betaFeedback

WorksPhysics and Astronomy

Fer expansion of driven spins: closed-form recursion and no-go theorems for termination

Abstract

The Fer expansion factorizes the propagator of a driven quantum system into exponentials of time-integrated effective Hamiltonians, which govern the evolution left over by the previous factors. Every truncation is unitary, and adding a factor squares the size of the remainder, but each effective Hamiltonian is an infinite series of nested commutators of the previous one. For su(2)\mathfrak{su}(2), in every representation, from qubits to spins of any dimension, we represent the effective Hamiltonians by time-dependent vectors in R3\mathbb{R}^3 and resum this series exactly. The result is a closed map from one vector to the next, in which the infinite series reduces to the previous vector crossed once and twice with its time integral, weighted by elementary functions. Using this map, we address the question of when the expansion can be stopped after finitely many factors. This is possible whenever the leftover evolution is known in closed form, for instance when the effective Hamiltonian commutes with itself at all times or precesses about a fixed axis with constant angular velocity. We prove that, for real-analytic drives, neither happens beyond the first factor unless the first exponential is already exact. Stopping the expansion therefore yields no new exactly solvable drives.

Introduction

A spin driven along a time-dependent axis is the working model of magnetic resonance [1] and, for qubits, of most quantum-computing platforms [2]. Its Hamiltonian is a time-dependent combination of the generators of su(2), up to a multiple of the identity that only contributes a global phase. Hamiltonians of the same form, in other representations of the algebra, describe a spin of any size, an ensemble of two-level atoms driven collectively, and, through Schwinger’s oscillator model [3], two bosonic modes coupled by a time-dependent beam splitter. In these settings, designing pulse sequences, decoupling schemes, and gates requires the evolution operator U(t)U(t) generated by the Hamiltonian H(t)H(t), or an effective Hamiltonian that approximates it [4–6]. Formally, U(t)U(t) is given by the time-ordered exponential Texp⁡[−i∫0tds H(s)]T\exp[-\mathrm{i}\int_{0}^{t}ds\,H(s)], whose expansion in powers of HH is the Dyson series [7, 8]. This is a formal solution only, and truncating it breaks unitarity at every finite order [9].

Several methods avoid this problem by keeping the propagator unitary at every order. The best known is the Magnus expansion [9–12], which writes the propagator as a single exponential, U(t)=eΩ(t)U(t)=e^{\Omega(t)}, with Ω(t)\Omega(t) skew-Hermitian and built from nested commutators of H(t)H(t) at different times. Its convergence and error estimates are well charted [9, 13, 14], where Ω(t)\Omega(t) is an infinite series, in which each new term adds one more commutator of the true Hamiltonian.

A second route exploits the algebraic structure of H(t)H(t) instead of expanding in nested commutators. If H(t)=∑a=1dim⁡gua(t) TaH(t)=\sum_{a=1}^{\dim\mathfrak{g}}u_a(t)\,T_a, with real coefficients ua(t)u_a(t), is

built from Hermitian generators TaT_a that close under commutation into a finite-dimensional Lie algebra g\mathfrak{g}, the Wei–Norman theorem [15–17] factorizes the propagator over a basis of g\mathfrak{g}, locally in time, into the ordered product U(t)=∏a=1dim⁡ge−iξa(t)TaU(t)=\prod_{a=1}^{\dim\mathfrak{g}}e^{-\mathrm{i}\xi_a(t)T_a}, and turns the Schrödinger equation into a closed system of scalar ordinary differential equations (ODEs) for the ξa(t)\xi_a(t). Except for special cases such as solvable algebras [16], the ξa(t)\xi_a(t) can diverge at finite times [18], beyond which the factorization has to be restarted. The method has been revisited in tutorial form for Gaussian and non-Gaussian quantumoptical models [19], underlies unitary integration [20], and its equations and their integrability have been studied in detail [21, 22]; the same factorizations are behind the Lie-group integrators of numerical analysis [23].

The Fer expansion [24–26] lies between these two approaches and has been explored much less, although it has been used to build numerical integrators [27] and compared with the Magnus and Floquet expansions in solid-state NMR [28]. Like Magnus, it is built from nested commutators alone and needs no prior knowledge of the algebra generated by H(t)H(t); like Wei–Norman, it writes the propagator as a product of exponentials. The factors are peeled off one at a time, each generated by the time integral of an effective Hamiltonian Hj(t)H_j(t) that governs the evolution left over by the previous ones. Every truncation is therefore unitary, and what it leaves out is not an uncontrolled error but the evolution generated by the next effective Hamiltonian Hj+1(t)H_{j+1}(t). Each factor also squares the size of this remainder, so that, within the convergence domain of the expansion [25], recently enlarged by about 30% [26], the error decreases doubly exponentially with the number of factors [9]. The price is that each effective Hamiltonian is an infinite series of nested commutators of the previous one, which in general has to be truncated or evaluated numerically at every order.

First we show that, for su(2)\mathfrak{su}(2), the Fer expansion can nevertheless be made explicit at every order: the infinite series of nested commutators that defines each effective Hamiltonian can be summed exactly, in every representation of the algebra. Each effective Hamiltonian is a combination of the three generators of su(2)\mathfrak{su}(2) with time-dependent coefficients that form a vector in R3\mathbb{R}^3, and commutators become cross products of these vectors. One step of the recursion thus becomes an explicit map from each vector to the next: the new vector is the previous one crossed once and twice with its time integral, and the two terms are weighted by elementary trigonometric functions. Every factor of the Fer product is then an explicit rotation of the spin.

We then use the closed map to answer a question that the structure of the Fer product raises: can the product stop by itself? If at some step the effective Hamiltonian commutes with itself at all times, the next one vanishes identically, there is nothing left to peel off, and finitely many factors give the exact propagator. More generally, the product can be stopped as soon as the leftover evolution is known in closed form, for instance when the effective Hamiltonian precesses about a fixed axis with constant angular velocity, as in Rabi’s rotating field [29]. If a generic drive could be steered onto one of these two classes after finitely many steps, the Fer expansion would produce new exactly solvable drives, starting from the drive rather than from the solution.

Such drives would be valuable, since closed-form propagators are rare. For a qubit, the classic examples are Rabi’s rotating field [29] and the Landau–Zener sweep [30]. Larger families have been obtained by reverse-engineering the two-level Schrödinger equation [31, 32], or through the dynamical invariants of Lewis and Riesenfeld [33], which trade the linear problem for an auxiliary nonlinear equation, such as the Ermakov–Pinney equation of the time-dependent harmonic oscillator [34], and underpin the inverse engineering of shortcuts to adiabaticity [35]. In every case the drive is built from the solution, and not the other way around: for a drive given in advance, the propagator is in general not known in closed form.

