VOLUME

61, NUMBER 9

PHYSICAL REVIEW LETTERS

29 AUGUST 1988

Water as a Free Electric Dipole Laser

Emilio Del Giudice
Istituto Nazionale di Fisica Nucleare, Sezione di Milano, Milano, Italy
Giuliano Preparata
di Milano, Milano, Italy, and
Sezione di Milano,
Istituto Nazionale di Fisica Nucleare, Milano, Italy

Dipartimento

di Fisica, Universita

and

Giuseppe Vitiello
Dipartimento di Fisica, Universita di Salerno, Salerno, Italy, and
Sezione di Napoli, Istituto Nazionale di Fisica Nucleare, Napoli, Italy
(Received 23 May 1988)

We show that the usually neglected interaction between the electric dipole of the water molecule and
the quantized electromagnetic radiation field can be treated in the context of a recent quantum field
theoretical formulation of collective dynamics. We find the emergence of collective modes and the appearance of permanent electric polarization around any electrically polarized impurity.
PACS nUmbers:

42. 55.Tb, 03.70.+k

It is well known that liquid water is a very complicated

and that it may show significant departures
system,
from its average bulk behavior in the presence of macroand polarized impuricolloidal particles,
molecules,
ties. In addition, there are some experimental indications of its important role in the dynamics of macromolecules.
In recent times very remarkable progress has been
made through detailed dynamical calculations in the
framework of a model which describes water as a network of H-bonded molecules.
However, it seems legitimate to us to wonder what elect on the structure of
field
liquid water could the quantized electromagnetic
That our question
have, which is usually neglected.
might be totally nontrivial is suggested by the surprisingly close analogy that one can establish with the freeThere the undulator field induces on a
electron laser.
bunch of free electrons an oscillating electric dipole,
transverse to their motion, which gets coherently coupled
to the appropriate modes of the electromagnetic radiation. On the other hand, it is well known that the water
molecules possess a considerable electric dipole.
The aim of this paper is to investigate whether under
suitable conditions the electric dipoles of water molecules
can interact coherently with selected modes of the radiation field, as it just happens in the free-electron laser.
We shall employ a theoretical approach that has already
been applied to describe successfully some important asThus our model
pects of free-electron-laser dynamics.
for liquid water in a volume V is defined by the following
Hamiltonian (A denotes the electromagnetic field in the

'

radiation gauge):
jv

H= QHJ,

H,

2I

LJ2+A d, ),

(2)

where the sum runs over the N molecules of the volume

V, and L, is the total angular momentum of the single
its (average) momentum of inertia, and d, its
molecule,
electric dipole vector. We set

I =2

dm2

(dg

=0.82 A)

(3)

(Ref. 10) and

Id, =2ed, (d, =0.24),

I

(4)

where m is the proton mass and e its charge.

As we are interested in the collective aspects of the interaction of the electric dipoles with the quantized electromagnetic field, we concentrate our attention only on
those electromagnetic modes whose wavelength is either
larger than or comparable with the size L of the system,
factor e'" " shall always be
so that the plane-wave
dropped out.
Following the development of Ref. 9, we can cast our
problem [Eqs. (1) and (2)] in the framework of a quantum field theory of a complex scalar field y(u, t), the
wave field, in interaction with the quantized electromagnetic field A.
quantum-field-theory
methods
it is
By standard
straightforward to write down the relevant path integral

1988 The American Physical Society

1085

VOLUME

PHYSICAL REVIEW LETTERS

61, NUMBER 9

that describes the dynamics of our system (we use the natural units h

Z=

[dydee

][da, da, *]exp(iW[y, y*, a, a

29 AUGUST 1988

=c =1),

]),

(5)

where the action integral (u is the unit vector in the dipole direction) is given by

W=d A, dt L(u, t);

L(u, t) =

ga,*(k,' t)a, (k,' t)+@*(u,' t)i

4n r, (.

at

(u,' t)

y*(u, t)
'

+,

2ed,
i'
V' ig
rk

1/2

4m'~

y(u, t)

(e, u) y (u, t) y(u, t) Ia, (k, t)e '"' a, (k, t)e'"'],

(7)

L is the square of the angular momentum operator

t

L2
sin8
and
tion

a8

a+a'

sin8

a0

The Noether theorem leads to the following conserved

quantity:

=dny '(u, t) y(u, t),

dn(*(u,t)g(u, t) = l.

