Product of Power Series

Key Questions

  • One example that I find useful is the use and manipulation of the products of power series to derive e^(ix) = cosx + isinx, which is an identity used many, many times to solve the Schroedinger Equation in Physical Chemistry, by substituting i for various different constants.


    What is accepted by Physical Chemists is that you can write out the general solution to the Equation as:

    y(x) = c_1 e^(alphax) + c_2 e^(-alphax)

    where alpha = iomegat and, after various proofs and empirical tests, it is agreed that we can use e^(alphax) as a working "trial function" when we guess the form of the overall solution in terms of a finite addition of these c*e^(alphax) functions so that we can predict molecular properties:

    psi(x) = sum_(i=1)^N c_i phi_i(x)
    where each phi could, for example, represent an atomic orbital, and psi(x) would in that case be the molecular orbital.

    A common example of solving the time-dependent Schroedinger equation is (example 2-4 in Physical Chemistry: A Molecular Approach):

    (d^2x(t))/(dt^2) + omega^2x(t) = 0

    subject to the boundary conditions x(0) = A and (dx(0))/(dt) = 0. These boundary conditions define the fact that a stationary transverse wave with one antinode has two endpoints, and these are at x = 0 and x = l, half of the wavelength.

    http://www.physicsclassroom.com/

    To solve this one, one would have to use identity written at the top, with alpha substituted for i like so:

    c_1e^(alphax) + c_2e^(-alphax)

    = c_1(cosx + alphasinx) + c_2(cosx - alphasinx)

    = c_1cosx + c_1alphasinx + c_2cosx - c_2alphasinx

    = (c_1 + c_2)cosx + (c_1alpha - c_2alpha)sinx

    and it is generally written out by absorbing the arbitrary constants c_1, alpha, and c_2 into new arbitrary constants c_3 and c_4, with c_1 + c_2 = c_3 and c_1alpha - c_2alpha = c_4:

    = c_3cosx + c_4sinx

    Then, substituting omegat for x, we get:

    c_3cos(omegat) + c_4sin(omegat)

    for the solution to the so-called common example.

    Looking at the boundary condition (dx(0))/(dt) = 0, we get:

    = c_3(-sin(omegat))*omega + c_4cos(omegat)*omega

    = cancel(c_3(-sin(omega(0)))*omega)^(0) + c_4cos(omega(0))*omega

    = c_4omega

    But we know that at t = 0, omega = 0 because time has not passed yet.

    => c_4omega = 0, thus satisfying the condition (dx(0))/(dt) = 0.

    Using the x(0) = A boundary condition we get:

    x(0) = c_3cos(omega(0)) + cancel(c_4sin(omega(0)))^(0)

    = c_3

    with c_3 taken as A---which is the amplitude of an initialized stationary wave---since it is the only contributor to the wave. Thus, since c_3 = A, we have satisfied the condition x(0) = A, and we just have:

    color(blue)(x(t) = Acos(wt))

    which is the familiar physics equation for a transverse wave, as depicted in the image above! :)

  • (sum_{n=0}^infty a_nx^n)cdot(sum_{n=0}^infty b_nx^n)=sum_{n=0}^infty(sum_{k=0}^na_kb_{n-k})x^n

    Let us look at some details.

    (sum_{n=0}^infty a_nx^n)cdot(sum_{n=0}^infty b_nx^n)

    by writing out the first few terms,

    =(a_0+a_1x+a_2x^2+cdots)cdot(b_0+b_1x+b_2x^2+cdots)

    by collecting the like terms,

    =a_0b_0x^0+(a_0b_1+a_1b_0)x^1+(a_0b_2+a_1b_1+a_2b_0)x^2+cdots

    by using sigma notation,

    =sum_{n=0}^infty(sum_{k=0}^na_kb_{n-k})x^n

Questions