Square-root binomial series at its endpoints 2026-10-18
Write , . The binomial series for has positive numbers as the magnitudes of its nonconstant coefficients. The product tends to zero, since . Thus by telescoping. The Weierstrass M-test gives uniform convergence and absolute convergence on . Continuity extends the interior identity to both endpoints. Its radius of convergence is one; outside this interval its terms do not tend to zero.
Over the complex numbers, if , thenHere the generalized binomial coefficient is , with . This is an exact finite identity, even when an infinite binomial series would not converge.
One can construct the required polynomial without any infinite series. At polynomial degree , the coefficient of is plus terms involving only earlier coefficients. Recursively choose it to agree with through polynomial degree ; the first coefficients are . The stated generalized binomial coefficients solve this recursion: near , the binomial series for cubes to , and truncating it changes no coefficient below polynomial degree . Thus is divisible by as a polynomial identity. Substituting the nilpotent matrix kills the remaining terms, all divisible by . For a Jordan block with , replace by and multiply by a scalar cube root of .
For a forward photon and a positive collinear Lorentz transformation, the Doppler factor is . The identity gives the exact expression in the title. Expanding the square root by the binomial series gives . A leading approximation has relative error of order ; retaining a purported subleading term requires expanding all dependent quantities consistently. For example, has subleading term and is not the exact longitudinal Doppler factor.