Split-Biquaternion Idempotents and Projections
Introduction
The algebra article defined the split biquaternion algebra $\mathbb{B}_{\mathrm{spl}} = \mathbb{C}_{\mathrm{spl}} \otimes_{\mathbb{R}} \mathbb{H}$, its four conjugations and its four distinguished real subspaces. This article treats the idempotents of $\mathbb{B}_{\mathrm{spl}}$ — the elements satisfying $\tilde\Pi^2 = \tilde\Pi$ — together with the direct sum decompositions they carry.
The treatment is purely mathematical. Every claim is either proved or stated as a definition. No physics is invoked. The split biquaternion algebra $\mathbb{B}_{\mathrm{spl}}$ is assumed from the basic algebra article, together with its four conjugations, its four fixed-point subspaces $\mathbb{C}_{\mathrm{spl},\mathbb{B}_{\mathrm{spl}}}$, $\mathbb{H}_{\mathbb{B}_{\mathrm{spl}}}$, $\mathbb{B}_{\Psi}$, $\mathbb{B}_{\Phi}$ and its idempotent decomposition. The split complex algebra $\mathbb{C}_{\mathrm{spl}}$ is assumed from the article on split complex algebra, together with its idempotents $\tilde\Pi_1 = \tfrac{1}{2}(1 + j)$ and $\tilde\Pi_2 = \tfrac{1}{2}(1 - j)$ and the isomorphism $\mathbb{C}_{\mathrm{spl}} \cong \mathbb{R} \oplus \mathbb{R}$. The quaternion algebra $\mathbb{H}$ is assumed to be a division algebra.
Throughout, a split biquaternion is written $\tilde{Q} = \sum_{\mu\in\{t,x,y,z\}} Q_\mu e_\mu$ with $Q_\mu = q_\mu + j q'_\mu$ and $q_\mu, q'_\mu \in \mathbb{R}$. The quaternion conjugate is $\tilde{Q}^{\natural} = Q_t e_t - \vec Q$, the split complex conjugate is $\bar{\tilde{Q}} = \bar{Q_t} e_t + \bar{\vec Q}$ with $Q_{\bar{\mu}} = q_\mu - j q'_\mu$, the Hermitian conjugate is $\tilde{Q}^{*} = \overline{\tilde{Q}^{\natural}}$, and the anti-Hermitian conjugate is $\tilde{Q}^\flat = -\tilde{Q}^{*}$. The split-biquaternion norm $\langle\tilde{Q}\rangle_{\natural} = \tilde{Q} \tilde{Q}^{\natural}$ is the form of Split-Biquaternion Norm and Invertibility, which this article names and does not use.
The result that shapes the whole article is that $\mathbb{B}_{\mathrm{spl}}$ has exactly four idempotents, and that all of them are central. This is a strict contrast with the biquaternion algebra $\mathbb{B} = \mathbb{C} \otimes_{\mathbb{R}} \mathbb{H}$, whose idempotents form a four-dimensional family in bijection with the roots of $-1$. The reason is structural: $\mathbb{B}_{\mathrm{spl}} \cong \mathbb{H} \oplus \mathbb{H}$ is a product of two division algebras, and a product of division algebras admits only the idempotents that are constant on each factor. This article records that rigidity and its consequences; the parallel study of the ideals of the algebra, and of the resulting Peirce decomposition and $\mathbb{H} \oplus \mathbb{H}$ splitting, is the subject of Split-Biquaternion Ideals and Peirce Decomposition.
