Академический Документы
Профессиональный Документы
Культура Документы
Citation: The Journal of Chemical Physics 136, 031102 (2012); doi: 10.1063/1.3680558
View online: https://doi.org/10.1063/1.3680558
View Table of Contents: http://aip.scitation.org/toc/jcp/136/3
Published by the American Institute of Physics
Bohmian mechanics with complex action: A new trajectory-based formulation of quantum mechanics
The Journal of Chemical Physics 125, 231103 (2006); 10.1063/1.2400851
(Received 22 November 2011; accepted 11 January 2012; published online 19 January 2012)
Eq. (1) obtained via Noether’s theorem does not take the form momentum,” analogous to the well-known “particle energy,”
of Eq. (2). Requiring this equivalence imposes rather special E = H.3, 4 Yet remarkably, p has barely been considered8
conditions on Q, e.g., that Q must be invariant under time in the previous literature, which generally considers s = mx˙
rescaling. The simplest, well-behaved, nontrivial Q giving the as particle momentum. Note that p is conserved for all free
form of Eq. (2) is particle TIQM states, including those exhibiting interference,
...
... 2 2 whereas s is conserved only for plane wave states, i.e., the
Q(x,˙ x,¨ x) ¯ x 5 x¨ . (3) classical special case for which p = s.
3 4
= 4m x˙ −2 x˙ The above Hamiltonian approach is proving extremely
Here ¯ is, in principle, an arbitrary positive constant. How- useful in numerical calculations of quantum reactive scatter-
11
ever, with the usual identification of ¯ as Planck’s constant, ing phenomena. For any 1D TIQM application, exact quan-
Eq. (3) is equivalent to the quantum potential of Bohmian tum results are obtained simply by propagating a single x(t)
mechanics2–4 for the 1D Cartesian TISE.7 This is surprising, trajectory using the classical Hamilton’s equations of Eq. (6),
given that the expression Eq. (3), being universal and kine- until p(t) flattens asymptotically to the desired level of numer-
matic, is determined entirely by the trajectory, x(t). In partic- ical accuracy. The final p value then provides a direct mea-
ular, no reference to the wavefunction, , or the TISE itself, sure of the quantum reaction probability. Note that Eq. (6)
is used in this derivation. Our quantum trajectories are nev- is amenable to efficient, symplectic numerical integrators as
ertheless Bohmian trajectories, although our formulation and are used in classical simulations, and the conservation of E
interpretation are not at all that of Bohmian mechanics, be- can be monitored as an on-the-fly measure of computed ac-
cause no is involved. curacy. This approach is extremely robust, accurate, and effi-
Action extremization applied to Eqs. (1) and (3) yields cient, leading to 15 digits of accuracy, even in the extremely
the following fourth-order autonomous ordinary differential deep tunneling regime (where absorbing potentials render
equation (ODE), describing 1D TIQM quantum trajectories: conventional exact quantum scattering methods intractible).
.... ... For many-D TIQM applications, classical-like sampling over
∂ V (x) ¯2 x x¨3
mx¨ 8 x x¨ 10 0. (4) quantum trajectory initial conditions leads to an approximate
simplifies very substantially to + limiting case (in general, only the total ensemble energy is
+ ∂x + 4m
x4 − x5 x6 = conserved). However, in the momentum conservation law
mx¨ ∂ V (x) ¯2 x 8x x 10x
3
0. Eq. (14), it does not seem to be possible to use a density that
reduces to the particle momentum p in Eq. (8).
(10) Also of note is the balance equation, Eq. (13). For any
Equation (10) is the perturbed Newton equation for the 1D autonomous Euler-Lagrange PDE, the Lagrangian density L is
TDQM case; it has no explicit coordinate dependences. It also determined only up to the addition of a divergence (i.e., the sum
bears an extremely close resemblance to Eq. (4), obtained by of C- and t-derivatives). We have chosen forms of the ac-tion
replacing C derivatives with t derivatives in the last term on [Eq. (11)] and the conservation laws such that L appears as
the left-hand side (representing the quantum force). More for- (minus) the flux in Eq. (13). Through gauge transforma-tions, it
mally, Eq. (4) is obtained on looking for travelling wave so- is possible to eliminate the third-order derivative from the
lutions of Eq. (10), i.e., solutions of the form x(C, t) = x(t Lagrangian and energy densities, but not from the flux of Eq.
