Convergent & Divergent Series — Defining Convergent and Divergent Infinite Series - Calculus 2

SECTION 1

Historical Context & Motivation

The question of whether an infinite process can produce a finite result has captivated mathematicians for over two millennia. Ancient Greek philosophers, most notably Zeno of Elea, posed paradoxes suggesting that summing infinitely many distances should be impossible — yet Achilles clearly does overtake the tortoise. These paradoxes revealed a deep tension between intuition and the infinite, a tension that would not be resolved until mathematicians developed a rigorous framework for infinite series. The journey from philosophical puzzlement to the precise ε-N definitions used in modern analysis spans contributions from Archimedes through Euler to Cauchy and Weierstrass, each building the scaffolding that allows us to distinguish series that settle from those that diverge.

~250 BCE • Archimedes and the Quadrature of the Parabola

Archimedes computed the area under a parabolic segment by summing a geometric series, effectively evaluating 1 + 1/4 + 1/16 + ⋯ = 4/3. This was one of the earliest explicit uses of an infinite sum to obtain a finite result.

1668 • Mercator's Series for the Logarithm

Nicolaus Mercator published the series ln(1 + x) = x − x²/2 + x³/3 − ⋯, demonstrating that transcendental functions could be represented as infinite sums — but only for certain values of x where the series converges.

1734 • Euler Solves the Basel Problem

Leonhard Euler showed that ∑ 1/n² = π²/6, proving that the sum of reciprocal squares converges to a specific transcendental number. His work popularized the study of series but often relied on manipulations that lacked rigorous justification.

1821 • Cauchy's Cours d'Analyse

Augustin-Louis Cauchy introduced the formal definition of convergence through the sequence of partial sums, establishing the framework still used today. He also formulated the Cauchy criterion, giving convergence a characterization independent of knowing the limit.

1860s • Weierstrass Rigorous Foundations

Karl Weierstrass completed the arithmetization of analysis, grounding Cauchy's ideas in the ε-δ and ε-N language that forms the backbone of modern real analysis and the precise definitions of convergent and divergent series.

The central question that motivated centuries of development is deceptively simple: given an infinite sequence of numbers a₁, a₂, a₃, …, does the expression a₁ + a₂ + a₃ + ⋯ represent a well-defined finite number? If so, how do we find it, and how do we prove it exists? The definitions of convergence and divergence for infinite series provide the precise answers to these questions, and they form the foundation upon which all subsequent convergence tests, power series, and Fourier analysis are built.

SECTION 2

Core Principles & Definitions

An infinite series is not simply an infinite addition problem performed all at once; rather, it is defined through the limit of a sequence of finite sums. This distinction is subtle but essential: we never literally add infinitely many terms. Instead, we examine the behavior of the partial sums S_N = a₁ + a₂ + ⋯ + a_N as N grows without bound. The series ∑ aₙ is merely a notational shorthand for the limit of this sequence {S_N}, provided the limit exists.

Partial Sum Sequence

Given a sequence {aₙ}, define S_N = ∑ₙ₌₁ᴺ aₙ. The sequence {S₁, S₂, S₃, …} is called the sequence of partial sums. This transforms the series problem into a sequence convergence problem.

Convergent Series

The series ∑ aₙ converges if and only if lim(N→∞) S_N = S for some finite real number S. We write ∑ₙ₌₁∞ aₙ = S and call S the sum of the series.

Divergent Series

If the sequence of partial sums {S_N} does not converge to a finite limit — whether it tends to ±∞ or oscillates without settling — then the series diverges. In this case, the expression ∑ aₙ does not represent a real number.

The Divergence Test (nth-Term Test)

If lim(n→∞) aₙ ≠ 0, then ∑ aₙ diverges. This is a necessary but not sufficient condition: the harmonic series shows that aₙ → 0 does not guarantee convergence.

✦

KEY TAKEAWAY

Think of an infinite series like filling a container by pouring in successively smaller amounts of water. If each pour adds a smaller and smaller quantity in just the right way, the water level approaches a definite height — the series converges. But if the pours don't diminish quickly enough (like the harmonic series), the container overflows no matter how large it is — the series diverges. The key insight is that convergence is about the cumulative behavior of partial sums, not about any single term.

SECTION 3

Visualizing Partial Sums

The most illuminating way to understand convergence and divergence is to plot the sequence of partial sums. For a convergent series, the partial sums approach a horizontal asymptote; for a divergent series, they either grow without bound or oscillate indefinitely. The diagram below contrasts the partial sums of two classic series: the convergent geometric series ∑ (1/2)ⁿ and the divergent harmonic series ∑ 1/n.

The green curve shows the partial sums of ∑ (1/2)ⁿ rapidly approaching the limit S = 1 (dashed line). The red curve shows the partial sums of the harmonic series ∑ 1/n, which continue to grow — albeit slowly — without bound.