Among the drives with a closed-form propagator, we consider two classes on which the Fer product can be stopped: effective Hamiltonians that commute with themselves at all times, and those that precess about a fixed axis with constant angular velocity, as in Rabi’s rotating field. Whether the product can reach either class at some order is not obvious in advance, since the recursion mixes the components of the drive nonlinearly at every step. We prove two no-go theorems showing that it cannot, except trivially. More precisely, for a real-analytic drive, an effective Hamiltonian beyond the first falls into either class only if it vanishes identically and the original drive already commutes with itself at all times, in which case the first exponential is already exact: the recursion creates no stopping point that was not already present in the original drive.

The paper is organized as follows. Section II recalls the Fer recursion, and Sec. III isolates what in it does not depend on the algebra, including the condition under which the product stops. Section IV derives the closed map and shows how it makes every order of the expansion explicit, Sec. V proves the two no-go theorems, and Sec. VI discusses where these results apply. The derivations are collected in Appendices A–C.

The Fer expansion

Throughout, we set ℏ=1\hbar=1 and a dot denotes the total time derivative d/dt\mathrm{d}/\mathrm{d}t. The evolution operator of a Hamiltonian H(t)H(t) obeys

ddtU(t)=−iH(t)U(t),U(0)=1.(1)\frac{\mathrm{d}}{\mathrm{d}t}U(t)=-\mathrm{i}H(t)U(t), \qquad U(0)=1 . \tag*{(1)}

Following Refs. [24, 25], and in the spirit of the average-Hamiltonian and Floquet–Magnus methods of magnetic resonance [36] [4] and of dynamical decoupling [5], one writes it as an ordered product of exponentials,

U(t)=eF1(t)eF2(t)⋯eFj(t)Uj+1(t),(2)U(t)=e^{F_1(t)}e^{F_2(t)}\cdots e^{F_j(t)}U_{j+1}(t) , \tag*{(2)}

in which the remainder obeys a Schrödinger equation of the same form,

ddtUj+1(t)=−iHj+1(t)Uj+1(t),(3)\frac{\mathrm{d}}{\mathrm{d}t}U_{j+1}(t)=-\mathrm{i}H_{j+1}(t)U_{j+1}(t) , \tag*{(3)}

with an effective Hamiltonian generated recursively by

Hj+1(t)=e−ad⁡Fj(t)Hj(t)−∫01dx e−xad⁡Fj(t)Hj(t),(4)H_{j+1}(t)=e^{-\operatorname{ad}_{F_j(t)}}H_j(t)-\int_0^1\mathrm{d}x\,e^{-x\operatorname{ad}_{F_j(t)}}H_j(t), \tag*{(4)}

where ad⁡Y(X)≡[Y,X]\operatorname{ad}_Y(X)\equiv[Y,X], conjugation by an exponential has been written as the exponential of the adjoint action (Appendix A), and the exponents are

Fj(t)=−i∫0tds Hj(s),(5)F_j(t)=-i\int_0^t\mathrm{d}s\,H_j(s), \tag*{(5)}

with H1(t)≡H(t)H_1(t)\equiv H(t). This is the defining step of the expansion: at order jj the integrated part of the dynamics is peeled off into eFj(t)e^{F_j(t)}, leaving a smaller remainder Hj+1(t)H_{j+1}(t) for the next step [24]. Each step squares the size of that remainder, which makes the expansion superconvergent, in the sense of the accelerated perturbation schemes of Refs. [37, 38]; its convergence domain is analyzed in Ref. [25], which for a Hermitian H(t)H(t) guarantees convergence while ∫0tds ∥H(s)∥<2\int_0^t\mathrm{d}s\,\lVert H(s)\rVert<2, with ∥⋅∥\lVert\cdot\rVert the operator norm; this radius was recently raised to about 2.6058 [26]. Expanding the two exponentials of ad⁡Fj(t)\operatorname{ad}_{F_j(t)} in Eq. (II.4) and doing the xx integral collapses it into a single series of nested commutators,

Hj+1(t)=∑n=0∞(−1)n+1(n+2)n!ad⁡Fj(t)n+1(Hj(t)),(6)H_{j+1}(t)=\sum_{n=0}^{\infty}\frac{(-1)^{n+1}}{(n+2)n!}\operatorname{ad}_{F_j(t)}^{n+1}\bigl(H_j(t)\bigr), \tag*{(6)}

with no auxiliary integrals left to perform.

The Fer map on a Lie algebra

Two features of the recursion hold for any algebra. We isolate them here, before specializing to su(2)\mathfrak{su}(2).

The recursion stays in the algebra

Let the initial Hamiltonian be a real combination of Hermitian generators {Ta}a=1dim⁡g\{T_a\}_{a=1}^{\dim\mathfrak{g}} of a Lie algebra g\mathfrak{g}, [Ta,Tb]=i∑cκabcTc[T_a,T_b]=i\sum_c\kappa_{abc}T_c,

H1(t)=h1(t)⋅T,h1(t)∈Rdim⁡g,(7)H_1(t)=\mathbf{h}_1(t)\cdot\mathbf{T},\qquad\mathbf{h}_1(t)\in\mathbb{R}^{\dim\mathfrak{g}}, \tag*{(7)}

and F1(t)=−if1(t)⋅TF_1(t)=-i\mathbf{f}_1(t)\cdot\mathbf{T}, with f1(t)=∫0tds h1(s)\mathbf{f}_1(t)=\int_0^t\mathrm{d}s\,\mathbf{h}_1(s). Every term of Eq. (II.6) is an iterated commutator of F1(t)F_1(t) with H1(t)H_1(t), so the next-order Hamiltonian is again a combination of the same generators, and by induction

Hj(t)=hj(t)⋅T,Fj(t)=−ifj(t)⋅T,(8)H_j(t)=\mathbf{h}_j(t)\cdot\mathbf{T},\qquad F_j(t)=-i\mathbf{f}_j(t)\cdot\mathbf{T}, \tag*{(8)}

for all j≥1j\ge1. The recursion never leaves g\mathfrak{g}; it only reshuffles the real coefficient vector. It therefore closes on the coefficient vectors themselves, taken as whole functions of time,

hj(t)⟼hj+1(t)=F[hj(t)],(9)\mathbf{h}_j(t)\longmapsto\mathbf{h}_{j+1}(t)=\mathcal{F}[\mathbf{h}_j(t)], \tag*{(9)}

with the Fer map F\mathcal{F} fixed by the structure constants κabc\kappa_{abc} of g\mathfrak{g} through Eq. (II.6). Iterating it generates the chain

h1(t)⟼h2(t)⟼⋯⟼hj(t)(10)\mathbf{h}_1(t)\longmapsto\mathbf{h}_2(t)\longmapsto\cdots\longmapsto\mathbf{h}_j(t) \tag*{(10)}

to any desired order. The Fer map is not invertible in general: if Hj(t)H_j(t) commutes with itself at all times, it also commutes with its own integral Fj(t)F_j(t), every nested commutator in Eq. (II.6) vanishes, and Hj+1(t)=0H_{j+1}(t)=0, so that F\mathcal{F} sends all the drives of this kind to the zero drive.

