Ben Goldacre
Thursday March 24, 2005
The Guardian

· Welcome to Bad Science, where we use GCSE physics to solve science problems set by sinister international companies. Thinking back to when you were 15, you will remember that heat is transferred from a warmer object to a cooler one along a temperature gradient. In fact, forget that, think back to any common sense situation you have encountered: ice cubes do not heat up cups of tea; when you hold a cool piece of metal in a fire, it gets warmer; and you cannot boil a bowl of milk by sitting it in a pan of boiling water, because milk boils at a temperature higher than the boiling point of water. Hold that thought. Water boils at about 100C, milk boils at something a bit higher than that, so heat will not transfer against the gradient from the water, that is 100C, to the milk, to make that milk even hotter than 100C.

· Where am I going with this? Straight to the “research and studies” section of the Penta “restructured drinking water” website where it has a jolly scientific paper showing that the boiling point of Penta is higher than the boiling point of de-ionised water, written by Dr Boris V Nemzer and Professor Andrew Dickson (tinyurl.com/527h8). Wow. Penta really must be different. So I couldn’t quite believe it when Mark Atkins wrote in to point out an interesting quirk in their apparatus. According to the study, the de-ionised water boils at 100.082C (+/- 0.008), and the Penta water boils at 100.125C (+/- 0.008), so the boiling point of Penta is 0.043C higher. The setup was this: they had a glass beaker of boiling de-ionised water, and in this they floated a smaller glass container filled with either Penta or de-ionised water, boiling, with its temperature measured with a probe. You can see where we’re going with this. If Penta has a higher boiling point, how did it boil at all, in a bath of de-ionised water that has a lower boiling point? I tracked down Prof Dickson at University of California, San Diego, where he suggested that this surprising phenomena might be explained by the water in the larger beaker being “slightly superheated”.

· I contacted Ed Tarleton, Bad Science emailer and research physicist, to run it by him. Not only did Ed agree with us, he also checked their thermometer and found that the Hart 1504 has an accuracy of only 0.02C at 100C. “The error is a lot larger than they make out,” said Ed, “which might explain how they observed the impossible and had heat transferring from a cooler to a warmer body.” Are we really right? We can’t quite believe it. Read the paper and let us know.