Von Neumann Entropy and the Biquaternion Norm
Introduction
The von Neumann entropy of a state is the fundamental quantitative measure of the information a quantum state fails to hold: it vanishes on the pure states, is maximal on the maximally mixed state, and quantifies the compression achievable in the quantum source coding theorem. In the biquaternion framework the state space of a qubit is the normalised hyperplane of the positive cone of $\mathbb{B}_{\Psi}$, and the cone is defined by the algebra's biquaternion norm. The question this article answers is what the entropy is, in that geometry.
The answer is a single clean statement: on the state space of the framework, the von Neumann entropy is a function of the biquaternion norm alone. For a state $\tilde{Ϛ}_{\Psi} = \tfrac12(e_t + i\vec s_{\Psi})$ one has
$$ \langle\tilde{Ϛ}_{\Psi}\rangle_{\natural} = \tfrac{1}{4}\bigl(1 - |\vec s_{\Psi}|^2\bigr)e_t , $$
so the biquaternion norm carries exactly one real number, the Bloch radius squared; and the spectrum of the state depends on the state only through that number, $\lambda_\pm = \tfrac12(1\pm|\vec s_{\Psi}|)$. The entropy is therefore
$$ \boxed{\;S(\tilde{Ϛ}_{\Psi}) = h\!\left(\frac{1+|\vec s_{\Psi}|}{2}\right) = h\!\left(\frac{1+\sqrt{1-4\langle\tilde{Ϛ}_{\Psi}\rangle_{\natural}}}{2}\right), \qquad h(p) = -p\log p - (1-p)\log(1-p)\;}, $$
a function of $\langle\cdot\rangle_{\natural}$ and nothing else. Equivalently, since the deviation of a state from idempotency is
$$ \boxed{\;\tilde{Ϛ}_{\Psi}^2 - \tilde{Ϛ}_{\Psi} = -\langle\tilde{Ϛ}_{\Psi}\rangle_{\natural}\,e_t\;}, $$
the entropy is a function of the single scalar by which the state fails to be idempotent. Pure states are the zero divisors of the biquaternion norm, maximally mixed is its maximum on the state space, and entropy measures the position of the state between them.
This is the informational interpretation of a geometric fact. The article does not claim that entropy is derived from the biquaternion norm in the sense of being forced by it — the function $h$ is the standard one, and its form is a theorem of information theory, not of the algebra — but it makes precise which algebraic invariant the entropy sees. The biquaternion norm supplies the number; information theory supplies the function. Two invariants of the algebra are in play throughout the framework and must not be conflated: the Hermitian form, $\langle\tilde{P}_{\Phi},\tilde H_\Psi\rangle_{*}=\mathrm{Sc}(\tilde{P}_{\Phi}\tilde H_\Psi^{*})$, and it gives the Born rule, and the biquaternion norm $\langle\tilde{Q}\rangle_{\natural} = \langle\tilde{Q},\tilde{Q}\rangle_{\natural} = \tilde{Q}\tilde{Q}^{\natural}$, which is indefinite of inertia $(1,3)$ and gives the cone. Entropy belongs to the second.
The article proceeds as follows. The state, its purity, and the biquaternion norm are recalled and the identities connecting them are derived. Then the spectral decomposition is used to compute the entropy and to write it as a function of the biquaternion norm, and the logarithm is exhibited as an element of the algebra. Then the consequences are collected: monotonicity, bounds, concavity, $*$-unitary invariance, and the behaviour at the two ends of the cone slice. Then the relation to the standard matrix formula is stated, and a section separates what the biquaternion-norm description adds from what it merely restates. Open questions and the usual closing sections follow.
The State, Its Purity, and the Biquaternion Norm
The state space
A state of the informational sector is an element of the Hermitian subspace
$$ \tilde{Ϛ}_{\Psi} = \tfrac{1}{2}\bigl(e_t + i\,\vec s_{\Psi}\bigr), \qquad \vec s_{\Psi} = r_1\vec e_x + r_2\vec e_y + r_3\vec e_z, \quad r_\eta\in\mathbb{R}, $$
normalised by $\langle\tilde{Ϛ}_{\Psi},e_t\rangle_{*}=\mathrm{Sc}(\tilde{Ϛ}_{\Psi})=\tfrac12$. Its matrix image is
$$ M(\tilde{Ϛ}_{\Psi}) = \tfrac{1}{2}\bigl(I_2 + \vec s_{\Psi}\cdot\boldsymbol{\sigma}\bigr), $$
whose eigenvalues are $\tfrac12(1\pm|\vec s_{\Psi}|)$. Positivity, $M(\tilde{Ϛ}_{\Psi})\geq0$, is therefore equivalent to $|\vec s_{\Psi}|\leq1$, and the states form the Bloch ball. The boundary $|\vec s_{\Psi}|=1$ is the set of pure states; the centre $\vec s_{\Psi}=0$ is the maximally mixed state $\tilde{Ϛ}_{\Psi} = \tfrac12 e_t$.