6LdQdt =0,
(io)

4md2,

i2ed,'

g(u,' t)

ab, (k, t)

N
k

g(e, u)
V

2ed, e(k)

((u, t) =g(

C(

(t)Yf"(u),

(t) denotes the number of molecules that populate the rotational state tl, m) of our rigid rotator,
which, neglecting all interactions, at thermal equilibrium
should follow a Boltzmann distribution
with F(=l(l
+ I )/ 4md~. Under normal conditions (unpolarized bulk
water) the levels l =0 and l = 1 have comparable populaN C(
t

1086

[b, (k, t)e '"'

equations:

b, (k, t)e'"']g
(u, t),

(i4)

dnug'(u, t)g(u, t).

In order to proceed any further we must get back to

the meaning of the complex functions g(u, t) and
represents the averb, (k, t). It is clear that N g(u, t)
age number of dipoles in the volume V that are polarized
in the u direction, while N tb, (k, t)
is the average
number of photons in the mode k, r. By expansion of
g(u, t) in spherical harmonics, i.e., writing
t

(13)

fluctuations" whose size is O(1/

modulo "quantum
N'(2). 9 The variational problem (13) yields the following Euler-Lagrange

t)

(12)

By changing variables in the path integral (5), we clearly

see that the solution of our quantum field theory consists
simply of the "classical" solution of the variational problem

t),

a(k, t) =N' b, (k, t).

. a((u
at

Lagrangian

where L(u, t) has, in terms of

and bprecisely the
same form as L but for the extremely important difference that the coupling strength 2ed, /V'( now gets
multiplied by the very large number N'(; furthermore,
Eq. (9) can now be rewritten as

(9)

where N is just the number of dipoles contained in the

volume V.
As emphasized in Ref. 9, Eq. (9), implying that the
paths contributing to Z must have the "macroscopic"
size N'(, suggests the relevance of the following rescalings:

&(u, t) =N'q(u,

b the

L(u, t) =NL(u, t),

a, (k, t) is the amplitude of the mode k with polarizar [k e, (k) 0, from the transversality condition].

In terms of the new functions g and

(7) scales as

(is)

"

tions. To simplify our problem,

from now on we shall
work in the subspace spanned by the four states t0, 0)
and l, m), and restrict the electromagnetic modes to the
resonating ones, i.e. , those for which k =coo =1/2mdg. '
Setting
t

yo(t) =Coo(t),

(t) =C( (t)e

and calling b(t) the amplitude of the electromagnetic

mode coupled to transition tl, m)
t0, 0), one can cast
the system (14) and (15) in the form

yo(t)=ng b*(t)y (t),

(t) = nb (t)yo(t),
b

(t) =2nyo(t)y, (t),

(i7)
.

PHYSICAL REVIEW LETTERS

61, NUMBER 9

VOLUME

and the small oscillations around it are controlled by the

pulsation

where

0= 4ed,
Q3

2copV

G~o.

