Big Test Coming Up For Kilogram Redefinition (ieee.org)

szotz writes: Electromechanical balances have got to be better than an aged lump of platinum and iridium right? Teams are working to get kilograms measured and shipped to Paris in time for a test to see whether the technology (along with another that uses ultrapure silicon spheres) is now ready to redefine the kilogram. Why is this redefinition interesting? Because it's about using physics to overcome one problem with weight standards based on tightly held exemplars in standards bodies' inner sanctums: the mass of those exemplars can change, however subtly, introducing uncertainty and confusion. From the article: The world's metrologists aim to change this state of affairs in 2018 by fixing the kilogram to the Planck constant, a fundamental physical constant. That shift would, at least in principle, allow any laboratory to "realize" the kilogram from scratch with a series of experiments and specialized equipment. But for that scheme to work, the kilogram derived by one laboratory must be the same as those derived by others.

Comments

Re:So really...

[sehlat]

So really it is just a global scientific test of who's is bigger.

No, it's a multiplayer game of "You show me yours and I'll show you mine."

Measurements

[Okian Warrior]

Why the planck constant then? Why not e, or (pi), or any other constant, for that matter? If you're going to change the definition, isn't it just a matter of choosing the close enough factor?

By the way, I'm asking. I am ignorant about this.

The fundamental distinction between math and physics is measurement. We need to base the physical constants on something measured from the universe we're interested in.

As a simple example, mathematics defines and explores 3 basic forms of geometry: Euclidean, hyperbolic, and elliptic.

The distinction between these is based on the curvature of space as defined by the behaviour of parallel lines: if parallel lines eventually meet, then space has positive curvature like the surface of a sphere. If parallel lines diverge, then space has a negative curvature like the surface of a saddle. And if parallel lines stay parallel, then space has zero curvature and is Euclidean.

Three equally valid forms of geometry, but which one does the universe have? To choose the correct model, we have to measure the actual universe.

The same is true with the fundamental physical constants. There's any number of ways to base our measurements on pure math, but these don't necessarily reflect the reality we live in.

To do that, we need to take a measurement.

Re:Measurements

[Anonymous Coward]

To bolster the argument, look at the fine-structure constant. When Arnold Sommerfeld introduced the constant in 1916, Arthur Eddington argued that you could get to it by pure math and found that for completely logical reasons, the constant should be exactly 1/136. When later measurements put the value closer to 1/137, he discovered an error in his deduction and published a new paper that the constant should be for even more logical reasons exactly 1/137. Currently measurements put the value of the fine-structure constant at about 1/137.036, and no numerological explanation so far has been accepted.

Re:Measurements

[tal_mud]

Though the above is true, it is *not* the reason why we can't base the kilo on some arbitrary multiple of Pi. The point is that we want to be able to actually reproduce the reference kilo in any lab. Take for example the definition of the meter as the distance travelled by light in 1/(299,792,458) of a second. A lab can actually measure the length light travels in that amount of time and thus reproduce the canonical meter. If we just defined the meter as 1/Pi, there would be no way to convert this number to an actual length.

When the article says that they define the kilo in terms of Planck's constant, they mean that you take the ratio of all sorts of measured quantities in the lab and the laws of physics say that the result should be the mass of what you are measuring times Planck's constant. The true emphasis is that the measurement is proportional to the mass of what you measured, not that the constant of proportionality is Planck's constant (except of course for the fact that we assume that the constant of proportionality, Planck's constant, being part of the fundamental laws of physics, is independent of where and when we do the measurement (at least in the time and distance scales that physics has managed to probe).

Re:Measurements

[serviscope_minor]

Well, this certainly qualifies as news for nerds! It's news, technical and amazingly esoteric.

Why the planck constant then? Why not e, or (pi), or any other constant, for that matter?

Neither e nor pi are physical constants. They are unitless mathematical constants, so you'd have to specify e or pi *somethings*. It's the somethings that are important at which point neither e nor pi would come into it all that much.

By way of example:

The second is defined in terms of a physical constant: "the duration of 9192631770 cycles of the radiation corresponding to the transition between the two hyperfine levels of the ground state of the caesium 133 atom." The important thing about this is that this was not the original definition of a second. The second was a bit more vague, so at some point a bunch of metrologists got together and declared that from now on this SHALL be the definition of a second and shall supercede all previous definitions.

The kit to measure a second is withing reach of well funded science labs and can be reproduced independently. You need the high frequency counter (capable of 10GHz operation), some pure caesium, and an assload of expensive support equipment and liquid helium and you can measure a second.

