---
title: 4D Sklyanin Algebras
url: https://www.emergentmind.com/topics/four-dimensional-sklyanin-algebras
type: topic
---

# 4D Sklyanin Algebras

Four-dimensional Sklyanin algebras are connected $\mathbb{N}$-graded quadratic algebras on four degree-one generators that occupy a central position in noncommutative projective geometry, elliptic algebras, and PI representation theory. In the standard presentation, a 4-dimensional Sklyanin algebra $S=S(\alpha,\beta,\gamma)$ is generated by $x_0,x_1,x_2,x_3$ subject to six quadratic relations involving commutators and anticommutators, with parameters satisfying $\alpha+\beta+\gamma+\alpha\beta\gamma=0$ and standard nondegeneracy exclusions; equivalently, one may describe it by geometric data $(E,\sigma)$, where $E\subset \mathbb{P}^3$ is an elliptic curve and $\sigma$ is translation by a point $\tau\in E$ [1802.06487]. For admissible parameters, these algebras are noetherian domains of global dimension $4$, Artin–Schelter regular/Gorenstein, Auslander-regular, Cohen–Macaulay, and have Hilbert series $(1-t)^{-4}$ [1802.06487]. Their structure is governed by two central quadrics, a twisted homogeneous coordinate factor ring attached to $E$, a pencil of noncommutative quadric surfaces, and, in the finite-order case, a rich PI center equipped with a nontrivial Poisson bracket whose symplectic stratification controls irreducible representations [1108.1552] [1802.06487].

## 1. Definition and parameterizations

A 4-dimensional Sklyanin algebra $S=S(\alpha,\beta,\gamma)$ is the connected $\mathbb{N}$-graded $k$-algebra generated by degree-one variables $x_0,x_1,x_2,x_3$ subject to
\[
x_0x_1-x_1x_0=\alpha(x_2x_3+x_3x_2),\qquad
x_0x_1+x_1x_0=x_2x_3-x_3x_2,
\]
\[
x_0x_2-x_2x_0=\beta(x_3x_1+x_1x_3),\qquad
x_0x_2+x_2x_0=x_3x_1-x_1x_3,
\]
\[
x_0x_3-x_3x_0=\gamma(x_1x_2+x_2x_1),\qquad
x_0x_3+x_3x_0=x_1x_2-x_2x_1,
\]
where the parameters satisfy
\[
\alpha+\beta+\gamma+\alpha\beta\gamma=0,
\]
equivalently
\[
(1+\alpha)(1+\beta)(1+\gamma)=(1-\alpha)(1-\beta)(1-\gamma),
\]
together with the exclusions
\[
(\alpha,\beta,\gamma)\notin\{(-1,1,\gamma),(\alpha,-1,1),(1,\beta,-1)\}
\]
in the regular setting [1802.06487]. In a common nondegenerate regime one also requires $\{\alpha,\beta,\gamma\}\cap\{0,\pm1\}=\varnothing$ [1512.05717] [2507.19719].

The associated projective geometry is encoded by an elliptic curve
\[
E\subset \mathbb{P}^3
\]
cut out by two quadrics
\[
\phi_1=v_0^2+v_1^2+v_2^2+v_3^2,\qquad
\phi_2=\frac{1-\gamma}{1+\alpha}v_1^2+\frac{1+\gamma}{1-\beta}v_2^2+v_3^2,
\]
together with an automorphism $\sigma\in \operatorname{Aut}(E)$ realized as translation by a point $\tau\in E$ [1802.06487]. The point scheme of $S$ is the union of $E$ and the four coordinate points
\[
e_0=[1:0:0:0],\ e_1=[0:1:0:0],\ e_2=[0:0:1:0],\ e_3=[0:0:0:1]
\]
[1802.06487].

A complementary analytic presentation writes the algebra as $A(E,\tau)$ over $k=\mathbb{C}$, with relations expressed in terms of Jacobi theta functions evaluated at $\tau$; in that form the standing assumption is that $\tau$ is not $4$-torsion [1802.06023]. The theta-function model and the $(\alpha,\beta,\gamma)$-presentation are two realizations of the same family.

## 2. Central quadrics, quotient geometry, and regularity

A decisive structural feature is the existence of two central degree-two elements
\[
g_1=-x_0^2+x_1^2+x_2^2+x_3^2,\qquad
g_2=x_1^2+\frac{1+\alpha}{1-\beta}x_2^2+\frac{1-\alpha}{1+\gamma}x_3^2,
\]
and a surjective homomorphism
\[
S\to B=B(E,\mathcal{O}(1)|_E,\sigma)
\]
whose kernel is generated by $g_1$ and $g_2$ [1802.06487]. Thus
\[
B\cong S/(Sg_1+Sg_2),
\]
the pair $\{g_1,g_2\}$ is a central regular sequence, and $B$ is the twisted homogeneous coordinate ring attached to $(E,\mathcal{O}(1)|_E,\sigma)$; its graded localization is isomorphic to $k(E)[t^{\pm1};\sigma]$ [1802.06487].

In this form the standard homological package is especially rigid: $S$ and $B$ are noetherian domains of global dimension $4$, AS-regular/Gorenstein, Auslander-regular, Cohen–Macaulay, with Hilbert series $(1-t)^{-4}$ [1802.06487]. Other treatments add that the algebra is Koszul and strongly noetherian in the nondegenerate case [1512.05717] [1612.06158].

The generic center is already visible from the central quadrics. For generic parameters, equivalently $\tau$ of infinite order, the center is generated by two algebraically independent central quadrics $\Omega_1,\Omega_2$, and
\[
Z(S)=\mathbb{C}[\Omega_1,\Omega_2]
\]
[1612.06158]. In the notation of the PI analysis this is the statement that if $|\sigma|=\infty$, then
\[
Z(S)=k[g_1,g_2]
\]
and $S$ is not module-finite over its center [1802.06487]. Moreover,
\[
(\Omega_1,\Omega_2)=\bigcap_{p\in E_\lambda}\operatorname{Ann}_S(M_p),
\]
so the central quadrics recover the annihilator ideal of all point modules on the elliptic point scheme [1612.06158].

A further connection with lower-dimensional Sklyanin geometry is provided by Van den Bergh’s noncommutative quadrics: the second Veronese of a cubic 3-dimensional Sklyanin algebra is a quotient of a 4-dimensional quadratic Sklyanin algebra, and the kernel is generated by a central quadratic lying in the span of the two central quadrics [1612.06158]. This places the 4-dimensional theory at an intersection point between noncommutative $\mathbb{P}^3$, twisted homogeneous coordinate rings, and noncommutative quadric hypersurfaces.

## 3. Noncommutative quadrics, line modules, and symmetry

The degree-two center defines a pencil of noncommutative quadrics. For a basis $\{g_1,g_2\}$ of the central two-dimensional subspace in degree $2$, one sets
\[
Q_{[\lambda:\mu]}=\operatorname{Proj}\big(S/(\lambda g_1+\mu g_2)\big).
\]
Its base locus is the elliptic curve $E=\operatorname{Proj}S/(g_1,g_2)$, directly paralleling a generic pencil of quadrics in commutative $\mathbb{P}^3$ [1108.1552]. Labeling the pencil by points $z\in E$ via the corresponding central quadratic $g(z)$, one obtains
\[
Q_z=Q_{-z-2\tau},
\]
and the smoothness criterion
\[
Q_z\ \text{is smooth if and only if}\ z+\tau\notin E[2].
\]
Equivalently, there are exactly four singular quadrics in the pencil, namely
\[
Q_{w-\tau},\qquad w\in E[2]
\]
[1108.1552].

Line modules supply the ruling structure. For $p,q\in E$, the secant line $\overline{pq}$ determines a line module $L(pq)$, and for fixed $z$ the two rulings on a smooth quadric are
\[
\{L(pq)\mid p+q=z\},\qquad \{L(pq)\mid p+q=-z-2\tau\}.
\]
When $z+\tau\in E[2]$, these two families coincide, so the singular quadric has only one ruling [1108.1552]. This recasts the classical geometry of rulings on a quadric surface in categorical terms.

For a smooth noncommutative quadric $Q$, one has
\[
\operatorname{Pic}(Q)\cong \mathbb{Z}^2,
\]
with generators $l,l'$ coming from the two rulings and hyperplane class $h=l+l'$ [1108.1552]. The intersection pairing defined from alternating sums of Ext-dimensions satisfies
\[
l\cdot l=0,\qquad l'\cdot l'=0,\qquad l\cdot l'=1=l'\cdot l,
\]
exactly as for $\mathbb{P}^1\times \mathbb{P}^1$ [1108.1552]. At the same time, the noncommutative setting exhibits an additional phenomenon: certain fat point objects have Euler self-pairing $2$, corresponding to a self-intersection number $-2$ under the geometric sign convention [1108.1552].

Heisenberg symmetry is another recurrent feature. The Heisenberg group $H_4$ acts by degree-preserving automorphisms on $Q_4(E,\tau)$, with degree-one part the $4$-dimensional simple $H_4$-representation [1612.06158]. This representation theory organizes the quadratic relation space, the proof of existence of two central quadrics, and the explicit quotient map to the second Veronese of a cubic 3-dimensional Sklyanin algebra [1612.06158]. In related constructions built from $S(E,\tau)$, finite quantum-group symmetries tied to the Heisenberg group act by autoequivalences and control point and line module incidence [1509.01634].

## 4. PI structure, enlarged centers, and Poisson geometry

Let $n=|\sigma|$ be the order of the elliptic translation automorphism. Then $S$ and its twisted homogeneous coordinate factor $B$ are module-finite over their centers, hence PI, if and only if $|\sigma|<\infty$; in that case
\[
\operatorname{PIdeg}(S)=\operatorname{PIdeg}(B)=n
\]
[1802.06487]. Writing
\[
s=\frac{n}{(n,2)},
\]
one obtains the parity parameter governing the even/odd dichotomy in the finite-order case [1802.06487].

Smith–Tate’s description of the center becomes explicit in this regime: $Z(S)$ is generated by four degree-$n$ elements $z_0,z_1,z_2,z_3$ together with the degree-two central elements $g_1,g_2$, subject to two relations $F_1,F_2$ of degree $2n$ [1802.06487]. For odd $n$, the relations take the form
\[
F_i=\Phi_i(z_0,z_1,z_2,z_3)+h_i(g_1,g_2),\qquad i=1,2,
\]
while for even $n$ they acquire additional linear terms in the $z$-variables [1802.06487]. The resulting affine variety
\[
Y=\operatorname{MaxSpec}Z(S)\subset \mathbb{A}^6
\]
is the central geometric object in the PI theory.

Walton–Wang–Yakimov construct on every PI 4-dimensional Sklyanin algebra a nontrivial Poisson $Z$-order structure in the sense of Brown–Gordon [1802.06487]. The construction proceeds by specialization from a formal deformation $S_\hbar$ and yields a nonzero induced Poisson bracket on $Z(S)$. Crucially,
\[
\partial_{g_i}=0,\qquad i=1,2,
\]
so $g_1$ and $g_2$ lie in the Poisson center [1802.06487]. Under the hypotheses verified in the paper, the bracket on $Z(S)$ is of Jacobian type:
\[
\{z_k,z_l\}=\eta\, J(F_1,F_2)_{k,l},\qquad \{g_i,\cdot\}=0,
\]
for some $\eta\in k^\times$ [1802.06487].

This Poisson structure controls the symplectic-core decomposition of $Y$. For fixed scalars $\gamma_1,\gamma_2$, the slices
\[
Y_{\gamma_1,\gamma_2}=Y\cap V(g_1-\gamma_1,g_2-\gamma_2)
\]
have smooth loci that are $2$-dimensional symplectic cores, while the singular points of all slices form the $0$-dimensional cores [1802.06487]. One obtains
\[
Y^{\mathrm{symp}}_0=\bigcup_{\gamma_1,\gamma_2}(Y_{\gamma_1,\gamma_2})^{\mathrm{sing}},
\]
and the Azumaya locus of $S$ coincides with the smooth locus
\[
Y\setminus Y^{\mathrm{sing}}
\]
[1802.06487].

The singular locus itself depends sharply on the parity of $n$. If $n$ is odd, then
\[
Y^{\mathrm{sing}}
\]
is the union of $2(n-1)$ cuspidal curves
\[
C(\omega+k\tau),\qquad \omega\in E_2,\ 0\le k\le n-2,
\]
meeting only at the origin, and $Y^{\mathrm{symp}}_0$ contains two additional cuspidal curves per $2$-torsion point $\omega$ [1802.06487]. If $n$ is even, then $Y^{\mathrm{sing}}$ is the union of two rational surfaces $Y^{\mathrm{sing}}_1,Y^{\mathrm{sing}}_2$, depending on a chosen $2$-torsion automorphism $\rho$, and over $\mathbb{C}$ these surfaces contain the cuspidal curves $\mathcal{C}(\omega+k\tau)$ that parametrize supports of fat point modules of multiplicity at most $s$ [1802.06487]. A useful contrast is that, unlike the 3-dimensional root-of-unity case, the symplectic cores of maximal dimension have codimension $2$ in the 4-dimensional PI setting, making Poisson geometry essential in the Azumaya analysis [1802.06487].

