Complex Split Biquaternions and the Clifford Algebra Cl(3)
Introduction
The split biquaternion algebra $\mathbb{B}_{\mathrm{spl}} = \mathbb{C}_{\mathrm{spl}}\otimes_{\mathbb{R}}\mathbb{H}$ is the Clifford algebra of three generators of square $-1$. Its three anticommuting generators are the elements $j \vec e_x, j \vec e_y, j \vec e_z$, where $\vec e_x, \vec e_y, \vec e_z$ are the quaternion units and $j$ is the central split unit of $\mathbb{C}_{\mathrm{spl}}$ with $j^2 = +1$; the product of the three is $-j$, the central split unit again. This article studies the complexification of that algebra, $$ \mathrm{Cl}(3) := \mathbb{C}\otimes_{\mathbb{R}}\mathbb{B}_{\mathrm{spl}} = \mathbb{C}\otimes_{\mathbb{R}}\mathbb{C}_{\mathrm{spl}}\otimes_{\mathbb{R}}\mathbb{H}, $$ the complex split biquaternions. Over $\mathbb{C}$ the algebra is eight-dimensional, and its structure is fixed by a single fact: the volume element is central with square $+1$ instead of square $-1$. The centrality splits the algebra into two copies, $$ \mathrm{Cl}(3) \cong \mathrm{Cl}(2)\oplus\mathrm{Cl}(2) \cong M_2(\mathbb{C})\oplus M_2(\mathbb{C}) \cong \mathbb{B}\oplus\mathbb{B}, $$ each copy isomorphic to the biquaternion algebra. Where $\mathrm{Cl}(2)$ is simple, $\mathrm{Cl}(3)$ is not; where $\mathrm{Cl}(2)$ has a single family of minimal left ideals, $\mathrm{Cl}(3)$ has two, and the two are interchanged by the parity automorphism $\vec e_\mu\mapsto -\vec e_\mu$ and distinguished by the sign of the volume element. That pair is the algebraic content of the two chiralities of the four-dimensional spinor representation, and it is the reason the algebra is of interest to the physics of the left-right symmetric extension of the Standard Model, treated in Left-Right Symmetric Fermions from Complex Split Biquaternions and Bioctonions.
The article is the Clifford entry of the split biquaternion system, the higher rung of the pair whose lower rung is Biquaternion Versors and the Orthogonal Group. The split biquaternion algebra, its conjugations, its idempotents $\tilde\Pi_{1,2}=\tfrac12(1\pm j)$ and its isomorphism $\mathbb{B}_{\mathrm{spl}}\cong\mathbb{H}\oplus\mathbb{H}$ are Split-Biquaternion Algebra and Split-Biquaternion Ideals and Peirce Decomposition; the biquaternion algebra $\mathbb{B}=\mathbb{C}\otimes_{\mathbb{R}}\mathbb{H}$, its idempotents and its off-diagonal elements are Introduction to the Algebra of Biquaternions and Peirce Decompositions of the Biquaternion Algebra, and its minimal left ideals are The Polarized Module of the Biquaternion Algebra; the complex Clifford algebra $\mathrm{Cl}(2)\cong\mathbb{C}\mathrm{l}_2\cong\mathbb{B}$, its volume element and its grade structure are Biquaternion Versors and the Orthogonal Group; and the maximal totally isotropic subspace, the ladder operators and the charge operator are Peirce Carriers and Inner Conjugation.
The Algebra and Its Generators
Proposition (the split biquaternion generators). In $\mathbb{B}_{\mathrm{spl}}$ the three elements $$ E_1 := j \vec e_x, \qquad E_2 := j \vec e_y, \qquad E_3 := j \vec e_z $$ satisfy $E_\mu^2=-1$ and $E_\mu E_\nu=-E_\nu E_\mu$ for $\mu\neq\nu$. Hence they generate a copy of $\mathrm{Cl}(0,3)$ inside $\mathbb{B}_{\mathrm{spl}}$, and $\mathbb{B}_{\mathrm{spl}}\cong\mathrm{Cl}(0,3)$.
