nLab sphere





topology (point-set topology, point-free topology)

see also differential topology, algebraic topology, functional analysis and topological homotopy theory


Basic concepts

Universal constructions

Extra stuff, structure, properties


Basic statements


Analysis Theorems

topological homotopy theory

Manifolds and cobordisms



Finite-dimensional spheres


The nn-dimensional unit sphere , or simply nn-sphere, is the topological space given by the subset of the (n+1)(n+1)-dimensional Cartesian space ℝ n+1\mathbb{R}^{n+1} consisting of all points xx whose distance from the origin is 11

S n={x:ℝ n+1|β€–xβ€–=1}. S^n = \{ x: \mathbb{R}^{n+1} \;|\; \|x\| = 1 \} \,.

The nn-dimensional sphere of radius rr is

S r n={x:ℝ n+1|β€–xβ€–=r}. S^n_r = \{ x: \mathbb{R}^{n+1} \;|\; \|x\| = r \} .

Topologically, this is equivalent (homeomorphic) to the unit sphere for r>0r \gt 0, or a point for r=0r = 0.

This is naturally a smooth manifold of dimension nn, with the smooth structure induced by the standard smooth structure on ℝ n\mathbb{R}^n.

Infinite dimensional spheres

One can also talk about the infinite-dimensional sphere in an arbitrary (possibly infinite-dimensional) normed vector space VV:

S(V)={x:Vsuch thatβ€–xβ€–=1}. S(V) = \{ x: V \;\text{such that}\; {\|x\|} = 1 \} .

If a locally convex topological vector space admits a continuous linear injection into a normed vector space, this can be used to define its sphere. If not, one can still define the sphere as a quotient of the space of non-zero vectors under the scalar action of (0,∞)(0,\infty).

Homotopy theorists (e.g. tom Dieck 2008, example 8.3.7) define S ∞S^\infty as the directed colimit of the S nS^n:

S βˆ’1β†ͺS 0β†ͺS 1β†ͺS 2β†ͺβ‹―S ∞. S^{-1} \hookrightarrow S^0 \hookrightarrow S^1 \hookrightarrow S^2 \hookrightarrow \cdots S^\infty .

Note that this S ∞S^\infty is not homeomorphic to the sphere of any metrizable space as defined above, since the metrizable CW-complexes are precisely the locally finite CW-complexes (Fritsch–Piccinini 1990: 48, prop. 1.5.17), which S ∞S^\infty is not (every open nn-cell intersects all closed mm-cells with mβ‰₯nm \ge n)

In themselves, infinite-dimensional spheres provide nothing new to homotopy theory, as they are at least weakly contractible and usually contractible. However, they are a very useful source of big contractible spaces and so are often used as a starting point for making concrete models of classifying spaces.

If the vector space is a shift space, then contractibility is straightforward to prove.


Let VV be a shift space of some order. Let SVS V be its sphere (either via a norm or as the quotient of non-zero vectors). Then SVS V is contractible.


Let T:V→VT \colon V \to V be a shift map. The idea is to homotop the sphere onto the image of TT, and then down to a point.

It is simplest to start with the non-zero vectors, Vβˆ–{0}V \setminus \{0\}. As TT is injective, it restricts to a map from this space to itself which commutes with the scalar action of (0,∞)(0,\infty). Define a homotopy H:[0,1]Γ—(Vβˆ–{0})β†’Vβˆ–{0}H \colon [0,1] \times (V \setminus \{0\}) \to V \setminus \{0\} by H t(v)=(1βˆ’t)v+tTvH_t(v) = (1 - t)v + t T v. It is clear that, assuming it is well-defined, it is a homotopy from the identity to TT. To see that it is well-defined, we need to show that H t(v)H_t(v) is never zero. The only place where it could be zero would be on an eigenvector of TT, but as TT is a shift map then it has none.

As TT is a shift map, it is not surjective and so we can pick some v 0v_0 not in its image. Then we define a homotopy G:[0,1]Γ—(Vβˆ–{0})β†’Vβˆ–{0}G \colon [0,1] \times (V \setminus \{0\}) \to V \setminus \{0\} by G t(v)=(1βˆ’t)Tv+tv 0G_t(v) = (1 - t)T v + t v_0. As v 0v_0 is not in the image of TT, this is well-defined on Vβˆ–{0}V \setminus \{0\}. Combining these two homotopies results in the desired contraction of Vβˆ–{0}V \setminus \{0\}.

