Papers
Topics
Authors
Recent
Search
2000 character limit reached

Logarithmic Comparison Theorem (LCT)

Updated 27 January 2026
  • Logarithmic Comparison Theorem (LCT) is a foundational result that equates the hypercohomology of logarithmic de Rham complexes with that of meromorphic forms under specific divisor conditions.
  • It bridges analytic, algebraic, and arithmetic geometry by leveraging free divisors, D-module techniques, and Bernstein–Sato polynomial properties.
  • Recent advances extend LCT to twisted local systems and prismatic cohomology, expanding its role in singularity, Hodge, and arithmetic theories.

The Logarithmic Comparison Theorem (LCT) is a fundamental result characterizing when the hypercohomology of the logarithmic de Rham complex along a divisor in a complex manifold coincides with that of the full meromorphic de Rham complex. The LCT provides crucial bridges between analytic, algebraic, and arithmetic geometry, with implications for D-module theory, singularity theory, Hodge theory, and the theory of motives. It has been generalized to contexts including twisted coefficients, log-structures, prismatic cohomology, and arithmetic comparison theorems.

1. Algebraic Formulation and Foundational Results

Given a complex manifold XX of dimension nn and a divisor DXD \subset X, the natural objects are:

  • The sheaf ΩXp(D)\Omega_X^p(*D) of meromorphic pp-forms on XX with arbitrary order poles along DD, forming the complex ΩX(D)\Omega_X^{\bullet}(*D).
  • The subsheaf ΩXp(logD)ΩXp(D)\Omega_X^p(\log D) \subset \Omega_X^p(*D) of pp-forms with at most logarithmic poles (i.e., locally, if D={h=0}D = \{h=0\}, then hωh\omega and hdωh d\omega are holomorphic), assembling into the logarithmic de Rham complex ΩX(logD)\Omega_X^{\bullet}(\log D).

The inclusion i:ΩX(logD)ΩX(D)i: \Omega_X^{\bullet}(\log D) \hookrightarrow \Omega_X^{\bullet}(*D) is canonical. The divisor DD is said to satisfy the Logarithmic Comparison Theorem if ii is a quasi-isomorphism, i.e.,

Hq(X,ΩX(logD))Hq(X,ΩX(D)),q.\mathbb{H}^q\left(X,\,\Omega_X^{\bullet}(\log D)\right) \cong \mathbb{H}^q\left(X,\,\Omega_X^{\bullet}(*D)\right),\quad \forall q.

For "free divisors" DD (i.e., ΩX1(logD)\Omega_X^1(\log D) is locally free of rank nn), and in particular for locally quasihomogeneous free divisors, it was proven by Castro-Jiménez, Mond, and Narváez-Macarro that the LCT always holds (Castro-Jiménez et al., 2023).

2. Key Hypotheses and Characterizations

The theorem’s scope and proofs depend on key properties of DD:

  • Free divisor (Saito): The sheaf DerX(logD)={ξΘXξ(h)h}\operatorname{Der}_X(-\log D) = \{\xi \in \Theta_X\,|\,\xi(h) \in \langle h\rangle\} is locally free. Equivalently, ΩX1(logD)\Omega_X^1(\log D) is locally free.
  • Locally quasihomogeneous (or strongly quasihomogeneous): For each pDp \in D, local equations can be chosen so the reduced defining equation hh is weighted homogeneous with strictly positive weights.
  • Spencer/Koszul-freeness: Relates to regularity properties of the associated sheaf of logarithmic differential operators VXV_X and is crucial for the D-module interpretations of LCT.

Significantly, these conditions dictate whether the inclusion of logarithmic forms provides a full cohomological account of the complement U=XDU = X \setminus D.

3. D-Module and Bernstein–Sato Framework

A cornerstone of the theory is the D-module perspective:

  • The inclusion of complexes can be realized as the de Rham functor applied to a morphism of DX\mathscr{D}_X-modules. The link between LCT and the behavior of these modules, especially their annihilators and generation by first-order operators, is central.
  • For locally quasihomogeneous free DD, the annihilator of f1f^{-1} (where D={f=0}D=\{f=0\}) is generated by first-order differential operators, specifically logarithmic derivations twisted by suitable constants (Castro-Jiménez et al., 2023, Bath et al., 2022).
  • The roots of the Bernstein–Sato polynomial bf(s)b_f(s) play a diagnostic role: order-one generation of the annihilator corresponds to bf(s)b_f(s) having roots with specific symmetries and locations, e.g., no integer roots less than 1-1. In the case of a homogeneous divisor with isolated singularities, LCT holds if and only if 1-1 is the only integral root of bf(s)b_f(s) (Bath et al., 2022).

4. Special Cases: Hyperplane Arrangements and Twisted LCT

Hyperplane arrangements constitute a primary class where the LCT has a complete and highly structured solution:

  • For any reduced hyperplane arrangement DD in Cn\mathbb{C}^n, both the analytic and algebraic untwisted LCT hold; this resolves Terao’s conjecture (Bath, 2022).
  • Twisted LCT: For arbitrary rank one local systems LL—corresponding to twisted logarithmic complexes—LCT holds provided precise combinatorial arithmetic "nonresonance" conditions are satisfied by the residues (weights) (Bath, 2022, Bath et al., 2022). The cohomology can then be computed explicitly and in finite terms via graded components of the logarithmic complex.
  • The DD-module approach extends to the twisted setting, and provides constraints on multivariate Bernstein–Sato ideals.