Idempotents in an Algebra
Let $A$ be an associative unital algebra. An element $e \in A$ is an idempotent if $e^2 = e$. Idempotents encode direct summands: for an idempotent $e$,
$$ A = Ae \oplus A(1 - e) \quad (\text{left}), \qquad A = eA \oplus (1 - e)A \quad (\text{right}), $$
and conversely every decomposition of $A$ into complementary left ideals (or complementary right ideals) arises from an idempotent in this way. Two idempotents $e, f$ are orthogonal if $ef = fe = 0$; then $e + f$ is again idempotent. A family $\{e_1, \dots, e_n\}$ is pairwise orthogonal if $e_i e_j = 0$ for $i \neq j$, and complete if in addition $\sum_i e_i = 1$. A nonzero idempotent $e$ is primitive if it is not a sum of two nonzero orthogonal idempotents. The criterion used throughout, valid for a semisimple algebra $A$, is
$$ e \text{ primitive} \iff Ae \text{ is a minimal left ideal} \iff eAe \text{ is a division ring}. $$
The algebra $\mathbb{B}_{\mathrm{spl}} \cong \mathbb{H} \oplus \mathbb{H}$ is a product of two division algebras and is therefore semisimple, so the criterion applies to it. The idempotents of a product of division algebras are easy to describe: if $e = (\vec e_x, \vec e_y) \in D_1 \oplus D_2$ with $D_1, D_2$ division rings, then $e^2 = e$ forces $\vec e_x^2 = \vec e_x$ and $\vec e_y^2 = \vec e_y$, and a division ring has only the idempotents $0$ and $1$. Hence a product of two division algebras has exactly four idempotents, all central. This is the phenomenon that the rest of the article realises in $\mathbb{B}_{\mathrm{spl}}$.
The Idempotents of the Split Complex Centre
The idempotents of the split complex algebra $\mathbb{C}_{\mathrm{spl}}$ are
$$ \tilde\Pi_1 = \tfrac{1}{2}(1 + j), \qquad \tilde\Pi_2 = \tfrac{1}{2}(1 - j). $$
They satisfy
$$ \tilde\Pi_1^2 = \tilde\Pi_1, \qquad \tilde\Pi_2^2 = \tilde\Pi_2, \qquad \tilde\Pi_1 \tilde\Pi_2 = \tilde\Pi_2 \tilde\Pi_1 = 0, \qquad \tilde\Pi_1 + \tilde\Pi_2 = 1. $$
So $\{\tilde\Pi_1, \tilde\Pi_2\}$ is a complete family of pairwise orthogonal idempotents of $\mathbb{C}_{\mathrm{spl}}$. Since $j$ is central in $\mathbb{B}_{\mathrm{spl}}$, so are $1$ and $j$, and therefore so are $\tilde\Pi_1$ and $\tilde\Pi_2$: they lie in the centre of $\mathbb{B}_{\mathrm{spl}}$, which is the split complex subspace $\mathbb{C}_{\mathrm{spl},\mathbb{B}_{\mathrm{spl}}}$. The pair $\{\tilde\Pi_1, \tilde\Pi_2\}$ is a basis of the centre over $\mathbb{R}$, and the map
$$ \mathbb{C}_{\mathrm{spl},\mathbb{B}_{\mathrm{spl}}} \to \mathbb{R} \oplus \mathbb{R}, \qquad a \tilde\Pi_1 + b \tilde\Pi_2 \mapsto (a, b) $$
is an isomorphism of real algebras. The two idempotents are the invertibles of the two copies of $\mathbb{R}$ in this decomposition, and they are the only nontrivial idempotents of the centre.
The idempotents $\tilde\Pi_{1,2}$ are exchanged by the split complex conjugation and fixed by the quaternion conjugation. Indeed $\bar{\tilde\Pi_{1,2}} = \tilde\Pi_{1,2}$ because the coefficients of $\tilde\Pi_{1,2}$ are real, while $j^* = -j$ gives
$$ \tilde\Pi_1^* = \tilde\Pi_2, \qquad \tilde\Pi_2^* = \tilde\Pi_1. $$
Consequently
$$ \tilde\Pi_1^{*} = \tilde\Pi_2, \qquad \tilde\Pi_2^{*} = \tilde\Pi_1, \qquad \tilde\Pi_1^\flat = -\tilde\Pi_2, \qquad \tilde\Pi_2^\flat = -\tilde\Pi_1, $$
so the Hermitian and anti-Hermitian conjugations swap the two idempotents (up to sign). The idempotents $\tilde\Pi_{1,2}$ are therefore not Hermitian: $\tilde\Pi_1^{*} = \tilde\Pi_2 \neq \tilde\Pi_1$, and likewise with the two exchanged.
