Zauner’s conjecture
Does a symmetric informationally complete quantum measurement (SIC-POVM) exist in every finite dimension?
Summary
Why it is open
In complex space of dimension d, at most d² lines through the origin can be equiangular, meaning that every pair has the same angle; a set that reaches d² lines must have squared overlap 1/(d + 1), and is called a SIC [3]. In quantum theory a SIC defines a measurement on a d-level system from whose outcome probabilities the probabilities of any other measurement can be computed, which is what "informationally complete" means [3].
By 1999 Zauner had found the connection between SICs and the Weyl–Heisenberg group, generated by a cyclic shift and a phase operator, and had proved that SICs exist up to d = 5 [3]. Independently, Carlton Caves developed SICs as quantum measurements, and a 2004 paper brought them to the wider quantum information community [1, 3]. The name covers two statements. Zauner's conjecture in the narrow sense is that SICs exist for all d [10]. Zauner, and independently Marcus Appleby, also conjectured that in every dimension d > 2 a Weyl–Heisenberg SIC exists whose starting vector, the fiducial, is an eigenvector of a particular unitary of order 3, now called the Zauner unitary [3].
No prize or institution sets the bar. A resolution would be a proof that SICs exist in every dimension, or a dimension shown to have none. Exact and numerical solutions cover only finitely many dimensions. A 2022 list of open problems in quantum information asks, as a step towards the conjecture, for SICs in any infinite sequence of dimensions, and notes that the known "SIC dimension tower" was still "of a finite size only" [8]. In a September 2026 preprint, not peer reviewed, Danylo Radchenko and Campbell Wheeler wrote that, after their work, "the only remaining obstacle to proving Zauner's conjecture in full" is to show that a suitable Galois automorphism exists [11].
What is known
Numerical searches (2004–2020). Joseph Renes, Robin Blume-Kohout, Andrew Scott and Caves constructed SICs in dimensions 2, 3 and 4, conjectured that a group-covariant SIC exists in every dimension, and gave numerical solutions up to dimension 45 [1]. Scott and Markus Grassl extended the numerical solutions to every d ≤ 67, with a putatively complete list of Weyl–Heisenberg covariant solutions for d ≤ 50 [2]. By 2017 numerical solutions were known in every dimension up to 151 and in some larger ones, up to 844 [3]. Paweł Horodecki, Łukasz Rudnicki and Karol Życzkowski report that by 2020 they were known for every d ≤ 193 and in further dimensions up to 2,208 [8]. All known SICs are group covariant, and all but one, the Hoggar SIC in dimension 8, come from the Weyl–Heisenberg group [3]. Every SIC found so far is invariant under an element of order 3 of the Clifford group, which maps the Weyl–Heisenberg group to itself [9].
Exact solutions (2017–2025). In 2017 exact solutions were known in dimensions 2–24, 28, 30, 31, 35, 37, 39, 43, 48 and 124 [3]. Grassl and Scott proposed a family with extra symmetry linked to Fibonacci and Lucas numbers, supported by exact solutions in dimensions 4, 8, 19, 48, 124 and 323 and a numerical one in 844 [4]. Ingemar Bengtsson, Grassl and Gary McConnell, and separately Appleby, Steven Flammia and Gene Kopp, state that exact solutions are known in every dimension up to 53 and in some higher ones [9, 10]. Bengtsson, Grassl and McConnell computed exact SICs in a series of dimensions of the form n² + 3 = 4p, with p prime, up to d = 39,604, and verified the SIC conditions for them exactly; they cannot prove that their algorithm works in dimensions where it has not been tested [9].
Number theory (2016–2022). For each d ≥ 4 there is a real quadratic field Q(√D), where D is the square-free part of (d + 1)(d − 3) [5]. Appleby, Flammia, McConnell and Jon Yard showed with the Magma computer algebra system that in dimensions 4–21, 24, 28, 30, 35, 39 and 48 some SIC generates exactly a specific ray class field of this field, a kind of abelian extension [5]. Explicit generators of such fields are known in general only for the rationals and imaginary quadratic fields; the search for them goes back to Hilbert's 12th problem [5]. The Stark conjectures predict special units, Stark units, from the derivatives of L-functions at s = 0 [6, 7]. Kopp showed that suitable units of this kind would give SICs in odd prime dimensions congruent to 2 modulo 3, verified the construction numerically in dimensions 5, 11, 17 and 23, and gave an exact solution in dimension 23 [6]. In SICs of prime dimension n² + 3, Appleby, Bengtsson, Grassl, Michael Harrison and McConnell found every fiducial entry but the first to be a common factor times the square root of a Stark unit, and constructed SICs in thirteen such dimensions up to 19,603 [7].
