Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

The complex refractive index vs. permittivity

  1. Nov 4, 2014 #1
    I have a process of thought and would like to run past some other minds to point out if I am incorrect in my thinking.

    I am looking into conductivity in high frequencies and a lot of papers I am looking up list a complex refractive index. They list something as in nAg = 0.1453 + j11.3587. (Excuse the imaginary symbol, j, I come from an electrical engineering background.)

    Various sources inform that [itex]n = \sqrt {ε_r} [/itex]. [Link]

    Also that [itex] ε = ε' - jε'' = ε_1 - j \frac {σ}{ω}[/itex] [Electromagnetics for Engineers; Fawwaz T Ulaby]

    We can get from [Wikipedia.org] that:
    [itex] ε = ε_1 + jε_2 = (n + j κ)^2 = n^2 + j 2nκ - κ^2 [/itex]
    [itex] ε_1 = n - κ^2; ε_2 = 2nκ [/itex]
    *Possible contradiction in Wikipedia vs. Ulaby*
    Ulaby states - [itex] ε = ε' - jε''[/itex]
    Wikipedia states - [itex] ε = ε_1 + jε_2 [/itex]
    So, does:
    [itex] 2nκ = \frac {σ}{ω} [/itex]
    n - real part refractive index
    κ - Complex part refractive index
    σ - conductivity
    ω - angular frequency
    I am kind of running this by so that someone can say "Yup" but also, I think, writing it down helps me to understand a little better. Also, this is the first time of me using LaTeX and wanted to keep trying it out.
    Last edited: Nov 4, 2014
  2. jcsd
  3. Nov 4, 2014 #2


    User Avatar
    Science Advisor

    I don't know why Ulaby uses a minus sign in the definition of epsilon''. Must be an engineering convention.
  4. Nov 4, 2014 #3


    User Avatar
    Science Advisor
    Gold Member

    Isn't that just an approximation?
    If you mean this equation, I believe it's a dash; not a negative sign.
  5. Nov 5, 2014 #4
    It was a proof from Maxwell's Equations
    [itex] \nabla \times H(t)= J(t) + jωεE(t) [/itex]
    [itex] J(t) = σE(t) [/itex]

    [itex] \nabla \times H(t) = (σ + jωε) E(t) [/itex]
    [itex] \nabla \times H(t) = jω (ε + \frac {σ}{jω}) E(t) [/itex]
    [itex] \nabla \times H(t) = jω (ε - j \frac {σ}{ω}) E(t) [/itex]
    [itex] \nabla \times H(t) = jω (ε' - j ε'') E(t) [/itex]
    H(t) - Time Varying Magnetic Field with direction (Still getting used to the [itex] coding)
    E(t) - Time Varying Electric Field with direction
    J(t) - Current Density
    σ - Conductance
    ω - Angular Frequency (2π f)
    ε - Permittivity​

    Though I could well be wrong in my assumptions as this provides a relationship between Time Varying (TV) Magnetic fields and TV Electric fields...
  6. Nov 6, 2014 #5
    I suppose the question would then be, is the permittivity gained through complex refractive index similar to that of the permittivity relating Magnetic Fields and Electric fields?
  7. Nov 6, 2014 #6


    User Avatar
    Science Advisor

    It certainly is, but usually, we use it in other frequency regions. There are also different conventions. E.g. in optics it is usual to set B=H. All potential magnetic effects are included in a dependence of the dielectric function on the wavevector, i.e. ##\epsilon(\mathbf{k},\omega)##.
  8. Nov 13, 2014 #7
    Previously I had given the proof:
    [itex]\varepsilon - j \frac{\sigma}{\omega} = n^2 - K^2 + j2nK[/itex]
    Should it be:
    [itex]\varepsilon_r = n^2 - K^2 + j2nK[/itex]
    And so:
    [itex]\varepsilon_0 \varepsilon_r = \varepsilon_0 \varepsilon_r - j \frac{\sigma}{\omega} [/itex]
    [itex]\varepsilon_r = \varepsilon_r - j\frac{\sigma}{\omega \varepsilon_0}[/itex]
    [itex]\varepsilon_r - j \frac{\sigma}{\omega \varepsilon_0} = n^2 - K^2 + j2nK[/itex]
    Does this sound correct?
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook