Ann Phys 97

Annals of Physics PH5737
Annals of Physics 261, 180218 (1997)

Article No. PH975737
Unfolding the Quartic Oscillator

Eric Delabaere and Frederic Pham
Laboratoire de Mathematiques, UMR CNRS 6621, Universite de Nice-Sophia Antipolis,
Parc Valrose, 06108 Nice Cedex 2 France
Received December 27, 1996; revised July 30, 1997
The `èxact WKB method'' is applied to the general quartic oscillator, yielding rigorous
results on the ramification properties of the energy levels when the coefficients of the fourth
degree polynomial are varied in the complex domain. Simple though exact ``model forms'' are
given for the avoided crossing phenomenon, easily interpreted in terms of complex branch
points in the `àsymmetry parameter.'' In the almost symmetrical situation, this gives a
generalization of the ZinnJustin quantization condition. The analogous ``model quantization
condition'' near unstable equilibrium is thoroughly analysed in the symmetrical case, yielding
complete confirmation of the branch point structure discovered by Bender and Wu. The
numerical results of this analysis are in excellent agreement with those computed by Shanley,
overtaking the most optimistic expectations of the realm of validity of semiclassical models.
1997 Academic Press
The aim of this article is to present exact results on how the energy levels of the
one-dimensional quartic oscillator,
& 2
d2
+V(q) 9=E9,
dq 2
&
(0)
where V is the general fourth degree polynomial 1

V(q)=q 4 +:q 2 +;q,
(V)
depend on the coefficients of the polynomial V.

By the ``quasihomogeneity'' property of Eq. (0), which is invariant under the
substitution
q [ *q,
[ * 3,
: [ * 2:,
; [ * 3;,
E [ * 4E,
(V)
no generality would be lost by setting =1. But will play a prominent role in our
techniques of proof, which will be based on the complex WKB method, or more
precisely its `èxact asymptotic'' version initiated by Voros [Vo1] and further
1
One can get rid of the constant term, and let the coefficient of q 4 be 1 (assuming it was positive),
by an affine rescaling of the energy. One gets rid of the q 3 term by a suitable change of origin in q.
180
0003-491697 25.00
Copyright 1997 by Academic Press
All rights of reproduction in any form reserved.
File: DISTIL 573701 . By:DS . Date:19:11:97 . Time:10:52 LOP8M. V8.0. Page 01:01
Codes: 4113 Signs: 2181 . Length: 51 pic 3 pts, 216 mm
UNFOLDING THE QUARTIC OSCILLATOR
181
worked out by the powerful methods of Ecalle's resurgence theory [E1, E2, E3,
DDP1].
First focusing on what happens for real values of the coefficients, we shall then
allow for complex deformations of V, which will shed light on the ramification
properties of the energy levels. For instance the energy levels of the double
oscillator have square root branch points near the real axis of the complexified
`àsymmetry parameter'' ;, which can be considered as the source of the `àvoided
crossing'' phenomenon. Most interesting are the situations close to equilibrium
(where double turning points ave involved in the WKB analysis). For instance in
the case of the symmetric double well, analyzing the quantization condition near
the bottom of the well (stable equilibrium) leads to a rigorous derivation of the
ZinnJustin quantization condition [ZJ1, ZJ2, ZJ3, ZJ4, ZJ5], the solutions of
which can be expressed as series of multi-instanton expansions (cf. [DDP2]). We
shall give here a more general version of this condition, allowing the asymmetry
parameter ; to be of order O() and not just 0. Near unstable equilibrium (the top
of the hump in the double well), our analysis will yield a more subtle branch point
structure in the complex plane of the variable | crit , where | crit is the action
integral along the critical (lemniscate shaped) trajectory. In the symmetrical case
this will give a rigorous justification of the complex branch point structure discovered a long time ago by Bender and Wu [BW1] and much commented upon
since (cf. [Si1, Si2, Sh1, Sh2, Wi, Al.Si, HS, BW2, BHS]).
The main idea of our approach can be summarized as follows: the (E, :, ;) space
can be covered by various analytic charts, in each of which the quantization condition is given exactly by an explicit analytic relation between the coordinates of the
chart. Such an analytic relation, which we call a model quantization condition, looks
very similar to what one would get by a lowest order semi-classical approximation
(for instance our ``model equation'' in Section 3.1 has exactly the same form as
Bender and Wu's ``secular equation'' in [BW1, Section 4]). But in contradistinction
to semiclassical approximations our quantization conditions are exact. The price
we have to pay is that instead of the (E, :, ;) variables we must work with new
variables (the coordinates of the chart), deduced from the former by resumming
WKB-type expansions in powers of (the resummability being ensured by
resurgence theory, as we explain elsewhere; cf. bibliographical comments hereafter).
All the necessary material for deriving our model quantization conditions is
extensively described in our common work [DDP2] with Herve Dillinger. 2 In the
present paper the reader is required to take these conditions for granted 3 and concentrate on their consequences:
1. Theoretical consequences, e.g., rigorous statements on the Riemann sheet
structure of the energy eigenvalues as multivalued analytic functions of the complex
variables (:, ;).
2
Two preliminary versions of [DDP2] have appeared in preprint form in April 1991 and Jan. 1996.
The final version will appear in J. Math. Phys.
3
Perhaps their similarity with usual WKB approximations will make them not too hard to accept.
182
DELABAERE AND PHAM
2. Numerical consequences, e.g., computing numerical values of the branch

points (of E as a function of (:, ;)). The numerical scheme is divided in two steps.
(a) Solve the ``model'' (transcendental) relations between the ``model
variables''; since the relations are explicit this is just classical analysis, and the
numerical scheme (Newton's method) converges rapidly.
(b) Go from the ``model'' variables to the ``physical'' ones. This step involves
resummation of divergent series. In the present paper we made the very crude
approximation of replacing the series by their lowest order terms. Surprisingly, the
results thus obtained are already in very good agreement with those obtained by
other methods (cf. [Sh1]). At this point one should mention that the ``resurgence
relations'' (cf. [DDP2]) which led to our model quantization conditions give
precisely the information required for resumming numerically the series by the ``hyperasymptotic methods'' of Berry and Howls [BeH]. One should thus be able to get
extremely accurate numerical values.
Genesis of this Paper and Acknowledgments
The proof of the ZinnJustin conjecture was announced in our 1990 note with
Herve Dillinger [DDP0] 4 and was detailed in Delabaere and Dillinger's thesis
(Nice, 1991). But only recently did we realize how these techniques could be used
in a geometrical spirit; the idea is to use the implicit resurgent function theorem to
reduce the exact quantization condition to a simple ``model form'' before trying
to solve it. In the simple oscillator case this was suggested to us by Y. Colin de
Verdiere. Further `ùnfolding'' of this idea was greatly stimulated by our participation to the `èxponential asymptotics'' semester at the Newton Institute in
Cambridge (Spring 1995). 5 We thank Carl Bender for most helpful discussions, and
J. Lascoux who first introduced us to the Bender and Wu problem.
Bibliographical Comments
For the more mathematically minded reader, detailed justifications of the techniques presented in [DDP2] are given in [DDP1] (an exposition of the genericenergy case, in a geometrical spirit), and in [DP]. The first part of [DP] starts
with a self-contained digest of Ecalle's resurgence theory and shows how it applies
to one-dimensional WKB expansions; the second part of [DP] is a detailed derivation of how to cope with double turning points, with a thorough study of the
``confluence'' phenomenon.
4
5
In Section 1 of that note the sign \ in the monodromy factor has been forgotten.
In particular our treatment of avoided crossings was inspired by a talk of S. Slavyanov.
183
0. CLASSICAL APPROXIMATION AND SEMI-CLASSICAL INVARIANTS

