---
title: Bernstein–Sato Ideals
url: https://www.emergentmind.com/topics/bernstein-sato-ideals
type: topic
---

# Bernstein–Sato Ideals

Bernstein–Sato ideals are multivariable \(D\)-module invariants attached to collections of functions or, more generally, to tuples of ideals. They extend the classical Bernstein–Sato polynomial \(b_f(s)\), which is defined for a single function by a functional equation \(b_f(s)\,f^s=P(s)\,f^{s+1}\), to settings in which several exponents vary simultaneously or in which a single ideal is encoded by auxiliary variables [1907.04010]. In the one-variable case the Bernstein–Sato ideal is principal and generated by the classical \(b\)-function; in the multivariable case one obtains an ideal in a polynomial ring \(\mathbb{C}[s_1,\dots,s_r]\), whose zero locus records singularity-theoretic, birational, and topological information, including multiplier-ideal jumping loci and monodromy support loci [1907.04010] [2111.03334] [2408.13560].

## 1. Classical origin and the passage from functions to ideals

For a nonzero function \(f\) on a smooth complex algebraic variety or complex manifold, the Bernstein–Sato polynomial \(b_f(s)\in\mathbb{C}[s]\) is the monic polynomial of minimal degree for which there exists a differential operator \(P(s)\in D_X[s]\) satisfying
\[
b_f(s)\,f^s=P(s)\,f^{s+1}.
\]
Equivalently, if \(M_f=D_X[s]\cdot f^s\subset \mathcal{O}_X[f^{-1}][s]\cdot f^s\), then \(b_f(s)\) is the monic generator of the annihilator in \(\mathbb{C}[s]\) of the cyclic element \(f^s\) [2109.00244]. In this classical case, roots of \(b_f(s)\) are negative rational numbers, and the negative of the largest root yields the log-canonical threshold or minimal exponent, depending on context [1906.03086] [2408.13560].

A first generalization replaces a principal ideal by an arbitrary ideal \(\mathfrak a=(f_1,\dots,f_r)\). Budur–Mustaţă–Saito defined a Bernstein–Sato polynomial \(b_{\mathfrak a}(s)\) using a multi-parameter \(D\)-module construction, and Mustaţă later showed that this polynomial can be recovered from a single auxiliary hypersurface
\[
g=f_1y_1+\cdots+f_ry_r
\]
on \(X\times \mathbf A^r\), via the identity
\[
b_{\mathfrak a}(s)=\tilde b_g(s)=\frac{b_g(s)}{s+1},
\]
where \(b_g(s)\) is the classical Bernstein–Sato polynomial of \(g\) and \(\tilde b_g(s)\) is its reduced Bernstein–Sato polynomial [1906.03086]. This reduction from an ideal to a single function is structurally important: it gives a new proof that \(b_{\mathfrak a}(s)\) exists and depends only on the ideal, not on the chosen generators [1906.03086].

The same construction underlies later developments for tuples of ideals. In particular, for a single ideal \(\mathfrak a=(f_1,\dots,f_r)\), Mustaţă’s hypersurface \(g=\sum f_i y_i\) is the model for defining reduced Bernstein–Sato invariants that are independent of generators and compatible with multiplier ideals [2109.00244].

## 2. Multivariable Bernstein–Sato ideals for tuples of functions

For a tuple of functions
\[
F=(f_1,\dots,f_r),
\]
the natural multivariable analogue is the Bernstein–Sato ideal \(B_F\subset \mathbb{C}[s_1,\dots,s_r]\). It is defined by requiring that \(b(s_1,\dots,s_r)\in B_F\) if there exists \(P\in D_X[s_1,\dots,s_r]\) such that
\[
b(s_1,\dots,s_r)\prod_{i=1}^r f_i^{s_i}
=
P(x,\partial_x,s_1,\dots,s_r)\prod_{i=1}^r f_i^{s_i+1}.
\]
When \(r=1\), this ideal is principal and generated by the classical Bernstein–Sato polynomial \(b_f(s)\) [1907.04010] [2509.09113]. For \(r>1\), the ideal is typically not principal, and its zero locus may have components of codimension \(>1\) [1907.04010].