## 5. Representation theory, fat points, and discriminants

The finite-order case admits a detailed representation theory. If $\tau$ has order $n$, then every simple $A(E,\tau)$-module has dimension at most $n$, and “almost all” have dimension precisely $n$; moreover, finite-dimensional simple modules separate elements, so for every nonzero $a\in A$ there exists a simple module $S$ with $a\cdot S\neq 0$ [1802.06023]. Consequently, $A$ satisfies the standard polynomial identity of degree $2n$ and none of lower degree, is a maximal order, and is finite over its center [1802.06023]. The associated degree-zero fraction division algebra has rational center and degree
\[
s=
\begin{cases}
n,& n\ \text{odd},\\
n/2,& n\ \text{even},
\end{cases}
\]
over that center [1802.06023].

Fat point modules organize the exceptional fibers. In the finite-order setting, for each $1\le k\le s-1$ and each $w\in E[2]$ there is a fat point $F(w+k\tau)$ of multiplicity $k+1$, characterized by exact sequences
\[
0\to M(p-(k+1)\tau,q-(k+1)\tau)(-k-1)\to M(p,q)\to F(w+k\tau)\to 0
\]
whenever $p+q=w+k\tau$ [1802.06023]. These are all intermediate-multiplicity fat points. There are also infinitely many multiplicity-$s$ fat points, and their parameter space is a rational $3$-fold [1802.06023].

Walton–Wang–Yakimov sharpen this picture by classifying irreducible representations via the Poisson geometry of $Y$, the algebro-geometric structure of the elliptic curve, and the theory of line and fat point modules [1802.06487]. If $y\in Y\setminus Y^{\mathrm{sing}}$, then
\[
S/\mathfrak{m}_yS\cong M_n(k),
\]
so there is a unique $n$-dimensional simple representation with central character $y$; if $y=0$, the unique simple is the trivial $1$-dimensional representation [1802.06487]. Over $\mathbb{C}$ the remaining simple modules are described in terms of Sklyanin’s theta-function representations $V(\omega+k\tau)$ and their scalar twists.

For odd $n$, the map from $n$-dimensional simples to $Y^{\mathrm{smooth}}$ is bijective, and for each $0\le k\le n-2$ there is a $2\!:\!1$ map
\[
\operatorname{Irr}_{k+1}S\ \sqcup\ \operatorname{Irr}_{n-1-k}S\longrightarrow C(\omega+k\tau)\setminus\{0\}
\]
[1802.06487]. Thus $S$ has irreducible representations of every dimension
\[
1,2,\dots,n.
\]
For even $n=2s$, the smooth locus again parametrizes the $n$-dimensional simples, while a generic point of the singular surfaces supports two $s$-dimensional simples, and the cuspidal curves $\mathcal{C}(\omega+k\tau)$ carry the intermediate-dimensional simples through $4\!:\!1$ maps
\[
\operatorname{Irr}_{k+1}S\ \sqcup\ \operatorname{Irr}_{s-1-k}S\longrightarrow \mathcal{C}(\omega+k\tau)\setminus\{0\}
\]
for $0\le k\le s-2$ [1802.06487]. In particular, over $\mathbb{C}$ the possible dimensions are
\[
1,2,\dots,s,n
\]
in the even case [1802.06487].

This representation-theoretic stratification is reflected exactly in the discriminant ideals over the center. If $\operatorname{tr}:S\to Z(S)$ is the reduced trace, then for the discriminant ideals $D_\ell(S/Z)$ and modified discriminants $MD_\ell(S/Z)$ one has
\[
V(D_\ell)=V(MD_\ell),
\qquad
V(D_{n^2})=Y^{\mathrm{sing}},
\]
and the lower zero sets are unions of the same cuspidal curves, rational surfaces, or the origin, depending on $\ell$ and the parity of $n$ [1802.06487]. The paper interprets these strata as recording the square-sum of dimensions of irreducible representations in the central fiber.

