Archive for the ‘Noncommutative Ring Theory Notes’ Category

We defined the n-th Weyl algebra A_n(R) over a ring R in here.  In this post we will find the GK dimension of A_n(R) in terms of the GK dimension of R. The result is similar to what we have already seen in commutative polynomial rings (see corollary 1 in here). We will assume that k is a field and R is a k-algebra.

Theorem. {\rm{GKdim}}(A_1(R))=2 + {\rm{GKdim}}(R).

Proof. Suppose first that R is finitely generated and let V be a frame of R. Let U=k+kx+ky. Since yx = xy +1, we have

\dim_k U^n = \frac{(n+1)(n+2)}{2}. \ \ \ \ \ \ \ \ \ (*)

Let W=U+V. Clearly W is a frame of A_1(R) and

W^n = \sum_{i+j=n} U^i V^j,

for all n, because every element of V commutes with every element of U. Therefore, since V^j \subseteq V^n and U^i \subseteq U^n for all i,j \leq n, we have W^n \subseteq U^nV^n and W^{2n} \supseteq U^nV^n. Thus W^n \subseteq U^nV^n \subseteq W^{2n} and hence

\log_n \dim_k W^n \leq \log_n \dim_k U^n + \log_n \dim_k V^n \leq \log_n \dim_k W^{2n}.

Therefore {\rm{GKdim}}(A_1(R)) \leq 2 + {\rm{GKdim}}(R) \leq {\rm{GKdim}}(A_1(R)), by (*), and we are done.

For the general case, let R_0 be any finitely generated k– subalgebra of R. Then, by what we just proved,

2 + {\rm{GKdim}}(R_0)={\rm{GKdim}}(A_1(R_0)) \leq {\rm{GKdim}}(A_1(R))

and hence 2+{\rm{GKdim}}(R) \leq {\rm{GKdim}}(A_1(R)). Now, let A_0 be a k-subalgebra of A_1(R) generated by a finite set \{f_1, \ldots , f_m\}. Let R_0 be the k-subalgebra of R generated by all the coefficients of f_1, \ldots , f_m. Then A_0 \subseteq A_1(R_0) and so

{\rm{GKdim}}(A_0) \leq {\rm{GKdim}}(A_1(R_0))=2 + {\rm{GKdim}}(R_0) \leq 2 + {\rm{GKdim}}(R).

Thus

{\rm{GKdim}}(A_1(R)) \leq 2 + {\rm{GKdim}}(R)

and the proof is complete. \Box

Corollary. {\rm{GKdim}}(A_n(R))=2n + {\rm{GKdim}}(R) for all n. In particular, {\rm{GKdim}}(A_n(k))=2n.

Proof. It follows from the theorem and the fact that A_n(R)=A_1(A_{n-1}(R)). \Box

As usual, I’ll assume that k is a field. Recall that if a k-algebra A is an Ore domain, then we can localize A at S:=A \setminus \{0\} and get the division algebra Q(A):=S^{-1}A. The algebra Q(A) is called the quotient division algebra of A.

Theorem (Borho and Kraft, 1976) Let A be a finitely generated k-algebra which is a domain of finite GK dimension. Let B be a k-subalgebra of A and suppose that {\rm{GKdim}}(A) < {\rm{GKdim}}(B) + 1. Let S:=B \setminus \{0\}. Then S is an Ore subset of A and S^{-1}A=Q(A). Also, Q(A) is finite dimensional as a (left or right) vector space over Q(B).

Proof. First note that, by the corollary in this post, A is an Ore domain and hence both Q(A) and Q(B) exist and they are division algebras. Now, suppose, to the contrary, that S is not (left) Ore. Then there exist x \in S and y \in A such that Sy \cap Ax = \emptyset. This implies that the sum By + Byx + \ldots + Byx^m is direct for any integer m. Let W be a frame of a finitely generated subalgebra B' of B. Let V=W+kx+ky and suppose that A' is the subalgebra of A which is generated by V. For any positive integer n we have

