Ref 1

Journal of Modern Optics
ISSN: 0950-0340 (Print) 1362-3044 (Online) Journal homepage: http://www.tandfonline.com/loi/tmop20
Electromagnetically induced transparency
J. P. Marangos
To cite this article: J. P. Marangos (1998) Electromagnetically induced transparency, Journal of

Modern Optics, 45:3, 471-503, DOI: 10.1080/09500349808231909
To link to this article: https://doi.org/10.1080/09500349808231909
Published online: 03 Jul 2009.
Submit your article to this journal
Article views: 1942
Citing articles: 401 View citing articles
Full Terms & Conditions of access and use can be found at

http://www.tandfonline.com/action/journalInformation?journalCode=tmop20
JOLTHNALOF MXII.:KN OP'I'ICS, 1998, VOL. 45, NO. 3, 471-503
Topical review
Electromagnetically induced transparency
J. P. MARANGOS
Laser Optics and Spectroscopy Group, Blackett Laboratory, Imperial
College, London SW7 2 B 2 , England
(Received 11 August 1997)
Abstract. The subject of electromagnetically induced transparency (EIT) is

reviewed in this paper. Emphasis is placed on the experimental work reported
in this field since 1990. Theoretical work is also covered, although it is not
intended to review all the very numerous recent theoretical treatments on this
topic. The basic physical ideas behind EIT are elucidated. The relation of EIT
to other processes involving laser-induced atomic coherence, such as coherent
population trapping, coherent adiabatic population transfer and lasing without
inversion, is discussed. Experimental work is described covering the following
topics: EIT with pulsed and continuous-wave sources, lasing without inversion,
pulse propagation in a laser dressed medium and EIT in nonlinear optical
processes. A full set of references and a bibliography are included.
1. Introduction
Interference between alternative pathways in quantum-mechanical processes is
an ubiquitous effect in physics (for example [l]). This interference is analogous to
constructive and destructive interference between classical waves but, whilst with
classical wave interference it is the field amplitudes that are interfering in the
quantum case, however, we must invoke intuitively less definite quantities (e.g.
probability amplitudes) to explain quantum interference phenomena. Matter wave
interferometry, especially atom beam interference between spatially separated
beam paths, has recently received considerable theoretical and experimental
attention [2-41, not least because of the potential for the construction of ultra-
sensitive interferometers that might be used to measure atomic and molecular
interactions with extreme sensitivity [S] and to test fundamental aspects of
quantum mechanics (see, for example, the proposal in [6,7]). Another kind of
interference phenomenon, between transition pathways induced within the inter-
nal quantum states of atoms and molecules coupled to one or more laser fields, is
the subject of the present review. In particular, three-level atomic and molecular
systems coupled to two laser fields exhibit interference effects that can result in
cancellation of absorption at a resonance transition frequency and other modifica-
tions of the optical response.
It has been known for some time [8] that, if states of an atom are coupled via
several possible alternative transition processes, interference between the ampli-
tudes of these processes may lead to an enhancement (constructive interference) or
a complete cancellation (destructive interference) of the total transition probabil-
ity. These effects arise because in quantum mechanics it is probability amplitudes
0950-0340/98 $12.00 0 1998 Taylor & Francis Ltd.

472 J . P . Marangos
12> T+F;k
channel (b)
)j continuum
hmuv\ Ihwuv
Figure 1. Interference between pathways in autoionization of an atom by ultraviolet
radiation. The ground state 11) is coupled to the continuum state IE2,k) via two
alternative pathways; channel (a) is direct photoionization by absorption of an
ultraviolet photon (WVUV), channel (b) is an indirect photoionization process in
which absorption of the ultraviolet photon (WVUV) results in excitation into bound
state 12) which is followed by a transition to the continuum state via the
interelectronic Coulomb interaction. The amplitudes due to these two channels
must be summed to give the overall photoionization amplitude; thus interference is
seen in the ionization rate.
(which may be positive or negative in sign), rather than probabilities, that must be
summed to obtain the total transition probability of a process.
An example of this in atomic systems is Fano [8, 91, interference seen for
radiative transitions to autoionizing states in atoms and also predicted in semi-
conductor quantum wells [lo], leading to asymmetric spectral profiles. Here
(figure 1 ) a doubly excited state of the atom 12) of energy E2, with a bound-state
character, lies in the continuum and is energetically degenerate with a continuum
state IE2,k) (at the same energy E2) to which it is coupled via the Coulomb
interaction between the two electrons. An ultraviolet electromagnetic field will
cause photoexcitation from the ground state 11) to the continuum state (E2,k ) . This
photoionization can proceed via two possible channels: firstly a direct photoioniza-
tion 11) + JE2,k ) or secondly an indirect process consisting of excitation 11) + 12)
followed by a rapid radiationless transition (caused by the interelectronic Coulomb
interaction) 12) + IE2, k ) . Interference occurs between these two channels with a
sign varying from constructive to destructive depending on the frequency of the
ultraviolet electromagnetic field.
This type of interference is deliberately induced by an applied laser field in
recent work concerned with the laser control of the optical properties of atomic
media. T h e cancellation of absorption for a probe field tuned in resonance to an
atomic transition, for which strong absorption would normally be expected, is
perhaps the most striking example. This process has been termed electromagneti-
cally induced transparency ( E I T ) [ l l , 121, the effect being caused by the inter-
ference between the coherences excited in the atom by the electromagnetic fields
and leading to an initially highly opaque medium being rendered almost trans-
parent. Similarly the refractive properties of the medium may be greatly modified
[13,14]; for instance the usual correlation of high refractive index with high
absorption can be broken, leading to the creation of media with very unusual
Electromagnetically induced transparency 47 3
optical properties. Recently classical field interference of an analogous type has

been proposed as a means for the cancellation of absorption of electromagnetic
radiation propagating through an ideal plasma at a frequency which would
otherwise be below the transmission cut-off [15].
One of the consequences of the coherence and interference effects related to
E I T that has been recognized is the possibility of building a short-wavelength laser
operating without the need to achieve population inversion in the atomic medium
[16-221. Lasing without inversion has been demonstrated in Na and R b in the
visible range [23,24]. T h e prospects that this might lead to the construction of
lasers able to circumvent the usual constraints placed on achieving inversion of
short-wavelength lasers (owing to the v 3 scaling of the spontaneous emission) has
been much discussed (see the reviews in [25, 261). Related lasing without inversion
and E I T effects in semiconductor quantum wells have also been theoretically
explored using laser-induced processes [27] or bandgap engineering [28] to create
the necessary coherences.
It is the aim of this review to provide a summary of the experimental work that
has so far been reported on E I T and to present an overview of the general progress
in this topic (including also theory) since 1990. An extensive review of all
theoretical and experimental work related to atomic coherence phenomena is
beyond the scope of the present article and the reader is advised to look at a
number of reviews on lasing without inversion [25,26], coherent population
trapping (CPT) [29-311 (a very comprehensive recent review on this subject has
been given in [32]) and laser-induced continuum structure [33,34] to find these
topics presented in detail. T h e theory pertinent to E I T is sketched in this text, but
again the reader is referred to the more detailed treatments published in the
literature which will be given as they arise. Although this review is intended to be
of use to the specialist reader, perhaps already engaged in research in this or related
fields, I hope that it will also be accessible to a wider cross-section of physicists and
that it will provide a useful introduction to what is proving to be a fascinating area
of research.
T h e review is organized in the following fashion. Section 2 examines the
physical basis for E I T and related phenomena. It attempts to provide a physical
picture of these effects; in the process some rigour will be sacrified but I hope that
this is compensated by providing an intuitively accessible picture that clarifies
some basic issues whilst providing the reader with comprehensive references to
more detailed treatments. Also in section 2.2 we shall introduce the most
significant physical consequences of E I T , namely modifications to the linear and
nonlinear optical response of the medium. T h e possible importance of this field to
future technological applications will also be touched upon at this point. Finally in
section 2.3 some of the very important physical constraints which can reduce or
extinguish observable effects in real systems will be discussed. This provides a link
between the predictions of basic theory and the situations likely to be encountered
in experiments.
Section 3 deals with experimental techniques and results. It is intended to
provide a solid overview of the field but it cannot hope to be comprehensive in its
scope owing to the large body of work on related topics (e.g. coherent population
transfer, dark resonances, and laser-induced modifications to spontaneous emis-
sion) that must necessarily be omitted or only partially covered. T h e discussion
will be organized along the lines of the scientific principles being investigated and
Figure 2. The basic energy level schemes for three-level atoms interacting with two
near-resonance electromagnetic fields. In all cases, 11) is dipole coupled to 12), and
12) to 13), but 13) is not dipole coupled with 11) (metastable). The schemes are
defined by the relative energies of the three states: (a)a ladder or cascade system;
(b) a A scheme; (c) a V scheme.
the experimental techniques used rather than in any strict chronological order.
Thus first (section 3.1) some of the earlier experimental investigations of atomic
coherence effects are briefly reviewed. Then the first E I T experiments that
employed pulsed (section 3.2) and continuous-wave (CW) (section 3.3) lasers are
discussed. Related experimental work on amplification without inversion and
lasing without inversion is presented in section 3.4. Pulse propagation and
matched pulse effects are a development that connects E I T with recent work on
coherent adiabatic population transfer, and both theoretical and experimental work
will be discussed in section 3.5. Finally, in section 3.6, recent experiments that
have used E I T and related processes to enhance nonlinear optical processes (a very
promising area) will be presented.
T h e review is completed with a discussion and conclusion in section 4 that
overviews the work done so far and the prospects for further developments in the
near future.
2. Physics of EIT
2.1. The physical basis of electromagnetically induced transparency
There is a close link between E I T and other atomic coherence phenomena such
as coherent population trapping [32,35-371 and coherent adiabatic population
transfer processes [38-41]. I n all these processes, three-level atomic systems are
involved (or at least systems that can be adequately reduced to three levels when
interaction with the pertinent electromagnetic fields are considered). T h e usual
atomic dipole selection rules normally require that two pairs of levels are dipole
coupled whilst the transition between the third pair is dipole forbidden. In figure 2
we show the three basic level schemes for three-level atoms; all the level schemes
involved in the experiments discussed in this paper can be reduced to one or other
of these schemes. We label the levels Il), 12) and 13), where 11) - 12) and 12) - 13)
are dipole allowed but 11) - 13) is not since 13) is a metastable state. Classification
Electromagnetically induced transparency 475
of the schemes then depends upon the relative energies of these three states [31]:
(a) in a ladder (or cascade) scheme El < E2 < E3;
(b) in a A scheme El < E3 < E2 (although, in a symmetric A scheme, El and
E3 are almost degenerate);
(c) in a V scheme E2 < El and E3 (although again in a symmetric V scheme, El
and E3 are almost degenerate).
In a A or ladder scheme, 11) is normally the ground state of the atom and is
where the majority of the population (initially) resides. E I T is associated with A
and ladder schemes since in this process population transfer is not required (in
contrast with C P T ) and states 12) and 13) can remain (essentially) unpopulated
throughout the process.
T o understand more clearly how laser fields may interact with a three-level
atom and create coherent superpositions of the atomic basis states we shall first
consider C P T in a A scheme. A three-level A system (figure 3) consisting of states
[ I ) , 12) and 13) is coupled b y two near-resonance laser fields of strength (defined in
terms of the Rabi frequency [31]) Q1 (at frequency w l ) and Q2 (at frequency w2).
Defining the frequency of transitions between states as wl2 = (E2 - E l ) / h ,
~ 2 = 3 (E2 - E3)/h and ~ 1 = 3 (E3 - E l ) / h we can further define one- and two-
photon (Raman) frequency detunings as A12 = w12 - w1, 423 = w23 - w2 and
A = (wl2 - ~ 2 3 -) (w1 - w z ) . T h e Hamiltonian HO of the bare atom should be
modified to include the interactions due to the two couplings, that is
+ +
H = Ho Vl V2 (where the interaction V, = hQj). T h e eigenstates of this new
Hamiltonian will be linear superpositions of the bare atomic states [ I ) , 12) and 13).
For the situation of exact two-photon resonance that is A = 0 (or rather exact two-
photon resonance taking into account any light shifts) two of the eigenstates of the
+ +
total Hamiltonian HO Vl V2 turn out to be symmetric and antisymmetric
coherent superpositions of the two lower states of the bare atom basis that are of the
form:
+
where Q’ = (a: Qi)”2. Importantly no component of the bare atom state 12)
appears in these superpositions. These form eigenstates of the atom-field system of
which one (I+)) is coupled to the intermediate state 12) via the electric dipole
interaction whilst the other state (I-)) becomes decoupled (a so-called ‘dark’ or
‘trapped’ state). This can be seen if we form the dipole moment between I+) or I-)
with the remaining bare atom state 12) [31]. If the magnitudes of the coupling fields
Q I and Qr are appropriately balanced, the negative sign in the superposition of 11)
and (3) which form I-) will ensure that the corresponding dipole moment will
vanish. In effect the two terms that are summed to give the transition amplitude
between 12) and I-) are of equal and opposite magnitude, and hence the total
amplitude will vanish. State I-) is often referred to as a non-coupled state (INC))
whilst I+) remains coupled to the electromagnetic fields (IC)).
In the above description the situation has been simplified by ignoring the fast
time development (at frequencies E l / h and E3/h) of the states in the superposition
in equation (1) (these terms will in fact disappear when the dipole moments are
476 J . P. Marangos
I 3)
Figure 3. The basic A (Raman) scheme leading to coherent population trapping. The
applied fields need not be in single-photon resonance, but the two-photon (Raman)
resonance condition should be met in this three-level system. In a CPT experiment
the laser fields Ql and Q2 are of comparable strength that this is sufficient to saturate
the two-photon transition.
formed). On the assumption that a steady-state situation has been reached, the
superposition state INC) will acquire all the population of the system through the
action of optical pumping (spontaneous emission from 12) will populate INC) but
there is no absorption process from J N C ) to 12) to depopulate it).
In the usual C P T scheme, 521 % 522, that is both coupling strengths are of
comparable magnitude, whilst the fields are also strong enough to reach the two-
photon transition saturation condition. In fact this does not necessarily require
that the fields are sufficient to saturate the single-photon transitions 11) - 12) and
12) - 13) since, under the two-photon resonance condition, state 12) can be
adiabatically eliminated and so does not enter into the consideration of the
coupling between atoms and fields.
In the C P T system, interference effects arise from both the coupling fields
since they are of comparable strength. If only one of the fields, that is 522, is strong
such that 521 << 522, then only interference effects due to processes driven by 522
will be important. This is the situation in E I T schemes and this close connection
between E I T and C P T has been discussed by a number of workers (for example
[32,42]). Typically in a C P T scheme the states 11) and 13) are Zeeman or hyperfine
sublevels of the ground state and are thus both initially populated. I n contrast, in
many E I T schemes, 13) is an excited state and has no population at any time during
the process. Unlike the case of C P T , where the time scale for population trapping
in the INC) state is several radiative lifetimes, in the case of E I T the effect is
established (within a single atom) in a time on the order of l / & which is generally
much faster. Comparison with coherent adiabatic population transfer schemes
(e.g. stimulated Raman adiabatic passage (STIRAP)) [38,39] also shows that the
E I T situation is equivalent to the initial stages in the population transfer process
when the counter-intuitively sequenced laser pulses satisfy the condition 521 << 522.
T h e interference associated with E I T arises because the transition amplitude
between 11) and 12) includes, as well as a term due to the resonant field 521, an
additional amplitude due to the presence of the other field 522. This additional term
has a negative sign with respect to the first and hence in an ideal situation will
13d>
12d>
Figure 4. (a)The basic scheme for an EIT experiment, illustrated here in a A system in
the bare states. The field coupling 11) to 12) is a relatively weak probe Qp (equivalent
to Q,), whilst the field coupling states 12) and 13) called the coupling field QC
(equivalent to 0 2 ) is strong. This makes the adoption of the dressed-state basis
logical as illustrated in (6). (b) The dressed states labelled here as (3d) and 12d)
(referred to as) . 1 and Ib) in the text); destructive interference between the probe
absorption amplitudes due to these two dressed states leads to EIT.
cancel the first term completely. In the case of EIT, since the field Q2 is large (in
EIT experiments, 0 22 is usually called the coupling field and labelled Qc, and Q1 is
a weaker probe field labelled Qp), it is logical to choose the dressed state basis [21]
to analyse this system (figure 4); on this basis, upper states form a coherent
superposition, which for a resonant coupling 4 2 3 = 0 is of the form
T h e transition amplitude at the (undressed) resonant frequency (E2 - E l ) & from

