Riemann Zeta Function Mode
- Background
- Special Values on the Real Axis
- Structural Features
- Non-Trivial Zeros
- Gram Points
- Historical Milestones in Computation
- Modern Computational Milestones
- References
The Riemann zeta function ζ(s) is a complex function defined for Re(s) > 1 by the Dirichlet series:
ζ(s) = 1 + 1/2ˢ + 1/3ˢ + 1/4ˢ + ...
Leonhard Euler studied this series for real s in the 1730s–40s. Bernhard Riemann extended it to all complex s (except s = 1) via analytic continuation in his landmark 1859 paper, where he also conjectured that all non-trivial zeros lie on the critical line Re(s) = 1/2 — the Riemann Hypothesis, still unproven today.
The zeros of ζ(s) are intimately connected to the distribution of prime numbers. The non-trivial zeros directly control the error term in the Prime Number Theorem.
A reference guide for all tour points you can find in the app. Organized by mathematical category, with historical context and primary sources.
Visualizer view: X Axis Location: Re(s) axis
The real axis hosts both the trivial zeros (negative even integers), special rational/transcendental values of ζ, and the sole singularity at s = 1. It is the most computationally accessible part of the function.
Visualizer view: Pole Location: s = 1 Discovery: Known since Euler; formalized by Riemann (1859)
ζ(s) has a simple pole at s = 1 with residue 1. This is the only singularity of the function in the entire complex plane. As s → 1, ζ(s) → ∞, corresponding to the divergence of the harmonic series. Riemann's analytic continuation defines the function everywhere else.
Visualizer view: Origin Location: s = 0 Discovery: Riemann (1859)
Via analytic continuation, ζ(0) = −1/2. This is the regularized value corresponding to the divergent series 1 + 1 + 1 + ···. The result follows directly from the functional equation relating ζ(s) to ζ(1 − s).
Source: Riemann (1859) [R1]
Visualizer view: Ramanujan Summation Location: s = −1 Discovery: Euler ~1749 (informal); Ramanujan (1913)
ζ(−1) = −1/12, the regularized value corresponding to the famously divergent sum 1 + 2 + 3 + 4 + ···. Euler hinted at related results in the 1740s–60s. Srinivasa Ramanujan independently derived this in Chapter 6 of his notebooks (~1913) using what is now called the Ramanujan summation method — a systematic technique for assigning finite values to divergent series via the Euler–MacLaurin formula.
The result is not a literal sum but arises from analytic continuation of ζ(s) to s = −1. It appears in theoretical physics in string theory (Casimir effect, zeta function regularization of vacuum energy).
Sources: Ramanujan notebooks (~1913) [Ram]; Riemann (1859) [R1]
Visualizer view: Basel Problem Location: s = 2 Discovery: Euler (1734/1735, published 1740)
ζ(2) = π²/6 ≈ 1.6449. The Basel Problem — finding the exact value of ∑ 1/n² — had stumped mathematicians for nearly a century. Euler solved it in 1734 (presented 1735, published 1740), to immediate international acclaim. He used the infinite product factorization of sin(x)/x equating to a polynomial in x² whose coefficients could be matched to the series. More rigorous proofs were later supplied by others.
More generally, ζ(2k) for positive integers k is always a rational multiple of π²ᵏ. The values ζ(2), ζ(4), ζ(6), ... are all well understood. Odd integer values remain much more mysterious.
Source: Euler (1740) [E1]
Visualizer view: Apéry's Constant Location: s = 3 Discovery: Apéry (1978/1979)
ζ(3) ≈ 1.2020569... The value had been computed numerically for centuries but its arithmetic nature was unknown. In 1978, Roger Apéry astonished the mathematical community by proving ζ(3) is irrational using an unexpectedly elementary argument involving rapidly converging integer sequences. The proof was initially met with skepticism but quickly verified.
Whether ζ(3) is transcendental, or a rational multiple of π³, remains unknown. It is the only odd-argument zeta value known to be irrational.
