Or using a small perfect sheet of graphene, and striking it with a known mass and velocity (similar to the above), then measuring the tensile force at the edge that results from the deflection. This force would be multiplied by the triangulation of the deflection and the distance to the point struck, so would be easier to measure. But it would have to take into account the amount of deformation and stretching that the graphene sheet would have.

I’m making these up as I go along, so they may not be feasible. I also keep thinking about how the Casimir force might be used to do this.

You have actually given the reason why the Kilogram cannot be defined in terms of the Newton in your own (second) post. Namely, that the Newton is defined from the Kilogram.
To make a long story short, units can never be defined in terms of each other. There are very valid scientific reasons for this.
Also, as to your point about “standard gravity”, gravity will differ by a very small yet very real amount at different points on Earth. Fundamental units such as the Kilogram need to be measured independently at different points on Earth (and at different times). Using a correction factor (to account for the different readings given by a scale due to elevation changes, for example) is _not_ sufficient.
Fundamental units _must_ be measured using real (non-corrected), _measured_ values that do _not_ change based on location (or time), so gravity cannot be used.

EDIT: Obviously, measured values will _always_ change by some tiny amount, regardless of what you do, but the very best measurements of the Kilogram are far less accurate than those of the meter and the speed of light. Trust me, the world’s best scientists haven’t simply been overlooking some simple solution to this problem.

Compare the average mass, of 15.9994(3)amu, to that of isotopically pure 16O, of 15.99491461956(16)amu. The (3) and (16) indicate the uncertainty. Clearly 1:1000000 from the average weight is not precise enough to be used as a mass standard since we are already able to measure some masses to one part in 10,0000,000,000,000

Still, it is possible to get isotopically pure water. The next question is, how do you get 1 cubic centimeter of it? Are you able to make a perfectly cubical container, such that the volume is off by no more than 1 part in 10,0000,000,000,000? Possibly. Though since it’s a liquid you’ll also have to worry about surface tension effects.

Then what about dissolved gasses in the water? The density of carbonated water is slightly higher than non-carbonated water, since it has the extra CO2 in the water. There’s CO2 in the air, so your water sample’s density will depend on the air around it.

The reason a sphere of isotopically pure silicon is proposed is in part because: 1) it’s possible to make isotopically pure, 2) it’s easier to make a sphere than a cube, and 3) it’s a solid, and 4) it’s easier to remove gasses and surface impurities.

The original title of the article was “How Much Does a Kilogram Weigh? Ask Again in 2014″ . Given that the Newton is defined from the kilogram I don’t see that weight value changing . And the standard Acceleration due to free-fall was defined in 1901 by the third conference of Weights And Measures, which also, coincidentally clarified that kilograms were units of mass.

@Jim H: The mass of a gram hasn’t been defined by water for over a century, as the article alludes. In fact, the mass of a cubic deciliter of Vienna Standard Mean Ocean Water at 4C is about 25 mg less than the IPK.