov

87 pointsoedmarap23 comments

On laser cooling:

> Very basically what they did was use a specially tuned array of lasers to bombard the atoms with photons from all angles. These photons are like pingpong balls compared to the bowling-ball-like atoms, but if you have enough of them, they can arrest the motion of the cesium atoms, slowing them from about 130 meters per second to a few centimeters per second

Such a cool idea deserves a better brief explanation: Atoms absorb light in a specific frequency, and are pushed by it. If your light is at a slightly lower frequency, due to doppler effect only atoms that are moving towards you will absorb and be slowed down. Atoms going in other directions will be undisturbed. The net effect is a reduction of movement, hence cooling.

Oh wow, neat explanation, thanks! For some reason I always thought laser cooling was about “cornering” atoms in between laser beams. But then apparently from your explanation it’s really only about finding the right laser color! Then you can beam it in pretty much any direction and it will result in a global reduction of the speed of atoms (aka temperature). Did I get that right?

Yes that's the general idea. Lasers are used because of their temporal coherence (i.e. they shine in a single, stable, frequency).

A critical part of a laser is coherency: all energy is at the same frequency and phase. This is really what sets laser light apart from other kinds of light and what makes it such a useful science tool.

Laser cooling only reduces the velocity component that is directed against the direction of the laser beam. For free atoms you therefore need three pairs of counter-propagating beams to achieve overall cooling. If the atoms are somehow confined (for example ions in a trap), their motion has to reverse periodically and a single cooling beam can be enough.

Bill Phillips (a NIST type who's also a Nobel laureate) gave a lecture called "Time, Einstein, and the Coolest Stuff in the Universe" at several locations a few years ago, and there are various videos of it online. It explains things rather thoroughly, and is a great example of an enthusiastic scientist going full nerd, including several classes of really bad jokes, etc. Worth a Google and a watch. If you can find the one he did in Toronto, that's probably the best one, but TVO (our provincial public broadcaster in Ontario) doesn't seem to have it up on their YouTube channel anymore.

One other neat use of really accurate clocks: measure if the fundamental constants drift relative to each other.

https://onlinelibrary.wiley.com/doi/full/10.1002/andp.201800...

I remember once seeing those NIST guys one evening at Dogfish Head Alehouse across the street talking physics and neutrinos. To measure time between two events, something needs to move faster than those events to perceive it. After a few beers, you'll be left with the check.

I have an older friend who has always wanted an "atomic clock", and he bought one off of e-bay, an HP-5061b.... it almost worked right. Between the two of us, we got it working, and tuned to the right Zeeman frequency. Once we got the hang of repairing them, two was even better.... we put each one into the XY channels of an oscilloscope.... the pattern never shifted over hours... so cool!

The optical clocks and optical frequency combs mean there are devices that you can buy, with internal atomic clocks, that are effectively Frequency Counters, for blue light... that read down to the Hertz!

I want one of those when I'm older. ;-)

I really wonder how we’ll use that mega-accurate clock that takes like a second every gazillion years. Like the article said, GPS is great but even then you don’t need such accuracy.

All the time in certain kinds of lab work.

Improved gravimetric geodesy? https://arxiv.org/pdf/1803.01585.pdf

That's a nice overview paper. There is much value in the gravimetric mapping realm and we're not even close to the limits of what people are imagining.

For instance, by combining multiple ultra-precise orbiting clocks with quantum teleportation networks linking them for time transfer one could gravimetrically map the planet in 3D at millimeter resolutions.

E.g.: these clocks would be able to carefully map the shape of Earth’s terrain down to the millimeter. “You’d be able to see someone digging a tunnel under the U.S.-Mexico border from space” [1]

One other capability of such a system: whoever possessed it would be able to locate every single significant concentration of dense matter, e.g., plutonium on the planet no matter where it was stored. They would know exactly how many nuclear weapons, etc. everyone had and exactly where they were keeping them.

[1] https://www.sciencenews.org/article/quantum-timekeeping (paywall)

Calibrating slightly less accurate clocks. What those clocks are used for I have no idea, but there’s times you need a good traceable measurement for some calibration of your own and your uncertainty stack can start to add up. The better, smaller, and cheaper NISTs equipment is the better your local cal lab will be and the easier your work is.

Try finding a thermometer setup that’s better than 1 degree accurate from 0-400C. It might be more than you expect. Now imagine you want to measure some property of a sensor or material as function of temperature using your new thermometer, add in the cost of an oil baths and precision multimeters. Now say you want to create your own golden standard with the best precision you can muster, your expensive oil bath is looking lacking and you’re buying Vienna mean standard ocean water...