The Classification of the Idempotents
Theorem. The idempotents of $\mathbb{B}_{\mathrm{spl}}$ are exactly $0$, $\tilde\Pi_1$, $\tilde\Pi_2$ and $1$. There are no others.
Proof. Write an idempotent in the idempotent basis as $\tilde\Pi = \tilde\pi_1 \tilde\Pi_1 + \tilde\pi_2 \tilde\Pi_2$ with $\tilde\pi_{1,2} = \tilde\Pi \tilde\Pi_{1,2} \in \mathbb{H}$. Since $\tilde\Pi_1^2 = \tilde\Pi_1$, $\tilde\Pi_2^2 = \tilde\Pi_2$ and $\tilde\Pi_1 \tilde\Pi_2 = 0$, the square is
$$ \tilde\Pi^2 = \tilde\pi_1^2 \tilde\Pi_1 + \tilde\pi_2^2 \tilde\Pi_2. $$
Hence $\tilde\Pi^2 = \tilde\Pi$ is equivalent to the pair of equations $\tilde\pi_1^2 = \tilde\pi_1$ and $\tilde\pi_2^2 = \tilde\pi_2$ in the quaternion algebra $\mathbb{H}$. Each $\tilde\pi_{1,2}$ is therefore an idempotent of $\mathbb{H}$.
The quaternion algebra is a division algebra, so its only idempotents are $0$ and $1$: if $\tilde\pi = a + \vec v$ with $a \in \mathbb{R}$ and $\vec v$ pure, then $\tilde\pi^2 = \tilde\pi$ gives $2a\vec v = \vec v$ and $a^2 - |\vec v|^2 = a$; if $\vec v = 0$ then $a^2 = a$, so $a \in \{0, 1\}$; if $\vec v \neq 0$ then $a = 1/2$ and then $|\vec v|^2 = -1/4$, which is impossible. Hence $\tilde\pi_{1,2} \in \{0, 1\}$, and the four combinations give
$$ \boxed{\; \tilde\Pi = 0, \qquad \tilde\Pi = \tilde\Pi_1, \qquad \tilde\Pi = \tilde\Pi_2, \qquad \tilde\Pi = \tilde\Pi_1 + \tilde\Pi_2 = 1 \;}. $$
These four are idempotent, and there are no others.
Corollary. Every idempotent of $\mathbb{B}_{\mathrm{spl}}$ is central, and every idempotent lies in the split complex subspace $\mathbb{C}_{\mathrm{spl},\mathbb{B}_{\mathrm{spl}}}$. The set of idempotents is $\{0, \tilde\Pi_1, \tilde\Pi_2, 1\}$, a discrete set of four points.
Proof. The four idempotents are real combinations of $1$ and $j$, hence lie in the centre $\mathbb{C}_{\mathrm{spl},\mathbb{B}_{\mathrm{spl}}}$.
A nonzero idempotent is primitive exactly when it cannot be written as a sum of two nonzero orthogonal idempotents. From the four idempotents, the only orthogonal pairs are $\{\tilde\Pi_1, \tilde\Pi_2\}$ and $\{0, \tilde\Pi_{1,2}\}$. Hence $\tilde\Pi_1$ and $\tilde\Pi_2$ are primitive, while $0$ is excluded (it is not nonzero) and $1 = \tilde\Pi_1 + \tilde\Pi_2$ is not primitive. Thus $\{\tilde\Pi_1, \tilde\Pi_2\}$ is a complete family of primitive orthogonal idempotents, and it is the only one.
The Absence of a Bijection With the Roots of Minus One
In the biquaternion algebra $\mathbb{B}$, the idempotents are in bijection with the roots of $-1$: every idempotent has the form $\tfrac{1}{2}(e_t + \tilde{Ζ} i)$ for a unique root $\tilde{Ζ}$ of $-1$, and the root set is large. That correspondence fails completely in $\mathbb{B}_{\mathrm{spl}}$, and the failure is quantitative.