Where the chain stops

Whatever the algebra, the evolution generated by H(t)=h(t)⋅TH(t)=\mathbf{h}(t)\cdot\mathbf{T} is solvable in closed form whenever the Hamiltonian commutes with itself at all times,

[H(t),H(t′)]=0for all t≠t′.(11)[H(t),H(t')]=0\qquad\text{for all }t\ne t'. \tag*{(11)}

We call family A\mathcal{A} the class of drives obeying Eq. (III.5), and we say that the coefficient vector h(t)\mathbf{h}(t) belongs to A\mathcal{A} when its Hamiltonian does.

Every Fer chain that reaches family A\mathcal{A} stops there, as anticipated in Sec. III A. Indeed, if hj(t)∈A\mathbf{h}_j(t)\in\mathcal{A}, the Hamiltonian Hj(t)H_j(t) commutes with its own integral, Eq. (II.5),

[Hj(t),Fj(t)]=−i∫0tds [Hj(t),Hj(s)]=0,(12)[H_j(t),F_j(t)]=-i\int_0^t\mathrm{d}s\,[H_j(t),H_j(s)]=0, \tag*{(12)}

because the integrand vanishes identically by Eq. (III.5). The innermost commutator of every term of Eq. (II.6) is ad⁡Fj(t)(Hj(t))=−[Hj(t),Fj(t)]\operatorname{ad}_{F_j(t)}(H_j(t))=-[H_j(t),F_j(t)], so the whole series vanishes term by term and

F(A)=O,O≡{0}.(13)\mathcal{F}(\mathcal{A})=\mathcal{O},\qquad\mathcal{O}\equiv\{\mathbf{0}\}. \tag*{(13)}

Reaching A\mathcal{A} at order jj means that the factorization closes there: Hj+1(t)=0H_{j+1}(t)=0 and U(t)=eF1(t)⋯eFj(t)U(t)=e^{F_1(t)}\cdots e^{F_j(t)} is exact, with no remainder left. Since O∈A\mathcal{O}\in\mathcal{A}, one has O⊆A\mathcal{O}\subseteq\mathcal{A}, and the chain stays at O\mathcal{O} forever after.

The question of § 1 is now precise: for which h1(t)∉A\mathbf{h}_1(t)\notin\mathcal{A} does the map of Eq. (III.3) reach A\mathcal{A} after finitely many steps? We exclude h1(t)∈A\mathbf{h}_1(t)\in\mathcal{A} because it is the trivial case: there [H1(t),F1(t)]=0[H_1(t),F_1(t)]=0 already at the first step, so Hj(t)=0H_j(t)=0 for all j>1j>1 and the expansion stops after a single factor, U(t)=eF1(t)U(t)=e^{F_1(t)}, which is the time-ordered exponential of § 1 with the time ordering dropped. Answering the question means going back along the chain, one step at a time: finding the drives that one step sends into A\mathcal{A}.

The Fer map on su(2)

For su(2)\mathfrak{su}(2) the map can be written down in closed form. A component of the Hamiltonian along the identity commutes with everything: it drops out of every commutator in Eq. (II.6) and only contributes a global phase to eF1(t)e^{F_1(t)}, so we take H(t)H(t) traceless. We take T\mathbf{T} to be a vector of Hermitian matrices of any dimension, obeying [3]

[Ta,Tb]=2iϵabcTc,(14)[T_a,T_b] = 2i\epsilon_{abc}T_c , \tag*{(14)}

that is, twice the spin operators of any nontrivial representation of su(2)\mathit{su}(2), for instance the Pauli matrices for a qubit, and write, as in Eq. (III.2),

Hj(t)=hj(t)⋅T,Fj(t)=−ifj(t)⋅T,(15)H_j(t) = \mathbf{h}_j(t)\cdot\mathbf{T}, \qquad F_j(t) = -i\mathbf{f}_j(t)\cdot\mathbf{T}, \tag*{(15)}

with hj(t)\mathbf{h}_j(t) a real three-dimensional vector, the drive vector about which the mean spin precesses, as the Bloch vector of a qubit in the representation of Feynman, Vernon, and Hellwarth [39], and fj(t)=∫0tds hj(s)\mathbf{f}_j(t) = \int_0^t \mathrm{d}s\,\mathbf{h}_j(s). The norms of the two vectors are abbreviated as fj(t)≡∥fj(t)∥f_j(t) \equiv\lVert\mathbf{f}_j(t)\rVert and hj(t)≡∥hj(t)∥h_j(t) \equiv\lVert\mathbf{h}_j(t)\rVert, with ∥⋅∥\lVert\cdot\rVert the Euclidean norm. By construction

f˙j(t)=hj(t),fj(0)=0.(16)\dot{\mathbf{f}}_j(t) = \mathbf{h}_j(t), \qquad\mathbf{f}_j(0) = 0 . \tag*{(16)}

Since the commutator is bilinear, Eq. (IV.1) turns the commutator of two combinations of the TaT_a into the cross product of their coefficient vectors,

[fj(t)⋅T,hj(t)⋅T]=2iΛfj(t)(hj(t))⋅T,(17)[\mathbf{f}_j(t)\cdot\mathbf{T},\mathbf{h}_j(t)\cdot\mathbf{T}] = 2i\Lambda_{\mathbf{f}_j(t)}(\mathbf{h}_j(t))\cdot\mathbf{T}, \tag*{(17)}

where Λx(y)≡x×y\Lambda_{\mathbf{x}}(\mathbf{y}) \equiv\mathbf{x}\times\mathbf{y}, so that the operator series of Eq. (II.6) becomes a recursion entirely in R3\mathbb{R}^3,

hj+1(t)=∑n=0∞ΦnΛfj(t)n(hj(t)),(18)\mathbf{h}_{j+1}(t) = \sum_{n=0}^{\infty}\Phi_n\Lambda_{\mathbf{f}_j(t)}^n(\mathbf{h}_j(t)), \tag*{(18)}

with numerical coefficients Φn=(−2)n+1/[(n+2)n!]\Phi_n = (-2)^{n+1}/[(n+2)n!]. This series resums into elementary functions: summing its odd and even powers separately gives the closed map

hj+1(t)=α(fj(t)) pj+1(t)+β(fj(t)) qj+1(t),(19)\mathbf{h}_{j+1}(t) = \alpha(f_j(t))\,\mathbf{p}_{j+1}(t) + \beta(f_j(t))\,\mathbf{q}_{j+1}(t), \tag*{(19)}

built from the two auxiliary vectors

pj+1(t)≡fj(t)×hj(t),qj+1(t)≡pj+1(t)×fj(t),(20)\mathbf{p}_{j+1}(t) \equiv\mathbf{f}_j(t)\times\mathbf{h}_j(t), \qquad \mathbf{q}_{j+1}(t) \equiv\mathbf{p}_{j+1}(t)\times\mathbf{f}_j(t), \tag*{(20)}

and weighted by the two scalar functions

α(x)=1−cos⁡(2x)2x2−sin⁡(2x)x,(21)\alpha(x) = \frac{1-\cos(2x)}{2x^2} - \frac{\sin(2x)}{x}, \tag*{(21)}
β(x)=cos⁡(2x)x2−sin⁡(2x)2x3.(22)\beta(x) = \frac{\cos(2x)}{x^2} - \frac{\sin(2x)}{2x^3}. \tag*{(22)}