0.1. The BohrSommerfeld sketch
For a given value of (:, ;, E ) we shall denote by L:, ;, E the classical `ènergy
shell'' in phase space, i.e., the curve in the ( p, q)-plane defined by the equation
p 2 +V(q)=E.
Answering the following questions is a well-known exercise for beginners in
Thom's ``catastrophy theory'' (cf., e.g., [PS]):
1. How does the shape of V depend on (:, ;)?
2. How does the shape of the energy shell L:, ;, E depend on (:, ;, E)?
The answers are the following:
1. The ``bifurcation curve'' 8: 3 +27; 2 =0 splits the (:, ;)-plane into two
regions (cf. Fig. 1), the simple well region and the double well region.
2. In the three-dimensional (:, ;, E )-space (with vertical E-axis, oriented
upwards), the set of points for which the energy shell L:, ;, E is singular is the wellknown ``swallowtail surface'' (V(q)&E=(dV(q)dq)=0); cf. Fig. 2a). Below that
surface lies the classically forbidden region. Above it lies the simple oscillator region,
where the energy shell L:, ;, E is a connected oval. The spearhead shaped region in
the middle is the double oscillator region, where the energy shell L:, ;, E splits into
two disconnected ovals.
Fig. 1. The bifurcation curve, separating the ``simple well region'' (on the right) from the ``double
well region'' (on the left).
File: 595J 573704 . By:XX . Date:14:11:97 . Time:15:45 LOP8M. V8.0. Page 01:01
184
DELABAERE AND PHAM
Fig. 2. (a) The swallowtail. (b) Plane sections of the swallowtail for :<0. The stable layer (1), the
unstable layer (2), the metastable layers (3), the Maxwell edge (4), the cuspidal edge (5).
The forbidden and simple oscillator regions are separated by the stable layer
(that part of the swallowtail where E is an absolute minimum of the potential),
Fig. 2b); the boundary of the double oscillator region has three two-dimensional
strata: the upper stratum is the unstable layer (where E is the local maximum value
of the double well potential); the two lower ones are the metastable layers (where
E is only a local minimum of the double well).
The stable and metastable layers meet along the Maxwell edge (where E is the
minimum of a symmetrical double well), which is a line of transverse self-intersection of the swallowtail. The unstable layer meets the metastable layers along the
cuspidal edge (where V has a horizontal stationary tangent at a triple turning point).
185
v In the simple oscillator region we get approximate energy levels E n as functions of (:, ;) by writing the BohrSommerfeld quantization condition
| # =(n+ 12 ) h,
n=0, 1, 2, ...,
(0.1)
where h=2? and | # := # p dq is the action integral over the oscillator cycle #,
defined as being the oval L:, ;, E , so oriented that | # >0.
v In the double oscillator region two analogous BohrSommerfeld quantization
conditions can be written
| #& =(n & + 12 ) h
(resp. | #+ =(n + + 12 ) h),
(0.2)
corresponding respectively to the ``left oscillator cycle'' # & and ``right oscillator
+
cycle'' # + . We shall denote by E &
n& and E n+ the resulting approximate energy levels.
Altogether these quantization conditions give a pattern of surfaces in (:, ;, E)space, which we shall call the BohrSommerfeld sketch. Typical plane sections of
this sketch are represented on Figs. 3a, b, and c).
The Relation between the double oscillator actions. Using the fact that the function p(q) has a pole at infinity, with residue \(i;2) one easily shows (deforming
contours on the Riemann sphere of the complex q-variable) that
| #& &| #+ =?;
(0.3)
(cf. [KL], where physical implications of this remark are discussed). As a conse+
quence, the value of (:, ;) at which two energy levels E &
n& and E n+ cross each other
depends only on the difference n & &n + of their quantum numbers.
0.2. WKB Invariants on the Complex Energy Shell
For a given value of (:, ;, E ) let L C:, ;, E be the complexified energy shell. It can
be considered as a twofold covering of the complex q-plane, ramified at the turning
points where p=0. Outside the turning points this covering carries formal solutions
of Eq. (0), the so-called WKB expansions: defined locally on the covering, up to an
arbitrary normalization factor, they read
q
P(q, 2 ) &12 e (i) q0 P(q$,
2 ) dq$
where
P(q, 2 )= p(q)+ p 1(q) 2 + p 2(q) 4 + } } }
is a formal powerseries in 2. Analytic continuation along any closed path #~ of the
covering therefore multiplies them by some factor which does not depend on the
initial normalization point q 0 : we call it the monodromy factor of #~.
186
DELABAERE AND PHAM
As an important example, let #~ be deduced from a real `òscillator cycle'' #

(cf. Section 0.1) by distorting it away from its turning points as shown on Fig. 4.
Noticing that analytic continuation along #~ multiplies P(q, 2 ) &12 by &1, one
sees that the monodromy factor of #~ reads &a #, where
a # =e (i) 0#
(0.4)
0# =
P(q, 2 ) dq=| # +0( 2 ).
The formal expansion a # is the Voros multiplier of the cycle # (cf. [Vo1, Vo2]).
Of course the same kind of construction applies to any cycle # tying two turning
points in L C
:, ;, E , provided these turning points are simple (i.e., simple zeros of
Fig. 3.
Plane sections of the BohrSommerfeld sketch: (a) :<0; (b) :>0; (c) ;=0.
187
Fig. 3Continued
E&V(q)). Among such cycles # (which geometers call vanishing cycles, because they
vanish when one makes the pair of simple turning points coalesce), an important
role will be played in the sequel by what we call fading cycles, i.e., those vanishing
cycles along which p dq is pure imaginary (positive), so that the Voros multiplier is
exponentially small. The simplest example of these is the tunnel cycle of the double
oscillator, which travels both ways along the classically forbidden segment of the
real q-axis, so that dq is real and p is pure imaginary.
1. EXACT QUANTIZATION OF SIMPLE OSCILLATORS

1.1. Quantization in the Simple Oscillator Region
In the simple oscillator region we just have one oscillator cycle #. Let us define
its monodromy exponent s by
s=
Fig. 4.
0# 1
&
2? 2
The distorted vanishing cycle.
(1.0)
188
DELABAERE AND PHAM
so that the monodromy factor reads

&a # =e 2i?s.
A careful study shows that one must distinguish two cases, splitting the simple
oscillator region into two subregions, the lower region and the higher region,
separated by the broken lines of Figs. 5a and 5b (``Stokes separatrix''; cf. also
Fig. 5c).
v The lower region is that region where the complexified energy shell has no
fading cycle. The Stokes pattern in the q-plane then looks as shown on Fig. 5c.I.
From the general results in [DDP1] it therefore follows that the Voros multiplier
a # (and its logarithm 0 # ) is Borel resummable. Understanding now a #, 0 # , s as the
true functions of (:, ;, E) deduced from the corresponding formal symbols by Borel
resummation, the exact quantization condition demands that the monodromy factor
should be equal to 1:
s(:, ;, E )=n # N
(the set of natural integers).
(1.1)
Reexpressing s in terms of 0 # , one recognizes the BohrSommerfeld quantization

condition with | # replaced by its (Borel resummed) semiclassical incarnation 0 # .
Remembering that (0 # E)| =0 =(| # E )=(T # 2) (half the time-period of the
classical particle along #), the implicit function theorem allows us to consider
(:, ;, E) [ (:, ;, s) as an analytic change of variables (analytic diffeomorphism); it
Fig. 5. Stokes splitting of the simple oscillator region: (a) :<0; (b) :>0. (c) Stokes pattern in the
q-plane, for (:, ;, E) in the lower region (I), upper region (II), or ``separatrix'' (III).
189
Fig. 5Continued
transforms the collection of energy levels into the lattice of horizontal planes
s=n # N.
v The higher region is that part where the complexified energy shell has one
complex fading cycle # 0 , which can be seen as the continuation of the tunnel cycle
across the unstable layer (crossing the unstable layer makes the middle pair of real
turning points become complex conjugate). The Stokes pattern in the q-plane then
looks as shown on Fig. 5c.II. Due to the fact that the oscillator cycle # has a nonzero intersection index with this fading cycle # 0 (cf. [DDP1]), the corresponding
190
DELABAERE AND PHAM
Voros multiplier a # (and therefore s) is not Borel-resummable any more; its Borel
transform has singularities on the real axis at integral multiples of u, where we
denote by iu the (pure imaginary) action integral along # 0 . But like any resurgent
series it can be resummed by ``right'' or ``left'' resummation (analogous to Borel
resummation but for the fact that the integration axis runs slightly above (resp.
below) the real axis), or more conveniently here by Ecalle's median resummation
(kind of a ``geometrical mean'' between the two), which has the advantage of
preserving reality properties. 6
Understanding s now as the median resummation of the corresponding symbol,
one can prove (using arguments analogous to those developed in [DDP2]) that
the exact quantization condition reads
sin ?s+
cos(?;2)
- 2 cosh(U2)
=0,
(1.1)$
where U=u+O( 2 ) is the (Borel resummed) real valued resurgent function defined
by
0 #0 =iU
(1.2)
(that the corresponding series is Borel-resummable follows from Theorem 2.5.1 of

[DP], because # 0 has a zero intersection index with itself!).
1.2. Regularity near the Stable Layer
The action integral | # is holomorphic near the stable layer, where it vanishes
with a simple zero (the time-period T # =2(| # E) has a strictly positive limit
T(:, ;) when # vanishes). The same is of course true for 0 # . In addition, the
dependence of 0 # on the parameters (:, ;, E ) is regular 7 near the stable layer. This
implies that such operations as
v rescaling the energy near the (minimal ) critical value V crit(:, ;) of the potential
function, i.e., making the substitution
E=V crit(:, ;)+E resc ,
(1.3)
v replacing the rescaled energy by a resurgent expansion

E resc =E 0 +E 1 +E 2 2 + } } } ,
(1.4)
when performed at the formal level, can be directly interpreted as operations on

functions through the resummation operations.
6
For the definition of median resummation see [DP, Section 0.5], for instance. Some examples of
computation of median symbols are given in [DDP2, Section IV].
7
In the sense of [DP] (regular dependence of a resurgent function on parameters). Here regularity
stems from the fact that the cycle #~ is not pinched by the pair of coalescing turning points.
191
Noticing that (sE resc )| =0 =(14?) T(:, ;) differs from zero (and is independent
on E resc ), it is easy to see that for every natural integer n Eq. (1.1) has a unique
formal solution of the form (1.4) with E 0 =(4?T(:, ;))(n+ 12 ) (the Rayleigh
Schrodinger series). By the `ìmplicit resurgent function theorem'' the above
regularity property allows us to conclude that the RayleighSchrodinger series is
resurgent (in our case, Borel resummable) and that its Borel sum is an exact solution
of the quantization condition; cf. [DDP2].
2. EXACT QUANTIZATION OF DOUBLE OSCILLATORS

The exact quantization condition for the double oscillator can be written in the
general form
(sin ?s + )(sin ?s & )=cos ?(s + +s & )
(1+a #0 ) 12 &1

