AAECC (2006) 17: 41–47

DOI 10.1007/s00200-006-0196-1

Tony C. Scott · Robert Mann

Roberto E. Martinez II

General relativity and quantum mechanics:

towards a generalization of the Lambert W
function
A Generalization of the Lambert W Function

Received: 13 September 2005 / Revised: 24 November 2005 / Published online: 2 March 2006
© Springer-Verlag 2006

Abstract We present a canonical form for a natural and necessary generalization

of the Lambert W function, natural in that it requires minimal mathematical defini-
tions for this generalization, and necessary in that it provides a means of expressing
solutions to a number of physical problems of fundamental nature. This generaliza-
tion expresses the exact solutions for general-relativistic self-gravitating N-body
systems in one spatial and one time dimension, and a previously unknown mathe-
matical link between the (1+1) gravity problem and the Schrödinger wave equation.
Keywords Implicit Functions · Schrödinger Equation · Relativity

1 Introduction

The W function satisfying W (t)eW (t) = t, introduced by J.H. Lambert (1728-1777)

is more than two hundred years old, but ironically its importance and universal

T.C. Scott (B)

Institut für Physikalishe Chemie, RWTH-Aachen, Fachbereich Chemie,
Universität Duisburg-Essen, and Zentralinstitut für Angewandte Mathematik (ZAM)
Forschungszentrum Jülich GmbH, Germany,
E-mail: t.scott@fz-juelich.de

R. Mann
Department of Physics, Professor of Physics and Applied Mathematics,
University of Waterloo, Ontario, Canada N2L 3G1,
E-mail: mann@avatar.uwaterloo.ca; Perimeter Institute for Theoretical Physics,
31 Caroline Street North, Waterloo, Ontario, Canada N2L 2Y5
E-mail: rmann@perimeterinstitute.ca

R.E. Martinez II
Harvard University, 9 Oxford Street, Cambridge, MA 02138, USA
E-mail: remartin@fas.harvard.edu
42 T.C. Scott et al.

application has only been appreciated since the last decade of the 20th century. It
had been injected into the Maple Computer Algebra system by Gaston Gonnet over
many objections because it did not appear as a “standard” special function known
in literature (e.g. see [1–3]). A presentation of the work of Scott et al.[4] in 1992
showed that it expressed an exact solution to a fundamental problem in quantum
mechanics. Corless et al. made a literature search of the W function to find that
it had been “invented” and “re-invented” at various moments in history. Its appli-
cations are numerous[5] ranging from Wien’s Displacement Law of blackbody
radiation (e.g. [6] ), to quantum chromodynamics, combinatorial number theory,
fuel consumption and population growth (see e.g. [7] and references herein) and
so forth. The W function seems be ubiquitous in nature.
More recently, the Lambert W function has expressed an exact solution to the
general relativistic gravity two-body problem[8] in (1 + 1) dimensions when the
two-bodies have exactly the same rest mass. However, the case of unequal masses
required a generalization of Lambert’s function[8, eq.(81)]. It was subsequently
found out that this generalization for unequal masses had a one-to-one relationship
with the problem of unequal charges for the quantum mechanical problem of Scott
et al.[4]
Something important and vital about the W function had been missed. For exam-
ple, long before D.E.G. Hare[5] extended the definition of the W function into the
complex plane, such an analysis had already been done by Byers Brown[9] not only
for the standard W function but also for its generalization discussed herein. The
present work examines this generalization and rescues results from the unfortunate
“fragmentation” of information and awareness of Lambert’s function. We present
a generalization that is “natural” i.e.
1. It minimizes the need for new mathematical definitions.
2. It has applications in nature like the standard W function itself.
3. It expresses solutions to a broad range of mathematical problems.
4. Its reduction to the standard Lambert W function is transparent.
Such a function satisfying these criteria is clearly a fundamental mathematical
structure worthy of consideration in the Mathematical/Physical literature. In this
work, we present a canonical form for a generalization of the W function which
satisfies these criteria according to the following mathematical physical thread of
development.