Proof. Since $j$ is central with $j^2=+1$ and the quaternion units anticommute with $\vec e_\mu^2=-1$, $$ E_\mu^2 = j^2\vec e_\mu^2 = -1, \qquad E_\mu E_\nu = j^2\vec e_\mu \vec e_\nu = \vec e_\mu \vec e_\nu = -\vec e_\nu \vec e_\mu = -E_\nu E_\mu \quad(\mu\neq\nu). $$ The eight monomials are independent and span $\mathbb{B}_{\mathrm{spl}}$, so the generated algebra is the whole of it. The split biquaternion algebra is therefore the Clifford algebra of a three-dimensional negative definite space.
Proposition (the complex split biquaternions). The complexification $$ \mathrm{Cl}(3) = \mathbb{C}\otimes_{\mathbb{R}}\mathbb{B}_{\mathrm{spl}} = \mathbb{C}\otimes_{\mathbb{R}}\mathbb{C}_{\mathrm{spl}}\otimes_{\mathbb{R}}\mathbb{H} $$ is the complex Clifford algebra generated by the same $E_\mu$, with complex coefficients; it has complex dimension $8$ and real dimension $16$, and $$ \mathrm{Cl}(3)\cong\mathrm{Cl}_3(\mathbb{C}). $$
Proof. Complexification extends scalars only, so the relations of the generators are unchanged and the $2^3$ monomials over $\mathbb{C}$ are a basis.
Proposition (the adjoint on the generators). With $E_\mu^{\dagger}=-E_\mu$ and $i^{\dagger}=-i$ the map $\dagger$ is an anti-involution of $\mathrm{Cl}(3)$, and the rescaled generators $iE_\mu$ are Hermitian, $(iE_\mu)^{\dagger}=iE_\mu$, with $\{iE_\mu,iE_\nu\}=+2\delta_{\mu\nu}$.
Proof. The adjoint of a product reverses the order, so $(E_\mu E_\nu)^{\dagger}=E_\nu^{\dagger}E_\mu^{\dagger}=(-E_\nu)(-E_\mu)=-E_\mu E_\nu$ for $\mu\neq\nu$, consistent; the rest is immediate.
The Volume Element and the Central Splitting
Proposition (the volume element). The element $$ \omega := E_1E_2E_3 $$ satisfies $$ \omega^2=1, \qquad \omega E_\mu = E_\mu\omega \quad (\mu=1,2,3), \qquad \omega^{\dagger}=\omega . $$ That is, $\omega$ is central, it is an involution, and it is self-adjoint.
Proof. For $\mu\neq\nu$ the generators anticommute, so $E_\mu\omega=\omega E_\mu$ for each $\mu$: each of the three factors crosses the other two once, at the cost of one sign, and the three signs cancel. Hence $\omega$ commutes with the generators, and therefore with the whole algebra. For the square, move the factors into adjacent pairs, $$ \omega^2 = E_1E_2E_3E_1E_2E_3 = E_1E_2(E_3E_1)E_2E_3 = -E_1E_2E_1E_3E_2E_3 = E_1^2E_2E_3E_2E_3 = -E_2E_3E_2E_3 = E_2^2E_3^2 = +1, $$ where $E_3E_1=-E_1E_3$, $E_2E_1=-E_1E_2$, $E_3E_2=-E_2E_3$ and $E_1^2=E_2^2=E_3^2=-1$ were used successively. This is the general formula $(-1)^{n(n-1)/2}c^n$ for the square of a product of $n$ anticommuting generators of square $c$, at $n=3$, $c=-1$. Finally, $\omega^{\dagger}=E_3^{\dagger}E_2^{\dagger}E_1^{\dagger}=(-E_3)(-E_2)(-E_1)=-E_3E_2E_1=+E_1E_2E_3=\omega$, since reversing three anticommuting factors costs one sign. Every identity was checked on the eight monomials of the algebra with zero residual.
The two properties, centrality and $\omega^2=+1$, are exactly what a split unit has, and they are the reason the complexified algebra behaves like the split biquaternions rather than like the biquaternions. Since $\omega=-j$ (because $E_1E_2E_3=(j\vec e_x)(j\vec e_y)(j\vec e_z)=j^3(\vec e_x \vec e_y)\vec e_z=-j$), the involution $\omega$ of $\mathrm{Cl}(3)$ is the central split unit of $\mathbb{B}_{\mathrm{spl}}$ up to sign.