,
2
(2.0)
where s & and s + are the monodromy exponents of the left and right oscillator
cycles # & , # + , whereas a #0 is the Voros multiplier of the tunnel cycle # 0 . Here again
s & and s + are not Borel resummable, their Borel transforms having singularities on
the real axis at integral multiples of u (the action integral along # 0 , divided by i).
It must be understood that the symbols in this formula should be interpreted
through median resummation (which, in the case of a #0, is just Borel resummation).
Such a relation is useful only inasmuch as all its ingredients depend regularly on
(:, ;, E). This is always the case for s + and s & . This is also the case for a #0 in the
double oscillator region (Section 2.1 hereafter), but not near the metastable layer
or the Maxwell edge, where our demand for regular dependence will force us to
reexpress a #0 in terms of other variables, not so simply related to it as the variable
U =0 #0 i=&ln(a #0 ) which we use in Section 2.1.
The Relation between left and right exponents. The same ``Cauchy integral''
reasoning as in Subsection 0.1 holds with | #\ replaced by 0 #\ , so that
0 #& &0 #+ =?;,
or equivalently,
s & &s + =;2.
(2.1)
2.1. Quantization in the Double Oscillator Region

In the double oscillator region none of the cycles # + , # & , # 0 vanishes. Defining
the (Borel-resummed) real valued power expansion U as previously by 0 #0 =iU,
and setting U =U, the quantization condition can be written as
(sin ?s + )(sin ?s & )=cos ?(s + +s & ) =(U ),
(2.2)
192
DELABAERE AND PHAM
Fig. 6.
The avoided crossing phenomenon in the double oscillator region.
where =(U ) is the exponentially small function defined by

=(U )=
(1+e &U ) 12 &1

1
t e &U .
2
4
Considering s + , s & , U as free variables, we call (2.2) the model form of the quantization condition. For fixed U it defines a doubly periodic lattice of curves exhibiting
the avoided crossing phenomenon shown on Fig. 6, where ; =;=2(s & &s + ), and
s=(s & +s + )2.
Notice that for large U this lattice of curves depends very slowly on U (when
differentiating Eq. (2.2), the differential of U is multiplied by the small exponential e &U ).
To translate this picture in terms of the (:, ;, E ) variables one just has to understand it as the image of the ``true'' picture through the following changes of
variables (analytic diffeomorphisms):
1. replace (:, ;, E) by (| #+ , | #& , u=| #0 i) (the Jacobian of this change of
variables is a nonzero constant; cf. [DDP1, Section 2.1]);
2. replace (| #+ , | #& , u) by their semiclassical incarnations (0 #+ , 0 #& , U),
interpreted as true functions of (:, ;, E ) by Borel resummation; by the implicit
resurgent function theorem this is a resurgent change of variables, tangent to identity as 0;
3. get (U, ;, s) by obvious linear rescaling.
Complex branch points of the energy levels. By straightforward trigonometry we
can rewrite Eq. (2.2) as
(1+=(U )) sin 2(?s)=sin 2
?;
+=(U ).
2
(2.2)$
193
It is clear from this equation that s can be seen as a multivalued analytic function
of (U, ; ) (periodic in ; of period 2), with square root branch points at
; =2&\
2i
sinh &1 - =(U )
?
(2.3)
(& # Z, the set of integers). We can interpret these branch points by saying that
the `àvoided crossing'' phenomenon is but the real trace of ``true
crossings'' at complex values of ;, exponentially close to the real axis.
Getting back to the (:, ;, E ) picture, a similar statement holds for E as a multivalued function of (:, ;) with the following caution.
Caution ! In comparing the branch points in the (U, ;, s) and (:, ;, E ) pictures
one must not forget that projecting on (U, ; ) parallel to the s-axis and projecting
on (:, ;) parallel to the E-axis are not exactly equivalent operations, because U
depends on E and not only on (:, ;). But as already noticed the quantization condition (2.2)$ depends on U only through the exponentially small function =(U )=
t 14 e &U, whose derivative with respect to E is also exponentially small:
dU
1 dU &U
d
=(U )=
=$(U )t &
e
.
dE
dE
4 dE
Taking this fact into account one checks that the branch points of E are given by
an expression which almost coincides with (2.3), but for the fact that to the small
=(U ) term in it must be added a still smaller correction, of the order of magnitude
&(116? 2 )(dU dE) 2 =$(U ) 2.
2.2. Quantization Near the Metastable Layer
What we call the metastable layer corresponds to the vanishing of one of the two
oscillator cycles in the double oscillator. Since the tunnel cycle is pinched by this
vanishing cycle, a #0 no longer depends regularly on (:, ;, E), so that the result of an
energy rescaling
E=V crit(:, ;)+E resc
is not so directly computable from the formal expression of a #0. As shown in
[DDP2] such a rescaling results in a formal expression for a #0 which is no longer
given by an integral power series of but by an expansion of the form (2.4) which
we call the critical Voros multiplier (we assume here that the vanishing cycle is the
one on the right, so that s + is no longer large),
a #0 =
&- 2? c +
1(1+s + )
\ +
s+ +12
e &ucrit (1+O()),
(2.4)
194
DELABAERE AND PHAM
where u crit =u crit(:, ;) is the value of u=| #0 i for E=V crit(:, ;) and c + =c +(:, ;)
is the ``critical action multiplier,'' exactly definable in terms of classical mechanics
(cf. [DDP2, Section IV]). The (1+O()) factor is a resurgent integral power series
expansion in , depending regularly on (:, ;, E resc ).
Consequently, defining the (large) parameter
V + =
u crit
1
u crit
+ s + + ln
+0()

2
c+
by the implicit equation

(1+a #0 ) 12 &1=
&- ?2 s +12 &V

V +
e +
1(1+s + ) +
(2.5)
and using Euler's complement formula we can rewrite the quantization condition
(2.0) in the form
1
- 2?
sin ?s & = cos ?(s + +s & ) V s++ +12 e &V+,
1(&s + )
2
(2.6)
where one should remember that s + and s & are linked by the relation s & &s + =;2.
Here again one gets an `àvoided crossing'' phenomenon (cf. Fig. 7), which can be
analysed in complete analogy with Subsection 2.0 (the details are left to the reader).
2.3. Quantization Near the Maxwell Edge
The Maxwell edge corresponds to the vanishing of both oscillator cycles. This
means that ; is assumed to be of order O(), and it is therefore convenient to work
Fig. 7. The avoided crossing phenomenon near the metastable layer. The oscillations around the
straight lines have been exagerated, just for the fun of making them visible.
195
with the rescaled asymmetry parameter ; =; (throughout this paper putting a
hat on a letter will mean making it a dimensionless quantity by dividing it by the
suitable power of e.g., by for an `àction-like'' quantity).
Since the tunnel cycle is pinched by both vanishing cycles, the critical Voros
multiplier a #0 is now given by an expansion of the form (cf. [DDP2])
a #0 =
2?
c
1(1+s + ) 1(1+s & )
\+
s+ +s& +1
e &ucrit (1+O()).
(2.7)
Consequently, defining the (large) parameter

W =
u crit
u crit
+(s + +s & +1) ln
+O()

c
(2.8)
by the implicit equation

(1+a #0 ) 12 &1=
?
W s+ +s& +1e &W
1(1+s + ) 1(1+s & )
(2.9)
and using Euler's complement formula we can rewrite the quantization condition
(2.0) under the form
2?
=cos ?(s + +s & ) W s+ +s& +1e &W
1(&s + ) 1(&s & )
(2.10)
which again exhibits the `àvoided crossing'' phenomenon (cf. Fig. 8). Since the
right-hand side is exponentially small (for large W ), the solutions of that equation
are close to the zeros of the left-hand side, which occur for all natural integral
values of s + or s & ; near a crossing point where both s + and s & are natural integers
(s + =n + , s & =n & ), a local model for the `àvoided crossing'' can be obtained by
expanding both sides of (2.10) in Taylor series of the variables
; n + &n &
x= +
4
2
y=s&
n + +n &
.
2
After factoring out the numerical coefficients, one gets

y 2 &x 2 =
W n+ +n& +1e 2y ln(W ) &W

e (1+O(x, y))
2?n + ! n & !
(where the O(x, y) term does not depend on W ).
(2.11)
196
DELABAERE AND PHAM
Fig. 8. The avoided crossing phenomenon near the Maxwell edge. The oscillations around the
straight lines have been exagerated, just for the fun of making them visible.
The ZinnJustin quantization condition. In the special case of the symmetrical

double oscillator ( ; =0), Eq. (2.10) becomes the ZinnJustin quantization condition
(cf. [ZJ5])
- 2?
=\- cos 2?s W s+12e &W 2.
1(&s)
(2.12)
Again expanding both sides in Taylor series of y=s&n (n a natural integer) the
resulting equation can be solved by iteration, yielding s as a convergent expansion

s=n\=^ n + : r n, k(\=^ n ) k,
(2.13)
k=2
where \ must be understood as & for even eigenstates and + for odd eigenstates,
and
=^ n =
W n+12e &W 2
- 2? n !
whereas r n, k =r n, k(ln W ) is a polynomial in ln(W ) of degree <k.

