Welcome to practical physicsPracticle physics - practical activities designed for use in the classroom with 11 to 19 year olds
 

Light – waves or particles?

What is light?

Two thinking models of light were argued about for many years. In the 17th century, Isaac Newton decided in favour of a particle theory because this would account for straight rays and sharp shadows. Around the same time, Huygens developed a wave theory. Much later (about 1800) the wave model of light gained strong experimental support from the work of Thomas Young. 
 
There were two serious difficulties with Newton’s particle theory. It failed to explain 

  • the fact that when a beam of light passes from one medium to another, some of it is reflected and some of it is refracted. 
  • the phenomena of interference and diffraction. 

 
These two difficulties forced Newton to suggest a strange scheme that endowed the interface with alternating ‘fits’ of easy reflection and easy transmission. He knew very well that this implied some periodic activity connected with moving particles, so that one could assign a wavelength - yet he did not change to a wave theory, because he considered sharp shadows too difficult to reconcile with waves. 
 
Long after Newton’s time, the speed of light in water was measured and compared with the speed of light in air. Light travels slower in water. That is generally regarded as a crucial experiment which decided clearly against the particle theory. Yet most crucial experiments, if not all, are only crucial - leading to an inescapable decision - if one sticks to the full details of the theories being tested. 
 
Newton’s prediction assumed that the mass of the light-particle remains constant, and that its component of momentum along the interface is conserved-by symmetry and Newton’s second law. On the basis of constant mass that predicts greater speed in water. If we allow the particle to change its mass but conserve its total kinetic energy, the prediction is reversed; smaller speed in water. Thus not even here is there a fully crucial test unless we choose a particular set of assumptions