Domenico Fiorenza representations for sl(2,C)

Representations of 𝔰𝔩 2()\mathfrak{sl}_2(\mathbb{C})

The Lie algebra 𝔰𝔩 2()\mathfrak{sl}_2(\mathbb{C}) can be introduced as the tangent space to the identity for SL 2()SL_2(\mathbb{C}). By an easy calculation it results that 𝔰𝔩 2()\mathfrak{sl}_2(\mathbb{C}) is three dimensional. Choose a basis h,e,fh,e,f as follows:

h=(1 0 0 1)e=(0 1 0 0)f=(0 0 1 0)h= \left( \begin{matrix}1 && 0 \\ 0 && -1 \end{matrix} \right) \quad e= \left( \begin{matrix}0 && 1 \\ 0&& 0 \end{matrix} \right) \quad f= \left( \begin{matrix}0 && 0 \\ 1&& 0 \end{matrix} \right)

with commutation rules:

[h,e]=2e,[h,f]=2f,[e,f]=h. [h,e]=2e, \quad [h,f]=-2f, \quad [e,f]=h.

First we want to identify a typical irreducible representation for 𝔰𝔩 2()\mathfrak{sl}_2(\mathbb{C}). We resume its properties in a theorem.

Theorem. For each positive integer mm there exists up to isomorphism a unique irreducible complex linear representation π\pi of 𝔰𝔩 2()\mathfrak{sl}_2(\mathbb{C}) on a complex vector space VV of dimension mm. Moreover, there are an integer kk and a basis {v k,v k+2,,v k2,v k}\{ v_{-k},v_{-k+2},\dots,v_{k-2},v_k\} of VV such that: 1. hv k2j=(k2j)v ih\cdot v_{k-2j} = (k-2j) v_i, 1. ev k2j=j(kj+1)v k(2j1)e\cdot v_{k-2j} = j(k-j+1)v_{k-(2j-1)}, with ev k=0e\cdot v_k=0, 1. fv k2j=v k2(j+1)f\cdot v_{k-2j}=v_{k-2(j+1)}, with fv k=0f\cdot v_{-k}=0.

where gv g\cdot v for vVv \in V and g𝔰𝔩 2()g \in \mathfrak{sl}_2(\mathbb{C}) is the action of 𝔰𝔩 2()\mathfrak{sl}_2(\mathbb{C}) on VV.

Proof. Let VV be a complex linear irreducible representation of 𝔰𝔩 2()\mathfrak{sl}_2(\mathbb{C}). We show that VV satisfy 1. to 3. Let v0v \ne 0 be an eigenvector for hh\cdot, say hv=λvh\cdot v = \lambda v. Then ev,eev,e \cdot v, e \cdot e \cdot v, \dots are also eigen vectors because:

hev=ehv+[h,e]v=eλv+2ev=(λ+2)ev he \cdot v = eh \cdot v + [h,e] \cdot v = e \cdot \lambda v + 2 e \cdot v = (\lambda + 2) e \cdot v

Since λ,λ+2, \lambda, \lambda +2, \dots are distinct, these eigenvectors are indipendent. By finite dimensionality of VV we can find v 0Vv_0 \in V with

a. v 00v_0 \ne 0 a. hv 0=λv 0h \cdot v_0 = \lambda v_0 a. ev 0=0e \cdot v_0 = 0

Define v i=f iv 0v_i = f^i \cdot v_0 . Then hv i=(λ2i)v ih \cdot v_i = (\lambda - 2i) v_i , by the same argument as above and so there is a minimum integer kk with f k+1v 0=v 0 f^{k+1} \cdot v_0 = v_0 . Then v 0v kv_0 \dots v_k are indipendent and

a. ev 0=0e \cdot v_0 = 0 a. hv i=(λ2i)v ih \cdot v_i = (\lambda -2i) v_i a. fv i=v i+1f \cdot v_i = v_{i+1} with v n+1=0v_{n+1}= 0.

Now we show that V=span{v 0,,v n}V = span \{ v_0, \dots, v_n \} . It is sufficient to show that span{v 0,,v n} span \{ v_0, \dots, v_n \} is stable under e e \cdot. This results from:

ev i=i(λi+1)v i1 e \cdot v_i = i(\lambda - i + 1) v_{i-1}

with v 1=0v_{-1} = 0 . We prove the calculation above by induction. Assume case for i. Then for i+1:

ev i+1=efv i=[e,f]v i+fev i=hv i+fev i=(λ2i)v i+f(i(λi+1))v i1=(i+1)(λi)v i e \cdot v_{i+1} = ef \cdot v_{i} = [e,f] \cdot v_i + fe \cdot v_{i} = h \cdot v_i + fe \cdot v_{i} = (\lambda - 2i) v_i + f \cdot(i(\lambda -i + 1 ))v_{i-1}= (i+1)(\lambda -i )v_i

For ending the proof of uniqueness, we show that λ=k \lambda = k. It results:

Tr(h)=Tr(ef)Tr(fe)=0 Tr(h) = Tr(ef) - Tr(fe) = 0

due to the cyclic property of the trace. Note that here we intend h,eh,e and ff, being elements of End(V)End(V). Then in the basis {v 0,v n} \{ v_0, \dots v_n \} the trace is explicitely:

i=0 k(λ2i)=0. \sum_{i=0}^k(\lambda - 2i)= 0.

Then λ=k \lambda=k .

Remark that by 1. in the statement of the theorem the smallest of the eigenvalues is the negative of the greatest.

The proof of the existence is easily achieved by defining a π:𝔰𝔩 2()End(W)\pi : \mathfrak{sl}_2(\mathbb{C}) \to End(W) on the generators of 𝔰𝔩 2() \mathfrak{sl}_2(\mathbb{C}) such that properties 1. 2. 3. are satisfied, and then by extending through linearity to whole 𝔰𝔩 2() \mathfrak{sl}_2(\mathbb{C}). This is a representation for 𝔰𝔩 2()\mathfrak{sl}_2(\mathbb{C}). To see irreducibility, let UU be a nonzero invariant subspace. Since UU is invariant under hh \cdot, UU is spanned by a subset of the basis v iv_i. Applying on such a v iv_i the action of ee \cdot several times, we can see that v 0v_0 is in UU. Repeated application of ff \cdot then shows that U=VU = V. Hence the representation is irreducible. This completes the proof.

Since we built directly a representation that is irreducible in a very concrete way, we want a result that assure us that we can decompose a representation in irreducible ones, that we report without proof.

Theorem. Let ϕ\phi be a complex linear representation of 𝔰𝔩 2()\mathfrak{sl}_2(\mathbb{C}) on a finite-dimensional vector space VV. Then V is completely reducible in the sense that there exist invariant subspaces U 1,..U rU_1,..U_r of VV such that V=U 1..U rV = U_1 \oplus .. U_r and such that the restriction of the representation to each U iU_i is irreducible

The importance of studying 𝔰𝔩 2()\mathfrak{sl}_2(\mathbb{C}) is that a semi-simple Lie algebra gg contains several copies of 𝔰𝔩 2()\mathfrak{sl}_2(\mathbb{C}). It follows that for a given VV representation of gg, there are several representantions of 𝔰𝔩 2()\mathfrak{sl}_2(\mathbb{C}) embedded in VV. This is shown shortly in Lie groups and algebras representations.

Revised on July 2, 2010 at 08:25:50 by giuseppe_malavolta