Purity and the deviation from idempotency
Squaring the state gives, using $\vec e_\mu \vec e_\nu = -\delta_{\mu\nu}e_t + \epsilon_{\mu\nu\lambda}\vec e_\lambda$ and the vanishing of the cross terms for the pure scalar part,
$$ \tilde{Ϛ}_{\Psi}^2 = \tfrac{1}{4}\bigl(e_t + i\vec s_{\Psi}\bigr)^2 = \tfrac{1}{4}\bigl((1+|\vec s_{\Psi}|^2)e_t + 2i\vec s_{\Psi}\bigr). $$
Hence the purity is
$$ P(\tilde{Ϛ}_{\Psi}) = 2\,\langle\tilde{Ϛ}_{\Psi},\tilde{Ϛ}_{\Psi}\rangle_{*} = 2\,\mathrm{Sc}\bigl(\tilde{Ϛ}_{\Psi}^2\bigr) = \tfrac{1}{2}\bigl(1 + |\vec s_{\Psi}|^2\bigr), $$
and the deviation from idempotency is a pure scalar multiple of the identity,
$$ \tilde{Ϛ}_{\Psi}^2 - \tilde{Ϛ}_{\Psi} = \tfrac{1}{4}\bigl(|\vec s_{\Psi}|^2 - 1\bigr)e_t . $$
The state is pure, $\tilde{Ϛ}_{\Psi}^2 = \tilde{Ϛ}_{\Psi}$, exactly when $|\vec s_{\Psi}|=1$; it is idempotent exactly when its purity is one. Everything that follows is a statement about the single real number $\mathrm{Sc}(\tilde{Ϛ}_{\Psi}^2-\tilde{Ϛ}_{\Psi}) = \tfrac14(|\vec s_{\Psi}|^2-1)\leq0$.
The biquaternion norm on $\mathbb{B}_{\Psi}$
The biquaternion norm of the algebra is $\langle\tilde{Q}\rangle_{\natural} = \langle\tilde{Q},\tilde{Q}\rangle_{\natural} = \tilde{Q}\tilde{Q}^{\natural} = \sum_\mu Q_\mu^2$. On a Hermitian element $\tilde H_\Psi = h_t\,e_t + i\vec h_\Psi$ with real coefficients it is the real scalar
$$ \langle\tilde H_\Psi\rangle_{\natural} = h_t^2 - |\vec h_\Psi|^2 , $$
so it has inertia $(1,3)$; the cone $\langle\tilde H_\Psi\rangle_{\natural}\geq0$ with $h_t\geq0$ is the positive cone of the Hermitian subspace, and positivity of the operator is exactly $\langle\tilde H_\Psi\rangle_{\natural}\geq0$ together with $h_t\geq0$. Equivalently, the biquaternion norm is the determinant of the matrix image,
$$ \langle\tilde H_\Psi\rangle_{\natural} = \langle\tilde H_\Psi,\tilde H_\Psi\rangle_{\natural} = \det M(\tilde H_\Psi) , $$
which is why it is the natural scalar for the cone: a Hermitian element is positive semidefinite precisely when its scalar part and its biquaternion norm are both non-negative.
For a state the biquaternion norm is
$$ \langle\tilde{Ϛ}_{\Psi}\rangle_{\natural} = \tfrac{1}{4}\bigl(1 - |\vec s_{\Psi}|^2\bigr)e_t , $$
a scalar element of the centre. Three interpretations of this single identity organize the rest of the article:
- $\langle\tilde{Ϛ}_{\Psi}\rangle_{\natural}\geq0$ is positivity, with equality exactly on the pure states; the biquaternion norm's zero set is the boundary of the Bloch ball, the zero-divisor cone of $\mathbb{B}_{\Psi}$.
