Chromatic Homotopy Theory

⚡ Key Takeaways

Chromatic homotopy theory organizes the stable homotopy category by "frequency" (height), using the algebraic geometry of formal group laws to stratify phenomena into chromatic layers. One-line statement: the stable homotopy category is filtered by the heights of formal groups, with the $n$ -th layer governed by Morava $K$ -theory $K(n)$ and Morava $E$ -theory $E_n$ , and reassembled by the chromatic convergence theorem. Where it applies: computations of stable homotopy groups of spheres, the structure of the $K(n)$ -local categories, periodic phenomena ( $v_n$ -self-maps), and connections to number theory and topological modular forms. Who/when: Quillen (1969), Morava (1970s), Miller–Ravenel–Wilson, Devinatz–Hopkins–Smith (nilpotence/periodicity, 1980s–90s), Hopkins–Miller, Ravenel's conjectures.

💡 The Core Insight

Homotopy theory has a "spectrum" in the algebro-geometric sense. The prime ideals of the ring of operations are organized by the heights of formal groups. Just as a scheme is studied locally at primes, the sphere spectrum is studied locally at the "chromatic primes" $K(n)$ . The thick subcategory theorem says these are all the primes (of finite spectra), and the chromatic tower

S^0 \to \cdots \to L_n S^0 \to L_{n-1} S^0 \to \cdots \to L_0 S^0

recovers $S^0$ as the inverse limit. Algebra (formal groups over $\mathbb{F}_p$ ) controls topology (periodic families in $\pi_* S^0$ ).

DefinitionFormal group law and height

A (one-dimensional, commutative) formal group law over a commutative ring $R$ is a power series $F(x,y) \in R[[x,y]]$ satisfying $F(x,0)=x$ , $F(x,y)=F(y,x)$ , and associativity $F(F(x,y),z)=F(x,F(y,z))$ . Over a field $k$ of characteristic $p>0$ , the $p$ -series is

1

\quad [p]_F(x) = \underbrace{x +_F x +_F \cdots +_F x}_{p} = v_n x^{p^n} + (\text{higher order}), \quad v_n \neq 0,

Derivation hint: iterate the formal sum $p$ times; over $\mathbb{F}_p$ the leading term vanishes until exponent $p^n$ , where $n$ is the height — invariant under isomorphism because the leading exponent of $[p]_F$ is preserved by coordinate changes. The integer $n$ is the height; if $[p]_F(x)=0$ formally (i.e. $F$ is additive after base change), the height is $\infty$ .

TheoremQuillen, 1969

The complex cobordism spectrum $MU$ carries the universal formal group law. Its coefficient ring is the Lazard ring:

2

\quad \pi_* MU \cong L \cong \mathbb{Z}[x_1, x_2, \ldots], \qquad |x_i| = 2i,

Derivation hint: Lazard's theorem identifies the universal FGL ring as a polynomial ring; Milnor–Novikov computes $MU_*$ as polynomial on classes in even degrees, and Quillen matches the universal FGL to the geometric one coming from the orientation $MU^2(\mathbb{CP}^\infty)$ . Consequently $\mathrm{Spec}\,\pi_* MU \to \mathcal{M}_{FG}$ presents the moduli stack of formal groups, and $MU_*MU$ corepresents isomorphisms of formal group laws (the Hopf algebroid $(MU_*, MU_*MU)$ ).

DefinitionMorava

K

-theory and

E

-theory

Fix a prime $p$ and $0 \le n \le \infty$ . Morava $K$ -theory $K(n)$ is a complex-oriented ring spectrum with

3

\quad K(n)_* = \mathbb{F}_p[v_n^{\pm 1}], \qquad |v_n| = 2(p^n - 1),

Derivation hint: $K(n)$ is built from $BP$ by killing $p, v_1, \ldots, \widehat{v_n}, \ldots$ and inverting $v_n$ ; its FGL is the Honda formal group of height $n$ over $\mathbb{F}_p$ with $[p](x)=v_n x^{p^n}$ . with conventions $K(0)=H\mathbb{Q}$ and $K(\infty)=H\mathbb{F}_p$ . Morava $E$ -theory (Lubin–Tate spectrum) $E_n$ has

4

\quad (E_n)_* = W(\mathbb{F}_{p^n})[[u_1, \ldots, u_{n-1}]][u^{\pm 1}], \qquad |u| = -2, \ |u_i| = 0,

Derivation hint: $(E_n)_0 = W(\mathbb{F}_{p^n})[[u_1,\ldots,u_{n-1}]]$ is the Lubin–Tate ring carrying the universal deformation of the height- $n$ Honda formal group; the Landweber exact functor theorem produces the spectrum, and Goerss–Hopkins–Miller give it an $E_\infty$ structure. $E_n$ is even periodic with residue field $\mathbb{F}_{p^n}$ and $E_n/\mathfrak{m} \simeq K(n)$ (up to extending scalars).