Theorem. The roots of $-1$ in $\mathbb{B}_{\mathrm{spl}}$ form the set
$$ \boxed{\; \{\tilde{Ζ} = \mu_+ \tilde\Pi_1 + \mu_- \tilde\Pi_2 : \mu_\pm \text{ unit pure real quaternions}\} \;}, $$
a set of real dimension $4$, parametrised by a pair of unit pure real quaternions. That the set is the topological product $S^2 \times S^2$, and its manifold structure, are established in Split-Biquaternion Analysis. The idempotents of $\mathbb{B}_{\mathrm{spl}}$ number $4$. Hence there is no bijection between the two sets.
Proof. The classification of the roots of $-1$ is the content of Split-Biquaternion Roots of Minus One: the equation $\tilde{Ζ}^2 = -1$, written in the idempotent basis as $\zeta_+^2 \tilde\Pi_1 + \zeta_-^2 \tilde\Pi_2 = -\tilde\Pi_1 - \tilde\Pi_2$, is equivalent to $\zeta_\pm^2 = -1$ in $\mathbb{H}$, whose solutions are the unit pure real quaternions, a two-sphere. The product is two-dimensional over each of the two components, so the root set has real dimension $2 + 2 = 4$. Since the idempotent set is the four-point set of the previous section, no bijection exists.
The reason for the failure is the same rigidity that produced the classification theorem. In $\mathbb{B}$ the construction $\tilde\Pi = \tfrac{1}{2} e_t \pm \tfrac{1}{2} \tilde{Ζ} i$ turns a root $\tilde{Ζ}$ of $-1$ into an idempotent because the central unit $i$ is available to convert the vector part into a scalar direction. In $\mathbb{B}_{\mathrm{spl}}$ there is no central scalar imaginary — the only central units are $1$ and $j$, with $j^2 = +1$ — and the obstruction $|\vec v|^2 = -1/4$ in the proof above is unavailable. The idempotents of $\mathbb{B}_{\mathrm{spl}}$ are therefore confined to the centre, and they cannot be parametrised by the roots of $-1$.
Idempotents and Direct Sum Decompositions
An idempotent $\tilde\Pi$ satisfies $\tilde\Pi^2 = \tilde\Pi$, and its complement $1 - \tilde\Pi$ is again an idempotent with
$$ \tilde\Pi(1 - \tilde\Pi) = \tilde\Pi - \tilde\Pi^2 = 0. $$
Because every idempotent of $\mathbb{B}_{\mathrm{spl}}$ is central, the pair $\{\tilde\Pi, 1 - \tilde\Pi\}$ gives a direct sum decomposition of the algebra as a two-sided decomposition:
$$ \mathbb{B}_{\mathrm{spl}} = \mathbb{B}_{\mathrm{spl}} \tilde\Pi \oplus \mathbb{B}_{\mathrm{spl}}(1 - \tilde\Pi) = \tilde\Pi \mathbb{B}_{\mathrm{spl}} \oplus (1 - \tilde\Pi) \mathbb{B}_{\mathrm{spl}}, $$
and the two summands are ideals. This is the algebraic content of the idempotent decomposition: the idempotent selects one summand, its complement the other, and the algebra splits into the image of one and the image of the other. The decomposition attached to $\tilde\Pi = \tilde\Pi_1$ is the idempotent decomposition
$$ \boxed{\; \mathbb{B}_{\mathrm{spl}} = \mathbb{H} \tilde\Pi_1 \oplus \mathbb{H} \tilde\Pi_2 \;}, $$
with $\mathbb{H} \tilde\Pi_1 = \mathbb{B}_{\mathrm{spl}} \tilde\Pi_1$ and $\mathbb{H} \tilde\Pi_2 = \mathbb{B}_{\mathrm{spl}} \tilde\Pi_2$.
A Hermitian idempotent is one satisfying $\tilde\Pi^{*} = \tilde\Pi$; in $\mathbb{B}$ these are the idempotents orthogonal with respect to the Hermitian form and the ones that occur in the spectral decomposition of a Hermitian element. In $\mathbb{B}_{\mathrm{spl}}$ there is no nontrivial Hermitian idempotent: the calculation $\tilde\Pi_1^{*} = \tilde\Pi_2$ of the second section shows that the two nontrivial idempotents are interchanged by ${}^{*}$, and neither is fixed. The idempotent decomposition of $\mathbb{B}_{\mathrm{spl}}$ is therefore a decomposition into two ideals that are exchanged by the Hermitian conjugation, not a decomposition into Hermitian idempotents.