− λC), where λ is a constant. In the 1D TIQM context, t (13). The choice of L we have made has the advantage that the
thus serves as an effective uniformizing coordinate. trajectory action, S(C, t ) = 0t L(C, t )d t , expressed in units of
Equation (10) also admits t-independent solutions that ¯, can be identified with the change in phase of .
correspond to the bound (fluxless) 1D TIQM quantum states We now consider Gaussian wavepacket evolution un-
(Sec. II con-cerns only the scattering states). der the free particle (V = 0) and harmonic oscillator (V
As in the 1D TIQM case, Eq. (10) is a variational equa- = 12 mω2x2) potentials. The respective x(C, t) solutions are
tion, obtained by extremizing the action,
2 − − 4m − 4 0+ m + − +¯ m2a4
x3 2 x
(t − t0)2 ,
2
1 ¯ 2 2 x p0(t − t0) a erfinv(2C 1) 1
x 5x
dCdt mx˙2 V (x) ,(11)
(15)
cf. Eqs. (1) and (3). This action is invariant under
translations of both coordinates t and C. By Noether’s and
theorem, this gives rise to two conservation laws, which are p sin ω(t − t0)
x0 cos ω(t − t0) + 0 + a erfinv(2C − 1) mω
easily found to be, respectively, − ¯
m2a4ω2
−
×
∂t2
0 +
+ + 4m x 3 2x4
2
∂ 1 2 x 2
sin2 ω(t − t0)
cos2 ω(t
5x
mx˙ 2 ¯ t ) , (16)
V (x)
+ − 4 −x3
4m ∂ C x4 x5 + x = where x0, p0, t0, and a are real wavefunction parameters, and 0
¯2 ∂ 2x ≤ C ≤ 1. Note these solutions diverge as C → 0 or 1. In general,
x 2
x˙ 2 x x˙ x˙ 0,
x(C, t) must diverge at the C endpoints; modulo this
requirement, any solution of Eq. (10) (or its arbitrary-C gen-
(12) eralization) can be used to reconstruct a normalized solution
+ (x, t) of the 1D TDSE.
∂t + ∂C −2 + 4m x3 −2x4 = We have successfully numerically integrated Eq. (10)
∂ ∂ 1 ¯2 x 5x 2
for an Eckart potential and initial Gaussian wavepacket,
[mxx˙ ] mx˙2 V (x) 0.
using the Stormer-Verlet and other methods. Initial results
com-pare favorably with those obtained using standard -
(13)
based methods, but further improvements in efficiency and
The first corresponds to conservation of energy. In the free accuracy are planned. In any case, these calculations
particle case, there is also a momentum conservation law, represent a mile-stone achievement, as the first successful
aris-ing from x-translation symmetry, synthetic Bohmian quantum trajectory calculations ever
∂t + 4m ∂ C
x4 − x5 = achieved for a system with substantial reflection
∂ 2 ∂ x 2x 2
[mx˙] ¯
0. (14) interference—a much-sought goal eluding chemical
dynamics researchers for over a decade, and a major hurdle
preventing exact quantum wavepacket calculations for large
Interpreted as hydrodynamical balance equations of the gen-eral molecular systems with few reaction pathways.4, 12
form ∂ A/∂ t + ∂ B/∂ C = 0, the first square bracket in Eq. (12)
[Eq. (14)] represents the energy [momentum] den-sity, and the
second term the corresponding flux. This desig-nation is only IV. THE MANY-D TIME-DEPENDENT CASE
determined up to addition of a C-derivative to the density A and
The 1D analysis generalizes to many-D. The single
subtraction of the corresponding t-derivative from the flux B. vari-able x is replaced with the n-dimensional configuration
Consequently, the energy density need not conform to the space vector x, with C likewise replaced with C, so that x(C,
standard TDQM “field” form,3 but instead may be chosen to be t) rep-resents an n-parameter family of trajectories. One
the particle energy, T + V + Q, as in Eq. (12). This choice is option9 is to take C = x0, though we wish to consider more
appropriate for the Lagrangian T general choices for which C and x0 are related via any
− V − Q [cf. Eqs. (1) and (2)], and has the great advantage invertible coordinate transformation.
of being conserved along individual trajectories in the TIQM
031102-4 J. Schiff and B. Poirier J. Chem. Phys. 136, 031102 (2012)
In analogy with the 1D case, it can be shown that for an been made by Holland and co-workers.3, 9, 10 The
arbitrary choice of the parameterization C, invariance of Eq. (21) under volume-preserving
ρC(C) diffeomorphisms suggests a connection with gravity, though
the physical significance of C is not yet clear.