DefinitionChromatic localization

Let $L_n = L_{E(n)}$ denote Bousfield localization at the Johnson–Wilson theory $E(n)$ (equivalently at $K(0)\vee\cdots\vee K(n)$ ). The chromatic tower of a spectrum $X$ is

5

\quad X \to \cdots \to L_n X \to L_{n-1} X \to \cdots \to L_0 X = X_\mathbb{Q},

Derivation hint: $E(n)$ -localization is smashing on finite spectra; the natural transformations $L_n \to L_{n-1}$ come from $E(n)\to E(n-1)$ via the inclusion of detected heights, and the homotopy fiber is the monochromatic layer $M_n X$ . The monochromatic layer $M_n X = \mathrm{fib}(L_n X \to L_{n-1} X)$ is equivalent, after $K(n)$ -localization data, to a $K(n)$ -local object. The fracture squares assemble these.

Figure

The chromatic tower and its monochromatic layers

```svg

```

TheoremNilpotence Theorem — Devinatz–Hopkins–Smith, 1988

A map $f: \Sigma^k X \to X$ of finite $p$ -local spectra is nilpotent (some smash power is null) if and only if $MU_* f$ (equivalently $K(n)_* f$ for all $n$ ) is nilpotent. More generally, an element $\alpha \in \pi_* R$ of a ring spectrum is nilpotent iff its Hurewicz image in $MU_*R$ is nilpotent:

6

\quad \alpha \text{ nilpotent in } \pi_*R \iff h(\alpha) \text{ nilpotent in } MU_*R.

Derivation hint: the proof reduces detection to $MU$ via a sequence of approximations ( $X(n)$ spectra interpolating $MU$ and the sphere) and a careful self-map/transfer argument; the Morava $K$ -theories arise as the only "field-like" homology theories detecting non-nilpotence.

TheoremThick Subcategory Theorem

Every thick subcategory of the category of finite $p$ -local spectra is one of the chains

7

\quad \mathcal{C}_0 \supseteq \mathcal{C}_1 \supseteq \mathcal{C}_2 \supseteq \cdots, \qquad \mathcal{C}_n = \{ X \text{ finite} : K(m)_*X = 0 \text{ for } m < n \},

Derivation hint: a thick subcategory is closed under cofibers, retracts, and suspension; the nilpotence theorem forces the $K(n)$ -acyclicity to be the only invariant, and the existence of finite type- $n$ complexes (via $v_n$ -self-maps) shows the chain is strict. A finite spectrum has type $n$ if $K(m)_* X = 0$ for $m<n$ and $K(n)_* X \neq 0$ .

TheoremPeriodicity Theorem — Hopkins–Smith

Every finite type- $n$ complex $X$ admits a $v_n$ -self-map: a map $v: \Sigma^d X \to X$ inducing multiplication by $v_n^k$ on $K(n)_* X$ and zero on $K(m)_*X$ for $m\neq n$ . Such a map is unique up to taking iterates and is asymptotically central.

8

\quad K(n)_*(v) = v_n^k \cdot (\text{unit}), \qquad K(m)_*(v) = 0 \ (m \neq n).

Derivation hint: construct the self-map by an Adams–Novikov obstruction argument; existence on type- $n$ complexes follows from the vanishing line in the $K(n)$ -based Adams spectral sequence, and uniqueness from the nilpotence theorem applied to differences of self-maps.

ProofSketch of the chromatic stratification

Step 1 (Algebraic geometry of $\mathcal{M}_{FG}$ ). By Quillen's theorem [2], $MU_*$ -homology of a spectrum is a quasicoherent sheaf on $\mathcal{M}_{FG}$ . The stack is stratified in characteristic $p$ by height: the height- $\geq n$ locus is a closed substack with open complement of height $<n$ .

Step 2 (Landweber exactness). The Landweber exact functor theorem says a graded $MU_*$ -module $M$ that is "Landweber exact" (the $v_i$ form a regular sequence) yields a homology theory $X \mapsto M \otimes_{MU_*} MU_* X$ . The Johnson–Wilson theories $E(n)$ and Morava $E$ -theories $E_n$ arise this way.

Step 3 (Bousfield localization). Localizing at $E(n)$ kills the height- $>n$ phenomena. The natural map $L_n S^0 \to L_{n-1} S^0$ has fiber $M_n S^0$ , supported on the height- $n$ stratum.

