Boffins chill out with new temperature measurement

Why would you get excited about a new measurement of the Boltzmann constant that took six years to achieve? In the case of the UK scientists who made the measurement at the UK's National Physical Laboratory (NPL), it's because the long-held standard for temperature measurement is problematic at extremes. The kelvin and the …

[Mikey]

Well, at least now people get another option to bitch about, instead of just having Imperial and Metric to choose from. I get the feeling this'll go up there with the whole MiB/KiB thing... technically correct, but no-one will really use it in day-to-day life, as we have things that work fine already.

Also, who came up with this... Boltzmann, or one of his many brains?

[Michael 31]

Fair point...

... but actually this stuff does affect day-to-day life: it affects every single temperature measurement made on Earth. The fact you are not aware of any issues simply means the system we have built works well - it is based around international agreements co-ordinated by the International Bureau of Weights and Measures.

This work is simply part of the process of updating an obsolete definition of the unit of temperature. Unlike the SI base units for Mass, Length and Time, when we learned to measure temperature, we didn't know what we were measuring. If we had known then what we know now, we would have defined temperature in terms of molecular energy.

Michael de Podesta, lead author.

[frank ly]

Definition/Measurement

If the triple point of water can be calculated to happen under certain conditions, thus providing a reference situation for checking instrument measurement (temperature in this case); can the required conditions for another physical situation, happening at a much lower temperature, be calculated? e.g. the triple point of some gaseous element or anything else you can think of that happens at very low temperatures.

[Number6]

Re: Definition/Measurement

While it is possible to calculate another triple point, I suspect that relating it to the water measurement would involve use of Boltzmann's constant. Therefore they do need to know the constant's value before it's worth doing the rest.

[Michael 31]

Re: Definition/Measurement

We can't calculate the temperature of the triple point of water from first principles. Molecular dynamic simulations would only an estimate within about 1 part in 10^4 or about 0.1 K at best. The measurements are at least 1000 times better.

M

[Cliff]

That constant for the rest of us...

"One and a bit * 10^-23 somethingorothers"

Luckily nobody's life depends on my research :)

[John Smith 19]

boltzman's constant has *lots* of uses

Roughly from global climate models to LED designs.

When you have very big numbers flying around (like say the number of atoms in the atmosphere) those last digits can make quite a change.

This subject may seem dull but the methods used are SoA. Making the reading is tough. Proving the accuracy (IE identifying, scoping and removal all sources of error) is tougher still. Thumbs up for this and I'll hope anyone who needs that level of accuracy (or should be using that level) updates their systems accordingly.

[Michael 31]

Re: boltzman's constant has *lots* of uses

Thank you John. I am lead author on the paper and you are right. The work itself is like renewing the foundations on a structure - no one notices until they crack! - and the techniques developed are indeed state of the art - let me give you an example.

We measured the speed of sound in a copper resonator about 62 mm in radius - but we determined the average radius with an uncertainty of 2 nm - or about 12 copper atomic layers. We certainly have other ideas about how to exploit this capability.

Anyway: Thanks for your positivity :-)

[John Smith 19]

Re: boltzman's constant has *lots* of uses

"We measured the speed of sound in a copper resonator about 62 mm in radius - but we determined the average radius with an uncertainty of 2 nm - or about 12 copper atomic layers. We certainly have other ideas about how to exploit this capability.

Anyway: Thanks for your positivity :-)"

I'd read Jones "Instruments and Experiences" and have some idea of the work on Gravity Probe B, so I knew the simple number hides a lot of detail.

Didn't stop the downvote though. Apparently someone holds a grudge against improving the accuracy of physical constants.

I'd love to know why

[Destroy All Monsters]

Something is wrong here

a particle's temperature and its energy

While a particle has a well-defined energy (subject to some caveats due to non-zero h), it doesn't have a temperature, which is a statistical property of large ensembles of particles.

[tojb]

Re: Something is wrong here

Au contraire: if we posit that the particle is coupled to a heat bath, then its temperature is well defined but energy will fluctuate.

You can work in an NVE ensemble (energy fixed) or an NVT ensemble (temperature fixed) as you prefer, providing you don't bugger off and try to actually realise either of these cases in an experiment.

[Michael 31]

Re: Something is wrong here

Well spotted: the reference should of course be to the >average< energy of the particle which for ensembles of this magnitude at this density is a well-defined.

Michael de Podesta - lead author

[Anonymous Coward]

Using the current system

The local "fast food" outlets in london cannot produce more than luke warm burgers or fries, what hope is there for them to get a more accurate representation of "ready to eat temperature" with the new system

[Michael 31]

Re: Using the current system

Sadly, this work is unlikely to affect that particular problem.

Sorry.

MIchael de Podesta, lead author.

[itzman]

Re: Using the current system

I suggest that if that is the case, you take your custom elsewhere and exercise such democratic rights as you have to get the place closed down

[His Royal Yodaship]

Fixed it for you

The triple point is a definition rather than a measurement: the three states of water can be calculated to exist at -1.99 Hiltons and a partial vapour pressure of 0.0000061173 Norris/mm2 – in kelvin, 273.16K. All other measurements of temperature are made relative to this value.

[brainwrong]

Marks deducted...

"The triple point is a definition rather than a measurement: the three states of water can be calculated to exist at 0.1°C and a partial vapour pressure of 611.73 pascals – in kelvin, 273.16K."

Err, you missed a zero, it's 0.01 celcius. Everyone knows that, yet even QI got it wrong once.

And I'm pretty sure it's not calculated to be anything, the triple point of any pure substance is just whatever it happens to be, but it is fixed and easily reproducible which makes it a very good reference point.

Here's a thought:

The newton is defined in terms of the kilogram, which is a problem unit to define. Triple points don't just have a precise temperature, they also have a precise pressure. Could that be used to define the newton, and hence the kilogram?

[Richard Chirgwin]

Re: Marks deducted...

Typo corrected, thanks.

Richard Chirgwin

[Anonymous Coward]

orders of magnitude

"the uncertainty, which amounts to 0.7 parts per million,"

parts per billion I believe