If VV admits a suitable function defining a spherical subset (such as a norm) then we can modify the above to a contraction of the spherical subset simply by dividing out by this function. If not, as the homotopies above all commute with the scalar action of (0,∞)(0,\infty), they descend to the definition of the sphere as the quotient of Vβˆ–{0}V \setminus \{0\}.




The topological complexity of the sphere is

(1)TC(S n)={2 ;nodd 3 ;neven TC \big( S^n \big) \;=\; \left\{ \begin{array}{ll} 2 & ; n \quad odd \\ 3 & ; n \quad even \end{array} \right.

(Farber 01, Theorem 8)

This proposition can be generalized:


The topological complexity of a product of spheres is

(2)TC((S m) n)={n+1 ;modd 2n+1 ;meven TC \big( (S^m)^n \big) \;=\; \left\{ \begin{array}{ll} n+1 & ; m \quad odd \\ 2n+1 & ; m \quad even \end{array} \right.

(Farber 01, Theorem 13)

A special case of this proposition is TC(T n)=n+1TC(T^n)=n+1 for the topological complexity of the torus.


The nn-sphere is an n n -dimensional CW complex in several ways:

  • The nn-sphere (nβ‰₯0n \ge 0) admits, for every point x 0∈S nx_0 \in S^n , a CW-structure with one 00-cell x 0x_0 and one nn-cell S nβˆ–{x 0}S^n \setminus \{ x_0 \}, by stereographic projection. (tom Dieck 2008, example 8.3.7)

  • The nn-sphere (nβ‰₯0n \ge 0) can also be constructed from the (nβˆ’1)(n-1)-sphere by attaching nn-cells (the north and south hemispheres) to the equator (nβˆ’1)(n-1)-sphere. Iteratively applying this construction starting with S βˆ’1=βˆ…S^{-1}=\varnothing yields a CW complex S nS^n with two kk-cells in each dimension k≀nk \le n, and subcomplex inclusions S nβŠ†S n+1S^n \subseteq S^{n+1} for all n∈Nn \in \mathbf{N}; the colimit of this sequence is (by definition) S ∞S^\infty. (tom Dieck 2008, example 8.3.7)

Coset space structure

As quotients of compact Lie groups


For nβˆˆβ„•n \in \mathbb{N} the n-spheres are coset spaces of orthogonal groups

S n≃O(n+1)/O(n). S^n \;\simeq\; O(n+1)/O(n) \,.

Similarly for the corresponding special orthogonal groups

S n≃SO(n+1)/SO(n) S^n \;\simeq\; SO(n+1)/SO(n)

and spin groups

S n≃Spin(n+1)/Spin(n) S^n \;\simeq\; Spin(n+1)/Spin(n)

and pin groups

S n≃Pin(n+1)/Pin(n). S^n \;\simeq\; Pin(n+1)/Pin(n) \,.

Fix a unit vector in ℝ n+1\mathbb{R}^{n+1}. Then its orbit under the defining O(n+1)O(n+1)-action on ℝ n+1\mathbb{R}^{n+1} is clearly the canonical embedding S nβ†ͺℝ n+1S^n \hookrightarrow \mathbb{R}^{n+1}. But precisely the subgroup of O(n+1)O(n+1) that consists of rotations around the axis formed by that unit vector stabilizes it, and that subgroup is isomorphic to O(n)O(n), hence S n≃O(n+1)/O(n)S^n \simeq O(n+1)/O(n).

Similarly, the analogous argument for unit spheres inside (the real vector spaces underlying) complex vector spaces, we have


For kβˆˆβ„•k \in \mathbb{N} the (2k+1)-sphere S 2k+1S^{2k+1} is the coset space of special unitary groups:

S 2k+1≃SU(k+1)/SU(k). S^{2k+1} \;\simeq\; SU(k+1)/SU(k) \,.

And still similarly, the analogous argument for unit spheres inside (the real vector spaces underlying) quaternionic vector spaces, we have


For kβˆˆβ„•k \in \mathbb{N}, kβ‰₯1k \geq 1 the (4k-1)-sphere S 4kβˆ’1S^{4k-1} is the coset space of quaternionic unitary groups:

S 4kβˆ’1≃Sp(k)/Sp(kβˆ’1). S^{4k-1} \;\simeq\; Sp(k)/Sp(k-1) \,.