- Comparing with the deviation from idempotency, $\tilde{Ϛ}_{\Psi}^2-\tilde{Ϛ}_{\Psi} = -\langle\tilde{Ϛ}_{\Psi}\rangle_{\natural}e_t$: the deviation from idempotency is minus the biquaternion norm. There is no other mixedness parameter.
- The purity is an affine function of the biquaternion norm, $P(\tilde{Ϛ}_{\Psi}) = 2\langle\tilde{Ϛ}_{\Psi},\tilde{Ϛ}_{\Psi}\rangle_{*} = 1 - 2\langle\tilde{Ϛ}_{\Psi}\rangle_{\natural}$.
The biquaternion norm therefore measures the failure of a state to be a zero divisor of the algebra, and it does so with one number. On the normalised hyperplane it takes its minimum $0$ on the pure states and its maximum $\tfrac14$ at the maximally mixed state.
The Spectral Decomposition and the Entropy
The spectrum
The state has a spectral decomposition into the two complementary idempotents along its Bloch direction. Writing $\vec u_s = \vec s_{\Psi}/|\vec s_{\Psi}|$ for $\vec s_{\Psi}\neq0$,
$$ \tilde{Ϛ}_{\Psi} = \lambda_+\,\tilde\Pi_1(\vec u_s) + \lambda_-\,\tilde\Pi_2(\vec u_s), \qquad \tilde\Pi_{1,2}(\vec u_s) = \tfrac{1}{2}\bigl(e_t \pm i\vec u_s\bigr), \qquad \lambda_\pm = \tfrac{1}{2}\bigl(1 \pm |\vec s_{\Psi}|\bigr). $$
The two idempotents are Hermitian, orthogonal, and sum to the identity: $\tilde\Pi_{1,2}^2 = \tilde\Pi_{1,2}$, $\tilde\Pi_1\tilde\Pi_2 = 0$, $\tilde\Pi_1 + \tilde\Pi_2 = e_t$. Their normalisation is $\langle\tilde\Pi_{1,2},e_t\rangle_{*}=\tfrac12$, and the Hermitian form gives $2\langle\tilde\Pi_{1,2},\tilde{Ϛ}_{\Psi}\rangle_{*}=\lambda_\pm$. This is the module-level spectral decomposition recalled in the article on the native qubit; the eigenvalues are read from the Bloch radius by $\lambda_+ - \lambda_- = |\vec s_{\Psi}|$ and $\lambda_+ + \lambda_- = 1$.
Entropy as a function of the biquaternion norm
The von Neumann entropy is
$$ S(\tilde{Ϛ}_{\Psi}) = -\lambda_+\log\lambda_+ - \lambda_-\log\lambda_- = h\!\left(\lambda_+\right). $$
Since $|\vec s_{\Psi}| = \sqrt{1-4\langle\tilde{Ϛ}_{\Psi}\rangle_{\natural}}$ on the normalised hyperplane, this is a function of the biquaternion norm alone:
$$ \boxed{\;S(\tilde{Ϛ}_{\Psi}) = H\!\left(\langle\tilde{Ϛ}_{\Psi}\rangle_{\natural}\right), \qquad H(N) = h\!\left(\frac{1+\sqrt{1-4N}}{2}\right)\;}. $$
The function $H$ is defined on the interval $N\in[0,\tfrac14]$ that the biquaternion norm takes on the state space. At the pure-state end $N=0$ it gives $H(0) = h(1) = 0$; at the maximally mixed end $N=\tfrac14$ it gives $H(\tfrac14) = h(\tfrac12) = \log 2$. On the interior it is smooth and strictly increasing, because $h$ is decreasing on $[\tfrac12,1]$ and $\sqrt{1-4N}$ is decreasing. Its derivative,
$$ \frac{dS}{dN} = -\frac{1}{\sqrt{1-4N}}\log\!\left(\frac{1-\sqrt{1-4N}}{1+\sqrt{1-4N}}\right) > 0 . $$
As $N\to0$ this derivative diverges to $+\infty$: the entropy leaves the pure states — the zero-divisor cone — with infinite slope in the biquaternion-norm coordinate. As $N\to\tfrac14$ it tends to the finite value $2$, matching the expansion $S\approx\log2-\tfrac12+2N$ near the maximally mixed state. The entropy is thus a monotone function of the biquaternion norm, and the biquaternion norm is its argument.
The logarithm as an element of the algebra
The standard formula can also be written algebraically. On the interior of the Bloch ball the state has full rank, so its real logarithm exists and lies in $\mathbb{B}_{\Psi}$:
$$ \log \tilde{Ϛ}_{\Psi} = \log\lambda_+\, \tilde\Pi_1(\vec u_s) + \log\lambda_-\, \tilde\Pi_2(\vec u_s) . $$
Multiplying by $\tilde{Ϛ}_{\Psi}$ and taking the scalar part,
$$ -2\,\mathrm{Sc}\bigl(\tilde{Ϛ}_{\Psi}\log\tilde{Ϛ}_{\Psi}\bigr) = -\lambda_+\log\lambda_+ - \lambda_-\log\lambda_- = S(\tilde{Ϛ}_{\Psi}) , $$
because the idempotents are orthogonal and each has scalar part $\tfrac12$. This is the ordinary matrix formula with the trace replaced by twice the scalar part. The derivation uses only the normalisation $\langle\tilde{Ϛ}_{\Psi},e_t\rangle_{*}=\tfrac12$ and the spectral decomposition; no property of the module beyond the spectral theorem is required.
Two caveats belong here. First, the expression requires the logarithm, whose value on the rank-one idempotents is unbounded below; the convention $0\log 0 = 0$ removes the boundary points and assigns the pure states entropy zero as a limit. In the interior the formula is exact; at the boundary it is a limiting statement. Second, the base of the logarithm is conventional; with the natural logarithm the maximal entropy is $\log 2$, and with base two it is one bit. The article uses the natural logarithm and states when a base-two value is quoted.
The entropy of a state as a function of its purity
Combining the two scalar invariants, the entropy depends on the state through the purity or through the biquaternion norm, and the two are affinely related:
$$ P(\tilde{Ϛ}_{\Psi}) = 2\,\langle\tilde{Ϛ}_{\Psi},\tilde{Ϛ}_{\Psi}\rangle_{*} = 1 - 2\langle\tilde{Ϛ}_{\Psi}\rangle_{\natural} , \qquad S = h\!\left(\frac{1+\sqrt{2\,P(\tilde{Ϛ}_{\Psi})-1}}{2}\right). $$
On the state space the purity ranges over $[\tfrac12,1]$ and the biquaternion norm over $[0,\tfrac14]$; the entropy is a concave function of the purity and an increasing function of the biquaternion norm. For a qubit the three invariants — Bloch radius, purity, biquaternion norm — carry exactly the same information, and the entropy is a function of any one of them. This degeneracy is special to two dimensions; for a $d$-level system the biquaternion norm is no longer a single number, and the entropy depends on the full spectrum rather than on one scalar.
Consequences of the Biquaternion-Norm Dependence
Bounds
Since $\langle\tilde{Ϛ}_{\Psi}\rangle_{\natural}\in[0,\tfrac14]$ and $H$ is increasing,
$$ \boxed{\;0 \;\leq\; S(\tilde{Ϛ}_{\Psi}) \;\leq\; \log 2\;}, $$
with $S=0$ exactly on the pure states (the zero divisors of the biquaternion norm) and $S=\log 2$ exactly at the maximally mixed state, the unique maximizer. The upper bound is the dimension of the module in nats, $\log\dim_{\mathbb{C}}S = \log2$, and it is attained at the centre of the Bloch ball, which is also the point where the biquaternion norm takes its maximal value $\tfrac14$ on the normalised hyperplane.
Monotonicity and the cone
Define the "radial" coordinate $\rho_{\mathrm{B}} = |\vec s_{\Psi}| = \sqrt{1-4N}$. Radial contraction of the state toward the centre — the operation $\vec s_{\Psi}\mapsto(1-p)\vec s_{\Psi}$ for $0\leq p\leq1$, which is depolarization — strictly increases $\langle\cdot\rangle_{\natural}$ and strictly increases $S$ for any state with $\vec s_{\Psi}\neq0$. Radial expansion toward the boundary strictly decreases both. In the language of the cone, moving the state deeper into the positive cone (larger $\langle\cdot\rangle_{\natural}$) increases its entropy; approaching the zero-divisor cone ($N\to0$) decreases it to zero. Entropy is thus a monotone coordinate on the cone slice, and the pure states form its boundary.