the ground state 11) to the dressed states will be the sum of the contributions to
states la) and Ib). If 13) is metastable, then the contributions from the 11) 4 (2)
transition cancel since they enter the sum with opposite signs. T h i s cancellation of
absorption on the 11) - 12) transition can also be viewed in terms of Fano-type
interference [21,42-44].
T h e connection between equation (2) and the coupled and non-coupled states
of equation (1) can readily be seen. T h e states INC) and IC) (i.e. I+) and I-) in
equations (1)) in the case of two-photon (Raman) resonance are related to the bare
atomic states by
or alternatively we can define the bare atomic states in terms of INC) and IC).
Thus, for example, state 11) (atom ground state) can be written
For the case << Q2 it is clear from equation (4) that 11) will be almost equivalent
to INC) and thus (for two-photon resonance and a metastable state 13)) absorption
vanishes.
Alternatively, if the E I T process is viewed within the atomic bare-state basis
(rather than the dressed states), the so-called ‘coherences’ can be seen as being the
quantities pertinent to the interference. These coherences can be thought of, in a
semiclassical picture, as associated with the oscillating electric dipoles driven by
the coupling fields applied between pairs of quantum states of the system, for
example );1 - I j ) . Strong excitation of these dipoles occurs whenever electromag-
netic fields are applied close to resonance with an electric dipole transition between
two states. If there are several ways to excite the oscillating dipole associated with
Ii) - l j ) , then it is possible for interference to arise between the various contribu-
tions to this dipole, and these must be summed to give the total electric dipole
oscillation between );1 and l j ) (see figure 7 later). This is directly analogous to the
Fano [8] effect in autoionization, to the classical interference proposed in ideal
plasmas [15] and to effects discussed in the context of laser-induced continuum
structure [34].
Of course, formally coherences are identified with the off-diagonal density
matrix elements p~ formed by taking bilinear combinations of probability ampli-
tudes of two quantum states of the system (i.e. the weighting factors associated
with the outer products such as l i ) ( j l ) [45]. Off-diagonal elements of the density
matrix play a critical role in the evolution of an atom coupled to electromagnetic
fields [46]. Many calculations of atomic coherence effects and of E I T (as well as
general nonlinear optics and laser action) in three-level systems are therefore
developed in terms of the density matrix [14,47-49]. This approach also naturally
lends itself to the inclusion of dampings that cause the decay of populations and
coherences (e.g. radiative decay and collisions) [46]. Although convenient it is by
no means essential to use a density matrix approach and many theoretical
treatments that give clear physical insight, have been performed in terms of
probability amplitudes (for example [12,20,21]). Additional physical insight has
also been obtained by adopting other alternative approaches, for instance by a
careful consideration of the Feynman diagrams representing the various processes
involved that lead to interference [42,44], or by applying a quantum jump
approach [SO]. In all cases the predictions are essentially identical.
T h e magnitudes of the relevant density matrix elements can be computed from
the basic coupled evolution equations (the Liouville equation) [31,45] and are
found to depend upon parameters that are controllable within an experiment (e.g.
detunings and laser intensities). In figure 5 we show a prototypical E I T scheme (in
this case a ladder configuration, but entirely equivalent to the A schemes discussed
so far if 12) is higher in energy than the metastable state 13)). T h e detunings
between the fields E l at w 1 and E 2 at w 2 with respect to the associated transition
frequencies w l 2 and ~ 2 are 3 defined by 4 1 2 = w 1 2 - wl and A 2 3 = ~ 2 =
3 w 2 as above.
T h e Rabi frequencies are Q l and s 2 2 respectively. I n addition to dependence on
these laser properties the E I T will also depend critically on atomic parameters. For
instance since perfect absorption cancellation depends on the metastability of (3)
any radiative or collisional decay of this state will lead to finite absorption even at
Figure 5. Illustration of a prototypical scheme for EIT, in this case a ladder (cascade)
system is shown. T h e single-photon detunings dl2 and 4 2 3 (defined in text) are shown.
zero detuning Awl = Aw2 = 0 . E I T will manifest itself in the value of the density
matrix element p12 whose real and imaginary parts should both vanish at zero
detuning (i.e. the coherence is cancelled by the interference of the pathways that
can excite it). A set of coupled equations connecting the density matrix elements,
for example p l 2 , p23 and ~ 1 3 and
, their temporal derivatives can be written down
and solved for various sets of conditions by either analytical or numerical means. It
should be noted that the interference that leads to E I T arises from the existence of
coherences p23 and pi3 that are coupled to p l 2 . T h e coherence pi3 between the
ground state 11) and metastable state 13) is present only because of the additional
laser coupling. T h e contribution to the coherence p i 2 from the coherences p23 and
p13 cancels with the direct contribution driving this coherence due to the applied
probe field (at frequency wl).
If a steady-state limit is assumed, all the derivatives in the density matrix
elements vanish; in this case the solution of coupled density matrix equations is
greatly simplified. This can be solved exactly usually under the additional
assumption that only the coupling field SZC is strong and hence is the only field
that must be retained to all orders. This solution is frequently obtained by using an
algebraic manipulation programme to invert the super-matrix resulting from the
set of equations [Sl]. These solutions yield the steady-state populations and
coherences (on- and off-diagonal density matrices respectively) and in particular
give the appropriate value of p12 (the coherence associated with the transition at the
probe frequency) that is of direct interest in describing E I T . T h e expression that i s
derived for p l 2 will include a number of terms in the various parameters (e.g. Qc:
and A ) that will lead to cancellation of its value (both real and imaginary parts)
when in two-photon resonance.
A time-dependent calculation is of course appropriate if there are time-
dependent laser pulses coupling with the atom. This is vital for modelling the
results of pulsed laser experiments and to account for transient effects correctly.
Some time-dependent calculations have been made but, unless restrictive simpli-
fying assumptions are applied [52], these calculations must be performed numeri-
cally rather than analytically. In most cases the results of the time-dependent
480 J. P . Marangos
calculation will be comparable with those of the steady-state method, at least in so

far as qualitative trends are concerned. In calculating the results of propagation of
pulses through an extended ensemble of atoms, the time-dependent density matrix
equations must also be coupled to Maxwell’s equations. This is necessary for
instance in computing the propagation of matched pulses [53-551 and in comput-
ing preparation losses and pulse shape modifications [56].
2.2. Physical consequences of EIT electromagnetically induced transparency

T o understand the modification of the optical properties (absorption and gain,
and dispersion and nonlinear response) of a laser dressed medium we need to
examine the linear (and nonlinear) optical susceptibilities rather than the density
matrix elements themselves. The macroscopic polarization at the transition
frequency w12 can be related to the microscopic coherence pl2 via the expression
Pl2 = Np12pl2, where N is the number of equivalent atoms in the ground-state
within the medium, and p12 is the dipole matrix element associated with the
(undressed) transition [46]. This relation holds for a medium sufficiently dilute
that dipole-dipole coupling between atoms can be ignored; otherwise these local
field effects must be incorporated into the treatment which complicates inter-
pretation [57,58]. In this way imaginary and real parts of the linear susceptibility
at frequency w can be directly related to pl2 (calculated by the methods outlined in
section 2.1), via the macroscopic polarization P12(w) = ~ o x ( w ) E[46]. Whilst the
microscopic coherences are treated quantum mechanically the electromagnetic
fields themselves are treated classically (in terms of Maxwell’s equations and
susceptibilities). This semiclassical approach is not essential and fully quantum
treatments for C P T (for example [30]) and E I T [59-611 have been developed.
These fully quantum approaches are appropriate for cases such as the coupling of
atoms to modes in cavities [59, 601, or when the statistical properties of the light
are of interest as they are in proposals to generate squeezed light using E I T [61].
For the relatively large fields envisaged in most laser experiments a semiclassical
treatment (with spontaneous decay added as a phenomenological damping) proves
adequate.
For the prototypical system introduced above, the real and imaginary parts of
the (dressed) linear susceptibility, associated with the dispersion and absorption of
the medium respectively, will be given by expressions of the form [12]:
(5)
where p12 is the electric dipole moment of the probe transition, A , j are the
detunings defined in section 2.1, and r2 and r3 are the radiative decays of states
12) and 13) respectively. In [12] these were derived using the equations describing
the time-dependent atomic bare-state amplitudes in a steady-state situation with
monochromatic coupling fields and with no collisional or Doppler broadening
included. Of course, identical expressions are obtained using a density matrix
approach. These susceptibilities are plotted in figure 6 as a function of detuning
4 r
-20 -1 0 0 10 20
12
Figure 6. T h e dressed susceptibilities are shbwn for a three-level ladder with the
coupling laser in exact resonance (423)and with the probe laser tuned over the
ranges shown (in units of the decay rate 712 between states 12) and 11)). T h e
coupling field Rabi frequency Szc = 1Oy12 in these plots. (a) T h e imaginary part
(Im x!;)) of the dressed linear susceptibility that determines the absorption in the
medium; note the vanishing value at exact resonance due to destructive interference.
( b ) Rex:)) that determines the dispersion in the medium. (c) T h e third-order
dressed susceptibility Ixg'I that characterizes the coupling strength in the four-wave
mixing process; note the enhancement of this quantity at exact resonance due t o
constructive interference.
412 with A23 = 0 (coupling field resonant). This is a striking result when compared
with the case when no coupling field is present (a2= 0) [12]; the absorption
vanishes at exact resonance (if 13) is perfectly metastable). Simultaneously the
dispersion is modified so that, although still zero at line centre as in the uncoupled
case, the group velocity (dependent upon the slope of R e x ( ' ) ) can become
anomalously low [62] where absorption has vanished.
T h e consequence of this transparency is that a medium that would otherwise be
optically thick is now rendered transparent (or at least the opacity is greatly
reduced). T h e reduction in absorption is not merely that caused by the effective
detuning induced by the Autler-Townes splitting of the dressed-state absorption
peaks, that is the absorption that would be measured if the probe field were
interrogating the absorption coefficient of the medium in the wing of the absorp-
482 J . P. Marangos
tion profile at a detuning &/2. Instead there is a destructive interference at this

frequency that leads to complete cancellation of all absorption if there are no
additional dephasing channels operating in the system. Even if the 11) + 13)
transition dipole moment is not zero (i.e. if there is a spontaneous decay rate r3)
the absorption will be reduced compared with the weak-field absorption at the
detuning L$/2 in the ratio r3/f2[12].
In the preceding case it was implicitly assumed that the upper state probability
amplitudes remain close to zero (i.e. the probe field is always very weak). If there is
an incoherent population pump into the upper states, such that these no longer
remain negligible, then gain on the 12) + 11) transition can result. T h e remarkable
feature of this gain is that under the circumstances in which E I T occurs (i.e. when
absorption is cancelled) the gain can be present without the requirement of
population inversion in the bare atomic states; this is an example of amplification
without inversion. This process has successfully been incorporated into a laser
without inversion in a V scheme in R b [23] and a A scheme in Na [24]. Much
theoretical work [16-221 has been reported on this effect (earlier predicted b y
Arkhipkin and Heller [19], and then further elucidated by Harris [20],
Kocharovskaya and Khanin [16] and Scully et al. [18]). A long-term objective in
this work is the prospect of overcoming the familiar difficulties of constructing
short-wavelength lasers (i.e. very-high-energy-density pump requirements).
As well as leading to gain without population inversion, any incoherent
pumping of population into the upper states also modifies the dispersive beha-
viour. In particular, it is then possible to have spectral regions for which the
refractive index is very high (with values comparable with those normally
encountered at the half-widths of the absorption profile) whilst the absorption
vanishes [14]. T h e prospects for engineering the refractive properties of media to
give novel combinations of absorption or gain, and dispersion have been explored
in a number of theoretical [63-691 and experimental studies (see below). An
application proposed earlier for refractive index modifications of this kind was for a
high-sensitivity magnetometer [70]. T h e large dispersion at the point of vanishing
absorption could it was suggested be used to detect, with high sensitivity, magnetic
level shifts via optical phase measurements in a Mach-Zehnder interferometer.
Modification of absorption relative to stimulated emission is a key element of
the lasing without inversion concept. Laser-induced modifications of the sponta-
neous emission have also been discussed as a potential route to reducing the pump
energy density requirements to achieve lasing in an atomic medium [18,20]. These
effects were first studied some time ago (for example [71]) and more recently
theoretical work has predicted spectral-linewidth elimination and spontaneous
emission cancellation [72,73]. These predictions were confirmed by experiments
on Na dimers [74]. A full review of this subject is outside the scope of this paper.
It is clear that the susceptibilities given by expressions such as equation (5) are
no longer the results of first-order perturbation theory since they implicitly include
the coupling field Q2 to all orders. In this sense they can be thought of as being the
dressed linear susceptibility of the system. In the same spirit it is possible to
consider a dressed nonlinear susceptibility in the presence of appropriate coupling
fields. Consider our prototypical system in figure 5 and imagine that this is now
incorporated into a four-wave mixing scheme such that, whilst there is now no
probe field present initially, one is generated by the mixing between the coupling
field and the additional fields (say wBand wb) that are in two-photon resonance with
Figure 7. (a), (b) T h e first two diagrams for a perturbative expansion of terms
associated with the linear susceptibility in the dressed system; there are an infinite
number of higher order terms of the same type as, ( 6 ) involving the exchange of an
even number of photons from the coupling field with the atom. (c), (d) T h e first two
terms associated with a perturbative expansion of the nonlinear susceptibility in the
four-wave mixing process. Note that in this case there will be an infinite number of
terms (like (d)) associated with the exchange of an odd number of photons from the
coupling field with the atom.
the 11) - 13) transition (this will be referred to as the two-photon coupling). In
addition to the dressed linear susceptibility, given by equations (S), which
describes the response of the medium to a weak probe at w1, there is now also a
dressed nonlinear susceptibility that describes the coupling of the atom to the four-
+ +
wave mixing process w1 = w, wb w2. This can be computed [12] along similar
lines to the nonlinear susceptibility and yields the result
(6)
where the sum on the right-hand side represents the contribution to the nonlinear
~
susceptibility of all the states of the atom; the fields at frequencies w, and wb are
close to two-photon resonant with 11) - 13) and are necessary to complete a four-
wave-mixing scheme with w2 in order to generate a field at frequency w1. This
nonlinear susceptibility is plotted in figure 6 (c); although similar to Im xD
(1) (figure
6 ( a ) )in the sense that the response displays the familiar Autler-Townes splitting,
there is a fundamental difference. Paying attention to the centre of the profile at