Theorem (the central splitting). Define $$ \Pi_x := \tfrac12(1+\omega), \qquad \Pi_y := \tfrac12(1-\omega). $$ Then $$ \Pi_x^2=\Pi_x, \qquad \Pi_y^2=\Pi_y, \qquad \Pi_x\Pi_y=\Pi_y\Pi_x=0, \qquad \Pi_x+\Pi_y=1, $$ and both are central. Consequently $$ \boxed{\; \mathrm{Cl}(3) = \mathrm{Cl}(3)\Pi_x \oplus \mathrm{Cl}(3)\Pi_y \;} $$ is a decomposition of $\mathrm{Cl}(3)$ into the direct sum of two proper nonzero two-sided ideals, each of complex dimension $4$, and each isomorphic to the biquaternion algebra: $$ \mathrm{Cl}(3)\Pi_{1,2} \cong \mathrm{Cl}(2) \cong \mathbb{C}\mathrm{l}_2 \cong \mathbb{B} \cong M_2(\mathbb{C}). $$
Proof. Orthogonality, idempotency and completeness are the algebra $\omega^2=1$: for instance $\Pi_x\Pi_y=\tfrac14(1-\omega^2)=0$ and $\Pi_x^2=\tfrac14(1+2\omega+\omega^2)=\tfrac12(1+\omega)=\Pi_x$. Centrality is centrality of $\omega$. The direct sum is the decomposition of a module by a complete family of orthogonal central idempotents. Each ideal has complex dimension $4$, since left multiplication by $\Pi_x$ is a rank-$4$ projector of the $8$-dimensional space; and a four-dimensional complex algebra that is a full matrix algebra is $M_2(\mathbb{C})$, which is $\mathbb{C}\mathrm{l}_2\cong\mathbb{B}$ by Introduction to the Algebra of Biquaternions. The dimension of each ideal was computed to be $4$ on the eight monomials.
Corollary (semisimple, not simple). $\mathrm{Cl}(3)$ is semisimple and is not simple; its two-sided ideals are $0$, $\mathrm{Cl}(3)\Pi_x$, $\mathrm{Cl}(3)\Pi_y$ and $\mathrm{Cl}(3)$. Its minimal left, right and two-sided ideals coincide, and its left ideals form the diamond $$ 0 \subset \mathrm{Cl}(3)\Pi_x,\ \mathrm{Cl}(3)\Pi_y \subset \mathrm{Cl}(3). $$
Proof. Because $\Pi_{1,2}$ are central, the left, right and two-sided ideals they generate agree; the summands are simple matrix algebras and have no proper nonzero ideal, so no further ideals occur. The lattice is then the diamond, as in Split-Biquaternion Ideals and Peirce Decomposition.
Theorem (the two summands are the two halves). The idempotents $\Pi_{1,2}$ of $\mathrm{Cl}(3)$ coincide, with the indices reversed, with the split biquaternion idempotents $\tilde\Pi_{1,2}=\tfrac12(1\mp j)$ of $\mathbb{B}_{\mathrm{spl}}$ promoted to the complexification, and the central splitting is the complexification of the splitting of the split biquaternion algebra: $$ \boxed{\; \mathrm{Cl}(3) = \mathbb{C}\otimes_{\mathbb{R}}\bigl(\mathbb{H}\tilde\Pi_1\bigr) \oplus \mathbb{C}\otimes_{\mathbb{R}}\bigl(\mathbb{H}\tilde\Pi_2\bigr) \cong \mathbb{B}\oplus\mathbb{B} \;}. $$
Proof. With $\omega=-j$ one has $\Pi_x=\tfrac12(1-j)=\tilde\Pi_2$ and $\Pi_y=\tfrac12(1+j)=\tilde\Pi_1$. The two-sided ideals $\mathbb{H}\tilde\Pi_{1,2}\cong\mathbb{H}$ of Split-Biquaternion Ideals and Peirce Decomposition complexify to $\mathbb{B}\cong\mathbb{C}\otimes_{\mathbb{R}}\mathbb{H}$.
Theorem (the parity automorphism). The map $$ \sigma : \mathrm{Cl}(3)\to\mathrm{Cl}(3), \qquad \sigma(E_\mu)=-E_\mu, $$ is an algebra automorphism; it fixes the even monomials and negates the odd ones; it satisfies $\sigma(\omega)=-\omega$, and it therefore interchanges the two summands, $$ \boxed{\; \sigma(\Pi_x)=\Pi_y, \qquad \sigma(\Pi_y)=\Pi_x \;}. $$ The two summands are the two pinor representations of $\mathrm{Cl}(3)$: the two simple modules of $M_2(\mathbb{C})\oplus M_2(\mathbb{C})$, inequivalent and distinguished by the eigenvalue $\pm1$ of the central volume element $\omega$.