Convergent expansion (2.13) is our ``model form'' for the ZinnJustin ``multiinstanton expansion.'' To obtain the latter from our ``model form'' all we have to
do is replace W by its (resurgent) expansion (2.8) and express E as a (resurgent)
power expansion in s by inverting the relation s=s(E, ) (here again we use the
`ìmplicit resurgent function theorem.'' For the details, cf. [DDP2, Section V]).
Fig. 9.
197
Complex branch points of the energy levels (bold lines are cuts).
Complex branch points of the energy levels. Fixing W at some large value, the
local branch point structure of s as a function of ; (near the crossing point of the
n + and n & levels) is easily read on Eq. (2.11). One finds that s has a pair of squareroot branch points at complex conjugate positions ; =b n&, n+ or b n&, n+ , where
Re(b n&, n+ ) & 2(n & &n + ) and Im(b n&, n+ ) & 4W (n+ +n& +1)2e &W 2- 2?n + ! n & !.
Looking at Fig. 8 it is easy to see how these local data patch together in a
neighbourhood of the real ;-axis. Start with the ground state for which st&| ; |4
for real ;; this (real analytic) function prolongs in the complex ; cut plane, with
a pair of complex conjugate cuts starting from b 0, 0 and b 0, 0 (cf. Fig. 9 0)). Crossing
any of these cuts leads us to a new determination of s, corresponding for real ; to
the first excited state, holomorphic in the cut plane shown on Fig. 9i with two new
pairs of cuts starting from b 0, 1 , b 0, 1 and b 1, 0 , b 1, 0 . Crossing again any of these new
cuts leads to the second excited state (cf. Fig. 9ii), etc. Getting back to the (:, ;, E)
picture, similar statements hold for E as a multivalued function of ; (for fixed :).
Of course, proving them requires the same precautions as explained at the end
of 2.0.
3. EXACT QUANTIZATION NEAR THE UNSTABLE LAYER

The unstable layer is characterized by the vanishing of the fading cycle # 0 . Equivalently, it is the surface where the real analytic function U=0 #0 i vanishes (with a
simple zero). For similar reasons as those expounded in 1.2 this (resurgent) function is the Borel sum of the corresponding formal symbol, and depends regularly on
(:, ;, E).
In a small neighbourhood of the unstable layer, of width O(), a convenient
+
+
=U), where | &
system of coordinates is (| &
crit , | crit , U
crit (resp. | crit ) is the action
198
DELABAERE AND PHAM
+
integral along the left part # &
crit (resp. right part # crit ) of the critical trajectory (a lemniscate, with an ordinary double point at the origin). The first two coordinates
+
(| &
crit , | crit ) deserve being called shape parameters, because they only depend on the
shape of the potential (i.e., on (:, ;)) and not on E. Of course they can be replaced
+
&
+
by ( ;, | crit =| &
crit +| crit ) (recall that ;=(| crit &| crit )?). It will turn out more
convenient to replace them by (; =;, X ), where X =(| crit )+ } } } is a (Borel
resummed) resurgent function to be defined below.
3.1. The Model Equation

&
As shown in [DDP2], the critical trajectory # +
crit (resp. # crit ) does not carry one
canonically defined ``critical Voros multiplier'' but two such, complex conjugates to
each other, which we shall denote by a +, a + (resp. a &, a & ). They are related to one
another by the relations
a +a + =a &a & =1+e U

a & =e i?;a +,
(3.0)
so that knowing a + is knowing all of them. One has

a+ =
- 2? e U 4
c
1(12+i(U 2?))
\+
i(U 2?)
e i|crit (1+O()),
(3.1)
where c=c(:, ;) is the ``critical action multiplier,'' and the (1+O()) factor is a
resurgent (Borel resummable) integral power series of , depending regularly on
(:, ;, E resc ).
The exact quantization condition reads
1+e &i?; +a + +a & =0.
(3.2)
To rewrite it in more workable form, let us introduce the parameter

X =
| crit U
| crit
& ln
+O()

?
c
defined implicitly by the relation:

a+ =
- 2? e U 4
X i(U 2?)e i((X &?; )2).
1(12+i(U 2?))
(3.3)
Setting
.(U )=arg 1
U
\2+i 2?+
(3.4)
199
(the determination of the argument being so chosen that .(0)=0) and noticing
that
- 2? e U4
= - 1+e U e i.(U ),
1(12+i(U 2?))
we thus get the ``model form'' for our quantization condition,
cos
?
; + - 1+e U cos(%&.(U ))=0,
2
(3.5)
where % stands for the function of (X, U ):

U
1
%= X + ln X.
2
2?
(3.6)
This is an analytic relation between three dimensionless parameters X, ;, U, the first
two of which we shall keep on calling shape parameters, although X depends on E
(but infinitely slowly, through O() terms).
Illustrations in the symmetrical case ( ;=0). Figure 10 illustrates the ``quantization of U,'' for fixed X >
>0 and ; =0: the ``quantized values'' can be read as the
abscissae of the intersection points of the two curves y=&1 - 1+e U and y=
cos(X +(U 2?) ln X &.(U )). Notice that for U >
>0 the values are evenly spaced
Fig. 10. Quantization near unstability, for the symmetrical oscillator. For large X the frequency of
the oscillations is of the order ln(X )2? for limited values of U ; it tends to a constant as |U | tends to
infinity (but this is not visible at the scale of the picture).
200
DELABAERE AND PHAM
Fig. 11.
Quantization in the (X, U )-plane.
(as could be expected for a simple oscillator), whereas for U <

<0 they tend to
cluster in pairs (as expected for a symmetrical double oscillator).
Figure 11 shows how these quantized values depend on X ; the corresponding
graph can be deduced from the periodic lattice of curves cos(%&.(U ))=
&1- 1+e U (cf. Fig. 12) by the transformation
(%, U ) W (X, U )
(an analytic diffeomorphism for large enough X ). Therefore it is invariant under the
discrete group of transformations generated by (X, U ) [ (g U (X ), U ), where g U is
the X variable interpretation of the translation % [ %+2? (for fixed U ).
Fig. 12.
Quantization in the (%, U )-plane.
201
Another interesting fact which can be read from Eq. (3.5) is the density of levels
for small . The angle inside the cos reads 3, where
X
U
X U
&.
.
3= + ln
2 2?

\ +
\ +
For small U (of the order O()) one has

1
X
3
& & ln
U
2?

\ +
so that the density diverges logarithmically (in conformity with Weyl's rule), as
emphasized by Colin de Verdiere and Parisse [CdVP]. When |U | is not small
Weyl's rule predicts that the density of level should have a finite limit, and it can
be checked indeed that the above logarithmic divergence is then cancelled by the
derivative of the &.(U) term (use the fact that .$(U )tln( |U | )2? when
|U | , by Stirling's formula).
3.2. Complex Branch Points of U as a Function of X ( for ;=0)
In order to study the complex analytic continuation of our ``model'' quantization
condition (3.5), it is convenient to rewrite it under the following equivalent form 8
e i(%&?(?4)) =G (U ) 12,
(3.7)
where
G (U )=
1(12+i(U2?))
1(12&i(U2?))
tan
U
?
\i
.
4
4
(3.8)
Here and in the sequel the symbol (resp. \) must be understood as + for even
states, & for odd states (resp. the other way round: mnemotechnically, just remember
that even states stand below odd states).
G is a meromorphic function, with double poles only, distributed along the
positive imaginary axis at the following positions u
l (indexed by l=1, 2, 3, ...):
3
u+
l =4i?(l& 4 )
&
l
1
4
u =4i?(l& )
(=i?, 5i?, 9i?, ...)

(=3i?, 7i?, 11i?, ...).
(3.9)
Since complex conjugation changes G into the complex conjugate of its inverse,
its zeros are double zeros, at the complex conjugate positions u
l .
Notice that G has modulus 1 along the real axis, with G (0)=1. If the
determination of the square root in (3.7) is so chosen that G (0) 12 =1, the nth
8
Remember that .(U )=(12i) ln[1(12+i(U2?))1(12&i(U2?))] and that % stands for the righthand side of (3.6).
202
DELABAERE AND PHAM
bound state is characterized by the following analytic relations between the real
quantities X, U :
%=
\n+4+ ?+2i ln G
1
1
\n+4+ ?+2i ln G
(U )
(even n)
(3.10) n
&
(U )
(odd n).
Notice that these relations for different even (resp. odd) values of n are connected
to one another by analytic continuation in the complex U-plane; more precisely,
turning clockwise around a pole of G (resp. anticlockwise around a zero of G )
results in adding 2 to the quantum number n. In other words, all the even n
branches of the curve in Fig. 11 are the real traces of one connected complex
analytic curve in C 2, and the same statement holds for odd n branches.
Problem. Describe this complex curve, as a Riemann surface over the X-complex
plane, for ``large enough'' X.
By ``large enough'' X we mean the following: for |IU | <4?L, with L any given
integer, we demand that |X | >C L (some constant depending on L), with |arg X | <
(?2).
Answer. The following statements give an almost complete answer (since our
Riemann surface is invariant by complex conjugation, we only describe it in the
lower half plane of X, which corresponds to the upper half plane of U ).
1. Position of the singularities.
Even case. The constant C L can be so chosen (as a function of L) that for
every L the above Riemann surface is a connected branched covering of the
(m)
X-plane, with only square root branch points (x (m)
l , u l ), doubly indexed by natural
integers l, m subject to the only following restrictions:
l=1, 2, ..., L;
m is larger than some integer depending on l; and
m has the same parity as l.
(+)
+
For every l as above, let m go to infinity; then the sequence (u (m)
l ) tends to u l , the
(m)
+
(m)
argument of u l &u l tending to ?2 from the >?2 side; the sequence (x l ) tends
to infinity in the positive real direction, with
Ix (m)
l
3
ln m
+O
4
m
\ + \ +
3
3
=&4 l& ln 2? m+
\ 4+ _ \ 4+& + } } }
Rx (m)
l =2? m+
(the beginning of a convergent asymptotic expansion for large m).
Fig. 13.
Odd case.
203
The n th inner sheet with n=8 (lines are cuts).
The statements are analogous, with the only following differences:
v condition (+) should be replaced by