Notice the qualitative difference in the two curves. The green partial sums flatten out, exhibiting the hallmark of convergence: the incremental contributions a_N = (1/2)^N shrink so rapidly that the cumulative sum stabilizes. In contrast, the red partial sums of the harmonic series continue climbing. Although each new term 1/N is small, the terms do not decay fast enough to prevent the total from eventually exceeding any prescribed bound. This visual distinction — partial sums that plateau versus partial sums that drift — is the geometric essence of the convergence/divergence dichotomy.

SECTION 4

Mathematical Framework

We now state the formal definitions with full analytical rigor. The definition of series convergence reduces directly to the definition of sequence convergence applied to the partial sum sequence, so familiarity with the ε-N definition of limits is essential.

PARTIAL SUM

S_N = ∑ₙ₌₁ᴺ aₙ = a₁ + a₂ + a₃ + ⋯ + a_N

S_N is the Nth partial sum: the sum of the first N terms of the sequence {aₙ}.

CONVERGENCE DEFINITION

∑ₙ₌₁∞ aₙ = S ⟺ ∀ε > 0, ∃N₀ ∈ ℕ such that N ≥ N₀ ⟹ |S_N − S| < ε

The series converges to S if the partial sums S_N can be made arbitrarily close to S by taking N sufficiently large. The number S is called the sum of the series.

DIVERGENCE

If lim_{N→∞} S_N does not exist (finite), then ∑ₙ₌₁∞ aₙ diverges.

Divergence includes three cases: S_N → +∞, S_N → −∞, or S_N oscillates (e.g., the series ∑ (−1)ⁿ has partial sums alternating between −1 and 0).

DIVERGENCE TEST (NTH-TERM TEST)

If lim_{n→∞} aₙ ≠ 0 ⟹ ∑ aₙ diverges

Contrapositive: If ∑ aₙ converges, then lim aₙ = 0. The converse is false — lim aₙ = 0 does not imply convergence (counterexample: the harmonic series).

ℹ

Proof Sketch: Divergence Test

Suppose ∑ aₙ converges to S. Then both S_N → S and S_N−1 → S as N → ∞. Since aₙ = S_N − S_N−1, we get lim aₙ = S − S = 0. Taking the contrapositive yields the Divergence Test.

SECTION 5

Classification of Series Behavior

Not all divergent series behave the same way, and not all convergent series converge for the same reasons. Understanding the taxonomy of series behavior helps you anticipate which convergence tests to apply and avoid common pitfalls. The diagram below organizes series into a decision tree based on partial-sum behavior.

Decision tree for classifying infinite series. A convergent series is further classified as absolutely convergent or conditionally convergent. Divergent series may tend to ±∞ or oscillate.

Series	Partial Sums $S_N$	lim aₙ	Classification
∑ (1/2)ⁿ	1 − (1/2)^N → 1	0	Convergent (abs.)
∑ (−1)ⁿ⁺¹/n	Oscillates, approaches ln 2	0	Convergent (cond.)
∑ 1/n	Grows without bound (slowly)	0	Divergent (to +∞)
∑ (−1)ⁿ	−1, 0, −1, 0, …	DNE	Divergent (oscillation)
∑ n	N(N+1)/2 → +∞	+∞	Divergent (to +∞)

Examples illustrating each classification of series behavior

SECTION 6

Worked Example: Geometric Series Convergence

Consider the series ∑ₙ₌₀∞ 3 × (1/4)ⁿ = 3 + 3/4 + 3/16 + 3/64 + ⋯. We will determine whether this series converges or diverges using the definition of partial sums and the formula for a finite geometric sum.

Determine Convergence: ∑ₙ₌₀∞ 3·(1/4)ⁿ

1
Step 1 — Identify the Series Type

This is a geometric series with first term a = 3 and common ratio r = 1/4. A geometric series ∑ₙ₌₀∞ arⁿ converges if and only if |r| < 1.

a = 3, r = 1/4, and |1/4| = 0.25 < 1 → convergence expected
2
Step 2 — Compute the Nth Partial Sum

The finite geometric sum formula gives S_N = a × (1 − rᴺ⁺¹)/(1 − r) = 3 × (1 − (1/4)^N+1) / (1 − 1/4) = 3 × (1 − (1/4)^N+1) / (3/4) = 4 × (1 − (1/4)^N+1).

S_N = 4(1 − (1/4)^N+1)
3
Step 3 — Take the Limit

As N → ∞, the term (1/4)^N+1 → 0 because |1/4| < 1. Therefore, lim(N→∞) S_N = 4 × (1 − 0) = 4.