Conditional and partial proofs (2025–2026). In a 2025 preprint, not peer reviewed, Appleby, Flammia and Kopp proved that the Stark conjecture for real quadratic fields, together with a new conjectured identity, the twisted convolution identity, implies Zauner's conjecture [10]. Their construction gives a putatively complete list of Weyl–Heisenberg SICs for every d > 3, and produced four inequivalent numerical SICs in dimension 100, three of them new [10]. Kopp and Jeffrey Lagarias showed that for 4 ≤ d ≤ 90 the number of known classes of such SICs equals the size of the ideal class monoid of the quadratic order of discriminant (d + 1)(d − 3), and conjectured that this holds for every d ≥ 4 [13]. In September 2026 Radchenko and Wheeler posted a preprint proving that the relevant Stark units are algebraic numbers and proving the quadratic relations that Appleby, Flammia and Kopp had conjectured [11]. Appleby, Flammia and Kopp then extended the proof to every rank and established, without assuming the Stark conjecture, that "ghost" SICs exist: non-Hermitian configurations that the Stark conjecture predicts are Galois conjugate to genuine SICs [12]. They write that these methods "have not established Zauner's conjecture", which would require a Galois automorphism sending √Δ to −√Δ that intertwines with complex conjugation in a particular way [12]. Both 2026 papers are preprints, not peer reviewed.
Main approaches
Numerical search with imposed symmetry. Assume Weyl–Heisenberg covariance and Zauner symmetry, which shrinks the search space [3]. Strongest evidence: numerical SICs in every dimension up to 193 [8, 9]. Main difficulty: dimension 121 was already "pushing the limits" of feasible computation without further assumptions [3], numerical solutions cover only finitely many dimensions [8], and searching only covariant SICs may itself bias what is found [3].
Symmetric families and dimension towers. Dimensions sharing the same quadratic field form towers, and SICs in later members are expected to carry larger symmetry, which makes them easier to find [9]. Strongest evidence: exact solutions up to dimension 323 in the Fibonacci–Lucas tower [4, 9] and up to 39,604 in dimensions n² + 3 = 4p [9]. Main difficulty: in 2022 no tower was known to continue indefinitely [8], and the algorithm is not proved to work in untested dimensions [9].
Stark units and class field theory. Build SICs from Stark units and the special function that Appleby, Flammia and Kopp call the Shintani–Faddeev modular cocycle [7, 10]. Strongest evidence: the conditional proof [10], with its second hypothesis now proved in a 2026 preprint [11, 12]. Main difficulty: the Galois step, and the real quadratic Stark conjecture on which it rests, remain unproved [11, 12].
Classification by quadratic orders. Match the number and fields of SICs in each dimension with arithmetic of real quadratic orders [13]. Strongest evidence: agreement for every dimension from 4 to 90 [13]. Main difficulty: the comparison relies on catalogues of known SICs, which are believed complete only up to dimension 90 [3, 13].
Open sub-questions
- The Galois step. For the ghost SICs that now exist unconditionally, is there in some infinite family of dimensions a Galois automorphism sending √Δ to −√Δ that turns them into genuine SICs [11, 12]?
- An infinite tower. Can the algorithm that Bengtsson, Grassl and McConnell tested in dimensions n² + 3 = 4p be proved to work for every prime p of this form [8, 9]?
- Counting classes. Does the number of Weyl–Heisenberg SIC classes equal the size of the ideal class monoid for the order of discriminant (d + 1)(d − 3) in dimensions between 91 and 193, where numerical SICs exist [3, 8, 13]?