m and l have opposite parity ;
(&)
&
v u+
l should be replaced by u l ;
v and
Ix (m)
l
ln m
\ 4+ +O \ m +
1
1
=&4 l& ln 2? m+
\ 4+ _ \ 4+& + } } } .
Rx (m)
l =2? m+
2. Sheet structure. Starting from the n th quantum level on the real X axis (as
given by Eq. (3.10) n ), analytic continuation in the lower half X-plane is possible in
the cut plane represented on Fig. 13. Let that cut plane be called the nth inner sheet.
downwards is a borderline 9
Then for every (l, m) the vertical cut starting from x (m)
l
between the (m&l)th inner sheet and the (m+l)th inner sheet.
Free from singularities.
204
DELABAERE AND PHAM
Proof of Statement 1. (Say, in the even case; G is short for G + ). Logarithmic

differentiation of (3.7) yields
U
1
ln X
1
+
dU = d ln(U ).
dX +
2 2?X
2?
2i
(3.11)
For large |X | this equation implies that dUdX is everywhere small, except perhaps
in a neighbourhood of the singular points of the right-hand side, namely u +
l
(hereafter we only consider the upper half U-plane). Near every such singular point
the ramification condition is obtained by equating the coefficients of dU in (3.11),
i.e.,
1 G$(U )
ln X
=
.
2? 2i G(U )
(3.12)
Since the right-hand side has a simple pole at u +

l (with residue i), relation (3.12)
and
R
ln
X

large
enough
defines U as a holomorphic
for U close enough to u +
l
function of ln X, which we shall denote by U l (ln X ):
U l (ln X )=u +
l +
2i?
ln X
+ }}}
(3.13)
(the sum of a convergent series of inverse powers of ln X ). For large enough C

relation (3.12) thus defines a conformal transformation between the domain
[ |X | >C, &(?2)<arg X 0], i.e., [R ln X >ln C, &(?2)<I ln X 0], and
&u +
some tapered neighbourhood of u +
l , contained in a small disc [|U
l | <\ l ] and
+
+
satisfying ?2arg(U &u l )<(?2)+o(|U &u l | ).
Substituting this function U l in (3.7) + (and taking the logarithms of both sides)
yields an implicit equation for X, which can be written as
U l (ln X )
3
1
1
X +
ln X = m+ ?+ ln G(U l (ln , X )),
2
2?
4
2i
(3.14)
where m is an even integer and the determination of the logarithm has been chosen
to be real along the real X-axis. 10
Key Lemma. For any given l and large enough C, relation (3.14) has a unique
in the domain (|X | C, &(?2)arg X 0).
solution X =x (m)
l
Proof.
Solving Eq. (3.14) is solving a fixed point problem X =F(X ), where
F(X )=2? m+
10
1
3
U l (ln X )
ln X + ln G(U l (ln , X ))
&
4
?
i
G(U ) is positive real all along the positive imaginary U-axis.
(3.15)
205
and we shall try to chose our domain in such a way that |F$(X )| <1. Remembering
(3.12), we check that
F$(X )=&U l (ln X )?X.
For large enough C we can assume that the image by U l of our domain satisfies
l | <4?l, and consequently,
|U l &u +
l | \ l <3? (say), so that |U
|F$(X )| <4lC.
For C>4L(Ll) this will be smaller than 1, allowing us to solve our fixed point
problem by successive iteration provided our domain is stable by F. To check that
this is the case, it is more convenient to work with the variable Y=ln X. Our
domain becomes the horizontal band [RYln C, &(?2)IY0], and the fixed
point problem now reads Y=ln 8(Y ), with
3
U l (Y )
\ 4+ & ? Y+ i ln G(U (Y ))
2
3
u
=2? m+ &
\ 4+ ? Y+ i ln Y+b (Y ),
8(Y )=2? m+
+
l
(3.16)
where b l (Y ) is a bounded function of Y in the band (remember that |U l &u +

l |
some small constant \ l and notice that this implies |ln G(U l (Y ))&2 ln Y | <some
constant c l ).
Denoting by b l some bound of the function b l (Y ), one thus easily checks that
a sufficient condition for our band to be stable is
2? m+
3
?
+(4l+1)+b l .
ln C+
4
ln C
(3.17)
(m)
Having thus found x (m)
of U is given by u (m)
l , the corresponding value u l
l =
U l (ln X ). The square root nature of the branch point follows from the fact that
2
deriving (3.12) with respect to U gives an equation 0=(1i(U &u +
l ) )+ } } } which
(m)
is not satisfied for U =u l .
Upper bounds on |(dUdX )| warrant that U cannot go to infinity for finite values
of X.
is a square root branch point of U(X ),

Proof of Statement 2. Saying that x (m)
l
(m)
is a quadratic critical point of the
with value u (m)
l , is equivalent to saying that u l
reciprocal function X(U ), with critical value x (m)
l . As we just saw, this critical point
+
is close to u +
l (for m large), and the direction in which it is seen from u l is close
to the pure imaginary direction. Consequently, since dX dU is real along the pure
imaginary U-axis, the steepest descent lines of IX going through the saddlepoint
will be close to the pure imaginary axis and will tend to the logarithmic
u (m)
l
+
goes to &; in other words, they make up a
singularity u +
l or u l+1 when IX
206
DELABAERE AND PHAM
+
smooth arc from u +
l to u l+1 , close to the pure imaginary axis, the image of which
downwards (counted
in the X-plane is the vertical half-line hanging from x (m)
l
twice).
We now claim that the cut half-plane of Fig. 13 is nothing but the conformal
image of the U-upper half plane cut along the union (for l=1, 2, ...) of all such
Fig. 14. The right (resp. left) path in the U-plane (top) is conformally mapped onto the left (resp.
right) path in the X-plane (bottom).
207
``steepest descent arcs.'' To understand why precisely those x (m)

shown on Fig. 13
l
mark the boundary of this conformal image, let U climb up the imaginary axis,
avoiding all logarithmic singularities by small half-circular detours (always on the
same side, never crossing the cuts); the corresponding paths in the X-plane are
shown on Fig. 14.
To understand how ``crossing to another sheet'' justs amounts to changing the
value of n, notice that crossing ]u l , u l+1[ from left to right in the U-plane increases
n by 2l.
Outer sheets. Completing the cut half-plane of Fig. 13 by symmetry, we get a
cut plane containing the real axis (which more fully deserves being called ``the nth
inner sheet,'' rather than just its lower half part). Consider in this cut plane the
<X <Rx (n+1)
, which we shall call the nth band.
vertical band Rx (n&1)
1
1
Along the right border of this band, consider the vertical cut hanging down from
, or the complex conjugate cut soaring above x (n+1)
. Crossing either cut
x (n+1)
1
1
leads us into the (n+2)th band, from which we can reiterate the procedure. We
thus see that the union of all the n th bands, when n runs over the set of integers
of the given parity, is the cut plane represented on Fig. 15 (please ignore the broken
lines). We shall call it the main (or zeroth) outer sheet.
and x (n+2)
yields
A similar construction with the vertical band between x (n+1)
1
2
the cut plane also shown on Fig. 15 (full and broken lines), which we shall call the
first outer sheet. More generally, the lth outer sheet will have two series of cuts
, x (n+l)
] separate
indexed by an even (resp. odd) integer n ; the smaller cuts [x (n+l)
l
l
(n+l+1)
(n+l+1)
, x l+1
] separate it
it from the (l&1)th outer sheet; the larger ones [x l+1
from the (l+1)th outer sheet.
Fig. 15.
The main and first outer sheets.
208
DELABAERE AND PHAM
3.3. Back to the Energy and ``Shape'' Variables

Still assuming that ;=0, let us examine which information the above study gives
us about quantization in the (:, E, ) variables. By the quasihomogeneity property
recalled in the Introduction, the quantization condition depends only on the couple
of variables
(:^ =: 23, E =E 43 ).
(3.18)
The relation between these variables and our ``model'' variables (X, U ) can be written in the form
U =?E(&:^ ) &12 + &

+
35?
\ 64 E (&:^)
3
3? 2
3?
E (&:^ ) &52 + (&:^ ) &32
8
8
&92
&
85?
E(&:^ ) &72 + } } } = : U p(E, :^ ),
64
p0
where U p is quasihomogeneous of weight p (:^ being given weight