Proof. Sending every generator to its negative preserves the quadratic relations $E_\mu^2=-1$, $E_\mu E_\nu=-E_\nu E_\mu$, and is invertible, so it is an automorphism; it is the grading automorphism of the Clifford algebra, equal to $+1$ on the even subalgebra and $-1$ on the odd part. On the volume element the three sign flips give $\sigma(\omega)=(-1)^3\omega=-\omega$, and the action on $\Pi_{1,2}$ follows. A central element of a product of two simple algebras acts as a scalar on each simple module; the idempotents $\Pi_{1,2}$ act as the two projections and their distinct images are the two non-isomorphic simple modules.
The Ladder Structure and the Minimal Left Ideal
The splitting above is the statement that $\mathrm{Cl}(3)$ carries two copies of the biquaternion algebra. Inside one copy the biquaternion machinery of the corpus applies unchanged, and it is collected here in the form used in the companion physics article.
Definition. Fix the $\mathrm{Cl}(2)$ subalgebra generated by $E_1,E_2$ and set $$ \alpha := \tfrac12\bigl(E_1+iE_2\bigr), \qquad \alpha^{\dagger}=\tfrac12\bigl(-E_1+iE_2\bigr). $$ Let $N := \alpha^{\dagger}\alpha$ be the number operator and $P := \alpha\alpha^{\dagger}$.
Proposition (the isotropic element and its conjugate). The element $\alpha$ is isotropic, $\alpha^2=0$, and the pair satisfies $$ \{\alpha,\alpha^{\dagger}\} = 1, \qquad \alpha\alpha^{\dagger}+\alpha^{\dagger}\alpha = 1 . $$ Hence $\operatorname{span}_{\mathbb{C}}\{\alpha\}$ is a maximal totally isotropic subspace of the complex two-dimensional space $\operatorname{span}\{E_1,E_2\}$, and $\operatorname{span}\{\alpha^{\dagger}\}$ is its conjugate.
Proof. Expansion in the anticommutation relations: $\alpha^2=\tfrac14(E_1^2+i(E_1E_2+E_2E_1)+i^2E_2^2)=\tfrac14(-1+0+1)=0$, and likewise for the mixed anticommutator, which equals $\tfrac14(E_1^2+E_2^2)$ with the cross terms cancelling. Both were computed with zero residual.
Proposition (the complementary Hermitian idempotents). The elements $$ P = \alpha\alpha^{\dagger} = \tfrac12\bigl(1+iE_1E_2\bigr), \qquad N = \alpha^{\dagger}\alpha = \tfrac12\bigl(1-iE_1E_2\bigr) $$ are Hermitian idempotents, $P^2=P$, $N^2=N$, $P^{\dagger}=P$, $N^{\dagger}=N$, and they are complementary, $P+N=1$. The element $iE_1E_2$ is a self-adjoint involution, $(iE_1E_2)^2=1$.
Proof. From $(E_1E_2)^2=-1$ and $(E_1E_2)^{\dagger}=E_2^{\dagger}E_1^{\dagger}=(+E_1E_2)=-E_1E_2$; hence $iE_1E_2$ is Hermitian with square $+1$, and the two elements $\tfrac12(1\pm iE_1E_2)$ are the Hermitian idempotents of its $\pm1$ eigenspaces. Idempotency was checked with zero residual.
Theorem (the minimal left ideal and its charges). The left ideal $$ S := \mathrm{Cl}(3)\,P $$ is minimal, of complex dimension two, with basis $$ S = \operatorname{span}_{\mathbb{C}}\bigl\{\,P,\ \alpha^{\dagger}P\,\bigr\}, $$ and on it the number operator $N$ acts diagonally: $$ N\,P = 0, \qquad N\,\alpha^{\dagger}P = \alpha^{\dagger}P . $$ The charge operator $Ϙ := N$ therefore takes the two values $0$ and $1$ on $S$, with the ground state $P$ of charge $0$ and the excited state $\alpha^{\dagger}P$ of charge $1$.
