Why This Matters
Spectral sequences are the computational workhorses of algebraic topologyโthey transform impossibly complex calculations into systematic, page-by-page approximations. When you're faced with computing the homology of a fibration, the stable homotopy groups of spheres, or the cohomology of a sheaf, spectral sequences provide the machinery to break these problems into manageable pieces. You're being tested on your ability to recognize which spectral sequence applies to which situation, understand how filtrations give rise to successive approximations, and trace differentials through the pages until convergence.
Don't just memorize the names and E2โ-pagesโknow what structural input each spectral sequence requires and what it computes. The key concepts here are fibrations and their associated long exact sequences, derived functors and their composition, filtrations on chain complexes, and the passage from algebra to topology. When an exam asks you to compute something, your first question should be: what's the underlying structure (fibration? group extension? sheaf?) and which spectral sequence exploits it?
Fibration-Based Spectral Sequences
These spectral sequences arise from fibrations FโEโB and systematically relate the homology or cohomology of the total space to that of the base and fiber. The key insight is that a fibration provides a filtration of the total space by preimages of skeleta of the base.
Serre Spectral Sequence
- Computes Hโโ(E) from Hโโ(B) and Hโโ(F)โthe E2โ-page is E2p,qโ=Hpโ(B;Hqโ(F)), assuming ฯ1โ(B) acts trivially on Hโโ(F)
- Differentials encode the twisting of the fibrationโdrโ:Erp,qโโErpโr,q+rโ1โ captures how the fiber "varies" over the base
- The workhorse for computing homology of Lie groups and classifying spacesโessential for results like Hโ(K(Z,n)) and the cohomology of BU(n)
Eilenberg-Moore Spectral Sequence
- Handles pullbacks of fibrationsโcomputes Hโโ(EรBโEโฒ) when you know the cohomology of E, Eโฒ, and B
- E2โ-page involves Tor over Hโ(B)โspecifically E2โ=TorHโ(B)โ(Hโ(E),Hโ(Eโฒ))
- Critical for loop space calculationsโsince ฮฉBโโรBโPB, this spectral sequence computes Hโโ(ฮฉB) from Hโ(B)
Compare: Serre vs. Eilenberg-Mooreโboth start with fibrations, but Serre computes the total space from base and fiber, while Eilenberg-Moore computes fiber products. If an FRQ asks about loop spaces, Eilenberg-Moore is typically your tool; for computing Hโ(BG), reach for Serre.
Stable Homotopy Spectral Sequences
These operate in the stable homotopy category, where spaces are replaced by spectra and ordinary homology gives way to generalized homology theories. The algebraic input involves Ext or Tor over structured ring spectra or their homotopy groups.
Adams Spectral Sequence
- Computes stable homotopy groups ฯโsโ(X)โthe E2โ-page is E2s,tโ=ExtAs,tโ(Hโ(X;Fpโ),Fpโ) where A is the Steenrod algebra
- Convergence is to p-completed homotopyโyou compute one prime at a time, then assemble via arithmetic
- The primary tool for ฯโsโ(S0)โmost of what we know about stable stems comes from Adams spectral sequence calculations
May Spectral Sequence
- Computes the E2โ-page of the Adams spectral sequenceโbreaks the Ext calculation over A into smaller pieces using the filtration by powers of the augmentation ideal
- E1โ-page involves Ext over associated gradedโexploits the fact that gr(A) is a primitively generated Hopf algebra
- Essential for practical Adams computationsโwithout May, computing ExtAโ directly would be intractable beyond low degrees
Compare: Adams vs. MayโAdams goes from algebra (Ext over Steenrod) to topology (stable homotopy), while May is purely algebraic, computing the input to Adams. Think of May as preprocessing for Adams.
Generalized Cohomology Spectral Sequences
These spectral sequences connect ordinary cohomology to generalized cohomology theories like K-theory or cobordism. The filtration typically comes from the skeletal filtration of a CW complex.
Atiyah-Hirzebruch Spectral Sequence
- Computes Eโ(X) from Hโ(X) and Eโ(pt)โfor any generalized cohomology theory Eโ, the E2โ-page is E2p,qโ=Hp(X;Eq(pt))
- Differentials measure the failure of Eโ to be ordinary cohomologyโfor K-theory, the first nontrivial differential is related to Steenrod operations
- The bridge between singular cohomology and K-theory/cobordismโlets you leverage known Hโ(X) calculations to get at Kโ(X) or MUโ(X)
Compare: Atiyah-Hirzebruch vs. Serreโboth have E2โ-pages that are "cohomology with coefficients," but Serre's coefficients are Hโโ(F) (from a fiber), while Atiyah-Hirzebruch's are Eโ(pt) (from a cohomology theory). Serre is geometric; Atiyah-Hirzebruch is about changing your cohomology theory.
Sheaf-Theoretic Spectral Sequences
These arise from the derived functor perspective on sheaf cohomology and are fundamental in algebraic geometry. The key principle is that composing left-exact functors leads to a spectral sequence relating their derived functors.