A broader family of generalized ideals is indexed by \(m=(m_1,\dots,m_r)\in\mathbb N^r\). The ideal \(B_F^m\subset \mathbb{C}[s_1,\dots,s_r]\) consists of all polynomials \(b\) for which
\[
b\cdot \prod_{i=1}^r f_i^{s_i}
=
P\cdot \prod_{i=1}^r f_i^{s_i+m_i}
\]
for some \(P\in D_X[s_1,\dots,s_r]\) [2509.09113]. The case \(m=(1,\dots,1)\) recovers the usual \(B_F\) [2509.09113].

The zero locus
\[
Z(B_F)\subset \mathbb{C}^r
\]
is a fundamental geometric object. In the one-variable case, it is simply the set of roots of \(b_f(s)\). In the multivariable case, codimension-one components are hyperplanes, and components of codimension \(>1\) can occur [1907.04010]. The multivariable theory is therefore genuinely richer than the principal case and is designed to capture interactions among several hypersurfaces rather than a single divisor.

A different but compatible multivariable formalism appears in the generalized Bernstein–Sato ideals used for arrangements and restricted functional equations. For a factorization \(f=f_1\cdots f_r\), a divisor \(f'\mid f\), and another divisor \(g\mid f\), one considers the ideal \(B_{g,f'}(F)\subset \mathbb{C}[s_1,\dots,s_r]\) of polynomials \(B(S)\) satisfying
\[
B(S)\,f'F^S\in \mathscr D_X[S]\cdot g f'F^S.
\]
The standard Bernstein–Sato ideal in the sense of Budur is recovered by taking \(f'=1\) and \(g=f\) [1909.00547]. This generality is particularly effective for hyperplane arrangements, where explicit combinatorial descriptions of zero loci become possible [1909.00547].

## 3. Bernstein–Sato ideals for tuples of ideals

A further extension, introduced in "A note on Bernstein-Sato ideals" [2109.00244], associates a Bernstein–Sato ideal to a tuple of ideals
\[
\mathfrak a=(\mathfrak a_1,\dots,\mathfrak a_\ell)
\]
in \(R=\mathbb{C}[x_1,\dots,x_n]\) or \(R=\mathbb{C}\{x_1,\dots,x_n\}\). For each \(i\), choosing generators
\[
\mathfrak a_i=(f_{i,1},\dots,f_{i,r_i}),
\]
one introduces new variables \(y_{i,j}\) and defines
\[
g_i=f_{i,1}y_{i,1}+\cdots+f_{i,r_i}y_{i,r_i}.
\]
This produces a tuple of hypersurfaces \(G=(g_1,\dots,g_\ell)\) in a larger ring \(A\), and one defines the Bernstein–Sato ideal \(B_G\subset \mathbb{C}[s_1,\dots,s_\ell]\) by the condition that
\[
\delta(s_1,\dots,s_\ell)\, g_1^{s_1+1}\cdots g_\ell^{s_\ell+1}
=
b(s_1,\dots,s_\ell)\,g_1^{s_1}\cdots g_\ell^{s_\ell}
\]
for some \(\delta\in D_A[s_1,\dots,s_\ell]\) [2109.00244].

Because the \(g_i\) are pairwise without common factors, one has
\[
B_G\subseteq ((s_1+1)\cdots (s_\ell+1)),
\]
and the reduced Bernstein–Sato ideal \(\widetilde B_G\) is defined by dividing out this universal factor [2109.00244]. The Bernstein–Sato ideal of the tuple of ideals is then
\[
B_{\mathfrak a}:=\widetilde B_G\subseteq \mathbb{C}[s_1,\dots,s_\ell).
\]
A key theorem states that \(\widetilde B_G\) is independent of the chosen generators of the \(\mathfrak a_i\), so \(B_{\mathfrak a}\) is a well-defined invariant of the tuple [2109.00244].

This construction specializes correctly in two basic directions. If \(\ell=1\), then \(B_{\mathfrak a}\subset \mathbb{C}[s]\) is generated by the reduced Bernstein–Sato polynomial of Mustaţă’s universal linear combination \(g\), hence recovers the one-variable Bernstein–Sato polynomial of an ideal [2109.00244]. If each \(\mathfrak a_i\) is principal, \(\mathfrak a_i=(f_i)\), then \(B_{\mathfrak a}\) coincides with the reduced Bernstein–Sato ideal of the tuple of functions \(F=(f_1,\dots,f_\ell)\) in Sabbah’s sense [2109.00244].