Proof. The idempotent $P$ is primitive (its rank as an endomorphism of the two-space is one), so $S$ is minimal; the two elements are independent; and the actions follow from $\alpha P=0$ (as $\alpha^2=0$) and $\alpha^{\dagger}\alpha\alpha^{\dagger}=\alpha^{\dagger}(1-\alpha\alpha^{\dagger})=\alpha^{\dagger}$. The two vacuum and excited eigenvalues were computed on $S$ and are exactly $0$ and $1$.
The conjugate ideal is obtained from the conjugated idempotent $P^{*}=\tfrac12(1-iE_1E_2)=N$; conversely the ideal $\mathrm{Cl}(3)N$ has basis $\{N,\alpha N\}$ with charges $1$ and $0$. The construction is the two-dimensional case of the general theorem that a maximal totally isotropic subspace of a complexified Euclidean space carries a $*$-unitary symmetry: on a one-dimensional isotropic subspace it is the abelian $\mathrm{u}(1)$ generated by the number operator, so in $\mathrm{Cl}(3)$ the intrinsic symmetry of one copy is $\mathrm{u}(1)$ alone, to be contrasted with the $\mathrm{su}(3)\oplus\mathrm{u}(1)$ of the six-dimensional case. The particle content attached to $S$ — a neutrino of charge $0$ and a charged lepton of charge $1$, with their conjugates on the conjugate ideal, and the doubling of the whole set into the two chiralities — is the reading of the companion physics article, not an algebraic statement of this one.
The Right Action and the Enveloping Algebra
The two-sided structure of the biquaternion algebra is The Enveloping Operator Algebra of the Biquaternion Algebra and the Bi-module Structure: the left multiplications $L_a:\tilde Q\mapsto a\tilde Q$ and the right multiplications $R_b:\tilde Q\mapsto \tilde Q b$ commute, and together they generate the enveloping algebra $$ \mathbb{B}^{\mathrm{e}}=\mathbb{B}\otimes_{\mathbb{C}}\mathbb{B}_{\mathrm{opp}}\cong\operatorname{End}_{\mathbb{C}}(\mathbb{B})\cong M_4(\mathbb{C}). $$ A complex Clifford algebra of even dimension is a full matrix algebra, $\mathrm{Cl}(2n)_{\mathbb{C}}\cong M_{2^n}(\mathbb{C})$, so the sixteen-dimensional enveloping algebra is the complex Clifford algebra of four generators, $$ \mathbb{B}^{\mathrm{e}}\cong M_4(\mathbb{C})\cong\mathrm{Cl}(4)_{\mathbb{C}}\cong\mathbb{C}\otimes_{\mathbb{R}}\mathrm{Cl}(1,3), $$ the complexification of the Dirac algebra. The one-rung-higher statement, in which the left action is enlarged from the biquaternion algebra to the whole of $\mathrm{Cl}(3)$, is the content of the theorem below.
Theorem (the right action of $\mathbb{B}$ on $\mathrm{Cl}(3)$). Let $\mathbb{B}\subset\mathrm{Cl}(3)$ be the biquaternion subalgebra spanned by $1,\vec e_x,\vec e_y,\vec e_z$, and let $L_a$ and $R_b$ be left multiplication by $a\in\mathrm{Cl}(3)$ and right multiplication by $b\in\mathbb{B}$ on the eight-dimensional space $\mathrm{Cl}(3)$. Then $$ (L_aR_b)(L_{a'}R_{b'})=L_{aa'}R_{b'b}, $$ so that the span of the products is closed under multiplication, and this algebra is $$ \mathrm{Cl}(3)\otimes_{\mathbb{C}}\mathbb{B}_{\mathrm{opp}}\cong\bigl(M_2(\mathbb{C})\oplus M_2(\mathbb{C})\bigr)\otimes_{\mathbb{C}}M_2(\mathbb{C})\cong M_4(\mathbb{C})\oplus M_4(\mathbb{C})\cong\mathrm{Cl}(5)_{\mathbb{C}}\cong\mathrm{Cl}(6)^+, $$ of complex dimension $32$ and with a two-dimensional centre.