Source: Apéry (1979) [A1]
Visualizer view: Trivial Zeroes Location: s = −2, −4, −6, ... Discovery: Riemann (1859)
ζ(s) = 0 at all negative even integers. These are the trivial zeros, so named because their existence is an easy consequence of the functional equation:
ζ(s) = 2ˢ πˢ⁻¹ sin(πs/2) Γ(1−s) ζ(1−s)
which Riemann established in his 1859 paper. The factor sin(πs/2) vanishes at s = −2, −4, −6, ... Since the other factors are non-zero there, ζ(s) = 0 at each. There are infinitely many trivial zeros and they are fully understood — in sharp contrast to the non-trivial zeros.
Source: Riemann (1859) [R1]
Visualizer view: Critical Strip Location: 0 < Re(s) < 1 Established: Riemann (1859)
The region of the complex plane where all non-trivial zeros of ζ(s) are known to reside. Riemann established in 1859 that ζ(s) has no zeros with Re(s) ≤ 0 or Re(s) ≥ 1 (outside the trivial zeros), so every non-trivial zero must have its real part strictly between 0 and 1. This is proven; the open question is whether they all lie on the center line Re(s) = 1/2.
Visualizer view: Critical Line Location: Re(s) = 1/2 Conjectured: Riemann (1859)
The vertical line Re(s) = 1/2 is the axis of symmetry of the critical strip (reflecting the functional equation ζ(s) ↔ ζ(1−s)). The Riemann Hypothesis asserts that every non-trivial zero lies on this line — that is, has Re(s) exactly 1/2. All zeros verified numerically so far (trillions of them) do lie on this line, but no general proof exists.
Visualizer view: Derivative Zero Location: s ≈ 2.46 (real axis)
The real zero of ζ'(s) to the right of the critical strip, in the region where ζ(s) is real-valued and monotonically decreasing. Studying where ζ'(s) = 0 in the complex plane illuminates the clustering and spacing of non-trivial zeros and connects to conjectures about simple zeros.
Visualizer view: Saddle Points Location: s ≈ 1 + 9.0858i and s ≈ 9.0858i
Points where ζ'(s) = 0 within or near the critical strip. Saddle points appear in conjugate pairs (due to the functional equation) and are flanked by pairs of non-trivial zeros of ζ(s). They are important in steepest-descent analysis of the zeta function's oscillatory behavior and in numerical methods for locating zeros.
These are zeros of ζ(s) with 0 < Re(s) < 1. All known non-trivial zeros lie on the critical line Re(s) = 1/2. They are indexed by their imaginary parts in ascending order; only the upper half (Im(s) > 0) is listed since zeros come in conjugate pairs.
Visualizer view: Gateway Zero Location: s = 1/2 + 14.1347251i Discovery: Riemann (1859)
The first and most famous non-trivial zero. Riemann identified approximate locations of several zeros in his 1859 paper. At t ≈ 14.135, it is the lowest-lying non-trivial zero and historically the key example confirming that the zeta function's zeros encode information about prime gaps and the distribution of primes.
Source: Riemann (1859) [R1]
Visualizer view: Spacing Benchmark Location: s = 1/2 + 21.0220396i First computed: Gram (1903)
The second non-trivial zero. The gap between the 1st and 2nd zeros (~6.89) was later used as a key data point by Hugh Montgomery in his 1973 work linking zero spacings to eigenvalue statistics of large random Hermitian matrices. Montgomery's pair correlation conjecture predicts that zero spacings follow the GUE (Gaussian Unitary Ensemble) distribution — a discovery connecting analytic number theory to quantum chaos.
Sources: Gram (1903) [G1]; Montgomery (1973) [M1]
Visualizer view: Stability Zero Location: s = 1/2 + 25.0108576i First computed: Gram (1903)
The third non-trivial zero. Together with the first few zeros, it formed the "original set" that gave Riemann and his successors confidence that all non-trivial zeros lie on the critical line. At such low heights, the zeros are relatively spread out and well-separated, making computation by hand feasible.
Source: Gram (1903) [G1]
Visualizer view: Chaos Indicator Location: s = 1/2 + 30.4248761i First computed: Gram (1903)
The fourth non-trivial zero. Its spacing relative to adjacent zeros already shows the irregular, pseudo-random pattern that characterizes the zero distribution. This statistical irregularity, when examined at large scale, is now known to mimic the eigenvalue statistics of random matrices from quantum chaos models (GUE universality).