The note also records basic properties inherited from \(B_G\): nontriviality, localization as an intersection of local Bernstein–Sato ideals, and divisibility by \(\prod_i(s_i+1)\) before reduction [2109.00244]. These properties place tuple-of-ideals Bernstein–Sato theory on the same formal footing as the already established theories for one function, one ideal, and tuples of functions.

## 4. Zero loci, monodromy, multiplier ideals, and jumping loci

The most developed geometric interpretation of Bernstein–Sato ideals concerns their zero loci. For a tuple of functions \(F=(f_1,\dots,f_r)\), the zero locus \(Z(B_F)\subset \mathbb{C}^r\) is related to the support \(S(F)\subset (\mathbb{C}^*)^r\) of Sabbah’s specialization complex and, equivalently, to cohomology support loci of rank-one local systems on local complements [1907.04010]. The decisive statement is
\[
\operatorname{Exp}(Z(B_F))=S(F),
\qquad
\operatorname{Exp}(a)=\bigl(e^{2\pi i a_1},\dots,e^{2\pi i a_r}\bigr),
\]
which generalizes the Malgrange–Kashiwara theorem from one function to several [1907.04010]. In particular, \(S(F)\) is a finite union of torsion-translated complex affine subtori of codimension \(1\), and every codimension-one irreducible component of \(Z(B_F)\) is a hyperplane
\[
a_1s_1+\cdots+a_rs_r+b=0
\]
with \(a_i\in\mathbb{Q}_{\ge 0}\) and \(b\in\mathbb{Q}_{>0}\) [1907.04010].

This picture extends to generalized ideals \(B_F^m\). If \(m\in\mathbb N^r\) and \(f^m=\prod_i f_i^{m_i}\) is not invertible, there is a corresponding restricted monodromy support \(\mathcal S_F^m\subset(\mathbb C^*)^r\), and one has
\[
\operatorname{Exp}(Z(B_F^m))=\mathcal S_F^m.
\]
Codimension-one components of \(Z(B_F^m)\) are again hyperplanes of the same rational type, with the additional positivity condition that some coefficient attached to an index \(i\) with \(m_i\neq 0\) is nonzero [2001.05728] [2408.13560].

A parallel geometric interpretation arises from multiplier ideals. In the one-variable case, jumping numbers of multiplier ideals in \((0,1)\) give roots of the Bernstein–Sato polynomial for a function or an ideal [2109.00244]. For tuples of ideals \(\mathfrak a=(\mathfrak a_1,\dots,\mathfrak a_\ell)\), mixed multiplier ideals are defined by
\[
\mathcal J(\mathfrak a^{\boldsymbol\lambda})
=
\pi_*\mathcal O_{X'}\bigl(\lceil K_\pi-\lambda_1F_1-\cdots-\lambda_\ell F_\ell\rceil\bigr),
\]
where \(\pi:X'\to X\) is a common log resolution and \(\mathfrak a_i\cdot \mathcal O_{X'}=\mathcal O_{X'}(-F_i)\) [2109.00244]. Their constancy regions and jumping walls are unions of rational hyperplanes determined by the numerical data of the resolution [2109.00244].

The main theorem of [2109.00244] states that if \(\boldsymbol\lambda\in\mathbb Q_{\ge 0}^\ell\) is a jumping point of \(\mathfrak a\) with \(\|\boldsymbol\lambda\|<1\), then
\[
-\boldsymbol\lambda\in Z(B_{\mathfrak a}).
\]
Thus every mixed jumping point near the origin yields a point on the zero locus of the Bernstein–Sato ideal [2109.00244]. The statement is an inclusion, not an equality, and the note explicitly does not claim the converse [2109.00244].

Further estimates for zero loci were obtained in [2111.03334]. For \(F=(f_1,\dots,f_r)\) and a multi-index \(a\in\mathbb N^r\), every codimension-one irreducible component of \(Z(B_F^a)\) has the form
\[
\operatorname{ord}_E(f_1)s_1+\cdots+\operatorname{ord}_E(f_r)s_r+k_E+c=0
\]
for some irreducible component \(E\) of a strong log resolution and some integer \(c>0\) [2111.03334]. Conversely, facets of jumping walls and of the \(\LCT\)-polytope that intersect the region \(\KLT_a(F)\) give actual codimension-one components of \(Z(B_F^a)\) [2111.03334]. This multivariate estimate generalizes Lichtin-type upper bounds and the classical principle that small jumping numbers give roots of the \(b\)-function [2111.03334].

