From coloured stars to colour flow imaging – or when is ‘Doppler’ not Doppler?

Abstract

In 1842 Christian Doppler wrote a paper entitled ‘On the Coloured Light of the Double Stars’ in which he sort to explain why some stars appear bluish and some reddish. His theory was that when a star is approaching an observer the wavelength of its light will be shorter and the star thus appears to be blue, whilst when a star is receding from an observer the wavelengths of its light will be longer and it would appear to be red. Doppler was actually wrong on a number of counts as to the explanation for the colour of stars (as will be explained during the talk) although his theory was correct. It is therefore perhaps slightly ironic that Doppler's theory is now widely used in astronomy to measure the speed with which stars are receding from the earth. Doppler's theory was challenged by the young Dutch scientist Christophorus Buys-Ballot, who believed that Doppler's theory could not explain the colours of double stars and set out to prove experimentally, using sound rather that light, that the Doppler effect did not exist. He carried out a number of experiments involving horn players travelling on the Amsterdam-Utrecht railway line, but instead of disproving the Doppler effect, showed that Doppler's theory was correct. The possibility of using the Doppler effect for measuring blood flow was introduced by Shigeo Satomura and colleagues in a number of papers published between 1955 and 1960, and a major advance took place around 1969 when Peter Wells, Pierre Peronneau, and Donald Baker independently described Pulsed ‘Doppler’ units which enabled blood flow signals from specific ranges to be acquired. Further important advances took place around 1982 when Kasai and Namekawa introduced real-time blood flow imaging. Curiously enough, however, whilst continuous wave Doppler units rely on the Doppler shift on the back-scattered ultrasound to measure blood flow velocity, instruments which use pulsed waves do not (and could not for reasons that will once again be explained during the talk) measure the Doppler shift on individual pulses, but rather measure the relative phase-shifts of returning signals with respect to a master oscillator in each inter-pulse interval. Fortunately for most practical purposes the resulting signal is equivalent to a Doppler shift signal. In summary, this talk will explain why the apparent colour of stars could not be due to the Doppler effect, and why pulsed ‘Doppler’ devices cannot detect Doppler frequency shifts.