Step 4 (Monochromatic = $K(n)$ -local data). The functor $M_n$ and $K(n)$ -localization $L_{K(n)}$ are equivalent on the relevant subcategory; in particular $M_n X \simeq \mathrm{fib}(L_n X \to L_{n-1}X)$ and the monochromatic category is equivalent to the $K(n)$ -local category.

Step 5 (Reassembly). The chromatic fracture square

9

\quad \begin{array}{ccc} L_n X & \to & L_{K(n)} X \\ \downarrow & & \downarrow \\ L_{n-1} X & \to & L_{n-1} L_{K(n)} X \end{array} \quad \text{is homotopy cartesian},

Derivation hint: this is the arithmetic-fracture analogue; it follows because $L_{n-1}$ and $M_n$ split $L_n$ , and $M_n X$ is determined by its $K(n)$ -localization. reconstructs $L_n X$ from the lower tower and the $K(n)$ -local piece. Iterating and taking limits gives convergence (Step 6).

∎

TheoremChromatic Convergence — Hopkins–Ravenel, 1992

For a finite $p$ -local spectrum $X$ , the chromatic tower converges:

10

\quad X_{(p)} \xrightarrow{\ \simeq\ } \operatorname*{holim}_n L_n X.

Derivation hint: bound the connectivity of the fibers of $X \to L_n X$ using the Hopkins–Ravenel "smash product theorem" ( $L_n$ is smashing) and a vanishing-line estimate showing the fibers become increasingly connected; the homotopy limit then recovers the $p$ -local spectrum.

TheoremDevinatz–Hopkins,

K(n)

-local sphere

The $K(n)$ -local sphere is the homotopy fixed points of Morava $E$ -theory under the (extended) Morava stabilizer group $\mathbb{G}_n = \mathrm{Aut}(\mathbb{F}_{p^n}, \Gamma_n) \rtimes \mathrm{Gal}(\mathbb{F}_{p^n}/\mathbb{F}_p)$ :

11

\quad L_{K(n)} S^0 \simeq E_n^{h\mathbb{G}_n}, \qquad \mathbb{G}_n = \mathcal{O}_D^\times \rtimes \mathrm{Gal},

Derivation hint: $\mathbb{G}_n$ acts on $E_n$ via automorphisms of the universal deformation (Lubin–Tate theory); descent along the pro-Galois $K(n)$ -local extension $S^0 \to E_n$ identifies the sphere with the fixed points, computed by a homotopy fixed point spectral sequence with $E_2 = H^*_c(\mathbb{G}_n; (E_n)_*)$ . Here $\mathcal{O}_D$ is the maximal order in the central division algebra over $\mathbb{Q}_p$ of invariant $1/n$ .

mathematician · Jack Morava | b. 1944

Jack Morava, working in the early-to-mid 1970s, recognized that the height stratification of formal groups should organize stable homotopy theory. Drawing on Quillen's identification of $MU$ with the universal formal group and on Lubin–Tate deformation theory from arithmetic, Morava introduced the spectra and stabilizer groups now bearing his name. His insight — that the automorphism groups of formal groups of height $n$ (the Morava stabilizer groups) act as "Galois groups" of the height- $n$ stratum — became the conceptual backbone of the field, later made precise by Devinatz–Hopkins. Though much of his early work circulated as preprints and lectures rather than published papers, his vision shaped Ravenel's conjectures and the entire chromatic program.

🧠 Intuition

Think of white light passing through a prism. The stable homotopy category is "white light"; the Morava $K$ -theories $K(n)$ are the monochromatic colors (frequencies), indexed by height $n=0,1,2,\ldots$ — hence "chromatic." Height $0$ is rational (the slowest, coarsest information); higher heights see finer, faster periodic phenomena. The chromatic tower is the prism: it separates $S^0$ into its frequency components $M_n S^0$ , and chromatic convergence says you can recombine all the colors to get the original $p$ -local light back. A second picture: $\mathcal{M}_{FG}$ is like $\mathrm{Spec}\,\mathbb{Z}$ , where heights play the role of primes. Localizing at $K(n)$ is like working $p$ -adically — you zoom into one stratum, where a single periodicity operator $v_n$ governs everything, exactly as a single prime governs $\mathbb{Z}_p$ . The Morava stabilizer group $\mathbb{G}_n$ is the "Galois group" of this completed local field, and the homotopy fixed point spectral sequence is the topological analogue of Galois descent.