detuning A 1 2 = 0 we see that instead of destructive interference there is construc-
tive interference at this point. T o get some physical insight into why this should be
so we can consider the diagrammatic representation of processes leading to xg)and
xD(3) (figure 7). What is clear from the figure is that, whilst in the case ofxg)
all the
diagrams representing higher-order interactions are even in the number of photons
exchanged with the coupling field 522, those for xg) involve an odd number of
photons being exchanged with this field.
In any four-wave mixing process the phase matching (propagation) of the
generated and driving fields is of critical importance. In the limit of an optically
deep medium considering plane-wave fields the figure of merit determining the
conversion efficiency is given IxD (3)
I/xD
(1)
; this is physically reasonable since xD
(1)
characterizes the reabsorption and wave-vector mismatch in the medium, and xD (3)
the coupling strength that generates the new field. Thus we can see the importance
of the destructive interference in the value of xg’ over the frequency range where
xg) experiences constructive interference. This enhances the generation efficiency
(in principle) in a Doppler broadened medium (see below) by many orders of
magnitude compared with the case with no coupling field causing a dressing of the
medium. Within the enhancement in &) are both the effect of reduced absorption
and the imposition of essentially perfect phase matching for all the resonant fields
(residual mismatch arising only through the dispersion caused by the remaining
off-resonance states). These effects were first suggested as a means to enhance
nonlinear optical frequency conversion efficiencies b y Tewari and Agarwal [75]
and by Harris et al. [12] and have been further discussed by several workers
[ S l , 76-79].
Propagation of two coupled pulses in a A (Raman) type of system has recently
been discussed by Harris et al. [53,54] and Eberly et al. [SS]. This Raman scheme
is equivalent to E I T in the limit where one of the fields is a weak probe and the
other a strong-coupling field. However, if both fields 521 and 522 are strong, the
dressed atomic system reacts back on the field modes in such a way as to result in
lossless propagation through the medium for both the fields. A proper insight into
this process is best obtained by thinking in terms of C P T . Essentially the coupling
laser fields cause the formation in the atomic system of coupled IC) and uncoupled
INC) coherent superpositions of the bare atom lower states. However, there is
additionally a strong coupling between these atomic states and the two driving
fields. Hence there are also formed stable normal modes of the driving fields, one
of which is uncoupled from the ‘uncoupled’ atomic state and the other of which is
‘uncoupled’ from the coupled atomic state. These new field modes result in the
lossless propagation of pulses through a normally ‘lossy’ medium once a certain
preparation energy has been extracted from the laser fields [.56]. Other work on
matched pulse propagation in a double-V system has also been reported recently [80].
2.3. Physical constraints to electromagnetically induced transparency

For laser-induced atomic coherence processes in a real medium the mainte-
nance of the phase of the coherence during the interaction is essential for effective
interference. Any dephasing of the coherence will reduce and, eventually, nullify
the interference effects. Dephasing can arise from a variety of different sources, for
example, the excitation of several closely spaced hyperfine or Zeeman components
(for example [23,31,81]), radiative decays of state 13) (e.g. on the 13) - 11)
channel), the existence of additional photoionization channels [82], collisions
[12,58] and phase diffusion of the laser field causing deviation from the transform
limit. Therefore in experiments these parameters need to be controlled by
choosing appropriate atomic systems, density ranges, laser intensities and laser
system parameters. In general experiments with C W lasers, in which phase
fluctuations can be made small with laser bandwidths reduced to the 1 M H z
level or less, most fully approach the steady-state monochromatic limit explored in
the elementary steady-state theory of these processes.
In many experiments, (inhomogeneous) Doppler broadening presents a serious
limit since it introduces a randomization in the effective laser detunings over the
ensemble of atoms in the sample [12,83,84]. Various methods have been employed
to eliminate this effect, that is working in Doppler-free configurations [85,86], or
using cooled atoms trapped in a magneto-optical trap [87,88]. Alternatively by
working with a coupling Rabi frequency larger than the width of the (Gaussian)
Doppler profile the influence of inhomogeneous broadening can be, in effect,
overcome [12]. T o generate Rabi couplings of greater than the Doppler width
(0.03 cm-') a C W laser, with a power typically of 1 W or less, must be focused
fairly tightly. This may lead to undesirable effects such as defocusing due to the
interplay between the dressed refractive index and the transverse intensity varia-
tion across the beam waist in the region of the focus [86]. For pulsed lasers, with
intrinsically high peak power, it is not generally necessary to focus the laser to
achieve the required intensity and thus the defocusing effects are not important
and there is also the important advantage that large numbers of atoms will be
within the dressed ensemble.
Pulsed laser E I T effects are not best modelled by a steady-state treatment.
Although pulsed laser bandwidths are inevitably larger than those of CW lasers
provided that the laser is transform limited the E I T effect is not reduced. Some
recent calculations [51,78] have been somewhat artificial in including relatively
large laser linewidths pertinent to the pulsed laser in a steady-state calculation. In
fact, provided that the laser is single mode and therefore transform limited (i.e.
without excess phase fluctuation), the dephasing introduced by the laser band-
width will not sufficiently disturb the phases of the coherences responsible for
interference within the interaction time scale (i.e. the duration r p of the pulse). I t
should also be appreciated that hyperfine sublevels will in general cause dephasing
of coherences on a time scale given by the inverse of their frequency separation
Awhf [31]. In a pulsed excitation the dephasing due to hyperfine levels within the
laser bandwidth will therefore be negligible provided that the interaction time
7p < l/Awhf (i.e. if the hyperfine splittings are sufficiently small).
In practice, incorporation of laser linewidths is most usually made via the
Wiener-Levy phase diffusion model (for example [51,78,89-911). In this way
the laser phase diffusion can be added directly to the relevant density matrix
coupled equations. This model, however, predicts Lorentzian laser linewidths, and
this may severely overestimate the line wing of the laser [91]. This leads to overly
pessimistic predictions about the effects of laser linewidth in the large G?c limit
where most pulsed laser E I T experiments need to be performed in order to
overcome the Doppler restrictions. More realistic phase diffusion models are
available [92-941 but are in general harder to implement in a calculation.
486 J . P. Marangos
3. Experimental work
3.1. Related atomic coherence experiments
T h e first experimental work on laser-induced atomic coherence dates back to
the 1970s. Evidence for atomic coherence induced by the illumination of atoms by
a modulated optical lamp in conjunction with resonant radio-frequency fields
was reported still earlier [95]. Other relevant early work includes the investigation
of dressing of two-level systems by strong microwave fields (Autler-
Townes splitting) [96], and work on photon echoes in two-level systems [97].
Mollow [98,99] reported novel features, subsequently termed the Mollow triplet,
of resonance fluorescence in a two-level system. Much work on two-level systems
has been carried out since (see for instance the review and books given in
[31,100,101] but, although two-level systems remain a subject of considerable
interest (e.g. in dressed-state lasers and quantum optics experiments), our concern
here is primarily with three-level systems (and in some cases with four-level
systems).
Atomic coherence and interference in three-level systems was first seen
experimentally in the work of Orriols and co-workers [35,102,103] in Pisa and
Gray et al. [lo41 in Rochester. T h e Pisa group performed experiments that
established coherence between the Zeeman split lower levels of a sodium atom
using a multimode laser. By employing a spatially varying magnetic field they were
able to observe a series of spatially separated dark lines (resonances), correspond-
ing to the locations where the Zeeman splitting matched the frequency difference
between modes of the coupling laser. This situation corresponds to a two-photon
resonant (Raman-resonant) A scheme and thus this is the first experimental
observation of CPT. T h e experiments of Gray et al. [lo41 are similar to this but
involved establishing coherence between the hyperfine lower levels of Na.
In essence in these experiments the laser fields couple to the atoms to create
superpositions of the two ground states. One of these superpositions can interact
with the fields (see section 2.1), a so-called ‘bright’ state (i.e. the coupled state IC)),
but the other does not because of cancellation of the two driven dipoles and is thus
termed a ‘dark’ state (i.e. the non-coupled state INC)). Once formed, the popula-
tion in the system will all be optically pumped into this dark state, in a matter of a
few times the radiative decay rate. Once in the dark state there is, by definition, no
process to remove the population and so there is a ‘trapping’ of the population in
this state. There has been a considerable body of experimental work carried out on
C P T and its applications (see [32] for a full review).
Recently the basic idea of C P T has been utilized in systems with time-varying
optical fields to yield very efficient excitation of atomic and molecular states
[38,40,41]. In these coherent adiabatic population transfer schemes (e.g.
STIRAP) the non-coupled ground-state superposition I NC) of C P T is evolved
in time from being a pure state of the lowest state 11) to being a pure state of the
upper state 13) as the relative strengths of the (strong) laser couplings vary. This is
achieved by employing counter-intuitive pulse sequences, typically with identical
Gaussian pulses, in which the coupling s22 reaches its peak value around the time
that 521 is just switching on and has decreased back to almost zero when the latter is
still just passing its peak. In his way the dark state INC) in equation (3) is initially
almost purely bare state 11) (the initial state of the system), then evolves through
being an admixture of 11) and 13) whilst both fields are strong and finally
corresponds to being purely state 13). T h e bright state IC) remains unpopulated
throughout the process. Recently work has been performed which demonstrates
the utility of these effects in laser cooling of trapped atoms in a technique called
velocity-selective coherent population trapping (VSCPT) [l OS-1071.
3.2. Electromagnetically induced transparency experiments with pulsed lasers