U
4
X = (&:^ ) 32 + ln 6+ : X p(U )(&:^ ) &3 p2.
3
?
p1
2
3
(3.19)
and E weight 43 ),
(3.20)
The series (3.19) and (3.20) 11 are ``resurgent'' in the following sense: for (3.20) we
mean ``resurgence'' in the usual (Ecalle) sense, the resurgence variable being
(&:^ ) 32 =(&:) 32, and the dependence on U being regular; for (3.19) what we
mean is that after the substitution (3.18) one gets a resurgent object in 1, depending
regularly on (:, E). 12
One should notice that X in formula (3.20) does not just depend on the ``shape''
parameter :^ but also on U, so that the ramification condition (3.12), which gives
the branch points of the quantized levels when projected on the complex X-plane,
does not exactly give the branch points of the projection on the complex :^-plane;
rewriting dX in Eq. (3.11) as a combination of d:^ and dU (obtained by differentiating
(3.20)), one sees that the condition for the dU terms to cancel out now reads under
the modified form
U
1 G$(U )
ln X ln 6 1
+
+
(1+ } } } )=
.
2?
? 2 2?X
2i G(U )
(3.21)
For large X this modification of the ramification condition (3.12) is numerically

small and does not affect the conclusions of Section 3.2.
11
To understand the first two terms in formula (3.20), notice that for ;=0 we have | crit = 43 (&:) 32
and c=6| crit , where c is the ``critical action multiplier'' in Eq. (3.1).
12
The domain in which this regularity property holds will be made precise in the next section. Apart
from this regularity property, which they do not mention, this is what Sternin and Shatalov call a
``resurgent function of two variables'' [StSh].
209
All our numerical computations (leading to Figs. 13, 14, and 15) have been made
with this modified equation (modifying Eq. (3.14) accordingly). Numerical evidence
shows that our iteration scheme for solving Eq. (3.14) converges not only for large
enough m (as proved in the ``key-lemma'' of Section 3.2) but for every l, m with
ml.
Translating these numerical results back to the (:^, E )-variables raises the difficult
task of computing numerically the Borel sums of the resurgent series (3.19) and
(3.20). For large enough |:^ |, the crude approximation consisting of just keeping the
first few terms in Eqs. (3.19), (3.20) should already give good results. As we shall
see, comparison with data from other sources shows that the agreement is good
indeed, not only qualitatively but to a large extent also quantitatively, even for small
values of |:^ |.
A good understanding of the multivaluedness of the energy levels should not be
limited to the domain in (:^, E )-space, where our `ùnstable layer model'' is valid. It
has to be set globally in C 2, using the various charts at our disposal. This will be
the aim of the next (and last) section.
4. GLOBAL STUDY OF THE SYMMETRICAL CASE

Up to this point our study was a purely local one and not applicable at places
involving turning points of multiplicity higher than 2, namely near the cuspidal
edge (Fig. 2) and, of course, near the origin of the swallowtail. In order to illustrate
how the results of this local study can be patched together, we shall restrict our
attention to the symmetrical case ;=0. Partly because of its relative simplicity and
partly for physical reasons, this case has received much attention from physicists,
so that one may find special interest in understanding how our methods give a
practically complete description of it.
By the quasihomogeneity property recalled in the Introduction, the quantization
condition can be considered in this case as a relation between two variables only,
namely
:^ =: 23,
E =E 43.
Using the methods of [DDP2] we can write this relation under various model
forms, valid in different domains of C 2 (the complex (:^, E )-space). As exemplified
in the previous sections, we can build two kinds of models:
v the generic models, which use WKB analysis with only simple turning
points; typical examples are Eqs. (1.1) and (1.1)$ of Section 1.1, and Eq. (2.2) of
Section 2.0 (specialized here by setting ; =0, since we deal with the symmetrical
case);
210
DELABAERE AND PHAM
v the critical models, which focus near critical values E crit of the energy such
that the WKB analysis involves double turning points: E crit =&: 24 as in the
``ZinnJustin'' model (Eq. (2.12) of Section 2.2); E crit =0 as in the `ùnstable layer''
model of Section 3.
At this stage it is natural to ask what are the respective domains of these models,
and which part of C 2 they cover altogether. One may also wonder whether our
initial emphasis on real values of the parameters did not lead us to forget some
important pieces in our collection of models. To answer the latter question, let us
recall that the collection of all models provided by [DDP2] is labelled by all
possible topological types of ``Stokes patterns'' 13 in q-space, having only simple or
double turning points. Since the critical Stokes patterns are the most easily drawn
(because of their symmetries), we shall focus our attention on the critical models,
which will turn out carrying enough information to give us a clear global picture
of the ramification properties of the energy levels.
4.1. A Complete List of the Critical Models
The critical models are of two kinds, depending on which component of the critical curve E(E+: 24)=0 one focuses.
1. For E crit =0, :{0 one has one double turning point q=0 (and two simple
ones, symmetrical with respect to the origin). It is easily checked that the topological type of the Stokes pattern depends only on how : is situated with respect to the
three half-lines arg(:)=0 mod(2?3): the ``left sector,'' which contains the negative
real axis, bears the `ùnstable layer model'' of Section 3; the two ``right sectors,''
separated by the positive real axis, bear two quantization conditions which are
formally similar to Eq. (1.1),
s(:, E resc )=n # N,
(4.1)
but which differ analytically in a subtle way. The Borel sum of the symbol s for
arg(:)>0 is not the analytic continuation of its Borel sum for arg(:)<0, because
of a Stokes phenomenon which occurs at arg(:)=0 (Borel resummability of s
breaks down there); still, no Stokes phenomena occur at arg(:)=0 for the solutions
of (4.1), which are Borel-resummable for &2?3<arg(:)<2?3 (cf. [DDP2,
Section V.A.2, Remark]).
2. For E crit =&: 24, :{0 one has two double turning points (symmetrical
with respect to the origin). The topological type of the Stokes pattern depends only
on how : is situated with respect to the three half-lines arg(:)=? mod(2?3): the
``right sector'' &?3<arg(:)<?3 bears an `èmpty'' model, where the quantization
13
We shall say that two Stokes patterns have the same topological type if one can transform one into
the other by a homeomorphism of the complex q-plane preserving the asymptotic directions of Stokes
lines at infinity (e.g., the positive real direction is transformed into itself ).
211
condition is impossible; the two ``left sectors,'' separated by the negative real axis,
bear two quantization conditions which read
c
- 2? 1
=\
1(&s) e i?s

\+
s+12
e &ucrit 2(1+O())
(4.2)
for the `ùpper left'' sector, and the complex conjugate condition for the ``lower left''
sector. The \ sign on the left-hand side selects the parity of quantum levels (cf.
[DDP2, Section V.B]). For real negative : a Stokes phenomenon occurs, and the
lack of Borel resummability suggests using median symbols, leading to Eq. (2.12).
4.2. Domains of the Critical Models
For every critical model the ``model variables'' are resurgent functions of 1,
depending regularly on the shape parameter : and the ``rescaled'' energy E resc =
(E&E crit ), for every E resc # C and every : in the sector ``bearing'' the model
(cf. Section 4.1). Our problem here is to translate this information in terms of the
variables :^ =: 23 and E =E 43. Resurgence with respect to 1 implies that Borel
resummation yields holomorphic functions of for &?2<arg()<+?2, |1| >
{(arg ), where the `ìndicatrix'' function { is the function of %=arg() giving the
exponential type of the Borel transform in the direction arg(!)=&% (we denote
by ! the dual variable of 1). In the case of WKB expansions there are good
reasons to believe that the Borel transforms are bounded at infinity, 14 so that { is
identically 0. This implies that our ``model variables,'' inasmuch as they are given
as entire functions of WKB expansions, are holomorphic functions of (:^, E ) for all
E # C and all :^ in a sector twice as big as the :-sector. For instance the expansion
s on the left-hand side of Eq. (4.1), Borel-resummed for 0<arg(:)<2?3, yields a
holomorphic function, defined for &?<arg(:^ )<?3; the expansion U of Section 3,
Borel-resummed for |arg(:)&?| <?3, yields a holomorphic function of (:^, E ) for
|arg(:^ )&?| <2?3. Notice that the domain of holomorphy may be smaller in the
case of a ``model variable'' whose definition is implicit, such as the X variable of
Section 3.1. Although the quantization condition (3.2) is well defined in the full
sector |arg(:^ )&?| <2?3, its ``model form'' (3.5), (3.6) is only valid in that sector
for large enough |:^ |, because the implicit definition (3.3) of X only makes sense for
small enough ||.
4.3. Matching the Models
It follows from the above discussion that apart from some bounded
neighbourhood of :^ =0 the whole complex (:^, E )-space is covered by the domains
of the three critical models focused around E crit =0, namely:
v the `ùnstable layer'' model (|arg(:^ )&?| <2?3);
v the two ``simple well'' models (&?<arg(:^ )<?3 ; &?3<arg(:^ )<?).
14
This was conjecture by Voros [Vo1] on the basis of numerical evidence. Then J. Ecalle outlined a
general proof, unfortunately unpublished yet.
212
DELABAERE AND PHAM
The domains of the latter two models contain the positive real :^ region, where the
potential function is indeed a ``simple well.'' Their union covers the complex :^ plane
minus the negative real-axis.
Looking at Eq. (4.1) it is not difficult to check that no ramification of the energy
levels can occur outside some neighbourhood of the negative real axis (whose size
may depend on E ). Just write
|
s
= # +O()=T # +O()
E resc E
and notice that T # (the time period of the oscillator cycle for E=0, i.e., at the
bottom of the well) is a nonvanishing function of :. Of course the rate of decrease
of the O() correction when 0 depends on : and E resc and can be controlled
uniformly only when : is kept inside a compact (closed bounded) subset of its
sector (and |E resc | is bounded).
On the other hand, our analysis in Section 3 precisely describes the ramification
of the energy levels in a neighbourhood of the negative real axis, at least for large
enough |:^ | (depending on |E | ).
Of course precise bounds should be computed in order to evaluate the sizes of
the above-mentioned neighbourhoods and to check that the neighbourhood, where
sE resc is not known to differ from 0, is contained in the neighbourhood where
our analysis of Section 3 applies. We have not done it, but as we shall now show,
taking this for granted leads to conclusions which are in complete agreement with
what is known from other sources, thus giving a posteriori support to our
hypothesis.
To get a complete portrait of the global ramification, all we have to do now is
``match'' the solutions of Eq. (4.1) with those of Section 3. This ``matching problem''
is easily solved by the following complexified version of the BohrSommerfeld
quantization rule. Near infinity in the :^ plane, not too close to the singularities, the
exact quantization condition in any chart can be rewritten under the form
|^ # =n+ 12 +O()
(n=0, 1, 2, ...),
(4.4)
where |^ # =| # is the (rescaled) classical action integral along a suitable complex

