Our first test for judging whether a series converges*
The divergence test will usually be the first test of whether or not a series converges. It's pretty simple; we're really just asking whether the nth term of a series converges to zero, but the divergence test has some important limitations that we should get out of the way right away.
*The divergence test can determine whether a series diverges, and if it does diverge, it can't possibly converge. But the divergence test is not a test for convergence.
A series can pass the divergence test (appear to converge), and still diverge when subjected to another test.
The divergence test
If the nth term, an of a series does not decrease to zero, then
X
Converse
In logic, we can encounter a case in which, if a thing (we'll call it P) is true, then something else (called Q) must be true. We write that as P → Q.
The converse of that statement is Q → P. The converse of a proposition like P → Q might be true, but it isn't necessarily true.
That is, just because A proves B, doesn't mean that B proves A.
Proof:
We start with the nth partial sum of a series, Sn,
$$S_n = (a_1 + a_2 + \dots + a_{n - 1}) + a_n$$
Note that we've kept the an term on the outside to give a sum of an and a sub-series of Sn, Sn-1:
Now we'll substitute Sn-1 for that sum in parentheses, then rearrange to solve for an:
$$ \begin{align} S_n &= S_{n - 1} + a_n \\ \\ a_n = S_n - S_{n - 1} \end{align}$$
Now we have an expression for an of which we can take advantage. Here is the substance of the proof:
$$ \begin{align} If \: \sum_{n = 1}^{\infty} a_n &= S \: converges, \: then \\ \\ \lim_{n\to\infty} a_n &= \lim_{n\to\infty} (S_n - S_{n - 1}) \\ \\ &= \lim_{n\to\infty} S_n - \lim_{n\to\infty} S_{n - 1} \\ \\ &= S - S = 0 \end{align}$$
In that last step we note that n-1 → ∞ as n → ∞, so the limits are the same (S). Another way to think of it is that for large enough partial sums of an infinite series, the difference between an and an-1 is small when n is large. Now if an does not converge to zero, then the sum of the terms, an, cannot converge.
Example 1: A divergent series
Show that $\sum_{n = 1}^{\infty} \frac{n}{2n + 1}$ diverges.
$$\lim_{n\to\infty} \frac{n}{2n + 1} = \lim_{n\to\infty} \frac{1}{2} = \frac{1}{2} \ne 0$$
The limit is not zero. That means
$$\sum_{n = 1}^{\infty} \frac{n}{2n + 1} = \frac{1}{3} + \frac{2}{5} + \frac{3}{7} + \dots$$
This series does not reach some finite limit, but continues to grow. For very large n, the series looks like ½ + ½ + ½ ...
This series will not be convergent, and is therefore not very useful as a series to represent some finite-valued function.
This plot shows the size of each term for n = 1 to 11 and how the series grows as terms are added. A convergent series would have an asymptote that the sum would approach as n → ∞.
If an infinite series fails the divergence test, we're done with it. It is not a convergent series, and therefore cannot be used to represent a finite-valued function.
Example 2
Show that the series $\sum_{n = 1}^{\infty}$ diverges.
$$\sum_{n = 1}^{\infty} \frac{1}{\sqrt{n}} = \frac{1}{\sqrt{1}} + \frac{1}{\sqrt{2}} + \frac{1}{\sqrt{3}} + \dots$$
Clearly the limit of the terms as n → ∞ is zero:
$$\lim_{n\to\infty} \frac{1}{\sqrt{n}} = 0$$
So it looks like this series converges: The limit of the terms is zero, so it doesn't fail the divergence test. But this is a test for divergence. That a series does not diverge by this test doesn't necessarily mean that it converges. It might diverge by another test, and this one does.
Look at the partial sum of N terms, SN, below:
Notice that that term-by-term, the element on the left sum is greater than the element on the right.
Now we can figure out (below) what the sum on the right is, and it's not zero, so this partial sum does not converge to zero, therefore neither does the series.
When a series diverges by the divergence test, it is a divergent series. When the divergence test is negative, as in this case, the series may converge, or it may diverge by another test, just as this one did. Just because the divergence test fails, does not mean that a series converges.
Example 3
Does the series $\sum_{n = 1}^{\infty} \frac{2 n^2 + n^3}{4n^3 - 3n}$ converge?
$$\lim_{n\to\infty} \frac{2 n^2 + n^3}{4n^3 - 3n} = \lim_{n\to\infty} \frac{4n + 3n^2}{12n^2 - 3}$$
and another couple of rounds (often needed with L'Hopital's rule) gives the limit:
$$= \lim_{n\to\infty} \frac{4 + 6n}{24n} = \frac{1}{4}$$
Because the limit of the terms is not zero, we can safely conclude that this series does not converge. This is where the divergence test is really useful. We can move on from this series for most uses because it is divergent.
xaktly.com by Dr. Jeff Cruzan is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 3.0 Unported License. © 2012, Jeff Cruzan. All text and images on this website not specifically attributed to another source were created by me and I reserve all rights as to their use. Any opinions expressed on this website are entirely mine, and do not necessarily reflect the views of any of my employers. Please feel free to send any questions or comments to [email protected].