Here is a straightforward solution. It is possible that there is a smarter and easier way that I did not see.
- First let's represent tan-1(x3) as a power series.
[tan-1(x3)] ' = 3x2/(1+x6) so tan-1(x3) = ∫0x 3t2/(1+t6)dt +tan-1(0) , and tan-1(0)=0
Using the power series 1/(1+x) = ∑n=0∞ (-1)n xn we get
tan-1(x3) = ∫0x 3t2 ∑n=0∞ (-1)n t 6n dt = 3 ∫0x ∑n=0∞ (-1)n t 6n+2 dt
Taking the integral, we get
tan-1(x3) = 3 ∑n=0∞ (-1)n x 6n+3 /(6n+3).
- Now, multiplying it by x2 we will get the power series representation for the original function:
x2 tan-1(x3) = 3 x2 ∑n=0∞ (-1)n x 6n+3 /(6n+3) = 3 ∑n=0∞ (-1)n x 6n+5 /(6n+3).
- Using ratio test to determine the radius of convergence we get:
lim n->∞ |x6(n+1)+5(6n+3) / x6n+5(6(n+1)+3) | = lim n->∞ |x6(6n+3) /(6n+9) | = |x6| <1
So |x|<1, the radius of convergence is 1.