cycle # (depending continuously on (:^, E ) in the given chart).
Comparing leading terms, it is thus easy to see that the solution of Eq. (3.5)
corresponding to what we called the ``n th outer sheet'' (Fig. 15) is the analytic continuation of the solution of Eq. (4.1) with the same quantum number n. The results
of Section 3 thus lead us to the following portrait of the ramification.
4.4. Global Ramification of the Energy Levels
v Outer continuation. Starting along the positive real :^ axis (the simple well),
consider the ground state energy (n=0). It prolongs as a holomorphic function in
the cut domain shown on Fig. 16, the left part of which (in the ( |arg(:^ )| >(2?3))
Fig. 16.
213
The Bender and Wu main sheet.
sector) is just the conformal image of the ``main outer sheet'' by the change of
variable X [ :^. Forgetting about the existence of an `ùnknown zone'' around :^ =0,
where our proofs do not apply, we see that our cut domain looks exactly like the
cut plane predicted by Bender and Wu. We shall call it the Bender and Wu main
sheet.
The cuts bordering the Bender and Wu main sheet form an infinite sequence
(indexed by the integer m of Subsection 3.2); the distance of the m th cut to the
origin grows like m 23, whereas its length tends slowly to zero like ln(m)m 13. Looking
at the picture near infinity by the change of variable :^ [ &1:^, what we see is the
``horn of singularities'' predicted by Bender and Wu.
Leaving the Bender and Wu main sheet through any of the cuts will take us into
a smaller domain, with an additional series of cuts intertwinned between the previous ones as shown on Fig. 17, the ``Bender and Wu first offspring,'' as we call it.
The left part of this domain is just the conformal image of what in Section 3 we
called the first outer sheet, whereas its right part (for positive real :) corresponds
to the first even excited state (quantum number n=2).
Fig. 17.
The Bender and Wu first offspring.
214
DELABAERE AND PHAM
Crossing any of the new cuts will lead us into the next even state n=4, etc.,
and similar statements hold for odd states if we start with the n=1 state. The
correspondence between the outer sheets of Subsection 3.3 (( |arg(:^ )| >(2?3) sector)
and the quantum number n of the simple well (positive real :^ axis) may be
summarized by the following table:
Main outer sheet

First outer sheet
Second outer sheet
}}}
Even states
Odd states
n=0
n=2
n=4
etc.
n=1
n=3
n=5
v Inner continuation. Suppose now that, instead of bypassing the `ùnknown

zone,'' we cross straight through it along the real axis (all energy levels are, of
course, known to be analytic functions of :^ along the real axis). Then our discussion in 3.2 leads to the following conclusion.
For any given quantum number n (starting on the positive real :^ side) the nth
energy level is holomorphic in a cut domain looking as shown on Fig. 18, the left
part of which is just the conformal image of the ``n th inner sheet'' of Section 3.
Figure 20 shows the union for all even n of the corresponding sets of branch
points; it looks like a regular lattice filling the ( |arg(:^ )| >(2?3)) sector, as predicted by Shanley (who determined it very precisely for |:^ | <20).
The fact that our picture for fixed n (Fig. 18) does not look like Shanley's [Sh1]
should not be a surprise, because our conventions for choosing the cuts are different
Fig. 18.
The domain of inner continuation.
File: 595J 573735 . By:SD . Date:17:11:97 . Time:09:30 LOP8M. V8.0. Page 01:01
Fig. 19.
Fig. 20.
215
An avatar of Fig. 13.
The lattice of all branch points.
216
DELABAERE AND PHAM

Table I. Comparison of P. Shanley's Results [Sh1, Section 6, Table VII] and
WKB Estimates for the Singularities with the Smallest Imaginary Part
Eigenvalues labels
n=0, 2
n=2, 4
n=10, 2
n=50, 52
n=100, 102
n=140, 142
Shanley's results
WKB estimates
&4.1937+2.1697i
&6.8432+1.8945i
&14.5470+1.4974i
&39.0374+1.0652i
&61.2629+0.9037i
&76.4159+0.8324i
&4.1177+2.2275i
&6.8101+1.9072i
&14.5395+1.4985i
&39.0363+1.0652i
&61.2625+0.9037i
&76.4156+0.8324i
from his (his cuts were chosen in the radial direction from the origin). The simplest
way to see that our result agrees with Shanley's is to get back to Fig. 13 and try
to figure out what happens if one changes the slopes of its cuts to &ln n (say)
instead of &. A little thought shows that the only singularities to be seen are
, x n&2
, x n&3
, ... and x n+1
, x n2 , x n&1
, ...; the resulting conformal image in the
x n&1
1
2
3
1
3
:^-plane looks as shown on Fig. 19, and this is just Shanley's picture.
4.5. Conclusion
Our results are therefore in complete qualitative agreement with Shanley's (as
well as with Bender and Wu's), and it would be interesting to check to which extent
this agreement is also quantitative. Of course Shanley's results [Sh1] are limited to
a finite range of values of n, whereas ours are asymptotic, and we have presently
no theoretical proof that their domains should overlap.
Nevertheless Table I shows surprisingly good agreement even for small values
of n. And yet, in translating our numerical results from the X-plane to the :^-plane,
we made the very crude approximation of forgetting the O() corrections in
Eq. (3.20) (Section 3.3).
One should get still better agreement by replacing this lowest order WKB
approximation by the exact WKB series (which can be computed by the `èxact
matching method'' of [DDP2, Section IV]), and applying to this formal series such
numerical resummation procedures as ``resummation to the least term,'' or its
``hyperasymptotic'' refinements `à la Berry and Howls'' [BeH].
Crude as they may be, our computations already suggest that the `ùnknown
zone'' where our methods do not apply consists only of an `ùninteresting'' neighbourhood of zero, containing no branch points. The subsistence of such an
`ùnknown zone'' is due to our technical inaptitude for handling turning points of
multiplicity higher than 2. 15 On the other hand, it is shown in [Ph] that this
inaptitude is not an inherent shortcoming of exact WKB analysis, but is due to our
15
Notice that the results of Voros [Vo1] on the potential q 4 and the completely new `èxact quantization'' scheme which he proposed more recently in [Vo3] for the potential q 2M stand, so to speak, on
the opposite side to ours, in the sense that they tell us what happens in the very center of what we called
the `ùnknown zone.''
217
poor knowledge of the special functions attached to turning points of higher multiplicity. If one believes (as conjectured in [Ph]) that these special functions are new
transcendentals, not expressible in terms of known special functions, it is not
astonishing that investigating the vicinity of :^ =0 would need to resort to completely different methods (theoretical ones like Simon's [Si1], L#ffel and Martin's
[LM], or numerical ones like Shanley's). Conversely the main theorem in [Ph]
implies that such investigations, pushed to a complete understanding of the general
quartic oscillator (not just the symmetrical one), would provide us with the
capacity of exactly handling turning points of multiplicity 4 in any asymptotic
problem. In other words, working on the general quartic oscillator is not just
investigating an interesting model in quantum physics, it is perfecting a tool for the
whole community of asymptoticians (in the same way as Stokes' works on the Airy
function went far beyond just improving our understanding of the rainbow!). But,
of course, special functions of three complex variables (:^, ;, E ) are not easy to
study, and the first task beforehand would be to study turning points of order 3,
i.e., the general cubic oscillator (where only two complex variables are involved).
This is also an interesting physical model, the simplest one exhibiting resonances;
some of the illustrative examples in [DDP2] deal with it.
REFERENCES
[AlSi]
[BW1]
[BW2]
[BHS]
[BeH]
[CNP0]
[CdVP]
[DDP0]
[DDP1]
[DDP2]
[DP]
[E1]
[E2]
[E3]
G. Alvarez and H. J. Silverstone, Large-field of the LoSurdoStark resonances in atomic