Well said. The ability to better realize primary standards and disseminate that to secondary metrology labs from national standards is so critical.

It’s the same reason I think the work done in changing the SI definition to physical constants is so critical. In a decade or so all secondary labs should have capabilities to measure directly to the definition of a unit without needing to do comparison measurements to national standards. To me it seems like the metrology equivalent of “open sourcing SI units”

It's important work and it can be surprisingly easy to run into the limits of your uncertainty. If you ask someone to measure a temperature in my experience 9/10 times they'll grab a type K thermocouple. Far fewer than 9/10 will tell you that the number they gave you is only guaranteed to be +/-3 degrees. That's good enough for a lot of applications but not all of them, and when something is really critical it helps to understand the full stack if your supplier/lab/coworker is making questionable claims about their accuracy.

Specific to my knowledge base, timing is critical to applications in inertial sensors (IMUs that guide missiles, AR/VR, smartphone apps, and so much more). To that end, here is interesting work that DARPA is doing in developing inertial sensors that can navigate without GPS for significant amounts of time. Critical to that is accurate chip level atomic clocks.

https://www.darpa.mil/program/micro-technology-for-positioni...

Other uses as other posters mentioned is in traceability of measurements. Improving the means of measurement of the primary standard has downstream impact to secondary standards in metrology labs which ultimately (through transfer standards) would result in better calibrations of production systems.

Finally, an interesting thought I have is that in the future this could be an interesting way to figure out your absolute position on Earth. At a conference one of the NIST Boulder folks mentioned that the frequency of clocks in the development pipeline was so precise/stable that it could pick up time dilation effects from moving the clock from a desk to the floor! An instrument like that combined with accurate gravimetery data could be used to map the earth, find concentrations of natural resources, or possibly act as a standalone positioning system.

It’s a chicken and egg problem though. I don’t know if we will know all the uses for it until we build it, so instead of saying “why build it?” I think it makes sense to go and just build it :)

> […] downstream impact to secondary standards in metrology labs which ultimately (through transfer standards) would result in better calibrations of production systems.

What is a "transfer standard"? Can you name an example(s)?

A transfer standard is an artifact that is produced at a standards laboratory to be sent to another laboratory for the purpose of calibration of the 2nd laboratory's processes and equipment.

"Measurement Assurance Programs are quality control programs for calibrating a customer's entire measurement system. In a typical MAP, a stable artifact or set of artifacts called transfer standards are first measured by NIST and then sent to a customer's laboratory for a series of measurements. The transfer standards are then returned to NIST for re-measurement, along with the participating laboratory's results. NIST reports its comparative findings to the customer and, when necessary, offers guidance on achieving and maintaining measurement quality.

Successful use of a NIST MAP requires that the customer make periodic measurements of in-house check standards to estimate their measurement process uncertainty and to ensure that the measurement process remains in a state of statistical control. Unless a laboratory has a measurement quality assurance program to monitor its own measurement process parameters continuously, there is no value in participating in a MAP. In fact, NIST recommends that its customers establish and use a measurement quality assurance program to monitor their measurement parameters, whether or not they participate in a MAP." [1]

https://www.nist.gov/calibrations/policies#:~:text=In%20a%20....

Time and frequency are to a large extent the basis of most of our measurement technology and our system of units.

The metre is defined based on the speed of light. It is realized using interferometry, for which you need lasers of precisely calibrated frequency.

Voltage is measured using Josephson junctions and a calibrated frequency.

The new definition of the kg is based on a Kibble balance (aka a watt balance) which relies on measuring position, current, and voltage, all of which rely on precise frequency.

If you look at the summary diagram of the new SI system https://en.wikipedia.org/wiki/2019_redefinition_of_the_SI_ba... you can see that all the base units depend on the second (except the mol).

Perhaps not Strontium, but the other clock types they talk about, Rubidium and Caesium, are readily available commercially at quite affordable prices. Rubidium clocks in particular are very widely available and every RF lab will have one.

I wonder how the "conversion" between cycles of different materials work. The spec says Cesium but it seems other clocks use different materials, how do you know X cycles in Cesium is Y cycles in Strontium for example

It is not possible to calculate the ratio between the transition frequencies in different atoms with anywhere near sufficient accuracy for metrological applications. Therefore, only Cesium clocks can serve as primary standards for the SI second. Other clocks can be calibrated to Cesium clocks in order to serve as secondary standards, but it is not possible to measure a frequency in Hz with an accuracy that is better than that of the best Cesium clocks. Probably the definition of the SI second will be changed at some point, but currently there are still too many possible candidate species and transitions without a clear favorite.

Reply on news.ycombinator.com