Leray Spectral Sequence
- Computes Hโ(X;F) via a map f:XโYโthe E2โ-page is E2p,qโ=Hp(Y;RqfโโF) where Rqfโโ is the higher direct image
- Generalizes Serre to the sheaf settingโwhen f is a fibration and F is constant, you recover Serre
- The tool for computing cohomology of fiber bundles in algebraic geometryโessential when the base has interesting sheaf cohomology
Grothendieck Spectral Sequence
- Relates derived functors of a composition GโFโif F sends injectives to G-acyclics, then E2p,qโ=RpG(RqF(โ)) converges to Rp+q(GโF)(โ)
- Leray is a special caseโtake F=fโโ and G=ฮ(Y;โ)
- The abstract machine behind most spectral sequences in algebraโonce you recognize a composition of functors, Grothendieck tells you there's a spectral sequence
Compare: Leray vs. GrothendieckโLeray is the geometric incarnation (maps of spaces, sheaves), while Grothendieck is the abstract categorical version. Know Grothendieck for the general principle; use Leray for concrete geometric computations.
Group Cohomology Spectral Sequences
These compute the cohomology of groups by exploiting exact sequences of groups. A short exact sequence 1โNโGโQโ1 gives a "fibration" of classifying spaces BNโBGโBQ.
Hochschild-Serre Spectral Sequence
- Computes Hโ(G;M) from a normal subgroup NโGโthe E2โ-page is E2p,qโ=Hp(G/N;Hq(N;M))
- Requires understanding the G/N-action on Hโ(N;M)โthis action encodes how conjugation in G affects N-cohomology
- The group-theoretic analog of SerreโBNโBGโB(G/N) is a fibration, and this is its Serre spectral sequence
Lyndon-Hochschild-Serre Spectral Sequence
- Same as Hochschild-Serre, different nameโsome authors distinguish by requiring M to be a trivial module, but the terms are often used interchangeably
- Particularly useful for computing Hโ(G;Z)โthe case of trivial integer coefficients appears constantly in geometric group theory
- First step in computing cohomology of solvable groupsโiterate over a composition series
Compare: Hochschild-Serre vs. Serreโthey're the same spectral sequence in different clothing. Hochschild-Serre is stated algebraically for group cohomology; Serre is stated topologically for fibrations. Recognizing this equivalence via BG is a key conceptual test.
Torsion and Coefficient Spectral Sequences
These spectral sequences arise from considering how cohomology changes with different coefficient groups. The Bockstein homomorphism ฮฒ:Hn(X;Z/p)โHn+1(X;Z/p) detects p-torsion in integral cohomology.
Bockstein Spectral Sequence
- Computes Hโ(X;Z)/torsionโZ/p from Hโ(X;Z/p)โthe E1โ-page is mod p cohomology, and d1โ=ฮฒ is the Bockstein
- Eโโ detects the "p-torsion-free part"โelements surviving to Eโโ lift to integral classes
- Key for understanding the Universal Coefficient Theorem dynamicallyโthe spectral sequence refines the UCT exact sequence
Compare: Bockstein vs. Adamsโboth involve mod p cohomology, but Bockstein stays in ordinary cohomology (detecting integral torsion), while Adams uses mod p cohomology as input to compute stable homotopy. Bockstein is about coefficients; Adams is about changing from cohomology to homotopy.
Quick Reference Table
|
| Computing homology of fibrations | Serre, Eilenberg-Moore |
| Stable homotopy groups | Adams, May |
| Generalized cohomology from ordinary | Atiyah-Hirzebruch |
| Sheaf cohomology via maps | Leray, Grothendieck |
| Group cohomology from extensions | Hochschild-Serre, Lyndon-Hochschild-Serre |
| Detecting torsion in coefficients | Bockstein |
| Loop space homology | Eilenberg-Moore |
| Derived functor composition | Grothendieck |
Self-Check Questions
-
Both the Serre spectral sequence and the Hochschild-Serre spectral sequence have E2โ-pages of the form "cohomology of the base with coefficients in cohomology of the fiber." What is the precise relationship between these two spectral sequences, and how does the classifying space functor mediate it?
-
You need to compute Kโ(X) for a CW complex X whose ordinary cohomology Hโ(X;Z) you know. Which spectral sequence do you use, and what information beyond Hโ(X) do you need?
-
Compare the Grothendieck spectral sequence and the Leray spectral sequence. Which is more general? Under what hypotheses does Grothendieck specialize to Leray?
-
An FRQ asks you to compute ExtAs,tโ(F2โ,F2โ) for small s and t. Which spectral sequence helps organize this computation, and what is its E1โ-page?
-
You're given a group extension 1โZ/pโGโZ/pโ1 and asked to compute Hโ(G;Fpโ). Which spectral sequence applies, what is the E2โ-page, and what additional information determines the differentials?