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

Classic and Relativity Kinetic Energy

  1. Nov 15, 2006 #1
    It is said that at low speeds the relativistic kinetic energy formula will give you the same answers as the classical KE formula. I tried this, and it doesn't work. I get different answers.
    I am just wondering what is amiss here.

    How would one derive the Classical KE from the Relativistic KE?

    Thanks for the help,
  2. jcsd
  3. Nov 15, 2006 #2


    User Avatar
    Science Advisor
    Homework Helper


    [tex] KE=m_{0}c^{2}\left(\frac{1}{\sqrt{1-\frac{v^{2}}{c^{2}}}} -1\right)\simeq m_{0}c^{2}\left(1+\frac{v^{2}}{2c^{2}}-1\right) =m_{0}\frac{v^{2}}{2} [/tex]

  4. Nov 15, 2006 #3


    User Avatar

    Staff: Mentor

    Show us an example where you get different results (show your work!) and someone can probably tell you where you went wrong.
  5. Nov 15, 2006 #4
    Just to clarify what dextercioby said, the approximation comes from the binomial expansion
    [tex](x+y)^r=\sum_{k=0}^{\infty} {r \choose k}x^ky^{r-k}[/tex]
    [tex](x+y)^r=\sum_{k=0}^{\infty} \frac{r!}{k!\,(r-k)!}x^ky^{r-k}[/tex]
    which we can use to expand

    The first few terms of the expansion are

    So we have
    [tex]\frac{1}{\sqrt{1-\frac{v^2}{c^2}}} \simeq 1+\frac{v^2}{2c^2}[/tex]
    which can be used, as shown above, to show that the classical expression for kinetic energy approximates the relativistic expression.
    Last edited: Nov 15, 2006
  6. Nov 15, 2006 #5


    User Avatar
    Science Advisor
    Homework Helper
    Gold Member

    As written, you'd probably have to explain how to use r= -1/2 in the general formulas you've written.

    It may be conceptually simpler to emphasize the more fundamental Taylor series expansion.
  7. Nov 15, 2006 #6
    ah, i forgot about the non-integer r in the 'r choose k'..

    [tex]{r \choose k}={1 \over k!}\prod_{n=0}^{k-1}(r-n)=\frac{r(r-1)(r-2)\cdots(r-(k-1))}{k!}[/tex]

    according to http://en.wikipedia.org/wiki/Binomial_theorem" [Broken]
    Last edited by a moderator: May 2, 2017
  8. Nov 15, 2006 #7


    User Avatar

    you have a small problem assuming an analytic continuation that a formula that is shown valid for integer r is also valid for a non-integer. it is true, you just didn't make the connection. one that doesn't start with the binomial expansion but uses Taylor series is

    [tex]\frac{1}{\sqrt{1-\frac{v^2}{c^2}}} = \left( {1-\frac{v^2}{c^2}} \right)^{-\frac{1}{2}}

    \quad = e^{-\frac{1}{2} \log\left( 1-\frac{v^2}{c^2} \right)}}

    \quad \approx 1 - \frac{1}{2} \log \left( 1-\frac{v^2}{c^2} \right)

    \quad \approx 1 - \frac{1}{2} \left( -\frac{v^2}{c^2} \right) [/tex]

    for [itex] |v| << |c| [/itex].

    get's the same thing.

    i was also gratified that dextercioby used the symbol [itex]m_0[/itex] which seems to be out of vogue.
    Last edited: Nov 15, 2006
  9. Nov 16, 2006 #8


    User Avatar
    Science Advisor
    Homework Helper

    There's an analytic continuation of the binomial formula using the Eulerian gamma function, but yeah, i still prefer Taylor's polynomial.

  10. Nov 16, 2006 #9
    Sorry, I didn't mean to add confusion. I just thought it seemed quite mysterious in the first post that [tex]\frac{1}{\sqrt{1-\frac{v^2}{c^2}}}\simeq1+\frac{v^2}{2c^2}[/tex] and wanted to explain where that came from.

    btw in the Taylor series example given above, why use [tex]e^{-\frac{1}{2}}log(1-\frac{v^2}{c^2})\simeq1-\frac{v^2}{2c^2}[/tex]? This seems quite mysterious to me since we can write [tex]e^{-\frac{1}{2}}e^{log(1-\frac{v^2}{c^2})} \equiv e^{-\frac{1}{2}}(1-\frac{v^2}{c^2})[/tex], which is exact.

    There is nothing wrong with doing it this way of course... I just don't like the idea of throwing away precision just to make the answer fit. [tex]e^{-\frac{1}{2}}[/tex] is more like 0.6 than [tex]\frac{1}{2}[/tex]
    Last edited: Nov 16, 2006
  11. Nov 16, 2006 #10


    User Avatar
    Science Advisor
    Homework Helper

    You've made a mistake on the e^(ab) expansion. It's not e^a e^b

  12. Nov 16, 2006 #11
    Ah! Thanks! Pretty fundamental mistake I made there, a^b+c = a^b a^c.

    I thought it was odd that someone would consider [tex]0.6\simeq\frac{1}{2}[/tex]. Still, it leaves the question (for me at least) of where that approximation comes from?

    EDIT: Never mind I just worked it out... I think it comes from the power series for [tex]e^x[/tex]
  13. Nov 16, 2006 #12


    User Avatar

    [tex]e^{-\frac{1}{2}}e^{log(1-\frac{v^2}{c^2})} [/tex]
    is not the same as

    allowing analytical extension from derivatives of integer power functions to non-integer, we know that

    [tex] (1+x)^\alpha \approx 1 + \alpha x [/tex]
    for small |x|.

    so there is a simple expression for what we need. i did the log - exp thing to avoid making that analytical extension since they were defined for noninteger arguments.

    yes, there were two applications of approximation instead of just one. but i didn't need to make an assumption of analytical extension and both approximations are good for small |x|.
    Last edited: Nov 16, 2006
  14. Nov 17, 2006 #13
    Sorry about that, I got confused about exponentiation rules, as stated above.
  15. Sep 9, 2011 #14
    Is this equation for Relativistic Kinetic Energy really proven?
  16. Sep 9, 2011 #15


    User Avatar
    Science Advisor

    Yes, millions of times each day at the LHC and other laboratories.
  17. Sep 9, 2011 #16


    User Avatar
    Science Advisor
    Gold Member

    Hey, I always just use as every day fact that for small x:

    sqrt(1+x) := 1 + x/2

    1/(1+x) := 1-x

    The former can be derived as trivially as:

    (1+x)^2 = 1 + 2x + x^2

    the latter from sum of geometric series, or even just subtracting and verifying the error is order x^2.
  18. Sep 10, 2011 #17
    And it results from Lorentz transformation (which is also proven experimentally).
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook