Riemann Hypothesis

The Riemann Hypothesis

Prime numbers look like scattered confetti, yet they obey a hidden musical rhythm. In 1859, Bernhard Riemann predicted that every single "note" governing the primes sits on one razor-thin vertical line. Nobody has proved him right — or caught a single note out of place.

$1,000,000
Clay Millennium Prize
1859
Year Riemann posed it
1013
Zeros checked on the line
0
Counterexamples found

Live preview: walking the critical line

As height t rises along Re(s) = 0.5, watch ζ(0.5 + it) loop straight through zero.

14.13
|ζ| 0.000
Part 1 of 7

Primes: wild up close, astonishingly smooth from afar

Every whole number above 1 is built by multiplying prime numbers together — numbers divisible only by 1 and themselves, such as 2, 3, 5, 7, 11, and 13. Because every number breaks down into primes in exactly one way, primes are the atoms of arithmetic.

Try filtering them out below with the 2,200-year-old Sieve of Eratosthenes. Tap a small prime to cross out all of its multiples. Whatever survives is prime.

Interactive sieve (1 to 100)

Eliminate multiples of 2, 3, 5, and 7 — every remaining number up to 100 is prime.

Survivors99 Confirmed primes ≤ 10025
Tap ×2 to remove all even numbers above 2.

Up close, primes behave erratically. Sometimes two primes sit nearly side by side (like 29 and 31); elsewhere you can walk through long barren deserts with no primes at all.

Yet in 1792, a fifteen-year-old Carl Friedrich Gauss stared at tables of primes and noticed something remarkable: around any number x, the chance that a number is prime is roughly 1 / ln(x). Adding up those smooth probabilities gives the logarithmic integral Li(x). Drag the zoom slider below from 30 all the way up to 10,000 and watch the jagged staircase of actual primes π(x) melt right into Gauss's smooth curve.

Lab 1: The Prime Staircase π(x) vs. Gauss's Guess

Hover or drag across the chart to inspect exact counts and the error underneath.

150
Error strip: π(x) − Li(x) inside the Riemann Hypothesis envelope Bound: ±√x
At x =150 Actual π(x) =35 Gauss Li(x) =37.6 Error =-2.6
The Riemann Hypothesis asks: does this error stay tightly pinned inside that green square-root trumpet forever?
The big question in one sentence: Gauss gave us the smooth average for the primes, and the Riemann Hypothesis predicts the tightest possible limit on how far the true prime count can ever drift from that average — roughly no more than √x.
Part 2 of 7

Euler's bridge: connecting every integer to the primes alone

How can you study spiky, discrete whole numbers using smooth curves and calculus? In 1737, Leonhard Euler found an astonishing secret passage. Take every positive whole number, raise it to a power s, flip it upside down, and add them all together. That sum is the zeta function:

ζ(s) = 1 + 1/2ˢ + 1/3ˢ + 1/4ˢ + 1/5ˢ + … = ∏p prime 1 / (1 − p⁻ˢ) Left side: an infinite sum over EVERY whole number. Right side: an infinite product over PRIMES ONLY.

Why on earth are those two sides equal? Because every whole number has a unique prime factorisation! Each prime factor 1 / (1 − p⁻ˢ) expands into a geometric series 1 + p⁻ˢ + p⁻²ˢ + p⁻³ˢ + …. When you multiply those brackets together across all primes, choosing one term from each prime builds every whole number 1/nˢ once and only once.

Lab 2: The Zeta Sum & Euler Product Workbench

Tune s and watch the all-integers sum and the primes-only product converge to the exact same target.

2.00
Exact ζ(s)1.644934
Sum over all integers: 1 + 1/2ˢ + 1/3ˢ + … 1.612
Product over primes only: ∏ 1/(1 − p⁻ˢ) 1.638
Why do they match? Pick any whole number to see how the prime brackets build it:

Try clicking s = 1 (blows up) above. The harmonic series 1 + 1/2 + 1/3 + 1/4 + … climbs to infinity. If there were only finitely many primes, the prime product on the right would stay finite — a contradiction! In a single line of algebra, Euler proved there must be infinitely many primes.

Part 3 of 7

Stepping off the number line: the complex zeta landscape

For over a century after Euler, mathematicians only fed ordinary real numbers greater than 1 into ζ(s), because at s ≤ 1 the simple sum explodes to infinity.