Most of the experiments on C P T systems have concentrated on the establish-
ment of the population trap state and optical pumping of the population into it.
Similarly coherent adiabatic population transfer experiments have concentrated on
the efficient movement of the population into an atomic or molecular excited state.
In E I T work the main emphasis has been on the linear (and nonlinear) optical
response of an extended ensemble of atoms. In these experiments, measurements
are made of the transmission of the (weak) probe field through an otherwise
optically deep medium that is made possible by the presence of E I T induced by a
strong-coupling field resonant with a pair of excited states. T h e first demonstra-
tions of this were by the group of Harris at Stanford in two atomic systems S r
[lo81 and P b [109]. In both these experiments narrow-band pulsed laser radiation
was used.
In the S r experiment, the first to be reported, the atoms are initially pumped
into an excited state via a pulsed laser. It is the transition (at 337.1 nm) between
this 5s5p 'PI excited state (11)) and the 4d5d 'D2 autoionizing (12)) state that was to
be rendered transparent (it should be noted that in our notation the state labels 12)
and 13) are swapped compared with those used by these workers). A coupling laser
(at 570.3 nm), derived from a single-mode Littman dye laser, was applied between
this autoionizing state and a metastable bound state 4d5p ID2 (13)). As in the
prototypical scheme the probe field excited the system to a state 12) with a large
decay, and in the case of a no coupling laser there was a strong absorption; the S r
vapour was completely opaque at resonance, with an inferred transmission of
exp (-20 f 1). When the coupling laser was applied, the transmission at line centre
increased by a large factor to exp (-1 k 0.1). It is pointed out by these workers
that, were it not for the interference effect, the detuning from line centre induced
by the Rabi splitting would only account for a reduction in transmission to a value
exp (-7.0). T h e latter point illustrates just how important the interference effect is
in ensuring a large reduction in the absorption of the medium.
In an experiment in Pb [lo91 transparency was demonstrated within the bound
states of a collisionally broadened medium. T h e three levels are in a ladder
configuration with the probe between the ground state 11) and the
6s26p7s 'P: excited state 12), with the coupling field between 12) and the 6s26p7p
'Dl state 13). This scheme was chosen because of the approximate coincidence
between the frequency of the 1064 nm Nd-doped yttrium aluminium garnet laser
(which was frequency stabilized by injection seeding and was thus a true trans-
form-limited pulse) and the 12)-13) transition frequency; there was, however, a
6cm-' detuning of the laser field from exact resonance. An important feature of
this experiment was the role of resonance broadening, which was the dominant
broadening channel for state 12) (about 40 times larger than the radiative width).
Owing to the destructive interference between the components of 12) in the two
dressed states these collisions have no effect on transparency. In contrast the
collisions that dephase state 13), and which will affect the degree of transparency,
are not resonance collisions and hence have a small effect on this scheme. T h e
48 8 J . P . Marangos
degree of transparency induced when the coupling laser was applied at a high
power in this scheme was very high (a factor of much greater than exp (10)).
These two experiments serve to demonstrate the principle of E I T in a three-
level system. They also indicate how E I T can occur for both autoionizing and
collisional broadening of level 12). Importantly, in these experiments, pulsed lasers
were employed; although in both cases the coupling laser was nearly transform
limited, the probe laser needed no special characteristics to demonstrate the effect
(other than that its bandwidth was less than the width of the transparency hole).
Experiments using pulsed lasers continue to be important, most especially in the
context of nonlinear optics (section 3.5) and matched pulse propagation (section
3.6). A related resonant E I T scheme in Pb has more recently been explored by
Kasapi [110] as a technique for enhanced isotope discrimination. This method
utilized the resonant opacity of a low-abundance isotope (207Pb)at the frequency
where the commonest isotope in their sample (208Pb)is transparent.
3.3. Continuous-wave, electromagnetically induced transparency experiments

As well as experimental work employing pulsed lasers there has also been much
work on E I T using CW lasers. This work is motivated in part because of the
attractive prospect of creating CW short-wavelength lasers without inversion.
Indeed often the work on CW E I T has been a part of an experiment investigating
gain without inversion, and these will be dealt with in more detail below.
Two groups, one in Arkansas and other in St Andrews, Scotland, have recently
been active in performing E I T experiments on Rb. A nearly ideal A scheme is
formed in the Rb atom between the ground state 5S1/2 F = 1 and F = 2 states and
the 5P1p F = 1 state; the transitions in this case are at 780nm (separated by the
6837 MHz ground-state hyperfine splitting). Likewise a ladder scheme is formed
in this atom for the 5S,/2-5P3/2-5D~/2 states with the transition wavelengths at 780
and 775 nm. Excitation of these transitions is easily achieved using narrow-band
tuneable laser systems such as CW Ti :sapphire ring lasers or grating-stabilized
laser diodes. These schemes in Rb have the additional advantage that they can be
operated as (nearly) Doppler-free cases. These researchers have studied transpar-
ency [83,111,112], refractive index modification [85] and propagation effects
[86,113].
Observation of E I T in the Rb ladder scheme showed excellent agreement with
a steady-state calculation in which the residual effects of inhomogeneous (Doppler)
broadening are included [83]. Owing to the near frequency coincidence between
the 780 nm and 775 nm transitions in this scheme the effects of inhomogeneous
broadening on the experiment could be almost eliminated by using counter-
propagating beams. This Doppler elimination has meant that relatively low-
power (less than 10 mW) grating-stabilized laser diodes could be used to provide
both probe and coupling fields in these experiments. Using the same scheme these
workers then went on to measure the dispersive properties of E I T by incorporat-
ing the Rb vapour within a Mach-Zehnder interferometer [85]. T h e dispersion
measured at the centre frequency is inferred to be equivalent to a group velocity of
the probe: wg = c/13.2. Related effects, giving rise to a wavelength-dependent
induced focusing or defocusing, have been reported by the St Andrews group
[86,113] employing the same atomic scheme. Enhanced refraction has also been
observed in the Rb A scheme when there is an additional pumping field between
hyperfine sublevels that results in a small (non-inverted) population in the upper
state 12) of the system [114]. In this experiment an enhanced index of refraction
was found at frequencies where the absorption was zero.
Results of experiments on the R b A scheme also prove to be consistent with
theory with transparency being observed at the probe line centre with a width and
depth in reasonable agreement with the steady-state calculation [l 111. In the case
of the A scheme, co-propagating beams leads to a Doppler-free situation, possible
again because of the near equality of the probe and coupling laser wavelengths. As
mentioned before, because of the elimination of Doppler effects, it was possible to
employ very-low-power coupling lasers for these experiments. This was exploited
by the same workers to perform an experiment that stresses the quantum
interference nature of the E I T effect. By employing a coupling laser strength
SZc < 7 1 2 (i.e. the Rabi splitting is too small to give rise on its own to any significant
absorption reduction) a well developed transparency, with a depth limited only
by the laser linewidths, was reported [112]. This experiment illustrates clearly how
the additional coherence due to the coupling laser causes interference that cancels
the effect of probe absorption.
Recently work has been reported on E I T (and C P T ) in Rb and C s A schemes
for cold atoms confined in a magneto-optical trap ( M O T ) [87,88]. In the R b
system a Zeeman scheme has also been studied [115]. T h e novelty here is that the
trapped atoms are very cool (i.e. Doppler broadening is almost absent) and the
system is collisionless, making it close to ideal. Work has been carried out that
exploits these features to study the nonlinear absorption and dispersion of cold R b
atoms [116]. In cold Cs atoms in a M O T the nonlinear sum rule [117] for E I T -
type situations has been experimentally verified [118].
3.4. Investigation of gain and lasing without inversion