The connected compact Lie groups with effective transitive actions on n-spheres are precisely (up to isomorphism) the following:

with coset spaces

SO(n)/SO(nβˆ’1) ≃S nβˆ’1 U(n)/U(nβˆ’1) ≃S 2nβˆ’1 SU(n)/SU(nβˆ’1) ≃S 2nβˆ’1 Sp(n)/Sp(nβˆ’1) ≃S 4nβˆ’1 Sp(n)β‹…SO(2)/Sp(nβˆ’1)β‹…SO(2) ≃S 4nβˆ’1 Sp(n)β‹…Sp(1)/Sp(nβˆ’1)β‹…Sp(1) ≃S 4nβˆ’1 G 2/SU(3) ≃S 6 Spin(7)/G 2 ≃S 7 Spin(9)/Spin(7) ≃S 15 \begin{aligned} SO(n)/SO(n-1) & \simeq S^{n-1} \\ U(n)/U(n-1) & \simeq S^{2n-1} \\ SU(n)/SU(n-1) & \simeq S^{2n-1} \\ Sp(n)/Sp(n-1) & \simeq S^{4n-1} \\ Sp(n)\cdot SO(2)/Sp(n-1)\cdot SO(2) & \simeq S^{4n-1} \\ Sp(n)\cdot Sp(1)/Sp(n-1)\cdot Sp(1) & \simeq S^{4n-1} \\ G_2/SU(3) & \simeq S^6 \\ Spin(7)/G_2 & \simeq S^7 \\ Spin(9)/Spin(7) & \simeq S^{15} \end{aligned}

This goes back to Montgomery & Samelson (1943), see Gray & Green (1970), p. 1-2, also Borel & Serre (1953), 17.1.


The isomorphisms in Prop. and Prop. above hold in the category of topological spaces (homeomorphisms), but in fact also in the category of smooth manifolds (diffeomorphisms) and even in the category of Riemannian manifolds (isometries).

The other coset space realizations of some n-spheres in Prop. are homeomorphisms, but not necessarily isometries (β€œsquashed spheres”). There is also a double coset space realization which is not even a diffeomorphisms (β€œexotic sphere”, the Gromoll-Meyer sphere).

For more see 7-sphere – Coset space realization.

coset space-structures on n-spheres:

S nβˆ’1≃ diffSO(n)/SO(nβˆ’1)S^{n-1} \simeq_{diff} SO(n)/SO(n-1)this Prop.
S 2nβˆ’1≃ diffSU(n)/SU(nβˆ’1)S^{2n-1} \simeq_{diff} SU(n)/SU(n-1)this Prop.
S 4nβˆ’1≃ diffSp(n)/Sp(nβˆ’1)S^{4n-1} \simeq_{diff} Sp(n)/Sp(n-1)this Prop.
S 7≃ diffSpin(7)/G 2S^7 \simeq_{diff} Spin(7)/G_2Spin(7)/Gβ‚‚ is the 7-sphere
S 7≃ diffSpin(6)/SU(3)S^7 \simeq_{diff} Spin(6)/SU(3)since Spin(6) ≃\simeq SU(4)
S 7≃ diffSpin(5)/SU(2)S^7 \simeq_{diff} Spin(5)/SU(2)since Sp(2) is Spin(5) and Sp(1) is SU(2), see Spin(5)/SU(2) is the 7-sphere
S 6≃ diffG 2/SU(3)S^6 \simeq_{diff} G_2/SU(3)Gβ‚‚/SU(3) is the 6-sphere
S 15≃ diffSpin(9)/Spin(7)S^15 \simeq_{diff} Spin(9)/Spin(7)Spin(9)/Spin(7) is the 15-sphere

see also Spin(8)-subgroups and reductions

homotopy fibers of homotopy pullbacks of classifying spaces:

(from FSS 19, 3.4)

As quotients of Lorentz groups

If one drops the assumption of compactness, then there are further coset space realizations of nn-spheres, notably as quotients of Lorentz groups by parabolic subgroups: celestial spheres, e.g.: Toller (2003, p. 18), Varlamov (2006, p. 6), Math.SE:a/4092474.