V^{2n} \supseteq W^n(kx+ky)^n \supseteq W^ny + W^nyx + \ldots + W^nyx^{n-1}

and thus \dim_k V^{2n} \geq n \dim_k W^n because the sum is direct. So \log_n \dim_k V^{2n} \geq 1 + \log_n \dim_k W^n and hence {\rm{GKdim}}(A) \geq {\rm{GKdim}}(A') \geq 1 + {\rm{GKdim}}(B'). Taking supremum of both sides over all finitely generated subalgebras B' of B will give us the contradiction {\rm{GKdim}}(A) \geq 1 + {\rm{GKdim}}(B). A similar argument shows that S is right Ore. So we have proved that S is an Ore subset of A. Before we show that S^{-1}A=Q(A), we will prove that Q(B)A=S^{-1}A is finite dimensional as a (left) vector space over Q(B). So let V be a frame of A. For any positive ineteger n, let r(n) = \dim_{Q(B)} Q(B)V^n. Clearly Q(B)V^n \subseteq Q(B)V^{n+1} for all n and

\bigcup_{n=0}^{\infty}Q(B)V^n =Q(B)A

because \bigcup_{n=0}^{\infty}V^n=A. So we have two possibilities: either Q(B)V^n=Q(B)A for some n or the sequence \{r(n)\} is strictly increasing. If Q(B)V^n = Q(B)A, then we are done because V^n is finite dimensional over k and hence Q(B)V^n is finite dimensional over Q(B). Now suppose that the sequence \{r(n)\} is strictly increasing. Then r(n) > n because r(0)=\dim_{Q(B)}Q(B)=1. Fix an integer n and let e_1, \ldots , e_{r(n)} be a Q(B)-basis for Q(B)V^n. Clearly we may assume that e_i \in V^n for all i. Let W be a frame of a finitely generated subalgebra of B. Then

(V+W)^{2n} \supseteq W^nV^n \supseteq W^ne_1 + \ldots + W^ne_{r(n)},

which gives us

\dim_k(V+W)^{2n} \geq r(n) \dim_k W^n > n \dim_k W^n,

because the sum W^ne_1 + \ldots + W^ne_{r(n)} is direct. Therefore {\rm{GKdim}}(A) \geq 1 + {\rm{GKdim}}(B), which is a contradiction. So we have proved that the second possibility is in fact impossible and hence Q(B)A is finite dimensional over Q(B). Finally, since, as we just proved, \dim_{Q(B)}Q(B)A < \infty, the domain Q(B)A is algebraic over Q(B) and thus it is a division algebra. Hence Q(B)A=Q(A) because A \subseteq Q(B)A \subseteq Q(A) and Q(A) is the smallest division algebra containing A. \Box

Here you can see part (2). We are now ready to prove a nice result.

Theorem. Let k be an algebraically closed field. Let A be a commutative k-domain and let B be a k-domain. Then A \otimes_k B is a k-domain.

Proof. Suppose that A \otimes_k B is not a domain. So there exist non-zero elements u, v \in A \otimes_k B such that uv=0. Let u = \sum_{i=1}^m a_i \otimes_k b_i and v = \sum_{i=1}^n a_i' \otimes_k b_i', where a_i,a_i' \in A, \ b_i,b_i' \in B and both sets \{b_1, \ldots , b_m\} and \{b_1', \ldots , b_n'\} are k-linearly independent. Let C=k[a_1, \ldots , a_m, a_1', \ldots , a_n'], which is a commutative domain because C \subseteq A. Also, note that u,v \in C \otimes_k B \subseteq A \otimes_k B. Now,  since u, v \neq0, there exist integers r,s such that a_r \neq 0 and a_s' \neq 0. Therefore, by the third part of the corollary in part (2), there exists a ring homomorphism \varphi : C \longrightarrow k such that \varphi(a_r) \neq 0 and \varphi(a_s') \neq 0. Let \psi = \varphi \otimes \text{id}_B. Then \psi : C \otimes_k B \longrightarrow k \otimes_k B \cong B is a ring homomorphism and hence \psi(u)\psi(v)=\psi(uv)=0. Therefore either \psi(u)=0 or \psi(v)=0, because B is a domain. But \psi(u)=\sum_{i=1}^m \varphi(a_i) \otimes_k b_i \neq 0, because \{b_1, \ldots , b_m \} is k-linearly independent and \varphi(a_r) \neq 0. Also, \psi(v)=\sum_{i=1}^n \varphi(a_i') \otimes_k b_i' \neq 0, because \{b_1', \ldots , b_n' \} is k-linearly independent and \varphi(a_s') \neq 0. This contradiction proves that A \otimes_k B is  a domain. \Box

Let k be an algebraically closed field. A trivial corollary of the theorem is a well-known result in field theory: if F_1,F_2 are two fields which contain k, then F_1 \otimes_k F_2 is a commutative domain. Another trivial result is this: if k is contained in both a field F and the center of a division algebra D, then F \otimes_k D is a domain.