In systems with no excited-state population the effects of absorption cancella-
tion lead to some degree of transparency (EIT). If, however, there is an incoherent
pump into the excited states leading to a small population in the upper state of the
probe transition, then gain of the probe may occur. Therefore some of the
experiments in which E I T was seen can also be modified to investigate inversion-
less gain. In studying gain without inversion, precautions must be taken to confirm
that the system is indeed not inverted. This in practice is rather difficult and
experimenters often need to go to some trouble to confirm that there is firstly clear
evidence for gain on the probe, and secondly obtain confirmation that the system is
truly non-inverted on this transition.
An early experiment that has stimulated some discussion and provoked
subsequent work was performed by Gao et al. [119] in Changchun. A Raman-
driven four-level scheme, based on that already previously studied theoretically by
Narducci et al. [120], is formed in Na between the F = 1 and F = 2 hyperfine
levels of the 3S1/2ground state and the 3P3/2 and 3P1/2 states. A pulsed laser beam
excites the transition between the ground state levels and 3P1/2, and this creates a
coherence between the two hyperfine states. A C W probe laser tuned around the
wavelength at 598 nm of the other transition sees amplification when it is tuned to
certain frequencies if a small amount of the population is pumped into the 3P312
state. T h e population pumping is achieved in this case by running a dc discharge
through the sodium vapour. Amplification is seen and these workers claimed that
this is inversionless. This claim was questioned by some workers since there was
no independent monitoring of the excited-state populations in the experiment. I n
subsequent work this issue was addressed and firmer evidence for inversionless
conditions were established by accompanying measurements of absorption [121].
A similar scheme in K vapour has been investigated more recently by Kleinfeld
and Streater [122]. Amplification of the probe laser is again seen but, although
these workers do not experimentally confirm the absence of inversion, they
support their claim that this is indeed amplification without inversion by perform-
ing careful numerical analysis of the system with results that are comparable with
the experiment.
Clear-cut evidence for amplification without inversion has been found in other
systems. Working with a A scheme in Na, with additional incoherent pumping into
the upper state, Fry et al. [123] have seen atomic coherence effects leading to
amplification without inversion. That the atomic coherence effect is responsible
was confirmed by switching the coupling field on or off and observing that
amplification is only possible when the field is present; when this field is absent
the large population always present in the lowest state leads to absorption of the
probe. Other schemes showing unambiguous amplification without inversion have
also been reported. Picosecond pulses were used to excite atomic coherence among
the Zeeman sublevels of the Na ground state and amplification without inversion
was seen and carefully confirmed Nottelmann et al. [124]. Amplification without
inversion has also been demonstrated in a transient scheme in '12Cd through the
formation of a linear superposition of coherently populated Zeeman sublevels by
van de Veer et al. [125]. In this case, nanosecond dye lasers were used and the
coherent nature of the process was proven by the dependence of the gain on
the delay between coherence preparation and probe pulse and the magnitude of the
Zeeman splitting which controls the period for coherent transfer of population in
the atom. Recently a double-A scheme in He, driven by 877.9nm radiation in a
He-Ne discharge, was used to observe amplification at both 1079.8 nm and 61 1 nm
(up-conversion); here the evidence for amplification without inversion rests on
comparison with calculation [126].
The next step beyond amplification without inversion is to demonstrate lasing
without inversion by incorporating the gain medium within an optical cavity. In
the two experiments [23,24] where this has so far been demonstrated, amplifica-
tion without inversion was first proved using a probe through the medium, and
then the cavity was formed and lasing observed even under conditions where no
inversion is possible. T h e first of these experiments was in a V scheme formed on
the D1 and D2 lines of 87Rb,with incoherent pumping from the F = 2 hyperfine
level into the upper state of the D1 transition (lasing transition). Laser diodes were
used for all the coherent and incoherent fields. T h e incoherent pump was
generated by injecting white noise into an acousto-optical modulator modulating
one of the diodes. This work was therefore also the first experiment to achieve
amplification without inversion using diode lasers. An important conceptual
advantage of the V scheme is that there is no dressed basis within which a hidden
inversion might reside; thus there is in principle true inversionless gain. This
scheme can be considered in simplified form as a four-level system (three levels
coherently coupled and a fourth coupled via the incoherent field). There are,
however, 32 hyperfine sublevels that must be considered in a detailed analysis of
this system, which was carried out by Zibrov et al. [23] to yield predictions in good
agreement with their experiment. Subsequently work was reported by
Padmabandu et al. [24] demonstrating lasing without inversion in the A scheme in

Na considered by the same workers in [123].
3.5, Propagation effects and matched pulses

T h e earliest experiments on E I T concentrated on the demonstration of
dramatically reduced absorption in pulsed [108, 1091 and C W experiments
[83,111,112]. This was then followed by investigations within C W schemes of
the dispersion (refractive index) modification induced by these effects [85,86,113].
T h e observation of electromagnetically induced focusing-defocusing [86,113] has
implications both for future C W E I T experiments, where it introduces constraints
to the tightness of focusing, and for potential applications where control of the
spatial properties of a beam at one frequency by a beam at another (perhaps widely
different) frequency might be feasible. Theoretical treatments of the dispersive
properties of E I T [62] and of the prospect for E I T with matched pulses [53] lead
to a new perspective on how two fields can propagate in a lossless fashion through a
medium that is initially optically 'opaque' to one of them (the probe). T h e crucial
idea here is that two normal modes are formed involving the field variables and the
probability amplitudes of the coupled and uncoupled states (note the similarity to
C P T systems) [54-561. These normal modes, one consisting of the INC) atomic
state and matched fields and the other of the IC) state and antimatched fields, will
each experience lossless propagation.
An importrant point demonstrated in this analysis [56] is that in establishing
transparency if adiabatic (i.e. sufficiently intense that the time scale of the envelope
evolution is slow compared with 1/(@ + Q;)''*) matched pulses are used, then the
time scale to establish transparency for the whole medium is fast. Indeed it can be
much faster than the time scale required to establish population trapped states in
an individual atom (since the latter process requires transfer of population, i.e.
irreversible exchange of energy between the field and the medium). In fact the
transparency preparation time is equivalent to a preparation energy condition. T h e
energy required to prepare the medium in a transparent state is essentially the
coupling photon energy multiplied by the oscillator-strength-weighted number of
atoms. Once the coupling laser pulse has been on long enough for this condition to
be achieved, the medium will be rendered transparent for all subsequent times.
This preparation energy is not transferred irreversibly to the medium but is stored
reversibly in the coherent excitation of state 13) and in the coupling field.
Predictions are therefore made of the possibility of applying these effects to the
propagation of even picosecond and femtosecond pulses.
T h e first experimental verification of the modification of dispersive properties
was made by the Stanford group in 1995 [127]. Again working in their 208PbA
scheme they were able to verify predictions for pulse propagation by measuring the
delay of the probe pulse for various coupling laser strengths. Pulse velocities as
slow as c/165 whilst still with high (55%) transmission through the medium were
measured. Moreover they showed how the presence of the coupling pulse leads to a
nearly diffraction-limited transmitted beam quality for a strong probe field under
conditions where, in the absence of the coupling laser, severe spatial distortion was
present. In the regime where probe and coupling pulses are both strong there is a
departure from the weak-probe E I T limit towards the CPT-type situation; this is
developed in a subsequent experiment that investigated further the elimination of
optical self-focusing and filamentation that afflict a strong probe field [128]. I t was
492 J . P. Marangos
also shown that the delay time is connected to the attenuation of the transmitted
pulse and the Lorentzian linewidth of the forbidden 11) - 13) transition via the
relation In (EOut/Ei,,)= -713Tdelay. This has subsequently been developed to sug-
gest a general method for measuring Lorentzian linewidths [129].
Harris [69] has examined theoretically the prospects to control refractive index
using strong off-resonance pulses. In this calculation it is shown that the off-
resonance bound and continuum states lead to Stark shifts of the 13) and 11) states
that can be compensated by detuning the lasers from exact (low-field) Raman
resonance. If this is done, the additional coherence pi3 will lead to EIT-like
modification of the refractive index experienced by both pulses. This extends to
the situation for which the probe pulse is also strong. With both pulses strong,
Harris predicts that off-resonance C P T is feasible. Formation of an off-resonance
population trapped state is an important aspect of the experiment investigating the
elimination of optical-self focusing [1281 in which nonlinear refractive index effects
would lead to self-focusing, filamentation and beam break-up of the strong probe
field were it not for the formation of the coherent population trapped state.
3.6. Electromagnetically induced transparency in nonlinear frequency mixing