## 5. Structural properties, symmetry, and functorial operations

Bernstein–Sato ideals exhibit several nontrivial structural operations. One concerns the whole family \(B_F^m\). Budur had proved an inclusion expressing \(B_F^m\) as contained in an intersection of shifted elementary ideals \(B_F^{e_j}\); "The Intersection Structure of Bernstein-Sato Ideals" [2509.09113] proves this inclusion is always an equality. For every permutation \(\pi\) of \(\{1,\dots,r\}\),
\[
B_F^m
=
\bigcap_{\substack{1\le j\le r\\ m_{\pi(j)}>0}}
\ \bigcap_{k=0}^{m_{\pi(j)}-1}
t_{\pi(1)}^{m_{\pi(1)}}\cdots
t_{\pi(j-1)}^{m_{\pi(j-1)}}t_{\pi(j)}^k
\cdot B_F^{e_{\pi(j)}},
\]
where \(t_j\) acts by the shift \(s_j\mapsto s_j+1\) [2509.09113]. This symmetric intersection property is proved by interpreting \(B_F^m\) as annihilators of logarithmic \(D\)-modules and decomposing monomial quotients via Gröbner bases [2509.09113]. In the one-variable case, it implies that \(B_f^{2e_1}\) is generated by the least common multiple of \(b_f(s)\) and \(b_f(s+1)\) [2509.09113].

A direct application is a formula for powers of a function:
\[
b_{f^n}(s)
=
\operatorname{lcm}\Bigl\{
\frac{b_f(ns+i)}{n^{\deg(b_f)}}
\ \Big|\ i=0,1,\dots,n-1
\Bigr\},
\]
derived from the identity \(B_{F^n}=B_F^n(n_1s_1,\dots,n_rs_r)\) and the intersection theorem [2509.09113].

Another operation is the tensor or multiplicative Thom–Sebastiani rule. For functions \(f\in A\) and \(g\in B\) on smooth varieties, one has
\[
b_{f\cdot g}(s)=b_f(s)\,b_g(s),
\]
and for effective divisors \(D_1,D_2\) on nonsingular varieties,
\[
b_{D_1\otimes D_2}(s)=b_{D_1}(s)\,b_{D_2}(s)
\]
[2406.04121]. For ideals, full multiplicativity fails in general: Example 3.8 in [2406.04121] gives monomial ideals \(\mathfrak a,\mathfrak b\) for which \(b_{\mathfrak a\mathfrak b}(s)\neq b_{\mathfrak a}(s)b_{\mathfrak b}(s)\). However, if one factor is principal then
\[
b_{\mathfrak a\cdot (g)}(s)=b_{\mathfrak a}(s)\,b_{(g)}(s),
\]
and for monomial ideals one has root-level inclusions and congruence modulo \(\mathbb Z\):
\[
W_{\mathfrak a}\cup W_{\mathfrak b}\subset W_{\mathfrak a\mathfrak b},
\qquad
W_{\mathfrak a}\cup W_{\mathfrak b}\equiv W_{\mathfrak a\mathfrak b}\ \mathrm{mod}\ \mathbb Z
\]
[2406.04121].

Hyperplane arrangements provide another source of structure. For strongly Euler-homogeneous, Saito-holonomic, and tame divisors, the modules underlying multivariate Bernstein–Sato ideals are \((n+1)\)-Cohen–Macaulay, which forces the corresponding ideals to be principal and their zero loci to be purely codimension one [2008.07447]. In that setting, different factorizations are related by a diagonal property: if \(H\) is a coarser factorization of the same divisor \(f\), then
\[
A\in Z(B_{H,\mathfrak x}^a)
\iff
\Delta_H(A)\in Z(B_{F,\mathfrak x}^{\Delta_H(a)}),
\]
where \(\Delta_H\) is the diagonal embedding induced by merging factors [2008.07447]. For hyperplane arrangements factored into linear forms, the corresponding Bernstein–Sato ideals are even reduced [2008.07447].