(18)

Note that in pure water (H20), G=-17. Taking advantage of the rotational symmetry of the problem, our system can be further simplified by setting y (t ) = yi (t )
and b (t) =b(t). One thus obtains
yp(t

) = 3 n b (t ) yl (t ),

= nb(t

yl(t)

b(t) =2nyp (t) yl(t),

yp

'+ 3

y|

=cos

yp(0)

yl(0)

I'=-,
'

yp,

specifies

the initial

co

(22)

sin'Op,

pp,

Pl, y the modulus and the phase

the system (19) admits a limit

pp- y=x/2,

(23)

rp(t) =rp

d, j

(28)

(29)

Ed,

which will mix the states

new eigenstates
I

I0, 0) and

1,0), to produce the

0) =cosa I0, 0)+sina 1,0),

I

I) = sina 0, 0)+ cosa 1, 0),

I

with eigenvalues

(I /$3)[1 +cos2O+( 1

rl(t) =r, =(1/ J3)(l r') '"

8(t) =8 =( ' ) '/'(I=o' cos'Op)
,

sin22O)

I/2] I/2

(24)
'/',

= co2 = (2rpr, /Et) n,

j3 ) [sin2a (I ) I

(25)
1

~p,

=(cop/2) [1 T-

= [cop (co)+4') '

(31)
(32)

]/2Vd.

For Vd not too strong (a & n/8) the dynamics of our system is practically unchanged, except for the rotation
(30), induced by the electrostatic mixing, which modifies
the form (28) of the polarization P3 as

) + cos2a 2I p(t ) I 1(t )cos([co2 (co(+ 4 V

Equations (24) to (28) and (31) to (33) display the

essence of this work. Let us see what their physical
meaning is. For diA'erent initial conditions (Op), the system will exhibit diA'erent dynamical behaviors; however,
for Op & x/4 the frequencies involved will fall in the interval [see (27)] 0& v&500 cm '; while for z/3&Op
& z/4 the frequencies populate three distinct bands
around 1600, 750, and 400 cm '. It should be recalled
that in our frequency range the main absorption bands of
pure water are located at 1640, 580, and 180 cm
But the most intriguing result of this work is the prediction (33) of the emergence of a permanent electric po-

(1+4'/co/)'/1,

and

tga

and

P3 (t ) = (1/

cop) t.

Let us now suppose that within the volume V we have

an impurity with a sizable electric dipole, which generates an electric field Ed oriented in the z direction.
The static part of the Hamiltonian
[Eqs. (1) and (2)]
will acquire a new term

Vd=

yi, b, respectively,

pp

(27)

P, =(gln ulg)
= (2/ J3)rp(t )r, (t ) cos(co2

cycle:
p =pl

=(2COS2Op) '/2n.

0,

Op,

Ib(0)l' =0
If we call I p, I 1, 8 and
of

r,

(21)

Op(0(Op~ x/2)

where the angle

conditions as

(26)

From our definitions it is straightforward to calculate the

polarization along any direction n; one finds

of motion:

Op

n.

sin 2Op]

It is easy to see that access to the limit cycle can be

obtained only for Op~ n/4. Note that the "thermal
start" [Boltzmann distribution, in normal conditions
(0)=r, (0)] corresponds to Op=+/3. On the other
hand, for Op & x/4 the system (19) goes through a quasiperiodic motion with average pulsation

' =I,

+ lbl = 3n

=242[1

col

(19)

) yp(t),

which admits the following constants

29 AUGUST 1988

j) '

] t )j.

(33)

larization around an impurity that carries a sizeable

electric dipole (organic macromolecules are known to
have this property). The amount of such polarization
turns out to be

(P3(t)) =(I/Q3) sin2a(I

r,'),

(34)

1 attains very
in the region where 2'/cop is
respectable values. Note that in the absence of the collective interaction analyzed in this paper, one must have
ro -r&', as prescribed by thermal equilibrium, and no
significant electric polarization can emerge even for rath-

which

1087

VoLUME61, NUMBER 9

PHYSICAL REVIEW LETTERS

er strong electric fields.