Once you have the second, you can move on. The meter is defined in terms of the second and the speed of light: as the distance travelled by light in a specific fraction (1/299 792 458) of a second. Much like before, this is a declaration by fiat, and is very very close to and supersedes the old platinum iriduim rod in Paris.

Now, here's where it gets interesting!

First, an aside:

The reason for Planck's constant comes down to what is colloquially known as E=mc^2, or more generally E= h v where v is momentum and h is plank's constant. In other words, Planck's constant connects energy, mass and time.

Here's a nice link:

www.bipm.org/utils/common/pdf/RoySoc/Michael_Stock.pdf

It's a bit more detailed, but essentially it relates the Kg, Planck's constant and a few others which are known. So, if you know what the Kg is exactly then you can measure Planck's constant with a Watt balance very accurately.

So what you do is calibrate the Watt balance with the prototype Kg, and measure Planck's constant. You then declare (by fiat) that Planck's constant is EXACTLY what you've written down and so now the Kg is defined in terms of that number, not the other way around.

In principal, now someone can build their own Watt balance, plug in the numbers which are now just numbers and measure their own chunk of metal to find out how much it weighs in Kg.

So that is a nutshell is why h is appropriate and pi and e aren't.

The other option is to build a very very pure, very very precise silicon sphere, in which case the Kg will be essentially determined by a single number which is the number of silicon atoms in a Kg. That will be measures in terms of the meter (for both the bond spacing of silicon and the radius of the sphere). In that case, Planck's constant will still be defined in terms of the Kg, not the reverse. In this case, pi would make it into the definition, via the volume of a sphere, of course, but in a somewhat peripheral role.

The question is whether we (collectively) can make silicon spheres more accurately than we can make Watt balances, or the reverse, right now.

A weighty matter

[penguinoid]

This is a massive development.

Interest groups

[worf_mo]

Weight Watchers International weighted in on the discussion requesting the new kg to be defined at twice the weight of the current kg ("Yes Sandy, I lost half of my weight in the blink of an eye!"), while grocers all over the planet petitioned to divide the current value by four.

All kilograms...

[SharpFang]

All kilograms are equal

but some are more equal than others.

Don't change the definition! 1 kg = 1024 g

[thisisauniqueid]

I bet they're going to change the definition from 1 kg = 1024 grams to 1 kg = 1000 grams. And we'll probably have to write "kig" too to make sure we don't get confused about the old definition.

Re:conventions and relativity

[Anonymous Coward]

Yes. How many atoms are there in 1 liter? I guess you could go more 'general' and say how many drops of water? But then, how big are the drops of water? What about the absorbtion rate of the material, can it only be measured in glass? is this deformed at 1 atmosphere pressure? evaporation rates?

Then you get to mineral/chemical impurities, atmospheric disturbances, etc.

Maybe 1 kg vs 0.999997 kg doesn't mater to you but there are many cases where it will. And calibrating our scales to allow that fine-grained approach is nothing but a Good Thing.

Re:conventions and relativity

[stevelinton]

Whatever standard you adopt needs to be reproducable within the limits of the best current measurements by any other technique. Otherwise when people want a stable reproducible result they will use the other technique and the standard won't have worked. Measuring volume of water, purity, temperature and pressure is just not precisely reproducible enough

Re:Will a Litre be Redefined?

[alexhs]

I thought it was a stupid conversion mistake, but investigating on the topic :

One litre of liquid water has a mass of almost exactly one kilogram, due to the gram being defined in 1795 as one cubic centimetre of water at the temperature of melting ice.

So, originally as I wrote.

From 1901 to 1964, the litre was defined as the volume of one kilogram of pure water at maximum density and standard pressure. The kilogram was in turn specified as the mass of a platinum/iridium cylinder held at Sèvres in France and was intended to be of the same mass as the 1 litre of water referred to above. It was subsequently discovered that the cylinder was around 28 parts per million too large and thus, during this time, a litre was about 1.000028 dm3.

Oops. Not too bad, given that at that time the metre was wrong too:

it was later determined that the first prototype metre bar was short by about 200 micrometres because of miscalculation of the flattening of the Earth, making the prototype about 0.02% shorter than the original proposed definition of the metre.

And all is fine again:

In 1964, the definition relating the litre to mass was abandoned in favour of the current one.

The litre [...] is an SI accepted metric system unit of volume equal to 1 cubic decimetre (dm3), 1,000 cubic centimetres (cm3) or 1/1,000 cubic metre.

Sources:
Litre
Metre

Since a Kg measures mass

[rossdee]

It should be defined by Pope Francis
He used to be a chemist, and is infallible.