In 1859, Bernhard Riemann made a bold leap: feed in complex numbers s = σ + it — points on a flat 2D map instead of a 1D line. When you raise a number to an imaginary power n⁻ⁱᵗ, it no longer shrinks or grows; it rotates like a clock hand at speed ln(n). Because those spinning arrows pull against one another and partially cancel, Riemann showed there is one unique smooth way to extend ζ(s) across the entire 2D plane (except for a single infinite spike, or "pole", at s = 1). Mathematicians call this analytic continuation.

Lab 3: The Complex Zeta Map & Spinning Arrows

Click or drag anywhere on the 2D plane on the left to probe ζ(σ + it) and watch its complex arrows wind on the right.

Input plane s = σ + it (dark wells = zeros, bright spike = pole) 0.50 + 14.13i
Live complex arrows (alternating η-spiral at selected s) ζ(s) ≈ 0.000 + 0.000i
Landmarks:
RegionCritical line |ζ(s)|0.0000
Right on the first non-trivial zero — the arrows curl and balance out to exact zero.
About that internet-famous "1 + 2 + 3 + 4 + … = −1/12" claim: click the s = −1 landmark above. You will see ζ(−1) = −0.0833… = −1/12. That does not mean adding positive integers literally gives a negative fraction! It means when you smoothly extend ζ(s) past the wall at s = 1 into territory where the original sum no longer works, the extended function passes cleanly through −1/12 at s = −1.
Part 4 of 7

The Critical Line: walking the razor's edge

Once ζ(s) is extended across the whole plane, one question becomes paramount: where does ζ(s) equal zero?

Some zeros are easy to account for: ζ(s) hits zero at every negative even integer (−2, −4, −6, …). These are called the trivial zeros. Mathematicians also proved that no zeros exist to the right of σ = 1, and by symmetry none exist to the left of σ = 0 (apart from the trivial ones).

Every remaining zero — the non-trivial zeros — must live inside the narrow vertical corridor 0 < σ < 1, called the critical strip. And Riemann made an astonishing conjecture: every single one of them sits dead centre on the critical line σ = 1/2.

Try it yourself below. At σ = 0.50, the looping output curve passes straight through the bullseye (0, 0) again and again. Now drag the σ slider slightly off centre — say to 0.58 or 0.42 — and watch every loop miss the target.

Lab 4: Break the Critical Line

Slide σ away from 0.50 to see the output loops miss zero; scrub t up to 60 to visit the first 12 zeros.

0.50
35.00
Output curve ζ(σ + it) as t runs from 0 to 35.00 On the critical line — hits origin at every zero
Hardy's Z(t) along σ = 0.50 — every crossing of the horizontal axis is an exact Riemann zero Z(35.00) = 0.00
First 12 zeros:
Part 5 of 7

The music of the primes: rebuilding primes from pure waves

Here is the revelation that turns an odd curiosity about complex numbers into the deepest mystery in mathematics.

In his 1859 paper, Riemann discovered an explicit formula linking the zeros of ζ(s) directly back to the prime numbers. Every non-trivial zero ρ = σ + iγ generates a smooth cosine wave:

• Its pitch (frequency) is set by the zero's height γ.
• Its loudness (amplitude) grows like xσ — set by the zero's horizontal position σ.

Add these smooth waves on top of Gauss's smooth curve, and something almost unbelievable happens: the waves reinforce one another at every prime number and cancel out in between, sculpting the sharp, jagged staircase of the primes out of pure smooth sine waves. Drag the slider below from 0 up to 100 zeros and watch the primes crystallise before your eyes.

Lab 5: The Prime Synthesiser

Stack real Riemann zeros to reconstruct the primes — or push a rogue zero off the line to see the chaos it causes.

18
Chebyshev prime-power staircase ψ(x) vs. Riemann wave reconstruction using 18 zeros x ∈ [2, 50]
First five individual zero waves (each zero contributes one pure harmonic) γ₁ = 14.13 … γ₅ = 32.94
Highest frequency γₙ72.067 AudioOff
Switch to Prime spikes to see the cosine waves fire bright vertical peaks right on top of 2, 3, 5, 7, 11, 13, 17, 19, 23…

Now you can see why σ = 1/2 is so special. Because of Riemann's mirror symmetry, if a zero ever sat at σ = 0.30, it would force a twin zero to sit at 1 − 0.30 = 0.70. The only way to keep every single wave's loudness down at the minimum possible size x1/2 = √x is for every single zero to balance right on the mirror line σ = 1/2.

Part 6 of 7

Fair coin flips, quantum chaos, and what is really at stake

If you want one mental image to carry away from this page, make it a fair coin flip.

Flip a fair coin x times, stepping +1 for heads and −1 for tails. You rarely stay at zero, yet you don't run straight away either — statistics guarantees your wandering distance stays roughly around ±√x.