Concavity
The entropy is concave. For two states and $0\leq t\leq1$,
$$ S\bigl(t\tilde{Ϛ}_{\Psi,1} + (1-t)\tilde{Ϛ}_{\Psi,2}\bigr) \;\geq\; t\,S(\tilde{Ϛ}_{\Psi,1}) + (1-t)\,S(\tilde{Ϛ}_{\Psi,2}). $$
In the biquaternion-norm description this is the statement that the composition $H\circ N$ is concave along convex combinations of states. Since $\langle\cdot\rangle_{\natural}$ is a quadratic (and hence not affine) function of the state, the concavity of $H\circ N$ is not a trivial consequence of the concavity of $H$; it is the statement that mixing cannot decrease information, verified spectrum by spectrum. The maximally mixed state is the unique fixed point of the mixing map $\tilde{Ϛ}_{\Psi}\mapsto\tfrac12\tilde{Ϛ}_{\Psi}+\tfrac14 e_t$ and is the state of largest entropy.
$*$-Unitary invariance
Rotor conjugation by a matrix-unitary biquaternion, $\tilde{Ϛ}_{\Psi}\mapsto\tilde{U}\tilde{Ϛ}_{\Psi}\tilde{U}^{*}$ with $\tilde{U}\tilde{U}^{*} = e_t$, rotates the Bloch vector without changing its length. It therefore preserves $\langle\tilde{Ϛ}_{\Psi}\rangle_{\natural}$, the purity, the spectrum, and the entropy. This is the statement that the entropy is a $*$-unitary invariant, as it must be; in the framework it is the statement that $S$ is a function of the biquaternion norm, which the inner automorphisms of the algebra preserve.
Relation to the Standard Formula
The standard von Neumann entropy of a density operator is $S = -\operatorname{tr}(\rho\log\rho)$, where $\rho$ is a matrix and the symbol is its matrix trace, with eigenvalues $\lambda_i$. For a qubit with Bloch vector $\vec s_{\Psi}$, the eigenvalues are $\tfrac12(1\pm|\vec s_{\Psi}|)$ and the entropy is $h\!\left(\tfrac{1+|\vec s_{\Psi}|}{2}\right)$. This is exactly the formula derived above, so the framework's entropy is the standard entropy. Nothing here changes a numerical prediction.
What the biquaternion-norm interpretation adds is the location of the entropy's argument. In the matrix formulation the entropy is a function of the spectrum, and the spectrum is obtained by diagonalizing a matrix. In the biquaternion formulation the state carries a biquaternion norm, and on the normalised hyperplane the biquaternion norm is the deviation from idempotency; the entropy is a function of that scalar. The operative content is that:
- purity is an affine function of the biquaternion norm, $2\langle\tilde{Ϛ}_{\Psi},\tilde{Ϛ}_{\Psi}\rangle_{*} = 1-2\langle\tilde{Ϛ}_{\Psi}\rangle_{\natural}$;
- the mixedness scalar is exactly the biquaternion norm up to sign, $\tilde{Ϛ}_{\Psi}^2-\tilde{Ϛ}_{\Psi} = -\langle\tilde{Ϛ}_{\Psi}\rangle_{\natural}e_t$;
- the entropy is a monotone function of the position of the state in the cone of the biquaternion norm.
None of this is available in the same form for a general $d$-level system, where the biquaternion norm is not a single number; the qubit is special because its state space is the cone slice of a single quadratic form.
What the Biquaternion-Norm Description Does and Does Not Claim
What it does.
- It exhibits the entropy as a function of the biquaternion norm, $S = H(N)$, on the whole state space, with $H$ explicit.
- It identifies the pure states as the zero divisors of the biquaternion norm and the maximally mixed state as its maximum on the normalised hyperplane.
- It gives the algebraic identity $\tilde{Ϛ}_{\Psi}^2-\tilde{Ϛ}_{\Psi} = -\langle\tilde{Ϛ}_{\Psi}\rangle_{\natural}e_t$ relating mixedness to the biquaternion norm.
- It reproduces the standard qubit entropy and its bounds $0\leq S\leq\log2$.
What it does not.
- It does not derive the Shannon function $h$; the functional form of the entropy is a theorem of information theory, imported as standard.
- It does not extend the "entropy is a function of the biquaternion norm" statement beyond the qubit. For larger systems the biquaternion norm is not a single real number, and the argument fails.
- It does not by itself give the entropy of a reduced state of a composite system except through the reduction over one factor, which is an additional structure on $\mathbb{B}\otimes\mathbb{B}$ and is treated in the companion articles on entanglement.
