NPR last week had a story about the changing kilogram:
More than a century ago, a small metal cylinder was forged in London and sent to a leafy suburb of Paris. The cylinder was about the size of a salt shaker and made of an alloy of platinum and iridium, an advanced material at the time.
In Paris, scientists polished and weighed it carefully, until they determined that it was exactly one kilogram, around 2.2 pounds. Then, by international treaty, they declared it to be the international standard.
Since 1889, the year the Eiffel Tower opened, that cylinder has been the standard against which every other kilogram on the planet has been judged. But that's creating problems. According to scientists, the cylinder's mass appears to be changing.
This is not a new problem-- it was old news when I was a student at NIST, and people at NIST and elsewhere have been working on alternatives to the physical kilogram standard for decades. The physical kilogram standard is really an artifact of a past age-- all the other major standard measures have been redefined in terms of more universal interactions, as explained by Physics Buzz.
Attempts to redefine the kilogram have yet to yield anything, though. The problem, as always, is the gravity is so damnably weak.
Gravity may not seem like a weak force, but it is. The simplest illustration of gravity's weakness is the old "rub-a-balloon-on-your-hair-and-stick-it-to-the-ceiling" trick. When you do that, the attractive force of maybe ten billion extra electrons on the balloon is enough to hold it up against the gravitational pull of the entire Earth pulling on a billion trillion atoms in the balloon. Gravity is preposterously weak compared to the electromagnetic force.
This is a nice feature for those of us who like to play basketball, but not such a good thing if you're trying to come up with a standard for mass. Tiny electric and magnetic fields are enough to throw off any attempt to measure mass by measuring the gravitational force, simply because of the huge disparity in the strengths of the two forces.
There are a couple of approaches currently being pursued for alternatives to the existing physical kilogram. The one highlighted in the NPR piece is the Watt Balance, which relates the gravitational force on an object to the current and voltage in an electrical circuit. Voltage and current can be defined in terms of fundamental physical constants through the Josephson and Hall effects, so this lets you express a mass in terms of Planck's constant and a bunch of other numbers. In effect, this would give you a mass standard whose precision is determined by the precision of the value of Planck's constant.
another approach, which I kind of get a kick out of, is to basically make a better physical standard. There are a couple different approaches, the best-developed of which uses extremely precise single-crystal silicon spheres. The idea is to redefine mass in terms of Avogadro's number, so one kilogram would be the mass of a specified number of silicon atoms, with the sphere being the physical realization of that number of atoms.
To date, neither of these approaches has managed to significantly improve on the existing mass standard, so for now, the platinum-irridium cylinder, flawed though it is, remains the standard. It remains to be seen whether any of the current efforts can replace it, or if it will be around until the kilogram is redefined in terms of the force between Planck-mass black holes produced in some sort of giant future accelerator.
- Log in to post comments
More like this
Can't you just define a kilogram as xteen fucktillion craptillion atoms of whatthefuckever pure element?
I like the atom-counting approaches as well. Define the Avogadro number exactly; define the kilogram in terms of that number of carbon or silicon or whatever atom you want.
It's technically difficult, of course, but honestly even with extra uncertainty it's still probably worth being able to jettison a standard based on a variable physical prototype.
So a 1kg cylinder of platinum and iridium is just chilling out and its mass is changing? (I guess, in theory, since it defines 1kg that its mass isn't changing but what 1kg represents is changing. Forgive me, being a biologist I'm not used to uncertainty coming from this direction :P)
Anyhow, what's going on with it? Like is it oxidizing or are little platinum or iridium atoms being raptured off to elemental heaven? Is the cleaning crew picking it up to dust and not wearing gloves?
It's a question of whether you can measure the number of atoms in a mole more precisely than any other reasonable standard. If you can, then by all means do it this way. Similar reasoning led to the redefinition of the meter in terms of the second, with the speed of light being by definition 299792458 meters per second. If the atom-counting technique is currently less precise than weighing a platinum-iridium artifact, then we have to live with the current standard, or adopt some other technique which is more precise than both. Presumably there is some feasible technique out there that would be more precise than weighing an artifact, or we wouldn't be having this discussion.