Now play the exact same game with prime factorisations. For every whole number n that has no repeated prime factor, take a step of +1 if it has an even number of prime factors (like 6 = 2 × 3), −1 if it has an odd number (like 30 = 2 × 3 × 5 or any prime), and 0 if a prime repeats (like 4 = 2²). The running total is called the Mertens function M(x). Remarkably, the Riemann Hypothesis is mathematically equivalent to saying that this prime walk behaves just as fairly as a coin toss — never growing faster than roughly x1/2.

Lab 6: Primes vs. Fair Coin Flips

Compare the completely deterministic prime-factor walk M(x) against three genuine random coin-flip walks.

Mertens M(x) at end:2 ±√x coin envelope:±31.6
Deterministic arithmetic looks astonishingly indistinguishable from pure chance.

Even stranger: in 1972, number theorist Hugh Montgomery happened to mention the spacing between Riemann zeros over afternoon tea at Princeton with physicist Freeman Dyson. Dyson froze. The exact formula Montgomery had derived for the gaps between Riemann zeros was already famous in nuclear physics — it describes the energy levels of heavy uranium nuclei!

Ordinary random events (like raindrops or Poisson arrivals) often clump closely together. Both quantum energy levels and Riemann zeros repel one another, rarely crowding close together. Below is the actual spacing histogram of our first 100 Riemann zeros plotted against pure randomness and quantum chaos (the GUE Wigner surmise).

Lab 7: Quantum level repulsion in the first 100 zeros

Actual normalised gaps between consecutive Riemann zeros avoid zero spacing — matching heavy atomic nuclei, not random clumping.

First 100 zeros Quantum GUE curve Random Poisson

Wait — would proving it break internet encryption?

No. That is one of the most persistent myths in popular science. Everyday RSA encryption relies on the difficulty of factoring a huge 600-digit number back into its two secret prime factors.

The Riemann Hypothesis governs how smoothly primes are distributed on average, not how to crack a specific composite number into factors. Your bank details are safe even if a proof lands tomorrow morning.

So what is actually at stake?

First, hundreds of theorems across mathematics begin with "Assuming the Riemann Hypothesis…" — including fast, completely deterministic primality tests (like Miller–Rabin). A proof turns all of them into unconditional facts overnight.

Second, computers have already checked the first 10 trillion zeros and every one lies on the critical line — yet in mathematics, ten trillion examples are still not a proof. Gauss's Li(x) stays above π(x) for every number we can easily compute, yet J. E. Littlewood proved it eventually flips somewhere past 1019 (Skewes's number). Only a proof settles infinity.

Part 7 of 7

Check your intuition

Five quick questions to lock in the big ideas. No timers, no trick algebra.

Answered: 0 / 5

Plain-English glossary

Analytic continuation

A way of uniquely extending a smooth complex function beyond the region where its original formula makes sense. Just as knowing one arc of a circle fixes the whole circle, knowing ζ(s) for s > 1 fixes its value across the entire complex plane except at s = 1.

Critical strip and critical line

The critical strip is the vertical band of complex numbers s = σ + it with 0 < σ < 1, where all non-trivial zeros must live. The critical line is its exact vertical midline at σ = 1/2.

Euler product

Euler's 1737 identity rewriting the sum of 1/nˢ over all integers as a product of 1/(1 − p⁻ˢ) over prime numbers only. It is the bridge that links calculus to prime numbers.

Explicit formula

Riemann's master equation showing that the prime staircase equals a smooth trend minus a sum of oscillating waves, one wave for each non-trivial zero of ζ(s).

Hardy's Z-function Z(t)

A clever rotation of ζ(1/2 + it) that removes its twisting phase and produces a real-valued wave whose zeros match the critical-line zeros of ζ(s) one for one.

Logarithmic integral Li(x)

The integral of 1/ln(t) from 2 to x. It represents adding up Gauss's local prime density 1/ln(t) and is a remarkably accurate smooth estimate of π(x).

Mertens function M(x)

The cumulative sum of the Möbius function μ(n), stepping +1 for square-free numbers with an even number of prime factors and −1 for an odd number. The Riemann Hypothesis is equivalent to M(x) growing no faster than x^(1/2 + ε).

Trivial vs. non-trivial zeros

Trivial zeros are the negative even integers −2, −4, −6, …, where a sine factor in Riemann's functional equation forces ζ(s) = 0. Non-trivial zeros are all other zeros; they lie in the critical strip and carry the musical frequencies of the primes.