Question. Let k be an algebraically closed field and let D_1,D_2 be two finite dimensional division k-algebras. Will D_1 \otimes_k D_2 always be a domain?

Answer. No! See the recent paper of Louis Rowen and David Saltman for an example.

Here you can see part (1). We are now going to prove a more interesting result than the one we proved in part (1). But we need to get prepared first. The following important result is known as Zariski’s lemma.

Lemma. (Zariski, 1946) Let k be a field and let A=k[a_1, \ldots , a_n] be a finitely generated commutative algebra. If A is a field, then A is algebraic over k  and thus \dim_k A < \infty.

Proof. The proof is by induction over n. If n=1, then A=k[a_1] and since A is a field, a_1^{-1} \in A. Thus a_1^{-1}=\sum_{i=0}^m \gamma_i a_1^i for some integer m \geq 0 and \gamma_i \in k. Then \sum_{i=0}^m \gamma_i a_1^{i+1}=0 and so a_1, and hence A, is algebraic over k. Now suppose that n \geq 2. If all a_i are algebraic over k, then A=k[a_1, \ldots , a_n] is algebraic over k and we are done. So we may assume that a_1 is transcendental over k. Since A is a field, K =k(a_1) \subseteq A and thus A=K[a_2, \ldots , a_n]. By the induction hypothesis, A is algebraic over K. So every a_i satisfies some monic polynomial f_i(x) \in K[x] of degree m_i. Let v_i \in k[a_1] be the product of the denominators of the coefficients of f_i(x) and put v=\prod_{i=1}^n v_i. Let m be the maximum of m_i. Then multiplying f(a_i)=0 through by v^m shows that each va_i is integral over k[a_1]. Note that since a_1 is transcendental over k, k[a_1] \cong k[x], the polynomial algebra over x. Thus I can choose an irreducible polynomial p(a_1) \in k[a_1] such that

\gcd(p(a_1), v)=1. \ \ \ \ \ \ \ \ \ (*)

Now (p(a_1))^{-1} \in A=k[a_1, \ldots , a_n], because A is a field. Thus for a large enough integer r, we have v^r(p(a_1))^{-1} \in k[va_1, \ldots , va_n] and hence v^r(p(a_1))^{-1} is integral over k[a_1]. But k[a_1] is a UFD and we know that every UFD is integerally closed (in its field of fraction). Therefore v^r(p(a_1))^{-1} \in k[a_1], which is absurd because then p(a_1) \mid v, contradicting (*). \ \Box

Corollary. Let k be a field and let A=k[a_1, \ldots , a_n] be a finitely generated commutative algebra.

1) If \mathfrak{m} is a maximal ideal of A, then \dim_k A/\mathfrak{m} < \infty.

2) If A is a field and k is algebraically closed, then A=k.

3) If A is a domain, k is algebraically closed and b_1, \ldots , b_m \in A \setminus \{0\}, then there exists a ring homomorphism \varphi : A \longrightarrow k such that \varphi(b_i) \neq 0 for all i=1, \ldots , m.

Proof. 1) Let \overline{a_i}=a_i + \mathfrak{m}, \ i=1, \ldots , n. Then A/\mathfrak{m} = k[\overline{a_1}, \ldots , \overline{a_n}] and we are done by the lemma.

2) By the lemma, A is algebraic over k and thus, since k is algebraically closed, A=k.

3) Let S be the set of all monomials in b_1, \ldots, b_m. Clearly S is multiplicatively closed and 0 \notin S because A is a domain. Consider B=S^{-1}A, the localization of A at S. Clearly B=k[a_1, \ldots , a_n, b_1^{-1}, \ldots , b_m^{-1}] and so B is finitely generated. Let \mathfrak{m} be a maximal ideal of B. Note that b_i \notin \mathfrak{m}, for all i, because each b_i is a unit in B. Now, we have B/\mathfrak{m} \cong k, by 2). Let f: B/\mathfrak{m} \longrightarrow k be a ring isomorphism. We also have the natural ring homomorphism g: B \longrightarrow B/\mathfrak{m} and an inejctive ring homomorphism h: A \longrightarrow B=S^{-1}A defined by f(a)=a/1. Let \varphi = fgh. Then \varphi : A \longrightarrow k and \varphi(b_i) \neq 0 for all i, because f is an isomorphism and b_i \notin \mathfrak{m} for all i. \ \Box

See part (3) here.

Lemma. Let R be a commutative ring and let G be a group. We define the map f : R[G] \longrightarrow R by

f(\sum_{g \in G} r_g g)=\sum_{g \in G} r_g.