Source: Gram (1903) [G1]
Visualizer view: Hardy's Milestone Location: s = 1/2 + 37.5861782i Significance: Hardy (1914)
In 1914, G.H. Hardy proved — for the first time — that infinitely many non-trivial zeros lie on the critical line Re(s) = 1/2. His proof applied to the Z-function (the real-valued version of ζ on the critical line) and showed it must change sign infinitely often, hence have infinitely many real zeros. This was a landmark result, though it did not resolve whether all non-trivial zeros lie on the critical line.
Source: Hardy (1914) [H1]
Visualizer view: Pre-Computer Limit Location: s = 1/2 + 40.9187186i Computed by: J.P. Gram (1903), extended by Backlund (~1912–1918)
Among the zeros first computed by J.P. Gram and subsequently extended by R.J. Backlund, who developed more systematic methods to count zeros in intervals. Backlund's work in the 1910s pushed the frontier of hand computation further and introduced rigorous interval-based zero-counting via the argument principle. This zero represents the outer edge of what was achievable by pure hand calculation in the early 20th century.
Sources: Gram (1903) [G1]; Backlund (1914) [B1]
Visualizer view: Repulsion Effect Location: s = 1/2 + 43.3270732i First computed: Gram (1903)
An example of level repulsion — the observed tendency of consecutive non-trivial zeros to avoid getting too close together. This mirrors the repulsion of energy levels in quantum systems (specifically, eigenvalues of random matrices from GUE). The statistical mechanism is the same: an effective "force" prevents levels from coinciding, leading to a distribution where very small gaps are rare.
Source: Gram (1903) [G1]
Visualizer view: Prime Regulator Location: s = 1/2 + 48.0051508i First computed: Gram (1903)
Higher zeros like this one contribute oscillatory correction terms to the explicit formula for π(x) (the prime counting function):
π(x) = Li(x) − ∑ₚ Li(xᵖ) + small terms
where the sum is over non-trivial zeros ρ = 1/2 + it. Each zero adds a wave at frequency log(x) · t that destructively interferes with other zeros to produce the irregular gaps between primes. The more zeros included, the more precisely π(x) is approximated.
Source: Gram (1903) [G1]
Visualizer view: Gram Point Link Location: s = 1/2 + 49.7738325i Computed by: J.P. Gram (1903)
Calculated in Gram's foundational 1903 paper, which introduced Gram points as a systematic tool for locating and verifying zeros. This zero is near a Gram point, illustrating the typical relationship between Gram points and zero positions that Gram's Law describes.
Source: Gram (1903) [G1]
Visualizer view: Zero Cluster Location: Im(s) ≈ 129.58 (zoomed out view)
A zoomed-out view of a cluster of consecutive zeros, illustrating the irregular spacing pattern of the zero distribution. While zeros have an average spacing of approximately 2π/log(t/2π) at height t (which decreases logarithmically as t grows), the actual spacings fluctuate widely and follow the GUE distribution statistically.
Gram points are the real values t > 0 where the argument of ζ(1/2 + it) is a multiple of π, equivalently where Im(ζ(1/2 + it)) = 0. They were introduced by J.P. Gram in his 1903 paper and used to systematically locate and count non-trivial zeros.
Gram's Law (Gram, 1903) states that each interval [gₙ, gₙ₊₁] between consecutive Gram points typically contains exactly one zero of ζ on the critical line. While true surprisingly often, it fails with positive density (approximately 1 in 4 Gram intervals). The first failure was discovered by J.W. Hutchinson in 1925.
Source: Gram (1903) [G1]
Visualizer view: Gram Point g₁ Location: Im(s) ≈ 17.8456 Introduced: Gram (1903)
The first Gram point. At t ≈ 17.846, Im(ζ(1/2 + it)) = 0. This point lies between the 1st and 2nd non-trivial zeros (t ≈ 14.13 and t ≈ 21.02), illustrating a typical case where Gram's Law holds: the Gram interval contains exactly one zero.