FA

Frequency analogy

A physicist may read the chromatic filtration as a frequency decomposition (Fourier-like) of homotopy theory. Height $n$ behaves like an energy scale: low height is the infrared (rational, long-wavelength) regime, high height the ultraviolet. The $v_n$ -periodicities are like resonant modes — once you tune into a height, the spectrum becomes periodic with period $|v_n| = 2(p^n-1)$ , much as a cavity supports discrete standing waves. Renormalization-group intuition is apt: $L_n$ "integrates out" the high-height (UV) degrees of freedom, leaving an effective theory at energy scale $n$ .

ExampleHeight 1: the image of

J

and

K

-theory

Given: $p$ odd, height $n=1$ . Morava $E$ -theory $E_1$ is $p$ -complete complex $K$ -theory $KU_p$ . The Morava stabilizer group is $\mathbb{G}_1 = \mathbb{Z}_p^\times$ , acting on $\pi_0 KU_p = \mathbb{Z}_p$ via the Adams operations.

Steps: 1. $L_{K(1)} S^0 \simeq KU_p^{h\mathbb{Z}_p^\times}$ by [11]. 2. The homotopy fixed point spectral sequence has $E_2 = H^*_c(\mathbb{Z}_p^\times; \pi_* KU_p)$ . Since $\mathbb{Z}_p^\times \cong \mathbb{Z}/(p-1) \times \mathbb{Z}_p$ , continuous cohomology is concentrated in degrees $0,1$ . 3. One computes $\pi_* L_{K(1)} S^0$ in low degrees: $\pi_0 = \mathbb{Z}_p$ , $\pi_{-1} = \mathbb{Z}_p$ , and the $v_1$ -periodic families. 4. The torsion in the $K(1)$ -local sphere realizes the image of the $J$ -homomorphism: $\mathrm{im}\, J$ in $\pi_{2k-1}^s$ has order the denominator of $B_k/2k$ (von Staudt–Clausen / Adams).

Boxed answer:

\boxed{\ \pi_{2k-1} L_{K(1)} S^0 \supseteq \mathbb{Z}/p^{v_p(k)+1} \text{ (the }v_1\text{-periodic im }J), \quad |v_1| = 2(p-1).\ }

The first nontrivial element is $\alpha_1 \in \pi_{2p-3}^s$ of order $p$ .

ExampleComputing height of a known formal group

Given: the formal group law of a supersingular elliptic curve $E$ over $\overline{\mathbb{F}}_p$ versus an ordinary one.

Steps: 1. For an ordinary elliptic curve, the formal group $\hat{E}$ has $[p](x) = u\, x^p + \cdots$ with $u\neq 0$ , so leading exponent $p^1$ — height $1$ . 2. For a supersingular curve, $[p](x)$ has no $x^p$ term; the leading exponent is $p^2$ — height $2$ . 3. This is why topological modular forms $\mathrm{tmf}$ realize chromatic height $\le 2$ : elliptic curves contribute heights $0,1,2$ through their formal groups. 4. Concretely, take $E: y^2 = x^3 + x$ over $\mathbb{F}_3$ (supersingular since $p=3 \equiv 3 \pmod 4$ ). Then $\hat E$ has height $2$ .

Boxed answer:

\boxed{\ \mathrm{ht}(\hat E) = \begin{cases} 1 & E \text{ ordinary} \\ 2 & E \text{ supersingular}\end{cases}, \qquad \mathrm{tmf} \text{ sees heights } 0,1,2.\ }

⚠️ Common mistake

" $L_n$ -localization equals $K(n)$ -localization." False. $L_n = L_{E(n)} = L_{K(0)\vee\cdots\vee K(n)}$ collects all heights up to $n$ , whereas $L_{K(n)}$ isolates the single height $n$ . They differ even for the sphere: $L_n S^0$ fits in a fracture square [9] relating it to $L_{n-1}S^0$ and $L_{K(n)}S^0$ . Also, $L_n$ is smashing (Hopkins–Ravenel) but $L_{K(n)}$ is not smashing in general.

⚠️ Common mistake

"Morava $K$ -theory is a field, so it behaves like $H\mathbb{F}_p$ for all purposes." $K(n)$ is a graded field (every module is free) and satisfies a Künneth isomorphism, but it is not an $E_\infty$ ring and is only homotopy-commutative for $p$ odd (at $p=2$ it is not even homotopy commutative). It does not see additive/unstable structure the way ordinary homology does.