Equation (IV.6) is the closed-form realization of the abstract map of Eq. (III.3) for su(2)\mathit{su}(2): it is what we iterate from now on. Appendix B derives it twice, once by resumming the series of Eq. (IV.5) and once, independently, by inserting the Rodrigues rotation formula [3] into the recursion of Eq. (II.4) and integrating. Each factor of the product of Eq. (II.2) is an explicit rotation: eFj(t)=e−ifj(t)⋅Te^{F_j(t)} = e^{-i\mathbf{f}_j(t)\cdot\mathbf{T}} rotates the spin by the angle 2fj(t)2f_j(t) about the unit vector f^j(t)≡fj(t)/fj(t)\hat{\mathbf{f}}_j(t) \equiv\mathbf{f}_j(t)/f_j(t), and in any representation it is given by the corresponding rotation matrix [3]. For a qubit, with T=σ\mathbf{T}=\boldsymbol{\sigma} the vector of Pauli matrices, eFj(t)=cos⁡fj(t)1−isin⁡fj(t)f^j(t)⋅σe^{F_j(t)} = \cos f_j(t)\mathbf{1} - i\sin f_j(t)\hat{\mathbf{f}}_j(t)\cdot\boldsymbol{\sigma}.

Two properties of the map will be used repeatedly. First, both auxiliary vectors lie in the plane Πj(t)\Pi_j(t) orthogonal to fj(t)\mathbf{f}_j(t), Figure 1: pj+1(t)\mathbf{p}_{j+1}(t) is orthogonal to both fj(t)\mathbf{f}_j(t) and hj(t)\mathbf{h}_j(t), and qj+1(t)\mathbf{q}_{j+1}(t) is fj2(t)f_j^2(t) times the projection of hj(t)\mathbf{h}_j(t) onto Πj(t)\Pi_j(t). Hence the whole output lies in Πj(t)\Pi_j(t),

Figure 1. One step of the Fer map on su(2), Eq. (IV.6), drawn in the space R3\mathbb{R}^{3} of the coefficient vectors. The integrated drive fj(t)\mathbf{f}_j(t), in green, is normal to the plane Πj(t)\Pi_j(t), in red, through the origin OO. The auxiliary vectors pj+1(t)=fj(t)×hj(t)\mathbf{p}_{j+1}(t)=\mathbf{f}_j(t)\times\mathbf{h}_j(t), in orange, and qj+1(t)=pj+1(t)×fj(t)\mathbf{q}_{j+1}(t)=\mathbf{p}_{j+1}(t)\times\mathbf{f}_j(t), in purple, span Πj(t)\Pi_j(t); the latter lies along the projection of the drive hj(t)\mathbf{h}_j(t), in brown, onto the plane (dotted parallelogram). The output hj+1(t)=α(fj(t))pj+1(t)+β(fj(t))qj+1(t)\mathbf{h}_{j+1}(t)=\alpha(\mathbf{f}_j(t))\mathbf{p}_{j+1}(t)+\beta(\mathbf{f}_j(t))\mathbf{q}_{j+1}(t), in blue, is assembled from them (dashed parallelogram) with the weights α\alpha and β\beta of Eqs. (IV.8) and (IV.9), both evaluated at the norm ∥fj(t)∥\lVert\mathbf{f}_j(t)\rVert, and never leaves Πj(t)\Pi_j(t).

hj+1(t)⊥fj(t).(23)\mathbf{h}_{j+1}(t)\perp\mathbf{f}_j(t). \tag*{(23)}

Second, fj(0)=0\mathbf{f}_j(0)=\mathbf{0} makes both auxiliary vectors vanish at the origin, so every Fer output starts there,

hj+1(0)=0,(24)\mathbf{h}_{j+1}(0)=\mathbf{0}, \tag*{(24)}

a constraint that the initial Hamiltonian h1(t)\mathbf{h}_1(t) does not carry.

Rigidity of the su(2)\mathit{su}(2) chain

With the map available in closed form, the question of Sec. III B can be answered by inspection of a single backward step. We do it first for family A\mathcal{A}, the drives that commute with themselves at all times, and then for the uniformly precessing drives of Rabi’s rotating field, a textbook solvable class outside A\mathcal{A}. In both cases the answer is negative: the recursion reaches the class at order j>1j>1 only when the original drive was already in A\mathcal{A}, and then only through the zero drive.

Throughout this section the original drive h1(t)\mathbf{h}_1(t) is taken to be real-analytic in tt. The weights α\alpha and β\beta of Eqs. (IV.8) and (IV.9) are even, and their singularities at the origin are removable, hence they are analytic functions of fj(t)⋅fj(t)\mathbf{f}_j(t)\cdot\mathbf{f}_j(t), so every hj(t)\mathbf{h}_j(t) produced by Eq. (IV.6) is real-analytic as well.

Drives with a fixed direction

Read through the commutator of Eq. (IV.4), the defining condition of family A\mathcal{A}, Eq. (III.5), becomes

h(t)×h(t′)=0for all t≠t′,(25)\mathbf{h}(t)\times\mathbf{h}(t')=\mathbf{0}\qquad\text{for all }t\ne t', \tag*{(25)}

i.e., the trajectory of h(t)\mathbf{h}(t) never leaves a fixed direction in R3\mathbb{R}^3: it moves along a single line through the origin. Since F(A)=O⊆A\mathcal{F}(\mathcal{A})=\mathcal{O}\subseteq\mathcal{A} by Eq. (III.7), the following theorem amounts to a single step back along the chain: a drive that one step sends into A\mathcal{A} was already in A\mathcal{A}.

Theorem 1. Let the original drive h1(t)\mathbf{h}_1(t) be real-analytic. In the Fer expansion associated with the su(2)\mathit{su}(2) algebra, if the coefficient vector hj(t)\mathbf{h}_j(t) at some order j>1j>1 belongs to family A\mathcal{A}, then hj(t)=0\mathbf{h}_j(t)=\mathbf{0}, and the coefficient vector h1(t)\mathbf{h}_1(t) of the original Hamiltonian H1(t)=h1(t)⋅TH_1(t)=\mathbf{h}_1(t)\cdot\mathbf{T} belongs to family A\mathcal{A} as well.

Proof of Theorem 1. We prove the one-step implication, valid for every j≥1j \ge1, from hj+1(t)∈Ah_{j+1}(t) \in\mathcal{A} to hj+1(t)≡0h_{j+1}(t) \equiv0 and hj(t)∈Ah_j(t) \in\mathcal{A}, and conclude by descending induction along the chain of Eq. (III.4).