Then f is an onto ring homomorphism and \{g - 1_G: \ g \in G, g \neq 1_G \} is a basis for the free R-module \ker f.

Proof. Obviously f is well-defined, additive and onto. Now, let x = \sum_{g \in G} r_g g, \ y = \sum_{g \in G} s_g g. Then

f(xy) = f(\sum_{g \in G} (\sum_{g_1g_2=g} r_{g_1}s_{g_2}) g)=\sum_{g \in G} \sum_{g_1g_2=g} r_{g_1}s_{g_2}=(\sum_{g \in G} r_g)(\sum_{g \in G} s_g)

=f(x)f(y).

So f is a ring homomorphism. Therefore \ker f is an ideal of R[G] and hence an R-module. Now, x = \sum_{g \in G} r_g g \in \ker f if and only if \sum_{g \in G}r_ g = 0 if and only if

x = x - (\sum_{g \in G}r_g)1_G=\sum_{g \in G}r_g(g-1_G).

Thus \ker f, as an R-module, is generated by the set \{g - 1_G : \ g \in G \}. Then, obviously, the set B=\{g-1_G: \ g \in G, g \neq 1_G\} still generates \ker f. To show that B is a basis for \ker f, as an R-module, we suppose that \sum_{g \in B}r_g (g-1_G)=0. Then

\sum_{g \in B}r_g g = (\sum_{g \in B}r_g)1_G.

But g \neq 1_G for all g \in B, and so r_g = 0 for all g \in B. \ \Box

Definition. The ring homomorphism f, as defined in the lemma, is called the augmentation map and \ker f is called the augmentation ideal of R[G].

Hurewicz Theorem. Let G be a group with the commutator subgroup G'. Let I be the augmentation ideal of \mathbb{Z}[G] and consider I as an additive group. Then \displaystyle \frac{G}{G'} \cong \frac{I}{I^2}.

Proof. Define the map \displaystyle \varphi : G \longrightarrow \frac{I}{I^2} by \varphi(g)=g-1_G + I^2. Clearly \varphi is well-defined because g - 1_G \in I for all g \in G. Also, since (g_1 - 1_G)(g_2-1_G) \in I^2, we have

\varphi(g_1g_2)=g_1g_2-1_G + I^2 = g_1g_2 -1_G - (g_1-1_G)(g_2-1_G) + I^2=

g_1 - 1_G + g_2 - 1_G + I^2 = \varphi(g_1) + \varphi(g_2).

Thus \varphi is a group homomorphism. So \varphi(g^{-1})=-\varphi(g) and therefore, since I is an abelian group, we have \varphi(g_1g_2g_1^{-1}g_2^{-1})=\varphi(g_1) + \varphi(g_1^{-1})+\varphi(g_2)+\varphi(g_2^{-1})=0. Thus G' \subseteq \ker \varphi and hence we have a group homomorphism \displaystyle \overline{\varphi}: \frac{G}{G'} \longrightarrow \frac{I}{I^2} defined by

\overline{\varphi}(gG')=g - 1_G +I^2.

Now, to show that \overline{\varphi} is an isomorphism, we will find an inverse for it. Define the map \psi : I \longrightarrow G/G' by \psi(\sum_{g \in G} n_g(g-1_G))=(\prod_{g \in G} g^{n_g})G'. Note that since g_1g_2G'=g_2g_1G' for all g_1,g_2 \in G, the map \psi is a well-defined group homomorphism. Now,

\psi((g_1 - 1_G)(g_2-1_G))=\psi(g_1g_2 -1_G -(g_1-1_G)-(g_2-1_G))=g_1g_2g_1^{-1}g_2^{-1}G' =G'.

So I^2 \subseteq \ker \psi because, by the lemma, the set \{(g_1-1_G)(g_2-1_G): \ g_1,g_2 \in G\} generates the additive group I^2. Hence the map \displaystyle \overline{\psi} : \frac{I}{I^2} \longrightarrow \frac{G}{G'} defined by

\overline{\psi}(\sum_{g \in G}n_g (g - 1_G)+ I^2) = (\prod_{g \in G} g^{n_g})G'

is a well-defined group homomorphism. It is now easy to see that both \overline{\varphi} \circ \overline{\psi} and \overline{\psi} \circ \overline{\varphi} are identity maps. \Box

Here you can see part (1).