CounterexampleChromatic convergence can fail for infinite spectra

The chromatic convergence theorem [10] is stated for finite $p$ -local spectra. For non-finite spectra it can fail: e.g. for $X = \bigvee_n \Sigma^{?} F(n)$ assembled from finite complexes of unbounded type, the natural map $X_{(p)} \to \operatorname{holim} L_n X$ need not be an equivalence, because the connectivity estimates of the fibers $\mathrm{fib}(X \to L_n X)$ rely on finiteness. Likewise, $\operatorname{holim}_n L_n(H\mathbb{Z})$ is not $H\mathbb{Z}_{(p)}$ — Eilenberg–MacLane spectra are dissonant (chromatically invisible above height $0$ ).

connection

Chromatic homotopy theory sits at a crossroads. It draws its algebraic input from Formal Group Laws and Lubin-Tate Theory (deformation spaces and the stabilizer group), and its categorical framework from Bousfield Localization and Stable Homotopy Category. The Adams-Novikov Spectral Sequence is the principal computational engine, with $E_2 = \mathrm{Ext}_{BP_*BP}(BP_*, BP_*)$ organized by the chromatic spectral sequence into height layers. At height $\le 2$ the theory meets Topological Modular Forms and the arithmetic of elliptic curves; the program continues into Topological Automorphic Forms at higher heights via Shimura varieties. The Morava stabilizer group connects to Local Class Field Theory (central division algebras over $\mathbb{Q}_p$ ), and the redshift conjecture links chromatic height to Algebraic K-Theory (raising height by one).

🔬 Where this appears

⏳ History

- 1958–1960: Atiyah–Hirzebruch and Bott establish

K

-theory and Bott periodicity; Adams computes the image of

J

, the first height-1 periodicity, using Adams operations.

- 1969: Quillen identifies

MU

with the universal formal group law (the foundational dictionary between topology and formal groups).

- 1970: Lazard's theorem and Landweber's exactness criterion provide the algebraic backbone; Johnson–Wilson construct

E(n)

.

- 1970s: Morava develops his "Morava modules," stabilizer groups, and the height stratification — largely unpublished but enormously influential.

- 1973–1977: Miller–Ravenel–Wilson compute the chromatic spectral sequence and the Greek-letter (alpha/beta) families via the algebraic chromatic resolution of

BP_*BP

.

- 1979: Ravenel formulates his conjectures (nilpotence, periodicity, thick subcategory, smashing, class invariance) organizing the entire field.

- 1984: Bousfield's localization framework matures, providing

L_n

and

L_{K(n)}

.

- 1988: Devinatz–Hopkins–Smith prove the Nilpotence Theorem; Hopkins–Smith prove the Periodicity and Thick Subcategory theorems — most of Ravenel's conjectures resolved.

- 1992–1994: Hopkins–Ravenel prove the Chromatic Convergence and Smash Product theorems; Goerss–Hopkins–Miller establish the

E_\infty

Lubin–Tate spectra and Devinatz–Hopkins identify

L_{K(n)}S^0 = E_n^{h\mathbb{G}_n}

.

- 2000s–present: Construction of

\mathrm{tmf}

(Hopkins–Miller–Lurie), topological automorphic forms (Behrens–Lawson), the Hill–Hopkins–Ravenel Kervaire solution (2009), and recent proofs of redshift (Hahn–Wilson; Burklund–Schlank–Yuan) and ambidexterity (Hopkins–Lurie).

open-problem

Major open questions and ongoing programs: 1. Telescope conjecture (height $\ge 2$ ). Ravenel's telescope conjecture asserts $L_n^f = L_n$ , i.e. the finite ( $v_n$ -telescope) localization agrees with $E(n)$ -localization. Recently disproved at height $2$ , $p\ge 5$ by Burklund–Hahn–Levy–Schlank (2023) using $K$ -theory and ambidexterity — a landmark result settling a long-standing problem in the negative. 2. Explicit $\pi_* L_{K(n)} S^0$ for $n\ge 3$ . The homotopy fixed point spectral sequence with $E_2 = H^*_c(\mathbb{G}_n; (E_n)_*)$ is computed only in low heights and small primes; the cohomology of $\mathbb{G}_n$ is largely intractable for $n\ge 3$ . 3. Higher chromatic redshift in full generality. While established in important cases, a complete structural theory of how $K$ -theory shifts chromatic height remains under development. 4. Geometric/physical meaning of $E_n^{h\mathbb{G}_n}$ . Connections to TQFT, $p$ -adic geometry (prismatic cohomology), and Lurie's ambidexterity suggest deep but incompletely understood structure. 5. Chromatic splitting conjecture (Hopkins). Predicts the precise structure of $L_{n-1}L_{K(n)}S^0$ ; known in low heights, open in general, and central to understanding the fracture squares.

Formal Statement

Why It Holds / Derivation

Building Intuition

Worked Examples

Common Misconceptions

Connections

Real-World Applications

Historical Development

Subtleties & Limits