Assume hj+1(t)∈Ah_{j+1}(t) \in\mathcal{A}: both hj+1(t)h_{j+1}(t) and h˙j+1(t)\dot{h}_{j+1}(t) lie on a fixed line through the origin. By Eq. (IV.10), fj(t)f_j(t) is orthogonal to this line wherever hj+1(t)≠0h_{j+1}(t) \ne0, hence at all times unless hj+1(t)≡0h_{j+1}(t) \equiv0, since the zeros of a real-analytic function are isolated. Either way hj+1(t)⊥fj(t)h_{j+1}(t) \perp f_j(t). Differentiating Eq. (IV.6) and projecting on fj(t)f_j(t) (Appendix C), every term is a cross product with fj(t)f_j(t) and drops out, except one of the two generated by q˙j+1(t)\dot{q}_{j+1}(t). What survives is the necessary condition

β(fj(t))fj(t)⋅[pj+1(t)×hj(t)]=0.(26)\beta(f_j(t)) f_j(t) \cdot\left[p_{j+1}(t) \times h_j(t)\right] = 0 . \tag*{(26)}

By the triple product identity and the Lagrange identity,

fj(t)⋅[pj+1(t)×hj(t)]=(fj(t)⋅hj(t))2−fj2(t)hj2(t)=−∥fj(t)×hj(t)∥2,(27)f_j(t) \cdot\left[p_{j+1}(t) \times h_j(t)\right] = \left(f_j(t) \cdot h_j(t)\right)^2 - f_j^2(t) h_j^2(t) = -\lVert f_j(t) \times h_j(t) \rVert^2 , \tag*{(27)}

so that Eq. (V.2) reads

β(fj(t))∥fj(t)×hj(t)∥2=0.(28)\beta(f_j(t)) \lVert f_j(t) \times h_j(t) \rVert^2 = 0 . \tag*{(28)}

At every instant this holds either because β(fj(t))=0\beta(f_j(t)) = 0 or because fj(t)×hj(t)=0f_j(t) \times h_j(t) = 0, and the first alternative can only occur at isolated times. Indeed, as noted at the beginning of this section, the composition t↦β(fj(t))t \mapsto\beta(f_j(t)) is real-analytic, even though the norm fj(t)f_j(t) itself need not be where fj(t)f_j(t) vanishes. Near the origin,

β(x)=−43+815x2+O(x4),β(0)=−43≠0,(29)\beta(x) = -\frac{4}{3} + \frac{8}{15}x^2 + O(x^4), \qquad\beta(0) = -\frac{4}{3} \ne0 , \tag*{(29)}

and since fj(0)=0f_j(0) = 0, the composition equals β(0)=−4/3\beta(0) = -4/3 at t=0t = 0, so it is not identically zero, and its zeros are isolated. Equation (V.4) then forces fj(t)×hj(t)=0f_j(t) \times h_j(t) = 0 on the complement of a discrete set, which is dense, and by continuity everywhere:

hj(t)∥fj(t)for all t,(30)h_j(t) \parallel f_j(t) \qquad\text{for all } t , \tag*{(30)}

that is, fj(t)×hj(t)≡0f_j(t) \times h_j(t) \equiv0.

Two consequences follow. First, both auxiliary vectors of Eq. (IV.7) vanish identically, and with them the output of Eq. (IV.6): hj+1(t)≡0h_{j+1}(t) \equiv0. Second, since f˙j(t)=hj(t)\dot{f}_j(t) = h_j(t), Eq. (V.6) reads fj(t)×f˙j(t)≡0f_j(t) \times\dot{f}_j(t) \equiv0. If fj(t)≡0f_j(t) \equiv0, then hj(t)≡0h_j(t) \equiv0. Otherwise, on any interval where fj(t)≠0f_j(t) \ne0, its derivative is parallel to it and its direction does not change; being real-analytic, fj(t)f_j(t) then stays on that line at all times, and so does hj(t)=f˙j(t)h_j(t) = \dot{f}_j(t). This is hj(t)∈Ah_j(t) \in\mathcal{A}.

Remark. The proof never inverts F\mathcal{F}, which in general cannot be done (Sec. III A). It does not reconstruct hj(t)h_j(t) from hj+1(t)h_{j+1}(t): it takes any drive hj(t)h_j(t) that one step sends into A\mathcal{A} and constrains it using only the forward map, Eq. (IV.6).

Together with Eq. (III.7), Theorem 1 says that the drives that one step sends into A\mathcal{A}, or into O\mathcal{O}, are exactly those already in A\mathcal{A}; in particular, no real-analytic drive outside A\mathcal{A} produces, at any later order, an effective Hamiltonian that commutes with itself at all times.

Precessing drives

Family A\mathcal{A} is not the only solvable class in su(2)\mathfrak{su}(2). After jj steps the propagator reads, by Eq. (II.2), U(t)=eF1(t)⋯eFj(t)Uj+1(t)U(t) = e^{F_1(t)} \cdots e^{F_j(t)} U_{j+1}(t), with Uj+1(t)U_{j+1}(t) generated by Hj+1(t)H_{j+1}(t). The product terminates when Hj+1(t)H_{j+1}(t) vanishes, which involves only family A\mathcal{A}, Eq. (III.7), and is settled by Theorem 1. The expansion can be stopped just as well whenever Hj+1(t)H_{j+1}(t) belongs to a class whose dynamics is solved exactly: Uj+1(t)U_{j+1}(t), and with it the whole propagator, is then known in closed form. Here we focus on the best-known such class, Rabi’s rotating field [29], which we call family A~\widetilde{\mathcal{A}}: it collects the drives whose vector precesses about a fixed unit axis n^\hat{\mathbf{n}} with constant angular velocity ω\omega,

h(t)=cos⁡(ωt)h(0)+sin⁡(ωt)(h(0)×n^)+[1−cos⁡(ωt)](n^⋅h(0))n^.(31)\mathbf{h}(t) = \cos(\omega t)\mathbf{h}(0) + \sin(\omega t)\left(\mathbf{h}(0) \times\hat{\mathbf{n}}\right) + \left[1-\cos(\omega t)\right]\left(\hat{\mathbf{n}}\cdot\mathbf{h}(0)\right)\hat{\mathbf{n}} . \tag*{(31)}

Despite the richer time dependence, these drives are solvable too: in the frame rotating about n^\hat{\mathbf{n}}, the propagator generated by H(t)=h(t)⋅TH(t) = \mathbf{h}(t)\cdot\mathbf{T} factorizes as

U(t)=ei(ωt/2)n^⋅T U~(t),(32)U(t) = e^{i(\omega t/2)\hat{\mathbf{n}}\cdot\mathbf{T}\,}\widetilde{U}(t), \tag*{(32)}

where U~\widetilde{U} is generated by the time-independent Hamiltonian H~=(h(0)+(ω/2)n^)⋅T\widetilde{H} = \left(\mathbf{h}(0) + (\omega/2)\hat{\mathbf{n}}\right)\cdot\mathbf{T}. Yet the chain cannot reach A~\widetilde{\mathcal{A}} either.