Idempotents and the Zero Divisors
A nontrivial idempotent is a zero divisor, because
$$ \tilde\Pi_1 (1 - \tilde\Pi_1) = \tilde\Pi_1 \tilde\Pi_2 = 0, $$
and both factors are nonzero. Explicitly, the annihilator of $\tilde\Pi_1$ contains $\tilde\Pi_2$, and the annihilator of $\tilde\Pi_2$ contains $\tilde\Pi_1$; since $\mathbb{B}_{\mathrm{spl}} = \mathbb{H} \tilde\Pi_1 \oplus \mathbb{H} \tilde\Pi_2$, the two annihilators are exactly
$$ \mathrm{Ann}(\tilde\Pi_1) = \mathbb{H} \tilde\Pi_2, \qquad \mathrm{Ann}(\tilde\Pi_2) = \mathbb{H} \tilde\Pi_1, $$
each of real dimension $4$. The idempotents are therefore the two most symmetric members of the zero divisor set: the zero divisors of $\mathbb{B}_{\mathrm{spl}}$ are precisely the elements with a vanishing idempotent component, $\tilde{Q}_+ = 0$ or $\tilde{Q}_- = 0$, that is, the union $\mathbb{H} \tilde\Pi_2 \cup \mathbb{H} \tilde\Pi_1$ with the origin deleted.
This is a different situation from the biquaternion case. In $\mathbb{B}$ every non-pure zero divisor is a complex multiple of an idempotent, so the idempotents in a sense parametrise the zero divisors. In $\mathbb{B}_{\mathrm{spl}}$ the zero divisor set is the union of two four-dimensional real subspaces, while the idempotents are only four points: almost no zero divisor is a scalar multiple of an idempotent. The classification and the structure of the zero divisor set are the subject of Split-Biquaternion Zero Divisors.
Idempotents and the Minimal Left Ideals
Because $\tilde\Pi_1$ and $\tilde\Pi_2$ are central, the left ideal generated by $\tilde\Pi_1$ coincides with the two-sided ideal it generates:
$$ \mathbb{B}_{\mathrm{spl}} \tilde\Pi_1 = \tilde\Pi_1 \mathbb{B}_{\mathrm{spl}} = \tilde\Pi_1 \mathbb{B}_{\mathrm{spl}} \tilde\Pi_1 = \mathbb{H} \tilde\Pi_1. $$
Proposition. The ideals $\mathbb{H} \tilde\Pi_1$ and $\mathbb{H} \tilde\Pi_2$ satisfy $\mathbb{B}_{\mathrm{spl}} = \mathbb{H} \tilde\Pi_1 \oplus \mathbb{H} \tilde\Pi_2$, each has real dimension $4$, and each is a minimal left ideal; moreover $\tilde\Pi_{1,2} \mathbb{B}_{\mathrm{spl}} \tilde\Pi_{1,2} = \mathbb{H} \tilde\Pi_{1,2} \cong \mathbb{H}$ as a ring.