S = lim $S_N$ = 4
4
Step 4 — State the Conclusion

Since lim(N→∞) S_N = 4, which is a finite real number, the series ∑ₙ₌₀∞ 3·(1/4)ⁿ converges by definition. The sum of the series is 4.

∑ₙ₌₀∞ 3·(1/4)ⁿ = 4 (convergent)
5
Step 5 — Verify with the Divergence Test

As a consistency check, lim(n→∞) aₙ = lim(n→∞) 3·(1/4)ⁿ = 0. This is consistent with convergence (the necessary condition is satisfied). Note that this check alone would not prove convergence — it only confirms we have no contradiction.

lim aₙ = 0 ✓ (necessary condition met)

SECTION 7

Common Pitfalls & Key Distinctions

Students frequently encounter conceptual traps when first working with infinite series. Understanding the logical structure of necessary versus sufficient conditions, and recognizing the boundaries of each test, is critical for avoiding errors in proofs and on exams.

Common Misconception	Why It's Wrong	Correct Statement
If aₙ → 0, then ∑ aₙ converges.	The harmonic series ∑ 1/n has aₙ → 0 but diverges. The condition aₙ → 0 is necessary but not sufficient.	aₙ → 0 is required for convergence but does not guarantee it. Further tests (comparison, ratio, integral, etc.) are needed.
If ∑ aₙ diverges, then aₙ → ∞.	Again, the harmonic series: each term 1/n → 0, yet the series diverges. Divergence can occur even with vanishing terms.	If aₙ does not tend to 0, the series diverges. But divergence can happen even when aₙ → 0.
Adding more terms always increases the partial sum.	This is true only for series with all positive terms. Alternating series have partial sums that oscillate above and below the limit.	For non-negative series, {S_N} is non-decreasing. For alternating or mixed-sign series, partial sums can decrease.
Changing finitely many terms can affect convergence.	Altering a finite number of terms changes the sum but not whether the series converges. Convergence depends on the tail of the series.	Convergence is a tail property: ∑ₙ₌₁∞ aₙ converges ⟺ ∑ₙ₌ₖ∞ aₙ converges for any fixed k.

Common misconceptions about series convergence and their corrections

⚠

KEY TAKEAWAY

The Divergence Test is like a filter at the entrance of a building: if you fail the filter (aₙ does not tend to 0), you are definitely denied entry (the series diverges). But passing the filter (aₙ → 0) does not guarantee you get in — you still need to show your credentials through a convergence test such as the comparison test, ratio test, or integral test. Never confuse a necessary condition with a sufficient one.

SECTION 8

Connections to Advanced Theory

The definition of series convergence you have now mastered is the gateway to a rich ecosystem of analytical tools and theoretical results. Understanding where this foundational concept leads will help you appreciate why rigor at this stage pays dividends throughout calculus, differential equations, and beyond.

This Lesson	Where It Leads
Definition of convergence via partial sums	Convergence tests (comparison, ratio, root, integral, alternating series) — systematic tools for determining convergence without computing partial sums directly
Absolute vs. conditional convergence	Riemann rearrangement theorem: conditionally convergent series can be rearranged to converge to any real number, or to diverge — a striking result that underscores the importance of absolute convergence
Geometric series as the prototypical convergent series	Power series and Taylor series: representing functions as ∑ cₙ(x − a)ⁿ, where the radius of convergence determines the domain
ε-N definition of series convergence	Uniform convergence of function series, Weierstrass M-test, and interchange of limits — essential in real analysis and functional analysis
Divergence of the harmonic series	p-series and the integral test; connections to the Riemann zeta function ζ(s) = ∑ 1/nˢ and analytic number theory

How foundational definitions connect to advanced topics

Perhaps the most important forward-looking idea is that the definition of convergence extends naturally to power series, where the "terms" are functions rather than constants. A power series ∑ cₙxⁿ converges for some values of x and diverges for others, and the boundary between these regions is governed by the radius of convergence. Understanding the convergence/divergence dichotomy for numerical series is therefore the indispensable prerequisite for working with Taylor expansions, Fourier series, and the analytic representations of functions that pervade applied mathematics, physics, and engineering.

SECTION 9

Practice Problems

Question 1: CONCEPTUAL

Explain why the statement "lim(n→∞) aₙ = 0 implies ∑ aₙ converges" is false. Provide a specific counterexample and explain, in terms of partial sums, why the series in your counterexample diverges despite the terms vanishing.

Show answer

The statement is false because lim aₙ = 0 is a necessary condition for convergence, not a sufficient one. The canonical counterexample is the harmonic series ∑ₙ₌₁∞ 1/n. Although 1/n → 0, the partial sums S_N = 1 + 1/2 + 1/3 + ⋯ + 1/N grow without bound. One classical proof groups terms: S_2ᵏ ≥ 1 + k/2, showing the partial sums exceed every real number. The terms shrink, but not fast enough — each doubling of terms adds at least 1/2 to the partial sum, so S_N → ∞.