Theorem 2. Let the original drive h1(t)h_1(t) be real-analytic. In the Fer expansion associated to the su(2)\mathfrak{su}(2) algebra, if the coefficient vector hj(t)h_j(t) at some order j>1j > 1 belongs to family A\mathcal{A}, then hj(t)≡0h_j(t) \equiv0, and the coefficient vector h1(t)h_1(t) of the original Hamiltonian H1(t)=h1(t)⋅TH_1(t) = h_1(t)\cdot\mathbf{T} belongs to family A\mathcal{A}.

Proof of Theorem 2. Since j>1j > 1, hj(t)h_j(t) is a genuine Fer output and therefore vanishes at t=0t = 0 by Eq. (IV.11). A member of A~\widetilde{\mathcal{A}} has constant magnitude ∥h(t)∥=∥h(0)∥\lVert h(t)\rVert= \lVert h(0)\rVert by Eq. (V.7), so one that vanishes at the origin vanishes identically. Hence hj(t)≡0∈Ah_j(t) \equiv0 \in\mathcal{A}, and Theorem 1 gives h1(t)∈Ah_1(t) \in\mathcal{A}.

Conclusion

For a spin of any size, and in every representation of su(2)\mathfrak{su}(2), we have summed the series of nested commutators that defines each step of the Fer expansion into a closed map on time-dependent vectors in R3\mathbb{R}^3, with two elementary trigonometric weights. Each factor of the product follows from the previous one through a single time integral and elementary functions of the norm of the previous integral, fj(t)=∥fj(t)∥f_j(t) = \lVert\mathbf{f}_j(t)\rVert. The known strengths of the expansion thus become directly usable. The product is unitary at every truncation, and each factor squares the size of the remainder, so that the error decreases doubly exponentially with the number of factors [9], within a convergence domain that has recently been enlarged by about 30% [26]. A few factors therefore give an accurate propagator for any spin drive, such as the pulse sequences of magnetic resonance [4] and dynamical decoupling [5], or the control pulses that implement quantum gates [6].

With the closed form obtained here and the recently enlarged convergence domain, the Fer product is thus well suited also to quantum algorithms, in which product formulas are among the most widely used methods to simulate time evolution [40, 41]: a truncated Fer product is a product formula of explicit rotations, which can be applied directly as gates.

We have also used the closed map to ask whether the Fer product can be stopped after finitely many factors on one of two classes of effective Hamiltonians whose evolution is known in closed form: those that commute with themselves at all times, on which the product terminates, and those that precess about a fixed axis with constant angular velocity, as in Rabi’s rotating field. For a real-analytic drive, no effective Hamiltonian beyond the first falls into either class unless the first exponential is already exact. In su(2)\mathfrak{su}(2), exactly solvable drives therefore still have to be engineered backward, from the solution to the drive [31, 32, 35], and whether the chain can reach other solvable classes remains open. Running the Fer map until the effective Hamiltonian becomes exactly solvable is nonetheless a direct method to search for solvable dynamics, since it starts from the drive rather than from the solution. For the families considered in this work, it fails for su(2)\mathfrak{su}(2), but the first part of our analysis, in Sec. III, holds for any finite-dimensional Lie algebra, so the method can be applied to other algebras, for instance those that describe squeezing and parametric amplification in quantum optics [19], where the answer may differ.

Acknowledgments

The authors thank Riccardo Castellano (Department of Applied Physics, University of Geneva) for helpful discussions on the presentation of the introduction and conclusions.

Use of AI

The authors used Claude Code (Anthropic) to assist with spelling corrections, graphical refinement of figure 1, and searches for relevant literature. The scientific content, including the theorems and their proofs, is the authors’ own. The authors reviewed the output and take full responsibility for the manuscript.

Data availability

No data were created or analyzed in this article.

Conjugation formula

Consider two operators X,YX,Y and the one-parameter family g(λ)≡eλYXe−λYg(\lambda) \equiv e^{\lambda Y}Xe^{-\lambda Y} with λ∈R\lambda\in\mathbb{R}. Differentiating nn times gives

dndλng(λ)=eλYad⁡Yn(X)e−λY,(33)\frac{\mathrm{d}^{n}}{\mathrm{d}\lambda^{n}}g(\lambda)=e^{\lambda Y}\operatorname{ad}_{Y}^{n}(X)e^{-\lambda Y}, \tag*{(33)}

as follows by induction, since one further derivative produces one further commutator with YY between the same two exponentials. Expanding g(λ)g(\lambda) in a Taylor series around λ=0\lambda=0 and evaluating it at λ=γ\lambda=\gamma then yields

eγYXe−γY=∑n=0∞γnn!ad⁡Yn(X),(34)e^{\gamma Y}Xe^{-\gamma Y}=\sum_{n=0}^{\infty}\frac{\gamma^{n}}{n!}\operatorname{ad}_{Y}^{n}(X), \tag*{(34)}

valid for arbitrary X,YX,Y. This is what allows conjugation by an exponential to be written as the exponential of the adjoint action, in the recursion of Eq. (II.4) as in its double-integral form, Eq. (B.7).

We now translate Eq. (A.2) into vector form, with X=x⋅TX=\mathbf{x}\cdot\mathbf{T} and Y=iy⋅TY=i\mathbf{y}\cdot\mathbf{T} for real x,y\mathbf{x},\mathbf{y} and Hermitian generators T\mathbf{T}. By the su(2)\mathfrak{su}(2) commutator of Eq. (IV.1), a single application of ad⁡Y\operatorname{ad}_{Y} gives

ad⁡Y(X)=i[y⋅T,x⋅T]=i⋅2iΛy(x)⋅T=−2Λy(x)⋅T.(35)\operatorname{ad}_{Y}(X)=i[\mathbf{y}\cdot\mathbf{T},\mathbf{x}\cdot\mathbf{T}]=i\cdot2i\Lambda_{\mathbf{y}}(\mathbf{x})\cdot\mathbf{T}=-2\Lambda_{\mathbf{y}}(\mathbf{x})\cdot\mathbf{T}. \tag*{(35)}

The result is again of the form z⋅T\mathbf{z}\cdot\mathbf{T}, so the same computation iterates: by linearity of Λy\Lambda_{\mathbf{y}}, each further application of ad⁡Y\operatorname{ad}_{Y} contributes one factor of −2-2 and one more power of Λy\Lambda_{\mathbf{y}}, whence ad⁡Yn(X)=(−2)nΛyn(x)⋅T\operatorname{ad}_{Y}^{n}(X)=(-2)^{n}\Lambda_{\mathbf{y}}^{n}(\mathbf{x})\cdot\mathbf{T} by induction. Substituting into Eq. (A.2) gives

eiγy⋅T(x⋅T)e−iγy⋅T=∑n=0∞(−2γ)nn!Λyn(x)⋅T,Λy(x)≡y×x.(36)e^{i\gamma\mathbf{y}\cdot\mathbf{T}}(\mathbf{x}\cdot\mathbf{T})e^{-i\gamma\mathbf{y}\cdot\mathbf{T}}=\sum_{n=0}^{\infty}\frac{(-2\gamma)^{n}}{n!}\Lambda_{\mathbf{y}}^{n}(\mathbf{x})\cdot\mathbf{T},\qquad\Lambda_{\mathbf{y}}(\mathbf{x})\equiv\mathbf{y}\times\mathbf{x}. \tag*{(36)}

Closed form of the Fer map on su(2)\mathfrak{su}(2)

Here we derive Eq. (IV.6) in two independent ways.