Proof. Every $\tilde{Q}$ satisfies $\tilde{Q} = \tilde{Q}(\tilde\Pi_1 + \tilde\Pi_2) = \tilde{Q} \tilde\Pi_1 + \tilde{Q} \tilde\Pi_2$, and the intersection is zero: if $\tilde{Q} \tilde\Pi_1 = \tilde Q_1 \tilde\Pi_2$, then multiplying on the right by $\tilde\Pi_1$ gives $\tilde{Q} \tilde\Pi_1 = 0$. In the idempotent basis $\tilde{Q} \tilde\Pi_1 = \tilde{Q}_+ \tilde\Pi_1$ with $\tilde{Q}_+ \in \mathbb{H}$, so $\mathbb{H} \tilde\Pi_1$ is the image of the projection $\tilde{Q} \mapsto \tilde{Q}_+ \tilde\Pi_1$, a real vector space of dimension $4$; the same holds for $\tilde\Pi_2$, and $4 + 4 = 8$ accounts for the whole algebra. For minimality, the map $\mathbb{B}_{\mathrm{spl}} \to \mathbb{H} \tilde\Pi_1$, $\tilde{Q} \mapsto \tilde{Q} \tilde\Pi_1$, is onto with kernel $\mathbb{H} \tilde\Pi_2$; any nonzero left ideal contained in $\mathbb{H} \tilde\Pi_1$ therefore has a preimage that is a left ideal strictly containing $\mathbb{H} \tilde\Pi_2$, and since the quotient $\mathbb{B}_{\mathrm{spl}} / \mathbb{H} \tilde\Pi_2 \cong \mathbb{H}$ is a division algebra, that ideal must be all of $\mathbb{H} \tilde\Pi_1$. Equivalently, $\tilde\Pi_1 \mathbb{B}_{\mathrm{spl}} \tilde\Pi_1 = \mathbb{H} \tilde\Pi_1 \cong \mathbb{H}$ is a division ring, so $\tilde\Pi_1$ is primitive.
The counterpart statement in the biquaternion algebra is different in kind. There the standard idempotents $\tilde\Pi_+ = \tfrac{1}{2}(e_t + i\vec e_z)$ and $\tilde\Pi_- = \tfrac{1}{2}(e_t - i\vec e_z)$ are not central, the ideals $\mathbb{B}\tilde\Pi_+$ and $\mathbb{B}\tilde\Pi_-$ are minimal left ideals only, and each is isomorphic to $\mathbb{C}^2$; the element $i\vec e_z$ is a root of $+1$, $(i\vec e_z)^2 = e_t$, and the two idempotents $\tilde\Pi_+$ and $\tilde\Pi_-$ are its associated idempotents. Here centrality upgrades the left ideals to two-sided ideals and replaces $\mathbb{C}^2$ by the division algebra $\mathbb{H}$. The ideals themselves, their lattice, the distinction between the two summands, and the Peirce decomposition associated with them are developed in Split-Biquaternion Ideals and Peirce Decomposition.
The Dimension of the Set of Idempotents
The set of idempotents of $\mathbb{B}_{\mathrm{spl}}$ is the four-point set $\{0, \tilde\Pi_1, \tilde\Pi_2, 1\}$, of real dimension $0$. It is a discrete subset of the centre $\mathbb{C}_{\mathrm{spl},\mathbb{B}_{\mathrm{spl}}}$, whose real dimension is $2$; the four points are its extremal points, and they are exactly the idempotents that the two-dimensional algebra $\mathbb{C}_{\mathrm{spl}} \cong \mathbb{R} \oplus \mathbb{R}$ possesses.
This is the sharpest form of the rigidity noted in the introduction. In $\mathbb{B}$ the non-trivial idempotents form a real four-dimensional manifold whose two-dimensional real locus is the sphere of Hermitian idempotents, inherited from the real roots of $-1$; in $\mathbb{B}_{\mathrm{spl}}$ the corresponding set has collapsed to two isolated points. The collapse is a consequence of the classification theorem: because the algebra is a product of two division algebras, its idempotent set is finite.
Summary
An idempotent of $\mathbb{B}_{\mathrm{spl}}$ is an element $\tilde\Pi$ with $\tilde\Pi^2 = \tilde\Pi$. The algebra has the two nontrivial idempotents
$$ \tilde\Pi_1 = \tfrac{1}{2}(1 + j), \qquad \tilde\Pi_2 = \tfrac{1}{2}(1 - j), \qquad \tilde\Pi_1 \tilde\Pi_2 = 0, \qquad \tilde\Pi_1 + \tilde\Pi_2 = 1, $$
which are central, lie in the split complex centre $\mathbb{C}_{\mathrm{spl},\mathbb{B}_{\mathrm{spl}}}$, and give the idempotent decomposition $\mathbb{B}_{\mathrm{spl}} = \mathbb{H} \tilde\Pi_1 \oplus \mathbb{H} \tilde\Pi_2$ into two minimal left ideals, each isomorphic to the quaternion division algebra $\mathbb{H}$.