hydrogen, Phys. Rev. A 50, No. 6 (1994), 46794699.
C. M. Bender and T. T. Wu, Anharmonic oscillator, Phys. Rev. 184 (1969), 12311260.
C. M. Bender and T. T. Wu, Anharmonic oscillator. II. A study of perturbation theory in
large order, Phys. Rev. D 7 (1973), 16201636.
C. M. Bender, H. J. Happ, and B. Svetitsky, Phys. Rev. D 9 (1974), 2324.
M. V. Berry and C. J. Howls, Hyperasymptotics for integral with saddles, Proc. Roy. Soc.
Lond. A 434 (1991), 657675.
B. Candelpergher, C. Nosmas, and F. Pham, Premiers pas en calcul etranger, Ann. Inst.
Fourier 43, No. 1 (1993), 201224.
Y. Colin de Verdiere and B. Parisse, Equilibre instable en regime semi-classique. II. Conditions de BohrSommerfeld, Ann. Inst. H. Poincare Phys. Theor. 61, No. 3 (1994), 347367.
E. Delabaere, H. Dillinger, and F. Pham, Developpements semi-classiques exacts des niveaux
d'energie d'un oscillateur a une dimension, C. R. Acad. Sci. Paris Ser. I 310 (1990), 141146.
E. Delabaere, H. Dillinger, and F. Pham, Resurgence de Voros et periodes des courbes
hyperelliptiques, Ann. Inst. Fourier 43, No. 1 (1993), 163199.
E. Delabaere, H. Dillinger, and F. Pham, Exact semi-classical expansions for one dimensional quantum oscillators, J. Math. Phys. (1997), to appear.
E. Delabaere and F. Pham, Resurgent methods in semi-classical asymptotics, Ann. Inst.
H. Poincare Phys. Theor., to appear.
J. Ecalle, ``Les fonctions resurgentes,'' Publ. Math. D'Orsay, Universite Paris-Sud 1981.05,
1981.06, and 1985.05.
J. Ecalle, ``Cinq applications des fonctions resurgentes,'' Publ. Math. D'Orsay, Universite
Paris-Sud 84T 62, Orsay.
J. Ecalle, Weighted products and parametric resurgence, in `Ànalyse algebrique des perturbations singulieres I: Methodes resurgentes,'' Travaux en cours, Hermann, Paris, 1994,
pp. 749.
218
DELABAERE AND PHAM
[Fr]
N. Froman, The energy levels of double-well potentials, Ark. Fys. 32 (1966), 7997.
[FHWW] K. W. Ford, D. L. Hill, M. Wakano, and J. A. Wheeler, Quantum effects near a barrier
maximum, Ann. Phys. 7 (1959), 239258.
[HS]
E. Harrel and B. Simon, The mathematical theory of resonances whose widths are exponentially small, Duke Math. J. 47 (1980), 845.
[KL]
D. Khuat-duy and P. Leboeuf, Multiresonance tunneling effect in double-well potentials,
Appl. Phys. Lett. 63, No. 14 (1993), 19031905.
[LM]
J. J. Loeffel and A. Martin, ``Proceedings of the R.C.P.,'' No. 25, Vol. II, Math. Department,
Strasbourg, 1970.
[Ph]
F. Pham, Confluence of turning points in exact WKB analysis, in ``The Stokes Phenomenon
and Hilbert's 16th Problem'' (B. L. J. Braaksma, G. K. Immink, and M. van der Put, Eds.),
pp. 215235, World Scientific, Singapore, 1996.
[PS]
J. Poston and I. Stewart, ``Catastrophe Theory and Its Applications,'' Pitnam, London, 1978.
[Si1]
B. Simon, Coupling constant analyticity for the anharmonic oscillator, Ann. of Phys. 58
(1970), 76136.
[Si2]
B. Simon, Large orders and summability of eigenvalue perturbation theory: A mathematical
overview, Int. J. Quant. Chem. 21 (1982), 325.
[Sh1]
P. E. Shanley, Spectral properties of the scaled quartic anharmonic oscillator, Ann. of Phys.
186 (1988), 292324.
[Sh2]
P. E. Shanley, Nodal properties of the scaled quartic anharmonic oscillator, Ann. of Phys.
186 (1988), 325354.
[StSh]
B. Sternin and V. Shatalov, ``BorelLaplace Transform and Asymptotic Theory (Introduction to Resurgent Analysis),'' CRC Press, Boca Raton, FL, 1995.
[Vo1]
A. Voros, The return of the quartic oscillator. The complex WKB method, Ann. Inst.
H. Poincare Phys. Theor. 39 (1983), 211338.
[Vo2]
A. Voros, Resurgence quantique, Ann. Inst. Fourier 43, No. 5 (1993), 15091543.
[Vo3]
A. Voros, Exact quantization condition for anharmonic oscillator (in one dimension),
J. Phys. A: Math. Gen. 27 (1994), 46534661.
[Wi]
A. S. Wightman, , 4& and generalized Borel summability, in ``Mathematical Quantum Field
Theory and Related Topics, Canadian Math. Soc. Conference Proceedings, 1988,'' Vol. 9,
pp. 113.
[ZJ1]
J. Zinn-Justin, The principles of instanton calculus: A few applications, in ``Recent Advances
in Field Theory and Statistical Mechanics, Les Houches, Session 39, 1982,'' North-Holland,
Amsterdam, 1984.
[ZJ2]
J. Zinn-Justin, Multi-instanton contributions in Quantum Mechanics (II), Nucl. Phys. B 218
(1983), 333348.
[ZJ3]
J. Zinn-Justin, Instantons in Quantum Mechanics: Numerical evidence for a conjecture,
J. Math. Phys. 25, No. 3 (1984), 549555.
[ZJ4]
J. Zinn-Justin, ``Quantum Field Theory and Critical Phenomena,'' Oxford Sci., Oxford, 1989.
[ZJ5]
J. Zinn-Justin, From instantons to exact results, in `Ànalyse algebrique des perturbations
singulieres I: Methodes resurgentes,'' Travaux en cours, pp. 5168, Hermann, Paris, 1994.

Ann Phys 97

Cargado por

Información del documento

Título original

Derechos de autor

Formatos disponibles

Compartir este documento

Compartir o incrustar documentos

Opciones para compartir

¿Le pareció útil este documento?

¿Este contenido es inapropiado?

Copyright:

Formatos disponibles

Ann Phys 97

Cargado por

Copyright:

Formatos disponibles

Annals of Physics  PH5737

Annals of Physics 261, 180218 (1997)

Unfolding the Quartic Oscillator

where V is the general fourth degree polynomial 1

depend on the coefficients of the polynomial V.

UNFOLDING THE QUARTIC OSCILLATOR

DELABAERE AND PHAM

2. Numerical consequences, e.g., computing numerical values of the branch

UNFOLDING THE QUARTIC OSCILLATOR

0. CLASSICAL APPROXIMATION AND SEMI-CLASSICAL INVARIANTS

DELABAERE AND PHAM

UNFOLDING THE QUARTIC OSCILLATOR

(resp. | #+ =(n + + 12 ) h),

P(q,  2 ) &12 e (i)  q0 P(q$, 

DELABAERE AND PHAM

As an important example, let #~ be deduced from a real ``oscillator cycle'' #

P(q,  2 ) dq=| # +0( 2 ).

UNFOLDING THE QUARTIC OSCILLATOR

1. EXACT QUANTIZATION OF SIMPLE OSCILLATORS

The distorted vanishing cycle.

DELABAERE AND PHAM

so that the monodromy factor reads

(the set of natural integers).

Reexpressing s in terms of 0 # , one recognizes the BohrSommerfeld quantization

UNFOLDING THE QUARTIC OSCILLATOR

DELABAERE AND PHAM

(that the corresponding series is Borel-resummable follows from Theorem 2.5.1 of

v replacing the rescaled energy by a resurgent expansion

when performed at the formal level, can be directly interpreted as operations on

UNFOLDING THE QUARTIC OSCILLATOR

2. EXACT QUANTIZATION OF DOUBLE OSCILLATORS

(1+a #0 ) 12 &1

2.1. Quantization in the Double Oscillator Region

DELABAERE AND PHAM

The avoided crossing phenomenon in the double oscillator region.

where =(U ) is the exponentially small function defined by

(1+e &U ) 12 &1

UNFOLDING THE QUARTIC OSCILLATOR

DELABAERE AND PHAM

by the implicit equation

&- ?2 s +12 &V

UNFOLDING THE QUARTIC OSCILLATOR

Consequently, defining the (large) parameter

by the implicit equation

After factoring out the numerical coefficients, one gets

W n+ +n& +1e 2y ln(W ) &W

(where the O(x, y) term does not depend on W ).

DELABAERE AND PHAM

The ZinnJustin quantization condition. In the special case of the symmetrical

whereas r n, k =r n, k(ln W ) is a polynomial in ln(W ) of degree <k.

UNFOLDING THE QUARTIC OSCILLATOR

3. EXACT QUANTIZATION NEAR THE UNSTABLE LAYER

DELABAERE AND PHAM

3.1. The Model Equation

a +a + =a &a & =1+e U

so that knowing a + is knowing all of them. One has

To rewrite it in more workable form, let us introduce the parameter

defined implicitly by the relation:

UNFOLDING THE QUARTIC OSCILLATOR

where % stands for the function of (X, U ):

DELABAERE AND PHAM

Quantization in the (X, U )-plane.

(as could be expected for a simple oscillator), whereas for U <

Annals of Physics PH5737

Annals of Physics 261, 180218 (1997)

P(q, 2 ) &12 e (i) q0 P(q$,

P(q, 2 ) dq=| # +0( 2 ).

Reexpressing s in terms of 0 # , one recognizes the BohrSommerfeld quantization

(1+a #0 ) 12 &1

where =(U ) is the exponentially small function defined by

(1+e &U ) 12 &1

&- ?2 s +12 &V

W n+ +n& +1e 2y ln(W ) &W

(where the O(x, y) term does not depend on W ).

The ZinnJustin quantization condition. In the special case of the symmetrical

whereas r n, k =r n, k(ln W ) is a polynomial in ln(W ) of degree <k.

a +a + =a &a & =1+e U

where % stands for the function of (X, U ):

Quantization in the (X, U )-plane.

(as could be expected for a simple oscillator), whereas for U <

Quantization in the (%, U )-plane.

For small U (of the order O()) one has

(the sum of a convergent series of inverse powers of ln X ). For large enough C

Solving Eq. (3.14) is solving a fixed point problem X =F(X ), where

G(U ) is positive real all along the positive imaginary U-axis.

where b l (Y ) is a bounded function of Y in the band (remember that |U l &u +

is a square root branch point of U(X ),

U =?E(&:^ ) &12 + &

where U p is quasihomogeneous of weight p (:^ being given weight

For large X this modification of the ramification condition (3.12) is numerically

where |^ # =| # is the (rescaled) classical action integral along a suitable complex

G. Alvarez and H. J. Silverstone, Large-field of the LoSurdoStark resonances in atomic