Source: Gram (1903) [G1]
Visualizer view: Gram Point g₂ Location: Im(s) ≈ 23.17 Introduced: Gram (1903)
The second Gram point. At t ≈ 23.17, Im(ζ(1/2 + it)) = 0. Also used in early computational verification of zeros on the critical line.
Source: Gram (1903) [G1]
Visualizer view: Gram's Law Violation Location: Im(s) ≈ 282.465 Discovered by: J.W. Hutchinson (1925)
The first known violation of Gram's Law. At this height, the expected alternating rhythm — one zero per Gram interval — breaks down: two zeros fall in the same interval and an adjacent interval is empty. Hutchinson discovered the first two violations (near t ≈ 282.5 and t ≈ 295.5) in his 1925 paper. It is now known that Gram's Law fails for approximately 1 in 4 Gram intervals at large heights. The violations become more frequent as t increases and are related to the increasing complexity of the Z-function's oscillations.
Source: Hutchinson (1925) [Hu1]
A chronology of landmark achievements in numerically verifying the Riemann Hypothesis.
Visualizer view: Pre-Computer Limit, Gram Point Link, and others Zeros verified: First 15 (up to t ≈ 58)
Gram's 1903 paper was the first systematic computational study of non-trivial zeros, introducing Gram points and what became known as Gram's Law. He computed the first 15 zeros entirely by hand, carefully checking that each lay on the critical line. His methods established the framework for all subsequent zero computations.
Source: Gram (1903) [G1]
Zeros verified: First ~79 (up to t ≈ 200)
Backlund developed rigorous methods using the argument principle to count zeros in intervals without computing each one individually. This allowed verification that no zeros had been missed, providing the first complete counted verification of a range of non-trivial zeros.
Source: Backlund (1914) [B1]
Visualizer view: Titchmarsh's Last Zero Location: s = 1/2 + 1468.82i (the 1041st zero) Discovery: Titchmarsh (1936)
E.C. Titchmarsh, aided by L.J. Comrie who used a mechanical Brunsviga calculator, computed and verified the first 1,041 non-trivial zeros, reaching t ≈ 1468.82. This marked the absolute limit of pre-electronic computation for this problem. The work appeared in the Proceedings of the Royal Society and was a milestone in analytic number theory.
Source: Titchmarsh (1936) [T1]
Visualizer view: Turing's Last Zero Location: s = 1/2 + 1540.572i (the 1104th zero) Discovery: Turing (1950, published 1953)
Alan Turing used the Manchester Mark 1 computer in 1950 to extend the verified zero list to the 1,104th zero, reaching t ≈ 1540.572. In addition to the raw computation, Turing developed a new method — now called Turing's method — for rigorously checking that no zeros had been missed in a given range using the functional equation and Gram's Law corrections. The paper was published posthumously in 1953.
This was the first use of an electronic computer to verify non-trivial zeros of the Riemann zeta function.
Source: Turing (1953) [Tu1]
Visualizer view: Lehmer's Phenomenon Location: s = 1/2 + 7005.063i and s = 1/2 + 7005.101i (zeros #6709 and #6710) Discovery: Lehmer (1956)
While computing non-trivial zeros on an early computer, D.H. Lehmer discovered that the 6,709th and 6,710th zeros are separated by only 0.038 — far smaller than the typical spacing of ~0.23 at that height. These two zeros are so close that the Z-function barely changes sign between them, making them extremely difficult to detect reliably.
This is now called Lehmer's phenomenon (or a Lehmer pair). Such close pairs are dangerous because a sufficiently close pair with a local extremum of the wrong sign would constitute a counterexample to the Riemann Hypothesis (a zero off the critical line). While no such counterexample has been found, Lehmer pairs remain the most challenging cases to verify computationally.
Source: Lehmer (1956) [L1]
Visualizer view: ZetaGrid Location: t ≈ 3.2 × 10¹³ (for reference only — beyond rendering precision)
ZetaGrid (2001–2005) was the first large-scale distributed computing project focused on the Riemann Hypothesis, comparable in structure to SETI@home. Volunteers donated idle CPU cycles to verify that the first ~10¹³ non-trivial zeros all lie on the critical line, checking over one billion zeros per day at peak capacity.