- It does not supply data-processing inequalities, monotonicity under channels, or strong subadditivity; those are properties of the entropy under maps and are developed with the channel formalism.
Open Questions
1. Entropy from the algebra alone. The entropy is a function of the biquaternion norm, but the function $h$ is not algebraic. Is there a characterization of the entropy as the unique (up to scale) quantity on the cone slice that is monotone, concave, and $*$-unitary invariant? For the qubit this is plausible; a derivation from algebraic axioms alone has not been given.
2. The many-qubit case. For $n$ qubits the state space is the positive cone slice of $\mathbb{B}_{\Psi}^{\otimes n}$, and the biquaternion norm is replaced by a family of invariants. Which algebraic data determine the entropy there, and does the biquaternion norm of the tensor product play the role the single norm plays here?
3. Entropy and the Hermitian form. Entropy is a function of the biquaternion norm; the Born rule is a function of the Hermitian form, whose value is read at normalisation. Is there an algebraic relation between the two pairings that makes this division of labour inevitable, or is it an accident of the qubit?
4. Relative entropy. The quantum relative entropy $S(\rho\|\sigma)$ measures distinguishability and underlies the Holevo bound. Its biquaternion form, and whether it admits a biquaternion-norm expression in the qubit case, is the natural continuation of this article.
5. Empirical content. As elsewhere in the framework, the entropy in biquaternion form is the standard entropy transcribed. It predicts nothing new; its value is structural.
Summary
The state space of the qubit is the normalised hyperplane of the positive cone of the Hermitian subspace $\mathbb{B}_{\Psi}$, and the cone is defined by the algebra's biquaternion norm
$$ \langle\tilde H_\Psi\rangle_{\natural} = \langle\tilde H_\Psi,\tilde H_\Psi\rangle_{\natural} = h_t^2 - |\vec h_\Psi|^2 = \det M(\tilde H_\Psi), \qquad \langle\tilde{Ϛ}_{\Psi}\rangle_{\natural} = \tfrac{1}{4}\bigl(1-|\vec s_{\Psi}|^2\bigr)e_t . $$
The biquaternion norm of a state is a single real number taking values in $[0,\tfrac14]$ on the state space. The deviation of the state from idempotency is minus that number, $\tilde{Ϛ}_{\Psi}^2-\tilde{Ϛ}_{\Psi} = -\langle\tilde{Ϛ}_{\Psi}\rangle_{\natural}e_t$, and the purity is affine in it, $P(\tilde{Ϛ}_{\Psi}) = 2\langle\tilde{Ϛ}_{\Psi},\tilde{Ϛ}_{\Psi}\rangle_{*} = 1-2\langle\tilde{Ϛ}_{\Psi}\rangle_{\natural}$. The spectrum is $\lambda_\pm = \tfrac12(1\pm\sqrt{1-4N})$, so the von Neumann entropy is a function of the biquaternion norm and nothing else,
$$ S(\tilde{Ϛ}_{\Psi}) = h\!\left(\frac{1+\sqrt{1-4\langle\tilde{Ϛ}_{\Psi}\rangle_{\natural}}}{2}\right) = -2\,\langle\tilde{Ϛ}_{\Psi},\log\tilde{Ϛ}_{\Psi}\rangle_{*} , $$
with $S=0$ exactly on the pure states (the zero divisors of the biquaternion norm) and $S=\log2$ exactly at the maximally mixed state (the maximum of the biquaternion norm on the slice). Entropy is monotone in the biquaternion norm, concave, and invariant under the $*$-unitary inner automorphisms that preserve the biquaternion norm.
The description is the standard qubit entropy, and it predicts nothing new. What it makes precise is which invariant the entropy measures: not the Hermitian form of the Born rule but the biquaternion norm whose cone is the state space, and, on that cone slice, the departure of the state from the zero-divisor boundary.