Enhancement of frequency mixing schemes by atomic coherence and inter-
ference was predicted by several workers [12,75,76]. T h e interference effects lead
to improved frequency mixing in resonant systems for the following reasons.
( a ) There is reduced resonant reabsorption in the medium due to the creation
of transparency.
( b ) The phase matching is optimized since near resonance the dispersion is
zero.
( c ) Although the coupling field causes some reduction in the value of the
nonlinear susceptibility at resonance, it is subject to constructive rather
than destructive interference.
For an infinitely long medium at exact resonance, effects ( a ) and ( c ) determine that
there is resonant enhancement of the four-wave mixing conversion efficiency. This
persists in a Doppler broadened medium provided that the coupling field strength
exceeds the Doppler width. In a finite medium a calculation including absorption
and phase matching must be performed to predict the enhancement and the
frequency dependence of the nonlinear response [12,51,76,78].
T h e experiments of Hakuta et al. [130,131] on sum frequency mixing in atomic
hydrogen were the first in which these types of effect were observed. I n the first of
these experiments, three-wave mixing (normally forbidden) was found to be
allowed when a dc electric field was applied to the sample. This experiment can
be viewed in the context of E I T by considering the dc field as a zero-frequency
electromagnetic field that dresses the atom, creating two Stark states between
which absorption cancellation occurs through interference. Second-harmonic
generation from a field at 243 nm, resonantly enhanced by the 1s-2s two-photon
resonance, can then occur without the generated radiation at 121.6 nm (Lyman a )
being reabsorbed.
Further experiments by the same group then demonstrated E I T enhancement
of a resonant four-wave mixing scheme in H [132]. This was achieved in a scheme
involving the Is, 2s and 3p levels, with the coupling laser at 656nm applied
between the 2s-3p states mixing with two photons from a field at 243 nm (in two-
photon resonance with ls-2s), generating a field at 103nm. Both the laser fields
used in the frequency mixing were derived from single-mode pulsed dye lasers. An
atomic H beam was specially constructed to increase the beam density and path
length so that these effects could be studied from the optically thin to the optically
thick limits [82]. T h e coexistence of undissociated H2 molecules in these systems
causes some additional effects that have also been studied in [133]. Photoionization
by the 243nm field was studied in the same experiments and shows an Autler-
Townes doublet structure. Both the sum-frequency mixing and photoionization
results agree qualitatively with the predictions of a steady-state calculation, but
this agreement is improved further if a time-dependent calculation is made [134].
In this scheme, photoionization of the 3p state was an order of magnitude greater
than that from the 2s state; this is important since the 1s-2s coherence must be
preserved for the E I T effect to survive and so decay due to photoionization is
critical. However, so long as this decay rate is significantly less than that affecting
the 3p state, the destructive interference in the absorption will persist and still lead
to enhancement of the frequency mixing.
Evidence for a quantum interference effect in the photoionization channel
arising from competition between the excitation of 2s atoms by either two photons
at 243 nm or a photon at 655 nm and another at 121.6 nm (the generated field) has
also been found [135]. This manifests itself when the product of N L (atom
density x path length) is increased from 10" to 10'' cm-I. At low values of N L
(optically thin) the Autler-Townes structure is seen in the photoionization
frequency response but, as N L is increased such that greater intensities of vacuum
ultraviolet radiation are generated, the channel for ionization at the position of the
Autler-Townes satellites cancels that driven by the laser field, and these features
disappear. Similar quantum interference effects affecting the generated field are
found to lead to a limiting density for resonantly enhanced four-wave mixing in a
scheme in R b [136]. This latter work, however, did not employ a single-mode
coupling laser and the low limiting R b concentrations (less than lo1' cmP3) found
may not be reflected in situations where E I T is present.
Electromagnetically induced phase matching was observed in an off-resonance
four-wave mixing scheme in 208Pb [137]. In this experiment the A = 6cm-'
detuning from resonance meant that transparency played little role in the en-
hancement (since the sample was optically thin). However, once the coupling laser
strength satisfied the inequality Slc > ($,opp,er + A2)"2 the dispersion (integrated
across the Doppler profile) became effectively zero, leading to perfect phase
matching and enhanced conversion efficiency.
Other recent work on this topic includes experiments on non-degenerate four-
wave mixing based on E I T in a A scheme in R b [138]. I n this experiment a
coupling field was applied resonantly (w1) whilst a second and third field (w2 and
w p ) were applied at a detuning of 450 MHz. A phase-conjugate field is created at a
+
frequency wc = wl w2 - wp. Measurements were made independently of I m x(')
and x ( ~under
) optically thin conditions, confirming that x ( ~was ) indeed enhanced
by constructive interference. If an optically dense medium was used a significant
enhancement in non-degenerate four-wave mixing was observed. A high phase-
conjugate gain achieved with very low laser powers arising from the presence of
population trapping in a double-A scheme in Na was recently reported [139, 1401.
CW resonant four-wave mixing [141] and frequency up-conversion [142] has been
observed in an experiment investigating a double-A scheme in Na dimers.
494 J. P. Marangos
T h e preparation of coherent population trapped states leading to the formation

of a large atomic coherence has been used by the group at Stanford to produce very
efficient nonlinear frequency conversion [143]. Again a A scheme in 208Pbwas used
with two strong lasers (coupling and probe) introduced into the sample; these
create nearly maximal coherence between the 11) and 13) states. T h e phase-
coherent atoms formed act as a local oscillator that is involved in mixing with a
third laser beam at 425 nm (1 112 cm-' off resonance). This generates a new field at
293 nm with an exceptionally high conversion efficiency of about 40%, as inferred
from the portion of energy from each beam known to be available for frequency
mixing. High conversion efficiency occurs since in this system the preparation of
the optimal coherence pi3 means that the nonlinear susceptibility is of the same
size as the linear susceptibility.
T h e same Pb system has recently been the subject of a proposal for a broad-
band high-efficiency optical parametric oscillator [144,145]. In this case, optimal
coherence between 11) - 13) is created in the same fashion as discussed in [143].
This coherence then acts as the local oscillator in an optical parametric down-
conversion scheme generating signal and idler waves in the infrared to far infrared.
In this system the nonlinear and linear responses of the medium are of the same
order, and high conversion efficiencies up to 10%are predicted for the middle of
the tuning range. Furthermore the device is predicted to be able to generate fields
over the entire spectrum from infrared out to almost dc fields.
Atomic coherence in solid hydrogen has recently been used to eliminate phase
mismatch in a Stokes or anti-Stokes stimulated Raman frequency conversion
scheme [146]. In this scheme the H2 molecule electronic ground state v = 0 and
v = 1 vibrational states form the lower states o a Raman scheme and, since the
11) - 13) dephasing rate is very small in appropriately prepared samples of solid
Hz, interference can occur that causes the dispersion to become negligible. Because
of the removal of the usual phase mismatch, efficient operation of this frequency
conversion scheme over a broad range of frequencies (infrared to vacuum ultra-
violet) has been inferred.
New E I T effects have also recently been predicted in four-level systems,
leading to giant Kerr nonlinearities viable even using extremely low-power laser
fields [147]. Another recent prediction is of broad-band spectral generation
associated with strong-field refractive index control [148]. In this case a pair of
lasers slightly two-photon detuned from Raman resonance are used to establish
adiabatically a superposition of two molecular states. This superposition then
mixes with the applied fields to form a broad spectrum of side bands.
4. Discussion and conclusion

E I T is one of an interrelated group of processes (including for example C P T
and coherent adiabatic population transfer) that result from externally induced
quantum-mechanical coherence and interference. These interference effects can be
significant whenever an atom or molecule is coupled to several resonant laser fields.
In this review we have chosen to highlight E I T as a process somewhat distinct
from CPT. This is a reasonable approach both in terms of the underlying physics
and because of the separate development of these topics in recent years. In contrast
with CPT, where states 11) and 13) are usually initially populated (since they are
typically sublevels of the atomic ground state), in E I T processes only state 11)
needs to be populated initially and can remain the only state with significant
population throughout the process. In C P T experiments the concern is primarily
with the changes within the state populations of individual atoms; for E I T the
interest is more especially in the optical response of the entire medium. This optical
response is determined by the coherences rather than by the populations. Within
the language of density matrices for C P T the pertinent measured quantities are
the on-diagonal elements (populations) whilst in E I T they are the off-diagonal
elements (coherences). Most importantly in the limit of a strong-coupling field
( 5 2 2 ) the coherences leading to E I T are almost instantaneously established (on a
time scale of about 1/01, which is typically a few picoseconds). For creation of
population trapping, time scales of several optical pumping times are required
(typically many nanoseconds).
A useful concept that has been used to emphasize the new aspects of these
effects is ‘phaseonium’ [13]. This term summarizes the idea of the formation of a
macroscopic medium consisting of phase-coherent quantum objects (atoms or
molecules). T h e optical properties (both linear and nonlinear) of this (coherent)
medium are very different from those of a normal (incoherent) medium. In these
laser dressed media the language of linear and nonlinear susceptibilities can be
retained only to the extent that it is recognized that all these resonant processes are
highly non-perturbative (i.e. even the ‘linear’ processes involve the coupling of
atoms with many photons). An important consequence of this is that the magni-
tudes of linear and nonlinear susceptibilities can reach equality in a phase-coherent
medium. This is in marked contrast with the normal situation in which the
nonlinear susceptibility is responsible for effects that are many orders of magni-
tude weaker than those arising from the linear susceptibility. As a consequence of
this strong nonlinearity, novel types of pulse propagation (matched pulses) and
high-efficiency frequency up-conversion can occur.
Since 1990 there has been a considerable level of activity devoted to research
into E I T and related topics. This has been motivated by the recognition of a
number of potential new applications, that is lasers without inversion, high-
efficiency nonlinear optical processes, lossless propagation of laser beams through
optically thick media, and high-efficiency population transfer via coherent adia-
batic processes. T h e earlier ideas associated with C P T (first observed in 1976) had
found application mostly as a tool of high-resolution spectroscopy rather than as a
new direction in nonlinear optics. Therefore the concept of E I T has contributed a
distinctive new thrust to work on atomic coherence and its applications.
Although it has been useful for the purposes of this review, for the reasons
stated above, to isolate E I T from other related processes it cannot be ignored that
some of the currently most exciting work on nonlinear optical generation [143,145]
and lossless pulse propagation [54,55] involve two strong laser fields (521 and 5 2 2 )
and the formation of CPT-type states. Therefore we see that it is in the integration
of ideas on atomic coherence, nonlinear optics and laser pulse propagation that
there are likely to be the most interesting future developments.
From the outset an intriguing question is whether E I T effects can be induced
in solid-state media. Experiments have recently been reported that claim to have
observed a modest degree of absorption reduction induced in ruby crystals in
ladder and V configurations using a microwave coupling field applied between
Zeeman shifted components of the ground state with optical frequency probing
[149]. In general the high dephasing rates in solid-state systems will prevent the
development of strong coherence unless a large coupling, sufficiently large that it