Proof. The left and right multiplications commute by associativity of $\mathrm{Cl}(3)$, so $(L_aR_b)(L_{a'}R_{b'})=(L_aL_{a'})(R_bR_{b'})$; the reversal in the second factor, $R_bR_{b'}=R_{b'b}$, is the opposite-algebra multiplication of The Enveloping Operator Algebra of the Biquaternion Algebra and the Bi-module Structure. Hence the span of the $L_aR_b$ is a subalgebra, and it is the image of the algebra homomorphism $\mathrm{Cl}(3)\otimes_{\mathbb{C}}\mathbb{B}_{\mathrm{opp}}\to\operatorname{End}_{\mathbb{C}}(\mathrm{Cl}(3))$, $a\otimes b_{\mathrm{opp}}\mapsto L_aR_b$. The domain is $\mathrm{Cl}(3)\otimes_{\mathbb{C}}\mathbb{B}_{\mathrm{opp}}\cong(M_2(\mathbb{C})\oplus M_2(\mathbb{C}))\otimes_{\mathbb{C}}M_2(\mathbb{C})\cong M_4(\mathbb{C})\oplus M_4(\mathbb{C})$, of complex dimension $8\cdot4=32$; the image was recomputed to be $32$-dimensional, so the map is injective and the image is the whole domain. The identifications $M_4(\mathbb{C})\cong\mathrm{Cl}(4)_{\mathbb{C}}$, $M_4(\mathbb{C})\oplus M_4(\mathbb{C})\cong\mathrm{Cl}(5)_{\mathbb{C}}$ and $\mathrm{Cl}(5)_{\mathbb{C}}\cong\mathrm{Cl}(6)^+$ are the low-dimensional classification of complex Clifford algebras. The closure identity, the dimension $32$ (thirty-two independent products) and the two-dimensional centre were all recomputed on the eight monomials with zero residual, and the opposite-algebra sign $R_{\vec e_x}R_{\vec e_y}=-R_{\vec e_z}$ was recomputed as well.
The theorem shows that the two-sided structure of the algebra has two rungs. On the biquaternion algebra alone the left and the right actions generate $\mathbb{B}^{\mathrm{e}}\cong\mathrm{Cl}(4)_{\mathbb{C}}$, of dimension $16$; enlarging the left action from $\mathbb{B}$ to $\mathrm{Cl}(3)$ and keeping the right action of $\mathbb{B}$ raises the generated algebra to $\mathrm{Cl}(3)\otimes_{\mathbb{C}}\mathbb{B}_{\mathrm{opp}}\cong\mathrm{Cl}(5)_{\mathbb{C}}$, of dimension $32$. The right action of $\mathbb{B}$ alone is four-dimensional and is the Clifford algebra $\mathrm{Cl}(2)$; it is the pair of actions together that produces the next complex Clifford algebra. Both rungs were recorded by the source of the companion physics article, which writes the first as "the right action of $\mathbb{C}\otimes\mathbb{H}$ on $\mathbb{C}\otimes\mathbb{H}$ gives $\mathrm{Cl}(4)$" and the second as "the right action of $\mathbb{C}\otimes\mathbb{H}$ on $\mathrm{Cl}(3)$ gives $\mathrm{Cl}(5)$"; the source's reading of the second — that the enlargement supplies the parity-reversal operator that $\mathrm{Cl}(3)$ does not contain, and that $\mathrm{Cl}(5)\cong\mathrm{Cl}(6)^+$ is the second of the two complex octonionic chain algebras of the division-algebra approach, one carrying $\mathrm{SO}(4)$ and one carrying $\mathrm{SU}(3)\times\mathrm{U}(1)$ — belongs to the companion physics article and is not an algebraic statement of this one.
The Legacy Pauli Gamma Formulation
The Pauli and gamma formulation is legacy: the reasons are explained in the article Why AXF Matrices instead of Pauli and Gamma Matrices. The generators carry no frame label: $\gamma_\mu$ is one set of matrices, $\gamma_0^2=I_4$ and $\gamma_k^2=-I_4$. The frame belongs to the coordinate, not to the array.
The gammas are a frame of $\mathrm{Cl}(3)=\mathbb{C}\otimes_{\mathbb{R}}\mathbb{B}_{\mathrm{spl}}\cong M_2(\mathbb{C})\oplus M_2(\mathbb{C})$, not a different algebra.
$$ \gamma_\mu=iE_\mu=i\,j\vec e_\mu,\qquad \gamma_\mu\gamma_\nu=-E_\mu E_\nu,\qquad \gamma_1\gamma_2\gamma_3=ij . $$
The native object is the triple $(\vec e_x,\vec e_y,\vec e_z)$ of quaternion units; a generator is one of them multiplied by $i\,j$, and a product of two is minus a quaternion unit. Each generator acts by left multiplication on the algebra, one basis operator of the two-sided family $L_aR_b$.
