An article has recently appeared in the press claiming that there is a problem with the world's standard of mass. What exactly is going on?
Very little. By definition the unit of mass, the kilogram, is the mass of an object known as the international prototype. Therefore, to know the mass in kilograms of any other object, it must be compared to the international prototype. Except in extraordinary cases, these comparisons are indirect and rely on a chain of standards ultimately derived from the international prototype. But for the highest accuracy in the chain, comparisons are made with respect to the international prototype. For this, we use special balances known as comparators. The highest-accuracy measurements are rarely required, the last series having been carried out between 1988 and 1992. The results of the 1988-1992 measurement campaign confirm a general trend that had already been discovered by the BIPM in the 1950s; namely that the copies are gaining slightly in mass with respect to the international prototype. But the comparator can only measure mass differences, so perhaps the international prototype is really losing mass with respect to the majority of copies. Whichever is happening, this feature has long been known. Furthermore, the effect is so small that it has been widely agreed by international bodies that it has no practical importance.
If this behaviour was known by 1992, why is it only now being reported?
Those with long memories will recall that this story has already appeared several times in the popular media. The first round of articles made its way around the world in 1990 and a second round followed in 2003. But one should also distinguish between stories in the popular media, which are interesting but by their nature incomplete, and detailed articles published in peer-reviewed scientific journals. The latter are essential in order to place results before a broad scientific community. The first of the scientific articles was written by the BIPM Director of the day and published in 1991. A full report of the 1988-1992 measurements was published in 1993 and BIPM staff continue to publish review articles on this subject.
Why has the "kilogram story" appeared again in the media?
The 1991 paper written by the BIPM Director of the day called upon the community of measurement experts to find in due course a way of comparing the mass of the international prototype to a fundamental constant of physics. This would be the only sure way to resolve the following question: is the international prototype getting lighter, or are most of the copies getting heavier, or are none of these objects stable in mass with respect to the fundamental constants? The suggestion was that the link between the present kilogram unit and a fundamental constant should be made to an accuracy of 20 parts per billion (20 micrograms in one kilogram). The challenge to the measurement community was made more formally in a 1995 Resolution of the General Conference for Weights and Measures. Since 1995, a second and distinct reason for doing this work has been generally accepted: the kilogram is the last remaining base unit still defined by an artefact rather than by a fundamental constant of nature. As a consequence, the values of some fundamental constants (Planck constant, elementary charge, Avogadro constant, etc.) are currently measured in terms of the international prototype, which has nothing fundamental about it. For this reason, many in the scientific community see an immediate benefit to defining the unit of mass, the kilogram, in terms of a fundamental constant of physics. But this redefinition will only be practical for the mass community if the present kilogram can be linked to the chosen constant with sufficient accuracy.
Thus the linking experiments are very important. They are also very difficult. The BIPM is heavily committed to supporting two such experiments. One promising approach aims to link the kilogram unit to the Avogadro constant. This is a complicated experiment but a major step along the way has recently been achieved. One of our partners in this world-wide collaborative work, justifiably proud of their contribution, issued a press release in June 2007 and this has generated considerable interest. [The second type of linking experiment relies on a device known as a watt balance and links the kilogram to the Planck constant. This approach was pioneered at the NPL (UK), and to date the lowest uncertainty is claimed by a group working at NIST in the USA. The product of the Planck constant and the Avogadro constant is a quantity called the molar Planck constant, whose value is already known to very high accuracy. The value of the molar Planck constant thus provides a valuable self-consistency check among the experiments linking the kilogram to the Planck or Avogadro constants.]
But why are the international prototype and its copies diverging in mass? There must be some explanation.
One can see from the graph that we are talking about a change of about 0.5 micrograms per year. An effect this small is tough to study over a time period of a few years. Actually, it should not surprising that manufactured objects or "artefacts" are not perfect. The wonder, perhaps, is that they have served us so well. The fact that this behaviour is of no practical consequence also means that it has not been studied intensively.
Nevertheless, measurement scientists have provided some suggestions. The international prototype and its copies are 90% platinum and this element is known to catalyse chemical reactions; perhaps there is some model that could explain the observed phenomena (but see below). In addition, a study has shown that mercury in the atmosphere will bond to a platinum surface. Ultimately, the BIPM is putting its primary effort into supporting the linking experiments that will allow a redefinition of the kilogram in terms of a fundamental constant of physics.
It is, nevertheless, relatively simple to rule out many hypotheses for the observed long-term behaviour of the prototypes. Quite a few such ideas are suggested to the BIPM by the public.
- As mentioned above, the data we obtain are derived from a mass comparator, which can only measure the difference in mass between two objects. The comparator is a special type of balance, but it operates in essentially the same way as an old-fashioned equal-arm balance (only better). We can rule out any effect that is common to both objects being compared, because the comparator simply cannot detect it. Among these undetectable effects are:
- gravity. Each object experiences the same pull of gravity so we do not need to account for gravity in determining the mass difference. Some of you may know that the pull of gravity and a given location undergoes small changes having a 12-hour cycle. But our results are unaffected by this effect. Remember, the kilogram is the unit of mass. Weight involves gravity but we are not measuring weight. We only require that the pull of gravity be the same for both artefacts at the time of the comparison.
- environment. The effect we want to explain is a change in mass difference over time between the international prototype and its copies. A subset of the latter is the group of "official copies", which are always stored with the international prototype in the same place and under the same conditions. Therefore, storage environment cannot explain the effect.
- Any viable explanation must correlate with time in order to produce an effect of 50 micrograms in 100 years. To rule out the possibility that the effect is due to the way we clean the international prototype and its copies, or to the number of times they are used in the comparator, or to any other such operations, we can see if the data really correlates with time. The alternative would be that it correlates with the number of times a given operation is performed. While we cannot control the passage of time, we can choose the number of operations that we carry out on a particular artefact. Such tests always confirm that time is the important parameter and not the number of times a particular operation is performed.