There are exactly four idempotents, namely $0, \tilde\Pi_1, \tilde\Pi_2, 1$, and all of them are central; equivalently, $\mathbb{B}_{\mathrm{spl}} \cong \mathbb{H} \oplus \mathbb{H}$ is a product of two division algebras, and a product of two division algebras has a four-point idempotent set. The nontrivial idempotents $\tilde\Pi_1$ and $\tilde\Pi_2$ are primitive and orthogonal, and together with $0$ and $1$ they exhaust the primitive idempotents. Because there is no bijection with the roots of $-1$ — that set is $S^2 \times S^2$, of dimension $4$ — the biquaternion correspondence between idempotents and roots of $-1$ does not transfer, and the idempotents are confined to the centre. The Hermitian conjugation interchanges $\tilde\Pi_1$ and $\tilde\Pi_2$, so no nontrivial idempotent is Hermitian and the idempotent decomposition is not a decomposition into Hermitian idempotents. Each nontrivial idempotent is a zero divisor, with annihilator the complementary ideal: $\mathrm{Ann}(\tilde\Pi_1) = \mathbb{H} \tilde\Pi_2$ and $\mathrm{Ann}(\tilde\Pi_2) = \mathbb{H} \tilde\Pi_1$, each of real dimension $4$; the zero divisors themselves form the union of the two ideals, and are treated in Split-Biquaternion Zero Divisors.
Summary of Notation
| Symbol | Meaning |
|---|---|
| $\tilde\Pi$; $e, f$ | Idempotents of $\mathbb{B}_{\mathrm{spl}}$, $\tilde\Pi^2 = \tilde\Pi$; generic idempotents of an arbitrary algebra |
| $\tilde\Pi_1 = \tfrac{1}{2}(1 + j)$, $\tilde\Pi_2 = \tfrac{1}{2}(1 - j)$ | The two nontrivial idempotents, $\tilde\Pi_1 + \tilde\Pi_2 = 1$ |
| $\mathbb{B}_{\mathrm{spl}} = \mathbb{C}_{\mathrm{spl}} \otimes_{\mathbb{R}} \mathbb{H}$ | Split biquaternion algebra, $\cong \mathbb{H} \oplus \mathbb{H}$ |
| $\mathbb{C}_{\mathrm{spl},\mathbb{B}_{\mathrm{spl}}}$ | Centre, the split complex subspace; contains every idempotent |
| $\mathbb{H} \tilde\Pi_1, \mathbb{H} \tilde\Pi_2$ | The two minimal left (and two-sided) ideals, each $\cong \mathbb{H}$ |
| $\tilde\Pi_1^{*} = \tilde\Pi_2$, $\tilde\Pi_2^{*} = \tilde\Pi_1$ | The Hermitian conjugation exchanges the idempotents |
| $\mathrm{Ann}(\tilde\Pi_1) = \mathbb{H} \tilde\Pi_2$, $\mathrm{Ann}(\tilde\Pi_2) = \mathbb{H} \tilde\Pi_1$ | Annihilator of an idempotent |
| $\tilde{Ζ}$, $\mu_\pm$ | Root of $-1$; unit pure real quaternion components |
| $S^2 \times S^2$ | The topological product form of the root set, established in Split-Biquaternion Analysis |
| $\langle\tilde{Q}\rangle_{\natural} = \tilde{Q}\tilde{Q}^{\natural}$ | Split-Biquaternion norm, named only; the subject of Split-Biquaternion Norm and Invertibility |
Further Reading
- William Kingdon Clifford, "Preliminary Sketch of Biquaternions" (1873), for the first systematic treatment of biquaternions and their relatives.
- Pertti Lounesto, Clifford Algebras and Spinors (Cambridge, 2001), for idempotents and minimal left ideals in Clifford algebras.
- J. P. Ward, Quaternions and Cayley Numbers: Algebra and Applications (Kluwer, 1997), for the algebraic structure of the split biquaternions and their idempotents.
- Richard S. Pierce, Associative Algebras (Springer, 1982), for idempotents, primitivity and the Peirce decomposition in semisimple algebras.