How to use the nth term test for divergence
What You Need to Know
The big idea (why this matters)
When you’re given an infinite series \sum_{n=1}^{\infty} a_n, the nth term test for divergence (also called the divergence test or term test) is the fastest “sanity check” you should try first.
It tells you a necessary condition for convergence:
- If a series converges, then its terms must approach 0.
- If the terms do not approach 0 (or the limit doesn’t exist), the series definitely diverges.
Warning: This test can only prove divergence. If the limit is 0, you learn nothing about convergence.
Core theorem (state it precisely)
For a series \sum_{n=1}^{\infty} a_n:
- If \lim_{n\to\infty} a_n \neq 0, then \sum_{n=1}^{\infty} a_n diverges.
- If \lim_{n\to\infty} a_n **does not exist**, then \sum_{n=1}^{\infty} a_n diverges.
- If \lim_{n\to\infty} a_n = 0, the test is inconclusive (series may converge or diverge).
When you use it
Use it:
- Immediately when you see a series (it’s quick).
- To catch “trick” series where terms don’t go to zero (common on AP Calc).
- As a first decision point before choosing other tests (geometric, p-series, comparison, ratio, alternating, etc.).
Step-by-Step Breakdown
Procedure (do this every time)
- Identify the term a_n in the series \sum a_n.
- Compute \lim_{n\to\infty} a_n.
- Use algebra simplification first.
- If it’s a rational expression in n, divide by the highest power of n.
- If it’s exponential/factorial, recall growth rates (factorials/exponentials dominate polynomials).
- Decide:
- If the limit is not 0 (including \pm\infty): diverges.
- If the limit does not exist (oscillates): diverges.
- If the limit is exactly 0: inconclusive → switch to another convergence test.
Mini worked walkthroughs (annotated)
Example A: Limit is not zero → diverges
Series: \sum_{n=1}^{\infty} \frac{n}{n+1}
- a_n = \frac{n}{n+1}
- \lim_{n\to\infty} \frac{n}{n+1} = \lim_{n\to\infty} \frac{1}{1+\frac{1}{n}} = 1
- Since the limit \neq 0, the series diverges.
Example B: Limit does not exist (oscillates) → diverges
Series: \sum_{n=1}^{\infty} (-1)^n
- a_n = (-1)^n
- \lim_{n\to\infty} (-1)^n does not exist (it flips between -1 and 1).
- So the series diverges.
Example C: Limit is zero → inconclusive
Series: \sum_{n=1}^{\infty} \frac{1}{n}
- a_n = \frac{1}{n}
- \lim_{n\to\infty} \frac{1}{n} = 0
- Test is inconclusive (this series actually diverges, but you need a different test).
Decision point: \lim a_n = 0 means “it might converge,” not “it does converge.”
Key Formulas, Rules & Facts
The test in one table
| Result of \lim_{n\to\infty} a_n | What you can conclude about \sum a_n | Notes |
|---|---|---|
| L \neq 0 | Diverges | Includes any nonzero finite number |
| Limit is \pm\infty | Diverges | Terms don’t approach 0 |
| Limit does not exist | Diverges | Often from oscillation or piecewise behavior |
| L = 0 | Inconclusive | Could converge or diverge |
Necessary condition for convergence (the “must” rule)
If \sum_{n=1}^{\infty} a_n converges, then:
- \lim_{n\to\infty} a_n = 0.
The contrapositive is what you use for the nth term test:
- If \lim_{n\to\infty} a_n \neq 0 or DNE, then \sum a_n diverges.
Quick limit evaluation facts that show up a lot
Rational functions: for a_n = \frac{p(n)}{q(n)}, compare degrees.
- If degree numerator = degree denominator, limit is ratio of leading coefficients.
- If degree numerator < degree denominator, limit =0.
- If degree numerator > degree denominator, limit is \pm\infty (or DNE by sign).
Growth rates (useful for deciding limits quickly):
- Factorials dominate exponentials dominate powers dominate logs, i.e.
n! \gg c^n \gg n^p \gg \ln(n)
So examples: - \lim_{n\to\infty} \frac{n^5}{2^n} = 0
- \lim_{n\to\infty} \frac{2^n}{n!} = 0
- Factorials dominate exponentials dominate powers dominate logs, i.e.
Oscillation warning: if a_n contains (-1)^n, \sin(n), \cos(n), etc., the limit might **not exist** unless there’s a “shrinking factor” like \frac{1}{n}.
- \lim_{n\to\infty} \frac{\sin(n)}{n} = 0 (inconclusive for the series)
- \lim_{n\to\infty} \sin(n) DNE (so the series diverges)
Examples & Applications
1) Classic “looks like it might converge” but fails immediately
\sum_{n=1}^{\infty} \frac{3n+1}{n}
- a_n = \frac{3n+1}{n} = 3 + \frac{1}{n}
- \lim_{n\to\infty} a_n = 3
- Since the limit \neq 0, the series diverges.
Exam vibe: They want you to notice you don’t even need a “real” convergence test.
2) Alternating-looking series that still diverges (because terms don’t go to 0)
\sum_{n=1}^{\infty} (-1)^n \frac{n}{n+1}
- a_n = (-1)^n \frac{n}{n+1}
- Magnitude approaches 1, and the sign flips.
- \lim_{n\to\infty} a_n does not exist (oscillates between values near -1 and 1).