The project was organized by Sebastian Wedeniwski at IBM. However, results were not formally peer-reviewed, which limits their standing as rigorous mathematical proof. The project ended in November 2005 due to hosting instability.
⚠ The height t ≈ 3.2 × 10¹³ is far beyond the rendering precision of the visualizer — this entry is shown for historical reference only.
Source: ZetaGrid (2001–2005) [Z1]
Visualizer view: Platt & Trudgian Location: t ≈ 3 × 10¹² (for reference only — beyond rendering precision)
Dave Platt (University of Bristol) and Tim Trudgian (UNSW Canberra) published a peer-reviewed, fully rigorous proof in 2021 that all non-trivial zeros of ζ(s) with 0 < Im(s) ≤ 3 × 10¹² lie on the critical line Re(s) = 1/2.
Unlike ZetaGrid's volunteer computation, Platt and Trudgian used rigorous interval arithmetic — a verified computation technique where every floating-point operation is accompanied by proven error bounds, making the result a genuine mathematical theorem rather than a numerical observation. This is currently the gold standard for computational verification of the Riemann Hypothesis.
⚠ The height t ≈ 3 × 10¹² is far beyond the rendering precision of the visualizer — this entry is shown for historical reference only.
Source: Platt & Trudgian (2021) [PT1]
| Tag | Citation |
|---|---|
| [R1] | Riemann, B. (1859). "Ueber die Anzahl der Primzahlen unter einer gegebenen Grösse." Monatsberichte der Königlichen Preußischen Akademie der Wissenschaften zu Berlin, 671–680. |
| [E1] | Euler, L. (1740). "De summis serierum reciprocarum." Commentarii Academiae Scientiarum Petropolitanae, 7, 123–134. (Result presented 1734/1735.) |
| [Ram] | Ramanujan, S. (~1913). Unpublished notebooks, Chapter 6. Published in: Hardy, G.H. et al. (eds.), Collected Papers of Srinivasa Ramanujan, Cambridge University Press, 1927. |
| [G1] | Gram, J.P. (1903). "Sur les zéros de la fonction ζ(s) de Riemann." Acta Mathematica, 27, 289–304. |
| [B1] | Backlund, R.J. (1914). "Sur les zéros de la fonction ζ(s) de Riemann." Comptes Rendus Acad. Sci. Paris, 158, 1979–1982. |
| [H1] | Hardy, G.H. (1914). "Sur les zéros de la fonction ζ(s) de Riemann." Comptes Rendus de l'Académie des Sciences, 158, 1012–1014. |
| [Hu1] | Hutchinson, J.W. (1925). "On the Roots of the Riemann Zeta-Function." Transactions of the American Mathematical Society, 27(1), 49–60. |
| [T1] | Titchmarsh, E.C. (1936). "The zeros of the Riemann zeta-function." Proceedings of the Royal Society of London. Series A, 151(873), 234–255. |
| [Tu1] | Turing, A.M. (1953). "Some calculations of the Riemann zeta-function." Proceedings of the London Mathematical Society, Series 3, 3(1), 99–117. |
| [L1] | Lehmer, D.H. (1956). "Extended computation of the Riemann zeta-function." Mathematika, 3(2), 102–108. Also: Lehmer, D.H. (1956). "On the Roots of the Riemann Zeta-Function." Acta Mathematica, 95, 291–298. |
| [M1] | Montgomery, H.L. (1973). "The pair correlation of zeros of the zeta function." In Analytic Number Theory, Proc. Symposia Pure Math., 24, 181–193. AMS. |
| [A1] | Apéry, R. (1979). "Irrationalité de ζ(2) et ζ(3)." Astérisque, 61, 11–13. |
| [PT1] | Platt, D. & Trudgian, T. (2021). "The Riemann hypothesis is true up to 3·10¹²." Bulletin of the London Mathematical Society, 53(3), 792–797. DOI: 10.1112/blms.12460 |
| [Z1] | Wedeniwski, S. ZetaGrid project (2001–2005). Source code and overview: https://github.com/Wedeniwski/ZetaGrid |
Document generated from the Fractal Traveler Riemann view data (src/data/riemann.json).