Introduction. Let’s take a look at a couple of properties of the trace map in matrix rings. Let k be a field and B=M_m(k). Let \{e_{ij}: \ 1 \leq i,j \leq m\} be the standard basis for B. Let b = \sum_{i,j}\beta_{ij}e_{ij} \in B, where \beta_{ij} \in k. Now, \sum_{r,s}e_{rs}be_{sr}=\sum_{i,j,r,s} \beta_{ij}e_{rs}e_{ij}e_{sr}=\sum_{i,j} \beta_{ii}e_{jj} = \text{Tr}(b)I, where I is the identity element of B. Now let’s define \nu_B : = \sum_{ij} e_{ij} \otimes_k e_{ji} \in B \otimes_k B. See that \nu_B^2=I \otimes_k I = 1_{B \otimes_k B} and (b_1 \otimes_k b_2)\nu_B=\nu_B(b_2 \otimes_k b_1) for all b_1,b_2 \in B. We are going to extend these facts to any finite dimensional central simple algebra.

Notation. I will assume that A is a finite dimensional central simple k-algebra.

Theorem. There exists a unique element \nu_A = \sum b_i \otimes_k c_i \in A \otimes_k A such that \text{Trd}_A(a)=\sum b_iac_i for all a \in A. Moreover, \nu_A^2=1 and (a_1 \otimes_k a_2) \nu_A = \nu_A(a_2 \otimes_k a_1) for all a_1, a_2 \in A.

Proof. As we saw in this theorem, the map g : A \otimes_k A^{op} \longrightarrow \text{End}_k(A) \cong M_n(k) defined by g(a,a')(b)=aba', \ a,a',b \in A, is a k-algebra isomorphism. Let’s forget about the ring structure of A^{op} for now and look at it just as a k-vector space.  Then A \cong A^{op} and so we have a k-vector space isomorphism g: A \otimes_k A \longrightarrow \text{End}_k(A) defined by g(a \otimes_k a')(b)=aba', \ a,a',b \in A. Since \text{Trd}_A \in \text{End}_k(A), there exists a unique element

\nu_A = \sum b_i \otimes_k c_i \in A \otimes_k A

such that g(\nu_A) = \text{Trd}_A. Then \text{Trd}_A(a)=g(\nu_A)(a) = \sum g(b_i \otimes_k c_i)(a) = \sum b_iac_i.

To prove \nu_A^2=1, we choose a splitting field of K of A. Then B=A \otimes_k K \cong M_n(K) for some integer n, which is the degree of A. Let’s identify A and K with A \otimes_k 1 and 1 \otimes_k K respectively. Then A and K become subalgebras of B and B=AK. Let a \in A and \gamma \in K. Recall that, by the last part of the theorem in part (1), \text{Trd}_B(a) = \text{Trd}_A(a), for all a \in A. Let a \in A and \gamma \in K. Then, since K is the center of B, we have

\sum b_i(\gamma a) c_i = \gamma \sum b_i a c_i= \gamma \text{Trd}_A(a)= \gamma \text{Trd}_B(a)= \text{Trd}_B(\gamma a).

Thus \nu_B = \sum b_i \otimes_k c_i = \nu_A, by the uniqueness of \nu_B. Hence \nu_B^2 = \nu_A^2. But B \cong M_n(K) is a matrix ring and so, as we mentioned in the introduction, \nu_B^2=1. So \nu_A^2=1.

To prove the last part of the theorem, let s,t,a \in A. Then

g((a_1 \otimes_k a_2) \nu_A)(a)=g(\sum a_1b_i \otimes_k a_2c_i)(a) = \sum g(a_1b_i \otimes_k a_2c_i)(a)

= \sum a_1b_iaa_2c_i = a_1 \text{Trd}_A(aa_2)= \text{Trd}_A(a_2a)a_1.

The last equality holds by the second part of the theorem in part (1). Also, the image of \text{Trd}_A is in k, the center of A, and so \text{Trd}_A(a_2a) commutes with a_1. Now, we also have

g(\nu_A(a_2 \otimes_k a_1))(a) = g(\sum b_i a_2 \otimes_k c_ia_1)(a) = \sum g(b_ia_2 \otimes_k c_i a_1)(a)

= \sum b_ia_2ac_ia_1 = \text{Trd}_A(a_2a)a_1.

Thus g( (a_1 \otimes_k a_2) \nu_A) = g(\nu_A( a_2 \otimes_k a_1)) and so (a_1 \otimes_k a_2) \nu_A= \nu_A(a_2 \otimes_k a_1). \ \Box

Definition. The element \nu_A \in A \otimes_k A in the theorem is called the Goldman element for A.