## 6. Twists, adjacent families, and later developments

Several later developments clarify both the robustness and the rigidity of the 4-dimensional Sklyanin class. Davies studied cocycle twists by the Klein four-group $G\cong C_2\times C_2$ and showed that the twist $A^{G,\mu}$ of a 4-dimensional Sklyanin algebra remains generated in degree $1$, strongly noetherian, AS-regular of global dimension $4$, Auslander regular, Cohen–Macaulay, Koszul, with Hilbert series $(1-t)^{-4}$ [1512.05717]. At the same time, the geometry changes drastically: when $|\sigma|=\infty$, the point scheme of $A^{G,\mu}$ consists of exactly $20$ points, while the twist has infinitely many fat point modules of multiplicity $2$ parametrized by the quotient elliptic curve $E/G$ [1512.05717]. The corresponding twisted homogeneous coordinate factor $B^{G,\mu}$ is described as a twisted ring attached to a sheaf of maximal orders on $E^G$, with
\[
\operatorname{qgr}(B^{G,\mu})\simeq \operatorname{coh}(\mathcal{E})
\]
[1512.05717].

Chirvasitu–Smith analyzed a larger 3-parameter class $A(\alpha,\beta,\gamma)$ with the same six formal relations but without imposing the Sklyanin constraint [1702.00377]. Within that ambient family, the genuine 4-dimensional Sklyanin algebras are characterized precisely by
\[
\alpha+\beta+\gamma+\alpha\beta\gamma=0
\]
together with the standard nondegeneracy assumptions [1702.00377]. Outside the Sklyanin locus, and under $\alpha\beta\gamma\neq 0$ and $\alpha+\beta+\gamma\neq 0$, the point scheme becomes a zero-dimensional scheme of degree $20$, in fact a graph of a bijection between two $20$-point subsets of $\mathbb{P}^3$, and all four squares
\[
x_0^2,\ x_1^2,\ x_2^2,\ x_3^2
\]
are central [1702.00377]. This provides an objective criterion for distinguishing true 4-dimensional Sklyanin algebras from Sklyanin-like $4$-generator, $6$-relator algebras.

The Sklyanin algebra also serves as input for other elliptic constructions. In the exotic elliptic algebra
\[
A(E,\tau)=\big(S(E,\tau)\otimes M_2(k)\big)^\Gamma,
\]
where $\Gamma=(\mathbb{Z}/2)^2$, one again obtains a noetherian AS-regular algebra of dimension $4$ with the same Hilbert series as the polynomial ring in four variables; its point modules form a $20$-point scheme and its line modules are parametrized by a degree-$20$ curve in $G(1,3)$ that is the union of four disjoint plane conics and three disjoint quartic elliptic curves isomorphic to $E/(t)$ for the three nonzero $2$-torsion points $t$ [1509.01634]. The finite quantum group related to the Heisenberg group of size $4^3$ acts by autoequivalences and governs much of the incidence theory [1509.01634].

A different specialization arises when $\sigma^4=1$. In that case, after normalizing to parameters $(\alpha,1,-1)$ and imposing a $\ast$-structure, Nikolaev proved that the quotient by a specific two-sided ideal $I_\mu$ is $\ast$-isomorphic to the algebraic noncommutative torus:
\[
S(\alpha)/I_\mu \cong A_\theta^{\mathrm{alg}},
\]
with $x_1\mapsto u$, $x_2\mapsto u^\ast$, $x_3\mapsto v$, $x_4\mapsto v^\ast$ [1109.6688]. This yields a covariant functor from elliptic curves to noncommutative tori in that normalization [1109.6688].

Recent work has also emphasized a limitation of the 4-dimensional theory from the viewpoint of noncommutative differential geometry. Although the nondegenerate 4-dimensional Sklyanin algebras are regular graded algebras of dimension $4$ with Hilbert series $(1-t)^{-4}$ and are noetherian domains outside the exceptional parameter values, none of them admits a connected integrable differential calculus of suitable dimension in the sense of Brzeziński–Sitarz [2507.19719]. This contrasts sharply with the nondegenerate 3-dimensional Sklyanin algebras, all of which are differentially smooth in that framework [2507.19719]. A plausible implication is that homological smoothness and AS-regularity are strictly weaker than differential smoothness for elliptic algebras of global dimension $4$.

Source: https://www.emergentmind.com/topics/four-dimensional-sklyanin-algebras