The regularity of the sheaves ΩXp(logD)\Omega_X^p(\log D) (specifically Castelnuovo–Mumford regularity at most zero) underpins the vanishing theorems and spectral sequence arguments that force the quasi-isomorphism (Bath, 2022, Bath et al., 2022).

5. Geometry: Free Divisors, Euler Homogeneity, and Counterexamples

The relationship between LCT and the geometric property of "strong Euler-homogeneity" is nuanced:

  • All known free divisors with LCT are strongly Euler-homogeneous, i.e., locally there exists a vector field δ\delta such that δ(f)=f\delta(f) = f.
  • Calderón-Moreno, Mond, Narváez, and Castro Jiménez (2002) conjectured that if a free divisor satisfies LCT, then it must be strongly Euler-homogeneous. This is now verified under broad circumstances, including dimensions 4\leq 4, Koszul-freeness, weak Koszul-freeness, or linearity in dimension $5$ (Rodríguez, 30 Apr 2025).
  • Counterexamples exist: explicit free divisors that are Koszul-free but not strongly Euler-homogeneous do not satisfy LCT, disproving conjectures that all linear free divisors satisfy LCT.
  • Techniques involved in recent advances exploit Jordan–Chevalley decomposition, trace and Fitting ideal analysis, and formal structure theorems for logarithmic vector fields.
Setting LCT characterization Comments
Plane curves, n=2n=2 LCT \Leftrightarrow strong EH All plane curves are free divisors
Free divisors, n=3n=3 LCT \Rightarrow strong EH Granger–Schulze result
Koszul-free divisors LCT \Leftrightarrow strong EH Narváez Macarro
Linear free divisors, n=5n=5 Example with LCT failing Not all linear free divisors satisfy LCT

6. Arithmetic and Logarithmic Comparison Theorems

Beyond the analytic and algebraic settings, LCT has deep arithmetic interpretations:

  • Faltings’ theorem (1989) compared pp-adic étale cohomology and crystalline cohomology via Fontaine–Faltings modules. This framework extends to the logarithmic setting, where log-structures on a smooth scheme Y/WY/W with a simple normal crossings divisor ZZ yield canonical isomorphisms

Hlogeˊti(UKˉ,Qp)Hlogcrisi((Y,Z)/W)[1/p].H^i_{\mathrm{log}\acute{e}t}\left(U_{\bar K},\,\mathbb{Q}_p\right) \cong H^i_{\mathrm{log}\mathrm{cris}}\left((Y,Z)/W\right)\left[1/p\right].

  • The construction and fully faithfulness of logarithmic Fontaine–Faltings modules and the log-DD-functor DlogD_{\log} provide the machinery for these comparison isomorphisms (Liu et al., 19 Apr 2025).
  • Such theorems play a central role in pp-adic Hodge theory, especially for non-proper or degenerating families, and underpin the compatibility between étale, crystalline, and de Rham cohomologies in arithmetic geometry.

7. Log-Prismatic and Prismatic Comparison

Recent advances extend LCT-type results to prismatic cohomology and its logarithmic variants:

  • For a log smooth pp-adic formal fs log scheme (X,MX)(X, M_X) over a log prism (A,I,M)(A, I, M), canonical filtered quasi-isomorphisms compare log prismatic cohomology to both derived log de Rham and log crystalline cohomology: $R\Gamma_\Prism((X, M_X)/(A, I, M))\,\widehat\otimes^L_{A}A_{\mathrm{dR}} \xrightarrow{\sim} R\Gamma_{\mathrm{dR}}((X, M_X)/A_{\mathrm{dR}})$

$R\Gamma_\Prism((X, M_X)/(A, I, M))\,\widehat\otimes^L_{A}W \xrightarrow{\sim} R\Gamma_{\mathrm{crys}}((X, M_X)/W)$

with explicit control via Nygaard filtrations and functorial spectral sequences (Binda et al., 2023).

Gysin maps and blow-up formulas for prismatic cohomology are constructed, and descent techniques ensure that LCTs in this context articulate the deep compatibility between logarithmic and non-logarithmic "motivic" theories.


The Logarithmic Comparison Theorem thus acts as a unifying principle: in all its variants, it expresses when the rich but "small" logarithmic complexes suffice to compute invariants (de Rham, étale, crystalline, or prismatic) associated to the open complement of a divisor, imposing surprisingly rigid algebraic and geometric conditions on the underlying singularities. These results connect D-module theory, Hodge and pp-adic Hodge theory, singularity theory, and the modern theory of log motives and prismatic cohomology (Castro-Jiménez et al., 2023, Rodríguez, 30 Apr 2025, Bath, 2022, Liu et al., 19 Apr 2025, Binda et al., 2023, Bath et al., 2022).

Topic to Video (Beta)

No one has generated a video about this topic yet.

Whiteboard

No one has generated a whiteboard explanation for this topic yet.

Follow Topic

Get notified by email when new papers are published related to Logarithmic Comparison Theorem (LCT).