## 6. Computation, examples, and variants

The computational theory of Bernstein–Sato polynomials and generalized Bernstein–Sato polynomials is based on Gröbner bases in Weyl algebras, graph embeddings, \(V\)-filtrations, and elimination [1002.1475]. For a hypersurface \(f\), the global \(b\)-function is obtained from the initial ideal \(\mathrm{in}_{(-w,w)}I_f\) and a minimal polynomial computation for \(\theta=-\partial_t t\) [1002.1475]. The same framework yields local Bernstein–Sato polynomials \(b_{f,P}(s)\), generalized Bernstein–Sato polynomials \(b_{f,g}(s)\) for an ideal \(I=(f_1,\dots,f_r)\), and Shibuta’s \(m\)-generalized Bernstein–Sato polynomials \(b_{f,g}^{(m)}(s)\), which characterize multiplier-ideal membership [1002.1475].

These algorithms were implemented in the D-modules package of Macaulay2 [1002.1475]. They compute global and local \(b\)-functions, generalized Bernstein–Sato polynomials of arbitrary polynomial ideals, log canonical thresholds, jumping coefficients, and multiplier ideals [1002.1475]. In the examples of [1002.1475], the roots of \(b_f(-s)\) and the resulting multiplier ideals show that not every root in \((0,1]\) is a jumping coefficient, so Bernstein–Sato data is finer than the jumping-locus data alone [1002.1475].

For arrangements, explicit formulas are available in several cases. For central generic arrangements in \(\mathbb C^n\) of degree \(d>n\), if \(F=(l_1,\dots,l_d)\) is the factorization into linear forms, then
\[
B_F^1
=
\mathbb C[S]\cdot
\Biggl(
\prod_{k=1}^d (s_k+1)\,
\prod_{i=0}^{2d-n-2}
\Bigl(\sum_{k=1}^d s_k+i+n\Bigr)
\Biggr)
\]
[2008.07447]. For other factorizations, the zero locus is
\[
Z(B_F^1)
=
\bigcup_{k=1}^r\{s_k+1=0\}
\ \cup\
\bigcup_{i=0}^{2d-n-2}
\Bigl\{\sum_{k=1}^r d_k s_k+i+n=0\Bigr\}
\]
[2008.07447]. These formulas refine earlier estimates of Maisonobe and illustrate how zero loci can often be described as explicit hyperplane arrangements.

Monomial ideals admit a different form of explicit analysis. In characteristic \(0\), Budur–Mustaţă–Saito described the roots of \(b_{\mathfrak a}(s)\) in terms of Newton polyhedra; in characteristic \(p\), [1907.11709] proves that for monomial ideals the roots of \(b_{\mathfrak a_\mathbb C}(s)\) coincide with the Bernstein–Sato roots of the mod-\(p\) reductions \(\mathfrak a_p\) for \(p\) large enough [1907.11709]. This result concerns roots rather than multivariable ideals, but it supports the view that Bernstein–Sato theory has a meaningful Frobenius-theoretic analogue in positive characteristic [1907.11709] [1402.1333].

A further extension replaces holomorphic functions by meromorphic functions \(f/g\). The paper [2112.08492] develops Bernstein–Sato polynomials \(b_{f/g}^\alpha(s)\) and \(b_{f/g}(s)\) from Sabbah’s two-variable Bernstein–Sato ideal for the pair \((f,g)\), by specializing along lines \(s_1=s\), \(s_2=-s-\alpha\) or \(s_2=-s\) [2112.08492]. The resulting roots are again negative rational numbers and are bounded by the log-resolution data of the meromorphic germ [2112.08492]. This suggests a broader principle: multivariable Bernstein–Sato ideals provide the natural ambient structure from which one-variable invariants of more specialized objects can often be extracted by linear specialization.

Across these variants, a common pattern emerges. Bernstein–Sato ideals are annihilator ideals in parameter rings attached to \(D\)-modules generated by symbols such as \(f^s\), \(F^S\), or their ideal-theoretic analogues; their zero loci are unions of rational affine hyperplanes or their translated intersections; and these loci interface directly with multiplier ideals, mixed jumping walls, local-system support loci, nearby cycles, and monodromy [1907.04010] [2109.00244] [2111.03334] [2408.13560]. This conjunction of \(D\)-module theory, birational geometry, and topology is the defining feature of Bernstein–Sato ideals as a research topic.

Source: https://www.emergentmind.com/topics/bernstein-sato-ideals