2 Quantum Problem and Link to Gravity

The one-dimensional hydrogen molecular ion H+

2 [9] is the double Dirac delta
model with ZA = q and ZB = λ q, 0 < R < ∞ and
1 ∂ 2ψ
− − q[δ(x) + λδ(x − R)]ψ = E(λ)ψ (1)
2 ∂x 2
and where λ is formally a perturbative parameter. The solution is well known[4]:
ψ = Ae−d|x| + Be−d|x−R| . Matching of ψ at the delta function peaks at x = 0, R
yields the determinental conditions:




q − d qe−dR





qλe−dR qλ − d
= 0 (2)
General relativity and quantum mechanics 43

and the energies are E± = −d±2 /2 where d± satisfies eq. (2). This determinant is
a pseudo-quadratic[4, eq.(17)] which can re-written as:

(1 − x) (λ − x)
e−2xR = where x = d± . (3)
λ

When λ = 1, both sides of (3) can be factored into:

d± = q[1 ± e−d± R ] (4)

where ± represent respectively the symmetric and anti-symmetric solutions. The

solution is [4]:

d± = q + W (0, ±qRe−qR )/R (5)

The first argument 0 indicates the W function is evaluated at its principal branch.
Henceforth we set q = 1. The anti-symmetric case, d− and consequently the eige-
nenergy appear to go to zero as R → 1 i.e. the corresponding eigenstate appears
to go into the continuum but W (−1, −Re−R ) has an order 2 branch at R = 1 and
yields a real number. Mathematical physicists of the 1970s[9], were well aware of
the branch structure of the W function.
The problem of (1+1) dimensional gravity[8] namely eq. (82) of ref. [8] relates
exactly to eq. (3) by the following transformation:

2x x−y
λ= − 1 where R = − (x + a) and d = (6)
x+a x+a

The reason why is this particular physical model produces a partial differen-
tial equation, namely eq. (30) of ref. [8] which can be treated formally as the
Schrödinger wave equation as written in (1).

3 Generalization

When λ = 1, symmetry is broken and thus the generalization we seek is a solution

to:

e−2xR = ao bo (x − r1 )(x − r2 ) (7)

where {r1 , r2 } = {1, λ} are the real roots of a quadratic polynomial and {ao , bo } =
{1, 1/λ}. We treat these parameters generally while making all necessary assump-
tions to ensure a real solution. Since a quadratic is merely a product of first order
polynomials, this guides us intuitively to consider that there exists a value y such
that:

e−Rxy = ao (x − r1 ) (8)
−Rx(2−y)
e = bo (x − r2 ) (9)
44 T.C. Scott et al.

Multiplication of the left sides and right sides of eqs.(8) and ( 9) yields (7) the
equation we desire to solve. However, individually eqs.(8) and (9) can be solved
using the standard Lambert W function:
W (yRe−r1 yR /ao )
ao (x1 − r1 ) = ao (10)
yR
W ((2 − y)Re−r2 (2−y)R /bo )
bo (x2 − r2 ) = bo (11)
(2 − y)R
The “separation” parameter y = 1 +  ensures x1 = x2 and therefore:
W ((1 − )Re−r2 (1−)R /bo ) W ((1 + ) Re−r1 (1+)R /ao )
(r1 − r2 ) = − (12)
(1 − )R (1 + ) R
Substituting eqs.(10) and (11) into (7) yields an equation governing x:
   
−2xR W (1 + )Re−r1 (1+)R /ao W (1 − )Re−r2 (1−)R /bo
e = ao bo (13)
(1 + )(1 − )R 2
Therefore, we infer that the generalized Lambert W function is:
2 = 2 (ao , bo , r1 , r2 , R) = W (z1 ) W (z2 ) (14)
where z1 = (1 + )Re−r1 (1+)R /ao and z2 = (1 − )Re−r2 (1−)R /bo This general-
ization fits into the framework of Martinez[10] which considered infinitely iterated
exponentiation or tetration, namely:
··
α·
lim n
α = αα ≡ ᑢ(α) = e−W (− ln α) α∈C (15)
n→∞

From z = W (z)eW (z) , the product on nW functions is given by:

