(Note: this is a longer version of a column which appeared in EE Times print edition. )
The kilogram, the standard unit of mass, is the only one of the seven key International System of Units (SI) units that is still based on a physical artifact. It's a cylinder of platinum-iridium alloy, about 39 mm high and 39 mm in diameter, housed in a temperature-controlled vault near Paris. Other standards, such as the meter and the second, are now defined in terms of reproducible standards. For the meter, it's the wavelength of a transition of krypton energy states; for the second, it's the period of a hyperfine transition of cesium.
The lonely kilogram has no such equivalent reproducible standards, notes an article in the April 2006 issue of Physics Today . Other metrology labs must use a tiered set of copies, and copies of copies, of the master kilogram, depending on where they are in the hierarchy. Even the master unit has its problems: it has lost weight, about 50 parts per billion (ppb), while being handled and used for calibration of those secondary standards. In addition, there are unavoidable measurement errors when comparing the master to other units. Scientists say that this error is also on the order of 50 ppb.
The goal of replacing this master artifact has challenged engineers and scientist for many years, and several alternatives are being explored at the International Bureau of Weights and Measures (BIPM, in French, www.bipm.fr) and the National Institute of Standards and Technology (NIST, www.nist.gov)) in the US. One approach involves a “watt balance” which has a test mass in a balance pan, which in turn is connected to a copper coil which creates an electromagnetic force which balances the mass. By measuring the coil current, the velocity and voltage of the coil (which is moving within a field produced by a superconducting magnet), and other factors, the mass in the pan is determined, after many calculations and correction factors are applied. .
While the unit of mass may seem removed from most electronic measurement, it's not. The ampere is formally defined in terms of the force between two infinitely long, parallel, current-carrying wires. It also affects, indirectly, precision metrology used in IC fabrication equipment, where billions of dollars are at stake.
There's another approach that sounds simple: just count the number of atoms of a pure substance, such as a silicon crystal, using x-ray and optical interferometry, to a specified count such as Avogadro's number. Though simple in theory, it's very tough in practice, as so many other things.
Then there's a radically different approach: forget about having the kilogram as a basic unit. Instead, define mass in terms of other physical constants such as Planck's constant h , the Josephson constant 2e/h , and the von Klitzing constant h/e2 .
However, the uncertainty using these constants is still around 50 ppb, so it's a lot of work for no real gain. Plus, there's something counterintuitive about defining something as tangible as mass in terms of such abstract constants. It looks like a cheat solution: “since we can't measure that item as well as we'd like to, we'll just conveniently define our way out of the problem.” This approach doesn't sit well with lots of folks, technical or not-so-technical.
It turns out that creating a reproducible mass standard (or “electronic kilogram”) is a very tough problem, as are all high-precision real-world measurements. At their core, they involve analog measurements and control, even when operating in the discrete world of counting atoms and assessing quantum states. Software and higher-processors won't help when you are diving down into these ppb regimes. It's with good reason that standards-settings bodies of the various countries are loath to change to a new definition, unless it can be clearly shown to offer better accuracy, easier measurement, and reduced uncertainty.
Bill Schweber , Site Editor, Planet Analog