The dictionary is complex-linear; over $\mathbb{R}$ the generators $E_\mu=j\vec e_\mu$ span $\mathrm{Cl}_{0,3}(\mathbb{R})\cong\mathbb{H}\oplus\mathbb{H}$.
Summary
The complex split biquaternions are the complexification of the split biquaternion algebra, $$ \mathrm{Cl}(3) = \mathbb{C}\otimes_{\mathbb{R}}\mathbb{B}_{\mathrm{spl}} = \mathbb{C}\otimes_{\mathbb{R}}\mathbb{C}_{\mathrm{spl}}\otimes_{\mathbb{R}}\mathbb{H} \cong \mathrm{Cl}_3(\mathbb{C}), $$ of complex dimension $8$. Its generators $E_\mu=j\vec e_\mu$ satisfy $\{E_\mu,E_\nu\}=-2\delta_{\mu\nu}$, and its volume element $\omega=E_1E_2E_3$ is central, self-adjoint and an involution, $\omega^2=1$. The centrality gives the complete family of orthogonal central idempotents $\Pi_{1,2}=\tfrac12(1\pm\omega)$ and the splitting $$ \mathrm{Cl}(3) = \mathrm{Cl}(3)\Pi_x \oplus \mathrm{Cl}(3)\Pi_y \cong \mathbb{B}\oplus\mathbb{B} \cong M_2(\mathbb{C})\oplus M_2(\mathbb{C}), $$ into two copies of the biquaternion algebra, each of complex dimension $4$, so that $\mathrm{Cl}(3)$ is semisimple and not simple, with a four-element ideal lattice. The parity automorphism $E_\mu\mapsto-E_\mu$ negates $\omega$ and interchanges the two summands; they are the two pinor representations of $\mathrm{Cl}(3)$, inequivalent and distinguished by the eigenvalue $\pm1$ of $\omega$. Inside each summand the ladder element $\alpha=\tfrac12(E_1+iE_2)$ is isotropic, its number operator $N=\alpha^{\dagger}\alpha$ has the complementary projector $\alpha\alpha^{\dagger}$ as partner, and the minimal left ideal $\mathrm{Cl}(3)( \alpha\alpha^{\dagger})$ has basis $\{P,\alpha^{\dagger}P\}$ with charges $0$ and $1$ under $Q=N$. The whole construction is the three-dimensional case of the doubling of $\mathrm{Cl}(2n)$ to $\mathrm{Cl}(2n+1)$ by a central involution. The two-sided structure has two rungs: the left and right actions of $\mathbb{B}$ on itself generate the enveloping algebra $\mathbb{B}^{\mathrm{e}}\cong M_4(\mathbb{C})\cong\mathrm{Cl}(4)_{\mathbb{C}}$, and the left action of $\mathrm{Cl}(3)$ together with the right action of $\mathbb{B}$ generates $\mathrm{Cl}(3)\otimes_{\mathbb{C}}\mathbb{B}_{\mathrm{opp}}\cong M_4(\mathbb{C})\oplus M_4(\mathbb{C})\cong\mathrm{Cl}(5)_{\mathbb{C}}\cong\mathrm{Cl}(6)^+$.