- Non-covariant SICs. Is every SIC group covariant? This is proved only for dimensions 2 and 3 [3].
- Sporadic SICs. Do sporadic SICs, exceptional examples such as the Hoggar SIC in dimension 8, exist in any further dimension [3, 10]?
Work on this in Papex
If you have a testable answer to one of the sub-questions above, you can draft it as a private hypothesis in Papex, or prepare the draft from Claude or ChatGPT with the Papex connector. You review the exact version, sign it and decide whether to publish it; a public record that names this problem is listed on this page. A Papex record shows who signed which version and when. It is not peer review and does not mean the hypothesis is correct.
Key references
- Renes, J. M., Blume-Kohout, R., Scott, A. J. & Caves, C. M. (2004). Symmetric informationally complete quantum measurements. Journal of Mathematical Physics 45(6), 2171–2180. 10.1063/1.1737053 · Checked 16 Oct 2026
- Scott, A. J. & Grassl, M. (2010). Symmetric informationally complete positive-operator-valued measures: a new computer study. Journal of Mathematical Physics 51(4), 042203. 10.1063/1.3374022 · Checked 16 Oct 2026
- Fuchs, C. A., Hoang, M. C. & Stacey, B. C. (2017). The SIC question: history and state of play. Axioms 6(3), 21. 10.3390/axioms6030021 · Checked 16 Oct 2026
- Grassl, M. & Scott, A. J. (2017). Fibonacci-Lucas SIC-POVMs. Journal of Mathematical Physics 58(12), 122201. 10.1063/1.4995444 · Checked 16 Oct 2026
- Appleby, M., Flammia, S., McConnell, G. & Yard, J. (2020). Generating ray class fields of real quadratic fields via complex equiangular lines. Acta Arithmetica 192(3), 211–233. 10.4064/aa180508-21-6 · Checked 16 Oct 2026
- Kopp, G. S. (2021). SIC-POVMs and the Stark conjectures. International Mathematics Research Notices 2021(18), 13812–13838 (published online 31 October 2019). 10.1093/imrn/rnz153 · Checked 16 Oct 2026
- Appleby, M., Bengtsson, I., Grassl, M., Harrison, M. & McConnell, G. (2022). SIC-POVMs from Stark units: prime dimensions n² + 3. Journal of Mathematical Physics 63(11), 112205. 10.1063/5.0083520 · Checked 16 Oct 2026
- Horodecki, P., Rudnicki, Ł. & Życzkowski, K. (2022). Five open problems in quantum information theory. PRX Quantum 3(1), 010101. 10.1103/PRXQuantum.3.010101 · Checked 16 Oct 2026
- Bengtsson, I., Grassl, M. & McConnell, G. (2025). SIC-POVMs from Stark units: dimensions n² + 3 = 4p, p prime. Journal of Mathematical Physics 66(8), 082202. 10.1063/5.0212617 · Checked 16 Oct 2026
- Appleby, M., Flammia, S. T. & Kopp, G. S. (2025). A constructive approach to Zauner's conjecture via the Stark conjectures. Preprint, arXiv:2501.03970 (version 2, 17 March 2025), not peer reviewed. Source · Checked 16 Oct 2026
- Radchenko, D. & Wheeler, C. (2026). Real quadratic fields and finite quantum dilogarithms I. Preprint, arXiv:2609.21892 (version 2, 29 September 2026), not peer reviewed. Source · Checked 16 Oct 2026
- Appleby, M., Flammia, S. T. & Kopp, G. S. (2026). The twisted convolution identity and ghost r-SICs from finite quantum dilogarithms. Preprint, arXiv:2609.39192 (version 1, 30 September 2026), not peer reviewed. Source · Checked 16 Oct 2026
- Kopp, G. S. & Lagarias, J. C. (2027). SIC-POVMs and orders of real quadratic fields. Journal of Number Theory 293, 128–181 (volume of April 2027). 10.1016/j.jnt.2026.05.013 · Checked 16 Oct 2026
Source
Open, as listed by Wikipedia · Unsolved problems in mathematics. Checked 08 Oct 2026.
The status is the source’s. Papex does not decide whether a problem is solved.
Records on this problem
No public record addresses this problem yet.