From the series

The key structural fact is that Λfj(t)\Lambda_{\mathbf{f}_j(t)} generates only two directions out of hj(t)\mathbf{h}_j(t): every power Λfj(t)n(hj(t))\Lambda_{\mathbf{f}_j(t)}^n(\mathbf{h}_j(t)) with n≥1n \ge1 is a multiple of either pj+1(t)=fj(t)×hj(t)\mathbf{p}_{j+1}(t) = \mathbf{f}_j(t) \times\mathbf{h}_j(t) or qj+1(t)=(fj(t)×hj(t))×fj(t)\mathbf{q}_{j+1}(t) = (\mathbf{f}_j(t) \times\mathbf{h}_j(t)) \times\mathbf{f}_j(t). Indeed, from Λfj(t)(hj(t))=fj(t)×hj(t)\Lambda_{\mathbf{f}_j(t)}(\mathbf{h}_j(t)) = \mathbf{f}_j(t) \times\mathbf{h}_j(t), a second application and the BAC–CAB identity give

Λfj(t)2(hj(t))=fj(t)×(fj(t)×hj(t))=fj(t)(fj(t)⋅hj(t))−fj2(t)hj(t) ,(37)\Lambda_{\mathbf{f}_j(t)}^2(\mathbf{h}_j(t)) = \mathbf{f}_j(t) \times(\mathbf{f}_j(t) \times\mathbf{h}_j(t)) = \mathbf{f}_j(t)(\mathbf{f}_j(t) \cdot\mathbf{h}_j(t)) - f_j^2(t)\mathbf{h}_j(t) \, , \tag*{(37)}

and a third, using fj(t)×fj(t)=0\mathbf{f}_j(t) \times\mathbf{f}_j(t) = 0,

Λfj(t)3(hj(t))=(fj(t)⋅hj(t))(fj(t)×fj(t))−fj2(t)(fj(t)×hj(t))=−fj2(t)(fj(t)×hj(t)) .(38)\Lambda_{\mathbf{f}_j(t)}^3(\mathbf{h}_j(t)) = (\mathbf{f}_j(t) \cdot\mathbf{h}_j(t))(\mathbf{f}_j(t) \times\mathbf{f}_j(t)) - f_j^2(t)(\mathbf{f}_j(t) \times\mathbf{h}_j(t)) = -f_j^2(t)(\mathbf{f}_j(t) \times\mathbf{h}_j(t)) \, . \tag*{(38)}

Each further application swaps the two structures, and every second one brings a factor −fj2(t)-f_j^2(t), so the pattern closes at all orders,

Λfj(t)2m+1(hj(t))=[−fj2(t)]mpj+1(t) ,(39)\Lambda_{\mathbf{f}_j(t)}^{2m+1}(\mathbf{h}_j(t)) = [-f_j^2(t)]^m \mathbf{p}_{j+1}(t) \, , \tag*{(39)}
Λfj(t)2m+2(hj(t))=[−fj2(t)]mfj(t)×pj+1(t)=−[−fj2(t)]mqj+1(t) ,(40)\Lambda_{\mathbf{f}_j(t)}^{2m+2}(\mathbf{h}_j(t)) = [-f_j^2(t)]^m \mathbf{f}_j(t) \times\mathbf{p}_{j+1}(t) = -[-f_j^2(t)]^m \mathbf{q}_{j+1}(t) \, , \tag*{(40)}

with pj+1(t)\mathbf{p}_{j+1}(t) and qj+1(t)\mathbf{q}_{j+1}(t) the auxiliary vectors of Eq. (IV.7). Splitting the series of Eq. (IV.5) into even and odd parts and inserting Eqs. (B.3) and (B.4),

hj+1(t)=(∑m=0∞Φ2m[−fj2(t)]m)pj+1(t)−(∑m=0∞Φ2m+1[−fj2(t)]m)qj+1(t) ,(41)\mathbf{h}_{j+1}(t) = \left(\sum_{m=0}^{\infty}\Phi_{2m}[-f_j^2(t)]^m\right)\mathbf{p}_{j+1}(t) - \left(\sum_{m=0}^{\infty}\Phi_{2m+1}[-f_j^2(t)]^m\right)\mathbf{q}_{j+1}(t) \, , \tag*{(41)}

and the two sums are precisely the Taylor series of the elementary functions of Eqs. (IV.8) and (IV.9),

∑m=0∞Φ2m[−fj2(t)]m=α(fj(t)) ,−∑m=0∞Φ2m+1[−fj2(t)]m=β(fj(t)) ,(42)\sum_{m=0}^{\infty}\Phi_{2m}[-f_j^2(t)]^m = \alpha(f_j(t)) \, , \qquad-\sum_{m=0}^{\infty}\Phi_{2m+1}[-f_j^2(t)]^m = \beta(f_j(t)) \, , \tag*{(42)}

which is Eq. (IV.6). In particular α\alpha and β\beta are the closed-form sums of the even and odd subseries of the recursion, and the apparent singularities at fj(t)=0f_j(t) = 0 are removable.

From the Rodrigues formula

The same result follows from the recursion of Eq. (II.4) without expanding in commutators. Written as a double integral, equivalent to Eq. (II.4) since both reduce to the series of Eq. (II.6), the recursion reads

Hj+1(t)=∫01dx∫0xdy e−(1−y)Fj(t)[Hj(t),Fj(t)]e(1−y)Fj(t) .(43)H_{j+1}(t) = \int_0^1 dx \int_0^x dy \, e^{-(1-y)F_j(t)}[H_j(t),F_j(t)]e^{(1-y)F_j(t)} \, . \tag*{(43)}