Summary of Notation
| Symbol | Meaning |
|---|---|
| $\mathbb{C}_{\mathrm{spl}}$ | Split complex algebra, unit $j$, $j^2=+1$ |
| $\mathbb{H}$ | Quaternion algebra, units $e_t=1,\vec e_x,\vec e_y,\vec e_z$, $\vec e_\mu^2=-1$, $\vec e_x \vec e_y=\vec e_z$ |
| $\mathbb{B}_{\mathrm{spl}}=\mathbb{C}_{\mathrm{spl}}\otimes_{\mathbb{R}}\mathbb{H}$ | Split biquaternion algebra, ${}\cong\mathrm{Cl}(0,3)\cong\mathbb{H}\oplus\mathbb{H}$ |
| $\mathbb{B}=\mathbb{C}\otimes_{\mathbb{R}}\mathbb{H}$ | Biquaternion algebra, ${}\cong M_2(\mathbb{C})\cong\mathbb{C}\mathrm{l}_2$ |
| $\mathrm{Cl}(3)=\mathbb{C}\otimes_{\mathbb{R}}\mathbb{B}_{\mathrm{spl}}$ | Complex split biquaternions, ${}\cong\mathrm{Cl}_3(\mathbb{C})$ |
| $E_1,E_2,E_3=j\vec e_x,j\vec e_y,j\vec e_z$ | Clifford generators, $E_\mu^2=-1$, $E_\mu E_\nu=-E_\nu E_\mu$ |
| $iE_\mu$ | Hermitian rescaled generators, $(iE_\mu)^{\dagger}=iE_\mu$, $\{iE_\mu,iE_\nu\}=2\delta_{\mu\nu}$ |
| $\dagger$ | Hermitian adjoint, $E_\mu^{\dagger}=-E_\mu$, $i^{\dagger}=-i$ |
| $\omega=E_1E_2E_3=-j$ | Volume element, central, self-adjoint, $\omega^2=1$ |
| $\Pi_x = \tilde\Pi_2 = \tfrac12(1+\omega)$, $\Pi_y = \tilde\Pi_1 = \tfrac12(1-\omega)$ | Central idempotents, complete and orthogonal |
| $\mathrm{Cl}(3)\Pi_{1,2}$ | The two simple summands, each ${}\cong\mathbb{B}\cong M_2(\mathbb{C})$ |
| $\sigma:E_\mu\mapsto-E_\mu$ | Parity (grading) automorphism, interchanges the summands |
| $\alpha=\tfrac12(E_1+iE_2)$ | Isotropic (ladder) element, $\alpha^2=0$ |
| $N=\alpha^{\dagger}\alpha$, $P=\alpha\alpha^{\dagger}$ | Number operator and its complementary projector, $P+N=1$ |
| $S=\mathrm{Cl}(3)P$ | Minimal left ideal, basis $\{P,\alpha^{\dagger}P\}$ |
| $Q=N$ | Charge operator, values $0,1$ on $S$ |
| $L_a,R_b$ | Left and right multiplications, $L_a\tilde Q=a\tilde Q$, $R_b\tilde Q=\tilde Q b$ |
| $\mathrm{Cl}(3)\otimes_{\mathbb{C}}\mathbb{B}_{\mathrm{opp}}$ | Generated by left $\mathrm{Cl}(3)$ and right $\mathbb{B}$, ${}\cong M_4(\mathbb{C})\oplus M_4(\mathbb{C})\cong\mathrm{Cl}(5)_{\mathbb{C}}\cong\mathrm{Cl}(6)^+$ |
Further Reading
- Why AXF Matrices instead of Pauli and Gamma Matrices, for the reasons the AXF matrices replace the Pauli and gamma matrices.
- V. Vaibhav and T. P. Singh, "Left-Right Symmetric Fermions and Sterile Neutrinos from Complex Split Biquaternions and Bioctonions," arXiv:2108.01858, for the construction of this article's algebra as a two-chirality fermion algebra and its physics reading.
- P. Lounesto, Clifford Algebras and Spinors (Cambridge, 2001), for the classification $\mathrm{Cl}_3(\mathbb{C})\cong M_2(\mathbb{C})\oplus M_2(\mathbb{C})$, the volume element of an odd complex Clifford algebra, and the centrality that produces the splitting.
- H. B. Lawson and M.-L. Michelsohn, Spin Geometry (Princeton, 1989), for the volume element, the chirality grading and the two pinor representations of an odd Clifford algebra.
- J. H. Conway and D. A. Smith, On Quaternions and Octonions (A K Peters, 2003), for the split biquaternions and their isomorphism $\mathbb{B}_{\mathrm{spl}}\cong\mathbb{H}\oplus\mathbb{H}$.
- Companion article Split-Biquaternion Algebra, for the algebra being complexified, its conjugations and its idempotents.
- Companion article Split-Biquaternion Ideals and Peirce Decomposition, for the splitting $\mathbb{B}_{\mathrm{spl}}\cong\mathbb{H}\oplus\mathbb{H}$ and its diamond of ideals.
- Companion article The Biquaternion Algebra and Peirce Decompositions of the Biquaternion Algebra, for $\mathbb{B}\cong\mathbb{C}\mathrm{l}_2$, its idempotents and its minimal left ideals.
- Companion article The Enveloping Operator Algebra of the Biquaternion Algebra and the Bi-module Structure, for the left and right multiplications, their commuting, and $\mathbb{B}^{\mathrm{e}}\cong\operatorname{End}_{\mathbb{C}}(\mathbb{B})\cong M_4(\mathbb{C})$.