So far no account has been made for the radiative energy losses from the volume V, which clearly is not an
electromagnetic cavity. By taking such losses into account one can easily see that the lifetimes of our collective modes are of the order of 2tr/coo, and thus considerably larger than the periods 2tr/tot, 2tr/toz, and 2tr/to of
the collective dynamics. For Oo) tr/4 the energy losses
will bring the I o and I 1 to their limit-cycle values, and
the system will resume "lasering" in such a regime only
when the thermal processes will have brought Oo above
~/4
Even though, it must be admitted, our analysis is in
many ways at a preliminary, rather rudimentary stage,
nevertheless, we believe that we can draw a number of
relevant conclusions: (i) In the study of the dynamics of
water, the neglect of the coherent interaction of water
molecules with the quantized radiation field is completely unjustified, for we have shown that its coupling to collective quantum states of size 2tr/too (a few hundreds of
microns) is indeed very large (G=12); (ii) because of
the latter fact, the time scale associated with the
coherent interaction is much shorter (=10 ' s) than
those connected with short-range interactions.
In such
coherent dynamics, our analysis has also recognized the
relevance of frequency bands which can be related to the
observed absorption bands of pure water; (iii) a macroscopic, permanent polarization can easily arise in water
in the presence of a small electric disturbance such as the
local field produced by a macromolecule, or the field at
the surface of a colloid grain or within clays.
As a result, one can envisage the possibility that the
coherent interaction between the water electric dipoles
and the radiation field fulfills the very important task of
generating ordered structures in macroscopic domains
(i.e., within a few hundred microns) which could then
have a fundamental role in the organization of inanimate
as well as living matter'
in the wonderful ways that
Of course
physical analysis is incessantly revealing.
much more work is needed in this direction.

1088

29 AUGUs~ 1988

We wish to express our thanks to Professor Silvia Doglia for her encouragement and for giving us useful information on the present experimental knowledge of wa-

ter.

'8'ater, 8 Comprehensive Treatise, edited by F. Fr" nks

(Plenum, New York, 1972-1982), 7 Vols.
ZC. A. Angell, Annu. Rev. Phys. Chem. 34, 593 (1983);
S. H. Chen and 3. Teixeira, Adv. Chem. Phys. 64, 1 (1986).
3Biophysics of Water, edited by F. Franks and S. Mathias
(Wiley, Chichester, 1982).
4D. Eagland, in Ref. 1, Vol. 4, Chap. 1, p. 305.
5J. B. Hasted, H. M. Millany, and D. Rosen, J. Chem. Soc.
Faraday Trans. 77, 2289 (1981).
V. Dahlborg, V. Dimic, and A. Rupprecht, Phys. Scr. 22,
179 (1980); G. Albanese, A. Deriu, F. Ugozzoli, and C. Vignali, Nuovo Cimento 9D, 319 (1987).
7H. E. Stanley, J. Teixeira, A. Geiger, and R. L. Blumbery,
106A, 260 (1981); E. Clementi, in
Physica (Amsterdam)
Structure and Dynamics of nucleic Acids, Proteins and Mem
branes, edited by E. Clementi and S. Chin (Plenum, New
York, 1986).
8G. Dattoli and A. Renieri, in Laser Handbook Volume 4,
New
edited by M. L. Stitch and M. Bass (North-Holland,
York, 1985).
9G. Preparata, Phys. Rev. A 3$, 233 (1988).
This value refers to the highest moment of inertia of the
water molecule. See Ref. 1.
One can see that in thermal equilibrium the ratios RI between the populations of the state
l, m) and the state ~0, 0)
are R I =0.887, R2 =0.698, R3 =0.487, R4 =0, 301, R5 =0. 165,
etc. For heavy water D20, the same ratios are 0.942, 0.835,
0.698, 0.549, 0.406, . . . , respectively.
~~This obviously implies that the size L of our V is now constrained to be less than 2tr/too =4.3x10
cm.
' G. E. Walrafen, in Ref. 1, Vol. 1, p. 151.
E. Del Giudice, S. Doglia, M. Milani, and G. Vitiello, in
and
edited by F. Guttmann
Modern Bioelectrochemistry,
H. Keyzer (Plenum, New York, 1986), and Nucl. Phys. B275
[FS17], 185 (1986).
~