W (z1 )W (z2 ) . . . W (zn ) = (z1 z2 . . . zn )e−[W (z1 )+W (z2 )+···+W (zn )]
The Lambert W function is governed by the addition law[11]:
W (a) + W (b) = W (ab [1/W (a) + 1/W (b)]) (16)
for (a) or (b) > 0. By induction, the product is then:
W (z1 ) W (z2 ) . . . W (zn ) = (z1 z2 . . . zn ) e−W (fn )
= (z1 z2 . . . zn ) W (fn )/fn (17)
where fn stands for an expression involving W functions of z1 , z2 . . . zn from the
addition law. Thus, with f2 = z1 z2 (1/W (z1 ) + 1/W (z2 )), then
W (z1 ) W (z2 ) = z1 z2 W (f2 )/(z1 z2 (1/W (z1 ) + 1/W (z2 )))
= W (f2 )/(1/W (z1 ) + 1/W (z2 )). (18)
In the context of the tetration function ᑢ(α) of (15), the product of Lambert W
functions is really just:
n (z1 , z2 , . . . , zn ) = W (z1 )W (z2 ) . . . W (zn ) = (z1 z2 . . . zn )ᑢ(e−fn ) (19)
General relativity and quantum mechanics 45

a multiple of the continued tetration function. Thus, no generalization in the sense

of additional mathematical definitions is needed: it is simply one single W func-
tion evaluated at a point involving other W functions. Moreover, this generalization
collapses into other special cases apart from  = 0, our original starting point. The
separation parameter  may appear as a burden but through experimental mathe-
matics, it has been used to find three previously unknown analytical solutions to
eq.(3) in the limit || → 1:
 
1 W (2Re−2R ) 1 W (2Re−2R )
λ = + and x = − ln (20)
2 4R 2R 2R
and for (7):
 
1 1 −2R(1 + bo r2 ) 1 + bo r2
r1 = , ao = exp , x= , (21)
bo r2 bo bo
and also
1 2R 1
r2 = , ao = − , x=− ln (r1 b0 ) (22)
ao 2r1 R + ln(r1 bo ) 2R
where the remaining parameters are arbitrary real numbers.

4 Applications

The solution of the three-body problem in (1 + 1) dimensions requires solving for

V which is governed by the following equation [12, eq.(32)]:
(V − m1 ) (V − m2 ) (V − m3 ) (23)
= m1 m2 (V − s1 s2 m3 ) exp(K V R | sin(q)|)
 π
+m1 m3 (V + sq s2 m2 ) exp(K V R | sin q + |)
 3 
π
+m2 m3 (V − sq s1 m1 ) exp(K V R | sin q − |)
3
(24)
where s1 = sign(sin(q + π
3
s2 = −sign (sin(q −
)), and sq = sign (sin(q).
π
3
))
This problem has no closed form solution. The hard part is in trying to extricate
the exponential terms from the rest of the expression. At q = 0 and q = ±π/3, V
is governed by exactly the same type of transcendental equation as eq. (7) whose
solution can be expressed in terms of our 2 function. The latter collapses further
into standard W functions if one of the masses is the sum of the other two. If we
consider the region 0 < q < π/3, we obtain the general form:
√ PN (V )
exp(−c 3 cos(q)V ) = where M, N → ∞ (25)
QM (V )
where c = (K R/2) is a constant and:
f (V − m3 )(V 2 − V (m1 + m2 ) − m1 m2 (f 2 − 1)) PN (V )
→
m3 (m1 m2 (f − 1) + m1 f + m2 V )
2 2 QM (V )
46 T.C. Scott et al.

where f = exp(c sin(q)V ) generates the rational polynomials. Here sin(q) ≈ q

is small and cos(q) ≈ 1. A similar story applies to other regimes [π/3, 2π/3],
[2π/3, π ] etc . . . Recently Imambekov and Demler[15] considered a bose-fermi
mixture in one dimension whose Hamiltonian is that of a linear molecule and whose
solution is also of this form.