Spin structure


The nn-sphere, for each nβˆˆβ„•n \in \mathbb{N}, carries a canonical spin structure, induced from its coset space-realization S n≃Spin(n+1)/Spin(n)S^n \simeq Spin(n+1)/Spin(n) (above), as a special case of the canonical HH-structure on G/HG/H (this example).

Other ways to see this:

  • Nikolai Nowaczyk, Theorem A.6.6 in: Dirac Eigenvalues of higher Multiplicity, Regensburg 2015 (arXiv:1501.04045)

  • S. Gutt, Killing spinors on spheres and projective spaces, p. 238-248 in: A. Trautman, G. Furlan (eds.) Spinors in Geometry and Physics – Trieste 11-13 September 1986, World Scientific 1988 (doi:10.1142/9789814541510, GBooks, p. 243)


Branched covers

Every nn-dimensional PL manifold admits a branched covering of the n-sphere (Alexander 20).

By the Riemann existence theorem, every connected compact Riemann surface admits the structure of a branched cover by a holomorphic function to the Riemann sphere. See there at branched cover of the Riemann sphere.

graphics grabbed from Chamseddine-Connes-Mukhanov 14, Figure 1, Connes 17, Figure 11

For 3-manifolds branched covering the 3-sphere see (Montesinos 74).

All PL 4-manifolds are simple branched covers of the 4-sphere (Piergallini 95, Iori-Piergallini 02).

But the n-torus for nβ‰₯3n \geq 3 is not a cyclic branched over of the n-sphere (Hirsch-Neumann 75)

Iterated loop spaces


(rational cohomology of iterated loop space of the 2k-sphere)


1≀D<n=2kβˆˆβ„• 1 \leq D \lt n = 2k \in \mathbb{N}

(hence two positive natural numbers, one of them required to be even and the other required to be smaller than the first) and consider the D-fold loop space Ξ© DS n\Omega^D S^n of the n-sphere.

Its rational cohomology ring is the free graded-commutative algebra over β„š\mathbb{Q} on one generator e nβˆ’De_{n-D} of degree nβˆ’Dn - D and one generator a 2nβˆ’Dβˆ’1a_{2n - D - 1} of degree 2nβˆ’Dβˆ’12n - D - 1:

H β€’(Ξ© DS n,β„š)β‰ƒβ„š[e nβˆ’D,a 2nβˆ’Dβˆ’1]. H^\bullet \big( \Omega^D S^n , \mathbb{Q} \big) \;\simeq\; \mathbb{Q}\big[ e_{n - D}, a_{2n - D - 1} \big] \,.

(Kallel-Sjerve 99, Prop. 4.10)




Axiomatization of the homotopy type of the 1-sphere (the circle) and the 2-sphere, as higher inductive types, is in

Visualization of the idea of the construction for the 2-sphere is in

Group actions on spheres

Discussion of free group actions on spheres by finite groups includes

The subgroups of SO(8) which act freely on S 7S^7 have been classified in

  • J. A. Wolf, Spaces of constant curvature, Publish or Perish, Boston, Third ed., 1974

and lifted to actions of Spin(8) in

  • Sunil Gadhia, Supersymmetric quotients of M-theory and supergravity backgrounds, PhD thesis, School of Mathematics, University of Edinburgh, 2007 (spire:1393845)

Discussion of transitive actions on nn-spheres by compact Lie groups:

Further discussion of these actions is in

where they are related to the black M2-brane BPS-solutions of 11-dimensional supergravity at ADE-singularities.

See also the ADE classification of such actions on the 7-sphere (as discussed there)

Discussion of actions of Lorentz groups on celestial spheres:

Geometric structures on spheres

Coset space structures on spheres:

The following to be handled with care:

Embeddings of spheres

The (isotopy class of an) embedding of a circle (1-sphere) into the 3-sphere is a knot. Discussion of embeddings of spheres of more general dimensions into each other:

  • AndrΓ© Haefliger, Differentiable Embeddings of S nS^n in S n+qS^{n+q} for q>2q \gt 2, Annals of Mathematics Second Series, Vol. 83, No. 3 (May, 1966), pp. 402-436 (jstor:1970475)

Iterated loop spaces

Topological complexity

On topological complexity of spheres and products of spheres (including tori as special case):

Last revised on June 3, 2024 at 01:19:41. See the history of this page for a list of all contributions to it.