Remark. David Saltman in a short paper used the properties of \nu_A to give a proof of Remark 2 in this post.

Let A be a finite dimensional central simple k-algebra. We proved in this theorem that \text{Nrd}_A: A^{\times} \longrightarrow k^{\times} is a group homomorphism. The image of \text{Nrd}_A is an abelian group because it lies in k^{\times}. So if a,b \in A^{\times}, then \text{Nrd}_A(aba^{-1}b^{-1})=\text{Nrd}_A(aa^{-1}) \text{Nrd}_A(bb^{-1})=1. Therefore (A^{\times})', the commutator subgroup of A^{\times}, is contained in \ker \text{Nrd}_A=\{a \in A: \ \text{Nrd}_A(a) = 1\} and so the following definition makes sense.

Definition. Let A be a finite dimensional central simple algebra. The reduced Whitehead group of A is the factor group \text{SK}_1(A) = (\ker \text{Nrd}_A)/(A^{\times})'.

Remark. Let D be a finite dimensional central division k-algebra of degree n. By the theorem in this post, for every a \in D there exists v \in (D^{\times})' such that \text{Nrd}_D(a)=a^nv. So if \text{Nrd}_D(a)=1, then a^n \in (D^{\times})'. Therefore g^n=1 for all g \in \text{SK}_1(D).

Example 1. \text{SK}_1(M_n(k))=\{1\} if k is a field and either n > 2 or n=2 and |k| > 3.

Proof. Let A=M_n(k). Then A^{\times} = \text{GL}(n,k) and

\ker \text{Nrd}_A = \{a \in M_n(k): \ \det(a) = 1 \}=\text{SL}(n,k).

We proved in here that (\text{GL}(n,k))' = \text{SL}(n,k). Thus \text{SK}_1(A)=\{1\}. \Box

Example 2. \text{SK}_1(\mathbb{H})=\{1\}, where \mathbb{H} is the division algebra of quaternions over \mathbb{R}.

Proof. Let z = \alpha + \beta i + \gamma j + \delta k \in \mathbb{H} and suppose that \text{Nrd}_{\mathbb{H}}(z)=\alpha^2+ \beta^2 + \gamma^2 + \delta^2=1. We need to prove that z is in the commutator subgroup of \mathbb{H}^{\times}. We are going to prove a stronger result, i.e. z = aba^{-1}b^{-1} for some a,b \in \mathbb{H}^{\times}. If z \in \mathbb{R}, then \beta = \gamma = \delta = 0 and z = \alpha = \pm 1. If z = 1, then we can choose a=b=1 and if z=-1, we can choose a=i, \ b=j. So we may assume that z \notin \mathbb{R}. We also have z^2 - 2 \alpha z + 1 = 0 (you may either directly check this or use this fact that every element of a finite dimensional central simple algebra satisfies its reduced characteristic polynomial).  Thus (z-\alpha)^2 = \alpha^2 - 1. Note that since \text{Nrd}_{\mathbb{H}}(z)=1 and z \notin \mathbb{R} we have \alpha^2 < 1. So 1 - \alpha^2 = \alpha_0^2 for some 0 \neq \alpha_0 \in \mathbb{R}. Let

w = \alpha_0^{-1}(z - \alpha). \ \ \ \ \ \ \ \ \ (1)

So, since the real part of z is \alpha, the real part of w is zero and hence

w^2 = -1, \ \ \overline{w} = -w. \ \ \ \ \ \ \ \ \ \ (2).

Since the center of \mathbb{H} is \mathbb{R} and w \notin \mathbb{R}, there exists c \in \mathbb{H} such that u = cw-wc \neq 0. Therefore

uw = -wu = \overline{w}u. \ \ \ \ \ \ \ \ \ \ (3)

Now, by (1), we have z = \alpha + \alpha_0 w and \alpha^2+\alpha_0^2=1. So we can write \alpha = \cos \theta and \alpha_0 = \sin \theta. Let v = \cos(\theta/2) + \sin(\theta/2)w. Then (2) gives us

v^2=z, \ \ \ \overline{v}=\cos(\theta/2) + \sin(\theta/2)\overline{w} = \cos(\theta/2) - \sin(\theta/2)w=v^{-1}. \ \ \ \ \ \ \ \ (4)

We also have u v^{-1} = \overline{v^{-1}}u=vu, by (3) and (4). Thus, by (4) again, uv^{-1}u^{-1}v = v^2=z. \ \Box

To be continued …