The conjugation of Eq. (A.4) resums, for su(2)\mathfrak{su}(2), into the Rodrigues rotation formula [3]

eiy⋅T(x⋅T)e−iy⋅T=[cos⁡(2∥y∥)x+sin⁡(2∥y∥)∥y∥(x×y)+1−cos⁡(2∥y∥)∥y∥2(x⋅y)y]⋅T ,(44)e^{i\mathbf{y}\cdot\mathbf{T}}(\mathbf{x}\cdot\mathbf{T})e^{-i\mathbf{y}\cdot\mathbf{T}} = \left[\cos(2\lVert\mathbf{y}\rVert)\mathbf{x} + \frac{\sin(2\lVert\mathbf{y}\rVert)}{\lVert\mathbf{y}\rVert}(\mathbf{x}\times\mathbf{y}) + \frac{1-\cos(2\lVert\mathbf{y}\rVert)}{\lVert\mathbf{y}\rVert^2}(\mathbf{x}\cdot\mathbf{y})\mathbf{y}\right]\cdot\mathbf{T} \, , \tag*{(44)}

where ∥⋅∥\lVert\cdot\rVert is the Euclidean norm in R3\mathbb{R}^3. Substituting Eq. (B.8) into Eq. (B.7), carrying out the integrations over the auxiliary variables and using the orthogonality (fj(t)×hj(t))⋅fj(t)=0(\mathbf{f}_j(t)\times\mathbf{h}_j(t))\cdot\mathbf{f}_j(t)=0 gives

Hj+1(t)=−2[pj+1(t)(sin⁡(2fj(t))2fj(t)+cos⁡(2fj(t))−14fj2(t))+qj+1(t)fj(t)(sin⁡(2fj(t))4fj2(t)−cos⁡(2fj(t))2fj(t))]⋅T .(45)H_{j+1}(t) = -2\left[\mathbf{p}_{j+1}(t)\left(\frac{\sin(2f_j(t))}{2f_j(t)}+\frac{\cos(2f_j(t))-1}{4f_j^2(t)}\right)+\frac{\mathbf{q}_{j+1}(t)}{f_j(t)}\left(\frac{\sin(2f_j(t))}{4f_j^2(t)}-\frac{\cos(2f_j(t))}{2f_j(t)}\right)\right]\cdot\mathbf{T} \, . \tag*{(45)}

whose vector of coefficients is again Eq. (IV.6) with the weights of Eqs. (IV.8) and (IV.9). This is an independent check of the resummation above.

Derivative of the Fer output

The proof of Theorem 1 needs the time derivative of the closed map of Eq. (IV.6). From f˙j(t)=hj(t)\dot{\mathbf{f}}_j(t)=\mathbf{h}_j(t) and hj(t)×hj(t)=0\mathbf{h}_j(t)\times\mathbf{h}_j(t)=0, the auxiliary vectors of Eq. (IV.7) obey

p˙j+1(t)=fj(t)×h˙j(t),q˙j+1(t)=(fj(t)×h˙j(t))×fj(t)+(fj(t)×h˙j(t))×hj(t),(46)\dot{\mathbf{p}}_{j+1}(t)=\mathbf{f}_j(t)\times\dot{\mathbf{h}}_j(t),\qquad\dot{\mathbf{q}}_{j+1}(t)=(\mathbf{f}_j(t)\times\dot{\mathbf{h}}_j(t))\times\mathbf{f}_j(t)+(\mathbf{f}_j(t)\times\dot{\mathbf{h}}_j(t))\times\mathbf{h}_j(t), \tag*{(46)}

so that differentiating Eq. (IV.6) gives

h˙j+1(t)=(fj(t)×h˙j(t))α(fj(t))+pj+1(t)α′(fj(t))fj(t)(fj(t)⋅hj(t))+[(fj(t)×h˙j(t))×fj(t)+(fj(t)×h˙j(t))×hj(t)]β(fj(t))+qj+1(t)β′(fj(t))fj(t)(fj(t)⋅hj(t)).(47)\begin{aligned} \dot{\mathbf{h}}_{j+1}(t)={}&(\mathbf{f}_j(t)\times\dot{\mathbf{h}}_j(t))\alpha(f_j(t))+\mathbf{p}_{j+1}(t)\frac{\alpha'(f_j(t))}{f_j(t)}(\mathbf{f}_j(t)\cdot\mathbf{h}_j(t))\\ &+\left[(\mathbf{f}_j(t)\times\dot{\mathbf{h}}_j(t))\times\mathbf{f}_j(t)+(\mathbf{f}_j(t)\times\dot{\mathbf{h}}_j(t))\times\mathbf{h}_j(t)\right]\beta(f_j(t))\\ &+\mathbf{q}_{j+1}(t)\frac{\beta'(f_j(t))}{f_j(t)}(\mathbf{f}_j(t)\cdot\mathbf{h}_j(t)). \tag*{(47)} \end{aligned}

Here a prime denotes the derivative of a function with respect to its argument, α′(x)≡dα(x)/dx\alpha'(x)\equiv\mathrm{d}\alpha(x)/\mathrm{d}x, and we used dα(fj(t))/dt=[α′(fj(t))/fj(t)]fj(t)⋅f˙j(t)\mathrm{d}\alpha(f_j(t))/\mathrm{d}t=[\alpha'(f_j(t))/f_j(t)]\mathbf{f}_j(t)\cdot\dot{\mathbf{f}}_j(t), and likewise for β\beta. Since α\alpha and β\beta are even and regular at the origin, so are α′(x)/x\alpha'(x)/x and β′(x)/x\beta'(x)/x, and Eq. (C.2) stays finite also where fj(t)f_j(t) vanishes. Every term on the right-hand side is a cross product with fj(t)\mathbf{f}_j(t), and therefore orthogonal to it, with the single exception of (fj(t)×hj(t))×hj(t)(\mathbf{f}_j(t)\times\mathbf{h}_j(t))\times\mathbf{h}_j(t). Projecting Eq. (C.2) on fj(t)\mathbf{f}_j(t) thus leaves

fj(t)⋅h˙j+1(t)=β(fj(t)) fj(t)⋅[(fj(t)×hj(t))×hj(t)],(48)\mathbf{f}_j(t)\cdot\dot{\mathbf{h}}_{j+1}(t)=\beta(f_j(t))\,\mathbf{f}_j(t)\cdot\left[(\mathbf{f}_j(t)\times\mathbf{h}_j(t))\times\mathbf{h}_j(t)\right], \tag*{(48)}

which vanishes whenever hj+1(t)∈A\mathbf{h}_{j+1}(t)\in\mathcal{A}, since then h˙j+1(t)⊥fj(t)\dot{\mathbf{h}}_{j+1}(t)\perp\mathbf{f}_j(t), as shown in the proof of Theorem 1. This is Eq. (V.2).

References

References

  1. [36]E. S. Mananga and T. Charpentier, J. Chem. Phys. 135, 044109 (2011).
  2. [37]W. Scherer, Phys. Rev. Lett. 74, 1495 (1995).
  3. [38]W. Scherer, J. Math. Phys. 39, 2597 (1998).
  4. [39]R. P. Feynman, F. L. Vernon, and R. W. Hellwarth, J. Appl. Phys. 28, 49 (1957).
  5. [40]S. Lloyd, Science 273, 1073 (1996).
  6. [41]A. M. Childs, Y. Su, M. C. Tran, N. Wiebe, and S. Zhu, Phys. Rev. X 11, 011020 (2021).

Paper details

vlnc:W_kznkq6eb
Version
v2 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 1

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Physics and Astronomy

Claude Code

View statement
Looking for related papers…
Contents