Do we really need to be able to measure it perfectly?
For instance, the AU is (or at least, at one time, was) defined as the semi-major axis of Earth's orbit around the Sun. We didn't know it perfectly, and we never held things up next to it to measure it, but it was a great measurement for what it was used for.
Could we do something like define the kilogram to be the energy in 2.26222x10^35 photons of wavelength 5000 angstroms, divided by the speed of light squared? That's not practical for measuring, but it is a definition in terms of other stuff that already has modern definitions.
I vote making the kilogram heavier so that I can weigh less.
Rob: It doesn't have to be perfect, but it does have to be (1) as precise as practical and (2) useful for measurement. The AU is/was convenient for measuring distances within the solar system, because it was known with relatively (compared to other available yardsticks near that scale) high precision, and distances measured in those units are of a convenient magnitude: O(1) for the inner planets up to O(100) for heliospheric boundaries. You wouldn't want to use it for interstellar distances; it's more than 250000 AU to Alpha Centauri. (Using it as an extragalactic yardstick is an even worse idea; the Earth's orbit changes on those time scales.)
As you say, your proposed definition is not practical for measuring. That's why it's not a good choice. At least we can weigh a platinum-iridium artifact.
@1: That is the idea of the sphere. Part of that proposal is to define Avogadro's number to be the cube of an integer.
I think they should count C-60 (buckyball) molecules made of isotopically pure C12, but only as a matter of principle.
@5: We need it to be able to measure the new standard MORE precisely and more accurately than we can measure the current standard. Otherwise, what is the point?
Ideally, it should be easily realized and stable, like the other standards that have been adopted in the last 30 years. As much as I like the atom counting method, the electromagnetic balance has many advantages when it comes down to reproducing the standard kilogram in each country and, for that matter, each state. Remember, the masses used to check the scales at each checkout lane in each grocery story are traceable back to the one in Paris.
@CCPhysicist, wouldn't there be a value in having a definition that is independent of a physical artefact, though? Define a kilogram as a given number of a given type of atoms, and then accept that for the time being the current best physical approximation of the definition is the standard weight. Once a more accurate weight - or other calibration method - has been created that will be better, but the definition itself no longer needs to change.
How can they tell that the mass of the standard kilogram is changing? It's not like they dumped it to a backup disk ten years ago and now the SHA2 hash is different.
I wonder if the atom-counting method can be re-defined to rely on some surrogate constant process like radioactivity. For example natural thorium is pure Th-232 with the half-life 14 billion years
Comrade PhysioProf at 1 and Janne at 9
This is a very attractive idea in principle and underlies the attempts to use a sphere. The hard part is counting the atoms. How are you going to know how many atoms are in your lump of stuff? You are not allowed to check by weighing it.
The best method anyone has come up with so far is to make a sphere of very pure stuff. It must be very pure to be accurately reproducible - a scoche of rogue atoms will affect the weight. The sphere must very accurately spherical. Depressions or plateaus will not be reproducible in other labs and will make the diameter ambiguous.
Kaleburg@10: Actually, you're pretty close; a large number of replica copies were made back in 1884, the year after the International Prototype Kilogram (IPK) was certified. The masses of these were carefully compared to each other and to the IPK, and these secondary standards were distributed around the world to serve as national standards. Comparisons between these standards over the years indicate that their masses are diverging from each other.
The Wikipedia article has a pretty good summary: http://en.wikipedia.org/wiki/Kilogram.
I've published (elsewhere) my proposal for a specific prime number as a standard Avogadro's constant. You then make a perfect cube of this number of atoms on edge to get the official Avogadro's constant number of atoms.
84446891 is the closest prime to the cube root of the present best estimate of Avogadro's constant. Frank J. Donahoe (Wilkes University) points out that its cube is only 5.53 x 10^-7 larger than the present best estimate of Avogadro's constant.
Reference: Letter to the Editor, American Scientist, May-June 2007, p.195.
Why not just define it in terms of eV/c^2?