- Therefore the series diverges by nth term test.
Key insight: “Alternating” does not automatically mean convergent. You still need a_n \to 0.
3) Limit is zero, but series diverges anyway (test is inconclusive)
\sum_{n=1}^{\infty} \frac{1}{n}
- \lim_{n\to\infty} \frac{1}{n} = 0
- nth term test says: inconclusive.
- Reality: diverges (harmonic series).
Exam vibe: They’re testing whether you falsely conclude convergence from a_n \to 0.
4) Limit is zero, and series converges (still inconclusive)
\sum_{n=1}^{\infty} \left(\frac{1}{2}\right)^n
- a_n = \left(\frac{1}{2}\right)^n
- \lim_{n\to\infty} a_n = 0
- nth term test: inconclusive.
- Reality: converges (geometric series).
Takeaway: Same nth-term result (limit 0) can correspond to both convergence and divergence.
5) A “hidden” nonzero limit inside a rational expression
\sum_{n=1}^{\infty} \frac{n^2 + 7}{2n^2 - 1}
- a_n = \frac{n^2 + 7}{2n^2 - 1}
- Divide by n^2: a_n = \frac{1 + \frac{7}{n^2}}{2 - \frac{1}{n^2}}
- \lim_{n\to\infty} a_n = \frac{1}{2}
- Since the limit \neq 0, the series diverges.
Exam vibe: Quick algebra → instant divergence.
Common Mistakes & Traps
Mistake: Thinking \lim_{n\to\infty} a_n = 0 means the series converges
- What goes wrong: You confuse a necessary condition with a sufficient one.
- Why wrong: Many divergent series still have terms that go to 0 (harmonic, p-series with p \le 1).
- Fix: Train your brain to say: “Limit 0 → inconclusive.”
Mistake: Applying the test to the wrong thing (partial sums instead of terms)
- What goes wrong: You compute \lim_{n\to\infty} S_n where S_n = \sum_{k=1}^n a_k.
- Why wrong: The nth term test is about a_n, not S_n.
- Fix: Always write “a_n = \dots” first.
Mistake: Forgetting that “limit does not exist” also forces divergence
- What goes wrong: You only look for a nonzero limit but ignore oscillation.
- Why wrong: If a_n doesn’t settle to a single number, it definitely doesn’t go to 0 in the needed way.
- Fix: If terms flip or wander (like (-1)^n), explicitly say “limit DNE → diverges.”
Mistake: Concluding convergence from a_n becoming small “quickly”
- What goes wrong: You see something like \frac{1}{\sqrt{n}} or \frac{\ln(n)}{n} and assume convergence.
- Why wrong: Small terms are not enough; you need an actual convergence test.
- Fix: Use nth term test only as a gatekeeper. If it passes (limit 0), move on.
Mistake: Not simplifying before taking the limit
- What goes wrong: You try to evaluate \lim \frac{n}{n+1} without dividing by n, or you mis-handle leading terms.
- Why wrong: Algebra is the difference between seeing 1 instantly vs. getting stuck.
- Fix: For rational expressions, divide numerator/denominator by the highest power of n.
Mistake: Misusing L’Hôpital’s Rule
- What goes wrong: You apply L’Hôpital’s Rule automatically to sequences.
- Why wrong: L’Hôpital is for functions with indeterminate forms; for sequences you can sometimes treat n as continuous, but that’s beyond what you typically need here.
- Fix: Stick to algebraic limit tricks and growth-rate reasoning unless your teacher explicitly allows otherwise.
Mistake: Ignoring domain issues and undefined terms
- What goes wrong: You assume a_n is defined for all n \ge 1, but something like \ln(n-3) starts later.
- Why wrong: A series must be well-defined from some index onward.
- Fix: If needed, shift the start index (finite number of initial terms doesn’t affect convergence), then apply the test.
Memory Aids & Quick Tricks
| Trick / mnemonic | What it helps you remember | When to use it |
|---|---|---|
| “Terms must go to zero to even have a chance.” | Convergence \Rightarrow a_n \to 0 (necessary condition) | Any series problem as a first check |
| “Divergence test = term test.” | Alternate name so you don’t think it’s a different tool | When you see either wording |
| “If the limit is not zero, it’s a no-go.” | \lim a_n \neq 0 (or DNE) guarantees divergence | Quick decision after computing a limit |
| “Zero means ‘zip’ information.” | \lim a_n = 0 gives no conclusion | To stop yourself from claiming convergence |
| Leading-term focus for rationals | For \frac{p(n)}{q(n)}, limit depends on highest powers | Rational-term series |
Quick self-check phrase: “I’m testing the terms, not the sum.”
Quick Review Checklist
- You’re given \sum a_n → first ask: what is \lim_{n\to\infty} a_n?
- If \lim a_n \neq 0, is \pm\infty, or DNE → series diverges.
- If \lim a_n = 0 → inconclusive (pick another test).
- Watch for oscillation: (-1)^n, \sin(n), \cos(n) can make limits DNE.
- For rational expressions in n: divide by the highest power of n.
- Don’t confuse a_n with partial sums S_n.
- Remember: Every convergent series has a_n \to 0, but not every series with a_n \to 0 converges.
You’ve got this—use the nth term test as your fast first filter, then move to a stronger test when it’s inconclusive.