ρ(x, t ) = det J , (17)
Numerically, the prospect of stable, synthetic quantum
i i j
where J is the Jacobi matrix, Jj = ∂ x /∂ C . As in the 1D case, trajectory calculations for many-D molecular applications will
we mostly work with a uniformizing C for which ρC (C) be fully explored, as the benefits here could prove profound. 4,
= 1. The resulting perturbed Newton equation, i.e., the 7, 11
many-D generalization of Eq. (10), can be written in various Our formalism offers flexibility for restricting action
differ-ent forms, the most compact being extremization to trajectory ensembles of a desired form (e.g.,
− reduced dimensions), thereby providing useful varia-tional
+ ∂ xi 4m ∂ Cm ij ∂ Ck ∂ Cl =
∂ V (x) ¯2 ∂ ∂ 2Kjl approximations. Our exact TDQM equations are PDEs, not
m x¨i KkKm 0. (18) single-trajectory ODEs—the entire ensemble must be de-
termined at once. But they provide the great advantage of
Here, K = J−1 denotes the inverse Jacobi matrix. The Einstein
making no reference to any external fields, such as densities or
wavefunctions. Alternatively, the many-D Hamiltonian ODE
summation convention is used, albeit with some mismatched
indices as we are currently assuming the Euclidean metric on approach is approximate, but evidently quite accurate. 11
x space. Regarding interpretation, we draw no definitive conclu-
Equation (18) is the variational PDE for the action sions here. However, it is clearly of great significance that the
dnC dt2 m x˙ · x˙ − V (x) − Q , (19) form of Q can be expressed in terms of x and its C
1 derivatives—implying the key idea that the interaction of
nearby trajectories, rather than particles, is the source of all
where + empirically observed quantum phenomena (suggesting a kind
= − 4m j∂ Cl ∂ Ck 2 ∂ Cl ∂ Ck
of “many worlds” theory, albeit one very different from Ref. 5).
¯2 2 l
k ∂ Kj 1 ∂ Kjl ∂ Kjk As such, it is locality in configuration space, rather than in the
Q K . (20)
usual position space per se, that is relevant. In effect, we have a
hidden variable theory that is local in con-figuration space, but
Equation (19) is the many-D generalization of Eq. (11),
nonlocal in position space—though the latter is hardly
which preserves the L = T − V − Q and phase properties.
“spooky” in the present non-relativistic context [even classical
Like Eq. (11), Eq. (19) has associated laws of conservation
theory is nonlocal in this sense, depending on
of en-ergy and momentum. Invariance under C translation is
V (x)]. Many ramifications are anticipated for a wavefunction-
re-placed by invariance under the infinite-dimensional group
free interpretation of measurement, entanglement, etc. One
of volume preserving diffeomorphisms.
For a general (not necessarily uniformizing) choice of wonders whether Bohm would have abandoned pilot waves,
C, and allowing a non-Euclidean metric gij(x) on x-space had he known such a formulation was possible—or, for that
√ matter, whether the notion of quantum trajectories might have
[note in this case a factor of g should be inserted in the actually appealed to Einstein.
denomi-nator of the RHS of Eq. (17)], the action, in a
simpler gauge involving only second derivatives, becomes
(21)
ACKNOWLEDGMENTS
This work was supported by grants from the
Robert A. Welch Foundation (D-1523) and the
National Science Foundation (NSF) (CHE-
0741321). Travel support from the US-Israel
Binational Science Foundation (BSF-2008023) is
also acknowledged.
1
J. von Neumann, Mathematical Foundations of Quantum
Mechanics
cf. Eq. (3.5) in Ref. 10. Note that a many-D solution of the (Princeton University Press, Princeton, NJ, 1932).
2
D. Bohm, Phys. Rev. 85, 166 (1952).
TDSE gives rise to a solution of Eq. (18), but the converse 3
P. R. Holland, The Quantum Theory of Motion (Cambridge University
holds only if the initial x˙ field is a gradient; this property is Press, Cambridge, England, 1993).
then preserved by the evolution Eq. (18).10 4
R. E. Wyatt, Quantum Dynamics with Trajectories: Introduction to Quan-
tum Hydrodynamics (Springer, New York, 2005).
5
H. Everett III, Rev. Mod. Phys. 29, 454 (1957).
6
M. F. González, X. Giménez, J. González, and J. M. Bofill, J. Math. Chem.
V. CONCLUDING REMARKS 43, 350 (2008).
7
B. Poirier, Chem. Phys. 370, 4 (2010).
We have developed a self-contained, trajectory-based 8
A. Bouda, Int. J. Mod. Phys. A 18, 3347 (2003).
formulation of spin-free non-relativistic TDQM, achieving 9
P. Holland, Ann. Phys. 315, 505 (2005).
all goals as outlined in the Introduction. Further develop- 10P. Holland, Proc. R. Soc. London, Ser. A 461, 3659 (2005).
11
ments are underway. Theoretical progress will require a cor- G. Parlant, Y.-C. Ou, K. Park, and B. Poirier, “Classical-like trajectory sim-
ulations for accurate computation of quantum reactive scattering probabil-
rect treatment of spin, relativity, particle indistinguishabil- ities,” Comput. Theor. Chem. (in press).
ity, and second quantization—with a promising start having 12
D. Babyuk and R. E. Wyatt, J. Chem. Phys. 124, 214109 (2006).