Summary of Notation
| Symbol | Meaning |
|---|---|
| $\mathbb{B} = \mathbb{C}\otimes_\mathbb{R}\mathbb{H}$ | Biquaternion algebra, $\cong M_2(\mathbb{C})$ |
| $\mathbb{B}_{\Psi}$ | Hermitian subspace (states and observables) |
| $e_t = 1, \vec e_x, \vec e_y, \vec e_z$ | Quaternion units, $\vec e_\eta^2 = -e_t$ |
| $i$ | Central scalar imaginary, $i^2 = -1$ |
| $\tilde{Ϛ}_{\Psi} = \tfrac12(e_t + i\vec s_{\Psi})$ | State with Bloch vector $\vec s_{\Psi}$, $|\vec s_{\Psi}|\le1$ |
| $\tilde\Pi_{1,2}(\vec u_s) = \tfrac12(e_t\pm i\vec u_s)$ | Spectral idempotents of the state |
| $\lambda_\pm = \tfrac12(1\pm|\vec s_{\Psi}|)$ | Eigenvalues |
| $2\,\mathrm{Sc}(\tilde{Q})$ | Twice the scalar part |
| $\langle\tilde{Q}\rangle_{\natural} = \langle\tilde{Q},\tilde{Q}\rangle_{\natural} = \tilde{Q}\tilde{Q}^{\natural} = \sum_\mu Q_\mu^2$ | Biquaternion norm |
| $\langle\tilde H_\Psi\rangle_{\natural} = \langle\tilde H_\Psi,\tilde H_\Psi\rangle_{\natural} = h_t^2 - |\vec h_\Psi|^2 = \det M(\tilde H_\Psi)$ | Biquaternion norm of a Hermitian element |
| $\langle\tilde{Ϛ}_{\Psi}\rangle_{\natural} = \tfrac14(1-|\vec s_{\Psi}|^2)e_t$ | Biquaternion norm of a state |
| $\tilde{Ϛ}_{\Psi}^2 - \tilde{Ϛ}_{\Psi} = -\langle\tilde{Ϛ}_{\Psi}\rangle_{\natural}e_t$ | Deviation from idempotency |
| $P(\tilde{Ϛ}_{\Psi}) = 2\langle\tilde{Ϛ}_{\Psi},\tilde{Ϛ}_{\Psi}\rangle_{*} = 1-2\langle\tilde{Ϛ}_{\Psi}\rangle_{\natural}$ | Purity |
| $h(p) = -p\log p-(1-p)\log(1-p)$ | Binary entropy |
| $S(\tilde{Ϛ}_{\Psi}) = -2\langle\tilde{Ϛ}_{\Psi},\log\tilde{Ϛ}_{\Psi}\rangle_{*}$ | Von Neumann entropy |
| $S = H(N) = h\!\left(\tfrac{1+\sqrt{1-4N}}{2}\right)$ | Entropy as a function of the biquaternion norm |
| $0 \le S \le \log 2$ | Entropy bounds |
| $\langle\tilde{Q}_1,\tilde{Q}_2\rangle_{*}$ | the general plain sesquilinear form, $\mathrm{Sc}(\tilde{Q}_1\tilde{Q}_2^{*})$; on the diagonal $\langle\tilde{Q}_2,\tilde{Q}_2\rangle_{*}=\sum_\mu\lvert Q_{2\mu}\rvert^{2}$ |
| $\langle\tilde{Q}_1,\tilde{Q}_2\rangle$ | the general plain bilinear form, the scalar part of the general plain bilinear product, $\mathrm{Sc}(\tilde{Q}_1\tilde{Q}_2)$ |
Further Reading
- J. von Neumann, Mathematical Foundations of Quantum Mechanics (Princeton, 1955), for the original definition of the entropy of a density operator.
- A. Wehrl, "General properties of entropy," Reviews of Modern Physics 50 (1978) 221–260, for concavity, monotonicity, and the standard properties of the von Neumann entropy.
- Michael A. Nielsen and Isaac L. Chuang, Quantum Computation and Quantum Information (Cambridge, 2000), for the entropy of a qubit, the Bloch-ball parametrization, and the bounds $0\leq S\leq\log2$.
- A. Peres, Quantum Theory: Concepts and Methods (Kluwer, 1995), for the spectral decomposition of a qubit state and the geometry of the Bloch ball.
- I. Bengtsson and K. Życzkowski, Geometry of Quantum States (Cambridge, 2006), for the state space as a cone slice and for the geometric meaning of positivity and purity.
- M. Ohya and D. Petz, Quantum Entropy and Its Use (Springer, 1993), for the axiomatic characterization of entropy and the relative entropy.
- T. M. Cover and J. A. Thomas, Elements of Information Theory (Wiley, 2006), for the Shannon function $h$ and the source-coding interpretation of the entropy.
- The companion articles of this series: Quantum Physics in Biquaternionic Form, States, Observables and the Born Weight, and Decoherence as Idempotent Projection.