Question 2: BASIC CALCULATION

Determine whether the geometric series ∑ₙ₌₁∞ 5·(−2/3)ⁿ converges or diverges. If it converges, find its sum.

Show answer

This is a geometric series with first term a₁ = 5·(−2/3) = −10/3 and common ratio r = −2/3. Since |r| = 2/3 < 1, the series converges. Using the formula S = a₁/(1 − r), we compute S = (−10/3)/(1 − (−2/3)) = (−10/3)/(5/3) = −10/5 = −2. Therefore the series converges and ∑ₙ₌₁∞ 5·(−2/3)ⁿ = −2. We can verify: the Divergence Test gives lim 5·(−2/3)ⁿ = 0, which is consistent with convergence.

Question 3: INTERMEDIATE

Consider the series ∑ₙ₌₁∞ 1/(n² + n). Use partial fractions to find a closed-form expression for S_N and determine whether the series converges. If so, find the sum.

Show answer

Decompose using partial fractions: 1/(n² + n) = 1/(n(n+1)) = 1/n − 1/(n+1). The partial sum telescopes: S_N = (1 − 1/2) + (1/2 − 1/3) + ⋯ + (1/N − 1/(N+1)) = 1 − 1/(N+1). Taking the limit: lim(N→∞) S_N = lim(N→∞) [1 − 1/(N+1)] = 1. Since this limit is finite, the series converges and its sum is 1. This is an example of a telescoping series, where most terms cancel in the partial sum, leaving a clean closed-form expression.

Question 4: APPLIED

A drug is administered intravenously every 8 hours. Each dose is 200 mg, and by the time the next dose is given, the body has eliminated 60% of the drug present. Model the amount of drug in the body just after the nth dose as a series and determine the long-term steady-state drug level.

Show answer

Just after the nth dose, the drug level is Dₙ = 200 + 200·(0.4) + 200·(0.4)² + ⋯ + 200·(0.4)ⁿ⁻¹, since 40% of each previous dose remains (100% − 60% eliminated = 40% retained). This is a geometric partial sum with a = 200 and r = 0.4. As n → ∞, the series ∑ₖ₌₀∞ 200·(0.4)ᵏ converges because |0.4| < 1, and the sum is S = 200/(1 − 0.4) = 200/0.6 = 333.33 mg. The steady-state drug level is approximately 333.3 mg. This means that after many doses, each new 200 mg dose raises the level by 200 mg, but the 60% elimination removes exactly 200 mg (60% of 333.3 ≈ 200), maintaining equilibrium.

Question 5: CRITICAL THINKING

Prove directly from the ε-N definition of convergence that the series ∑ₙ₌₁∞ 1/2ⁿ converges to 1. That is, given any ε > 0, find an explicit N₀ (in terms of ε) such that for all N ≥ N₀, |S_N − 1| < ε.

Show answer

The Nth partial sum is S_N = ∑ₖ₌₁ᴺ 1/2ᵏ = 1 − 1/2ᴺ (by the geometric sum formula). Therefore |S_N − 1| = |−1/2ᴺ| = 1/2ᴺ. We need 1/2ᴺ < ε, which gives 2ᴺ > 1/ε, hence N > log₂(1/ε) = −log₂(ε). Choose N₀ = ⌈−log₂(ε)⌉ + 1 (the smallest integer exceeding −log₂(ε)). Then for all N ≥ N₀, we have 2ᴺ ≥

2^{N₀}

> 1/ε, so 1/2ᴺ < ε, and thus |S_N − 1| < ε. Since ε > 0 was arbitrary, the series converges to 1 by definition. For example, with ε = 0.001, we need N₀ = ⌈log₂(1000)⌉ + 1 = ⌈9.97⌉ + 1 = 11.

SUMMARY

Summary

An infinite series ∑ aₙ is defined through the limit of its sequence of partial sums S_N = a₁ + a₂ + ⋯ + a_N. The series converges to a sum S if lim(N→∞) S_N = S exists and is finite, formalized by the ε-N definition: for every ε > 0, there exists N₀ such that N ≥ N₀ implies |S_N − S| < ε. If no such finite limit exists — whether the partial sums tend to ±∞ or oscillate — the series diverges.

The Divergence Test provides a quick necessary check: if lim aₙ ≠ 0, the series diverges. However, the converse fails — the harmonic series ∑ 1/n diverges despite 1/n → 0. Convergent series are further classified as absolutely convergent (∑ |aₙ| also converges) or conditionally convergent (∑ aₙ converges but ∑ |aₙ| does not). These foundational definitions underpin every subsequent convergence test, power series analysis, and application of infinite series in mathematics and the sciences.