5 Conclusions

We have identified a generalization of the Lambert W function namely the solution

to (26), denoted n fitting into the analytic framework of R. Martinez[10] and
expressing solutions to a huge class of transcendal equations and delayed differen-
tial equations[13] including applications in medicine[14]. In particular, the general
solution of the two-body problem in (1+1) dimensional gravity and the quantum
double well dirac delta function model have the solutions 2 .
The special functions in the known literature (e.g. [1]) are special cases of the
hypergeometric functions and/or the Meijer G-function [3]. The Lambert W func-
tion apparently bears no relationship to these functions and belongs to a class of
its own. The generalization presented here is a first step in identifying that class.
The next phase would naturally consist of establishing Taylor series and asymp-
totic series expansions of (26) but given today’s computational means, getting float-
ing-point numbers is relatively trivial. Naturally, it is expected that our “separation”
parameter  in ( 13) likely has a restricted domain of applicability (its primary use
being to infer the general result starting from the standard W function).
Furthermore, this work on the Lambert W function helped find analytic solu-
tions to the quantum mechanical 3-body problem known as the hydrogen molecu-
lar ion in the case of clamped nuclei of equal charges [17]. All this vindicates our
proposed generalization of the W function as being of fundamental and physical
importance.

Acknowledgements One of us (T.C.S.) would like to thank Johannes Grotendorst of the Fors-
chungszentrum Jülich, Professor Arne Lüchow of the Institut für Physikalishe Chemie, RWTH-
Aachen and Professor Georg Jansen of the Institut für Organische Chemie of the University of
Essen for their wonderful hospitality and support for allowing this work to be possible. Special
thanks to R.A. Moore of the Physics department for his diligence in many-body relativistic theory.
This work was supported in part by the Natural Sciences and Engineering Research Council of
Canada.

References

1. Abramowitz, M., Stegun, I.A.: Handbook of mathematical functions. (9th printing), Dover,
New York, 1970
2. Arfken, G.B.: Mathematical methods for physicists, 2nd ed., Academic Press, New York,
1970
3. Luke, Y.L.: The special functions and their approximations. Vol. 1, Academic Press, New
York, 1969
4. Scott, T.C., Babb, J.F., Dalgarno, A., Morgan III, J.D.: J. Chem. Phys. 99, 2841–2854 (1993)
5. Corless, R., Gonnet, D., Hare, E.G., Jeffrey, D.: Lambert’s W function in maple. In: Scott,
T.C. (ed) MapleTech 9, (Spring 1993); Corless, R., Gonnet, G., Hare, D.E.G., Jeffrey, D.,
Knuth, D.: Adv. Comp. Math. 5, 329–359 (1996)
General relativity and quantum mechanics 47

6. Valluri, S.R., Jeffrey, D.J., Corless, R.M.: Some applications of the Lambert W function
to Physics. Can. J. Phys. 78, 823–831 (2000); see also Eric Weisstein’s World of Physics,
(22.02.2005)
7. Cranmer, S.R.: Am. J. Phys. 72, 1397–1403 (2004). http://cfa-www.harvard.
edu/scranmer/News2004/
8. Mann, R.B., Ohta, T.: Phys. Rev. D. 55, 4723–4747 (1997)
9. Certain, P.R., Byers Brown, W.: Branch Point Singularities in the Energy of the Delta-Func-
tion Model of One-Electron Diatoms. Intern. J. Quantum Chem. 6, 131–142 (1972); Whitton,
W.N., Byers Brown, W., Int. J. Quantum Chem. 10, 71–86 (1976)
10. Martinez II, R.E.: Irrationality from tetration. in preparation, (2005)
11. http://functions.wolfram.com/ElementaryFunctions/ProductLog/
16/02/01/
12. Malecki, J.J., Mann, R.B.: Phys. Rev. E. 69, 1–26 (2004)
13. Campbell, S.A.: Dynamics of continuous discrete and impulsive systems. 5, 225–235 (1999)
14. MacDonald, N.: Biological delay systems: linear stability theory. Cambridge University
Press, 1969; an der Heiden, U.: J. Math. Biol. 8, 345–364 (1979)
15. Imambekov, A., Demler, E.: Exactly solvable one-dimensional Bose-Fermi mixture. private
communication, 2005
16. Mann, R.B., Scott, T.C., Martinez, R.E.: Lineal gravity and the Schroedinger wave equation.
in preparation, 2005
17. Scott, T.C., Aubert-Frécon, M., Grotendorst, J.: New approach for the electronic energies of
the hydrogen molecular ion, Chem. Phys., in press (2005)