may risk causing optical damage to the medium, is employed [ll]. Given this
limitation it remains to be seen how solid-state E I T effects can be further
developed.
T o an extent the beauty of E I T is that it leads to exceptionally high-efficiency
nonlinear optical processes and potentially high gains (non-inverted) in a gas-phase
medium. The high nonlinear conversion efficiencies are of a magnitude normally
associated with nonlinear frequency mixing in optical crystals. T h u s a renewed
interest in gas-phase nonlinear optical devices, possessing unique capabilities (e.g.
high conversion efficiencies into the vacuum ultraviolet and far infrared), seems
likely. It is somewhat premature, however, to make firm predictions as to where
this work will lead. There remain significant questions over the fundamental
limitations of these processes (see section 2.3) that are so far only partially
resolved. Nonetheless it is reasonable to point out that in the few years since the
first experimental demonstration of E I T we have seen very substantial develop-
ments, and there seems to be no reason to suppose that this growth in the subject
will not continue for the foreseeable future.
Acknowledgments
I would like to acknowledge the very useful discussions with Christopher
Dorman, Stephen Harris, Peter Knight and Ulrich Rathe that have guided me in
writing this review. In particular I am indebted to Christoph Keitel for his reading
of the manuscript and his helpful criticism. I would also like to thank Neal Powell
for the preparation of the figures. Despite this invaluable help from colleagues, I
do of course accept full responsibility for any shortcomings, errors or omissions in
this work. This work was supported b y the U K Engineering and Physical Science
Research Council.
Bibliography
In this bibliography I have attempted to collect what I regard as the most
salient recent publications on the topic of EIT. This therefore is not intended as a
comprehensive literature review (see the references to the main article for a more
complete listing). In order to provide a concise (and therefore useful) bibliography
to the interested reader I have necessarily omitted a significant number of relevant
publications. What I have aimed to include are those papers which reported for the
first time some distinct new theoretical or experimental result that has shaped the
development of the subject.
This bibliography is divided into two sections; firstly recommended books and
reviews (arranged alphabetically by author); secondly research papers (arranged in
approximately chronological order). T h e reference number relevant to the present
text is given at the end of each reference for completeness.
Recommended books and reviews

E., 1996, Prog. Optics, 35 [32].
ARIMONDO,
HARRIS,
S. E., 1997, Phys. Today, June, 36 [ll].
KOCHAROVSKAYA,O., 1992, Phys. Rep., 219, 175 [25].
SARGENT, M., 111, SCLiI.LY, M. O., and LAMB,W. E., JR., 1974, Laser Physics (Reading,
Massachusetts: Addison-Wesley).
M. O., 1992, Phys. Rep., 219, 191 [13].
SCI:I.LY,
SHORE, B. W., 1990, Theory of Coherent Atomic Excitation (New York: Wiley) [31]; 1995,
Contempt. Phys., 36, 15 [49].
Research papers
HARRIS, S. E., 1989, Phys. Rev. Lett., 62, 1033 [20].
IniAitoc;i.t:, A,, and HARRIS, S. E., 1989, Optics Lett., 14, 1344 [21].
HARRIS, S. E., FIELD, J. E., and INAMOGLU,A., 1990, Phys. Rev. Lett., 64, 1107 [12].
SCXII,I,Y, M . O., 1991, Phys. Rev. Lett., 67, 1855 [14].
BOLLHI,K. J., I M A ~ I O G LA., L I , and HARRIS, S. E., 1991, Phys. Rev. Lett., 66, 2593 [108].
FIELD, J. E., HAHN,K. H., and HARRIS, S. E., 1991, Phys. Rev. Lett., 67, 3062 [109].
Hxiir.~r;\,K., MARMET, L., and STOICHE, B. P., 1991, Phys. Rev. Lett., 66, 596 [130].
J.AIS, M., YIS, G. Y., FIELD, J. E., and HARRIS, S. E., 1993, Optics Lett., 18, 998 [137].
Zriasc;, G. Z., H.wuiA, K., and STOICHE, B. P., 1993, Phys. Rev. Lett., 71, 3099 [132].
HARRIS, S. E., 1993, Phys. Rev. Lett., 72, 552 [53].
HARRIS, S. E., 1994, Phys. Rev. Lett., 72, 5 2 [54].
ERERI.Y, J . H., POM, M. L., and HAQ,H. R., 1994, Phys. Rev. Lett., 72, 56 [55].
FLEISCIIIIACER, M., and Sciri.12Y,M. O., 1994, Phys. Rev. A, 49, 1973 [70].
KASAPI, A., J A I NM.,
, Y I N ,G. Y., and HARRIS, S. E., 1995, Phys. Rev. Lett., 74, 2447 [127].
J A I N , M., MtxRia>i, A. J., KASAPI, A., YIN,G. Y., and HARRIS, S. E., 1995, Phys. Rev. Lett.,
75, 4385 [128].
LI, Y.-Q., and XIAO,M., 1995, Phys. Rev. A, 51, R2703.
LI, Y.-Q., and XIAO,M., 1995, Phys. Rev. A, 51, 4959 [112].
XIAO,M., LI, Y.-Q., and GEA-BANACLOCHI.:, J., 1995, Phys. Rev. Lett., 74, 666 [85].
MOSEI.EY, R. R., S H E P H E R D , S., FULTON, D. J., SINCLAIR, B. D., and DUNS,M. H., 1995,
Phys. Rev. Lett., 74, 670 [86].
HEMMER, P. R., KAW, D. P., DONOGHUE, J., CRONIS-GOLOIL~R, M., SHAHRIAR, M. S., and
KLNAR,P., 1995, Optics Lett., 20, 982 [139].
ZIBROV, A. S., L l I l i I N , M. D., HOLLBERG, L., NIKONOV, D. E., SCULI-Y, M. O., ROBINSON,
H. G., and VI.:LICIIANSKY, V. L., 1996, Phys. Rev. Lett., 76, 3935 [114].
L I , Y.-Q., and X I A ~M., , 1996, Optics Lett., 21, 1064 [138].
JAW,M. Xia, H., YIN,G. Y., M E R R i A n i , A. J., and HARRIS, S. E., 1996, Phys. Rev. Lett., 77,
4326 [143].
HARRIS, S. E., and JAIN,M., 1997, Optics Lett., 22, 636 [144].
HAKUTA, K., SUZLIKI,M., KA’TSURAGAWA, M., and L I , J. Z . , 1997, Phys. Rev. Lett., 79, 209
[146].
ZHAO,Y., Wu, C., H m , B.-Y., Kini, M. K., and AWAD,E., 1997, Phys. Rev. Lett., 79, 641
[149].
References
[l] FEYSJI.AS, R. B., and SANDS,
R. P., LEIGHTOS, M., 1966, ‘The Feynman Lectures on
Physics’, Quantum Mechanics, Vol. I11 (Reading, Massachusetts: Addison-Wesley).
[2] CARSAI., 0.. and MLYNEK, J., 1991, ‘Young’s double-slit experiment with atoms: A
simple atom interferometer’, Phys. Rev. Lett., 66, 2689.
[3] KEITH,D. W., EKWIWL~, C. K., TURCHETTE, Q. A., and PRITCHAKD, D. E., 1991, ‘An
interferometer for atoms’, Phys. Rev. Lett., 66, 2693.
[4] MLYSEK, J., BALYKIN, V., and MEYSTRE, P., 1992, ‘Optics and interferometry with
atoms’, Appl. Phys. B, 54, 319.
[5] S ~ ’ l i I L I I E I ~ R I E Y E RJ.,
, CHAPMAN, M. S., EKSTROM, C. R., HAMMOND, T. D., WEHINGER, S.,
and PRITCHARU, D. E., 1995, ‘Index of refraction of various gases for sodium matter
waves’, Phys. Rev. Lett., 74, 1043.
[6] PERCIVAL, I. C., and S T H U N Z , W. T . , 1997, ‘Detection of space-time fluctuations by a
model matter interferometer’, Proc. R. Sac., 453, 431.
498 J. P. Marangos
[7] PERCIVAL, I. C., 1997, ‘Atom interferometry, space-time and reality’, Phys. World, 10,
43.
[8] FANO, U., 1961, ‘Effects of configuration interaction on intensities and phase-shifts’,
Phys. Rev., 124, 1866.
[9] MADDEN, R. B., and CODLING, K., 1965, ‘Two-electron excitation states in helium’,
Astrophys. J . , 141, 364.
[lo] MASCHKE, K., THOMAS, P., and G ~ B E E.
L , O., 1991, ‘Fano interference in Type-I1
semiconductor quantum-well structures’, Phys. Rev. Lett., 67, 2646.
[I 11 HARRIS, S. E., 1997, ‘Electromagnetically induced transparency’, Phys. Today, June,
36.
[12] HARRIS, S. E., FIELD,J. E., and IMAMOGLU, A,, 1990, “on-linear optical processes
using electromagnetically induced transparency’, Phys. Rev. Lett., 64, 1107.
[13] SCULLY, M. O., 1992, ‘From lasers and masers to phaseonium and phasers’, Phys. Rep.,
219, 191.
[14] SCULLY, M. O., 1991, ‘Enhancement of the index of refraction via quantum coherence’,
Phys. Rev. Lett., 67, 1855.
[15] HARRIS, S. E., 1996, ‘Electromagnetically induced transparency in an ideal plasma’,
[16] KOCHAROVSKAYA, O., and KHANIN, YA. I., 1988, ‘Coherent amplification of an ultrashort
pulse in a three level medium without population inversion’, JETP Lett., 48, 630.
[17] KOCHAROVSKAYA, O., and MANDEL, P., 1990, ‘Amplification without inversion: The
double Ascheme’, Phys. Rev. A, 42, 523.
[18] S C U L L Y , M. O., ZHU,s.-Y., and GAVRIELIEDES, A., 1989, ‘Degenerate quantum-beat
laser: Lasing without inversion and inversion without lasing’, Phys. Rev. Lett., 62,
2813.
[19] ARKHIPKIN, V . G., and HELLER, Yu. I., 1983, ‘Radiation amplification without
population inversion at transitions to autoionising states’, Phys. Lett. A, 98, 12.
[20] HARRIS, S. E., 1989, ‘Lasers without inversion: Interference of a lifetime broadened
resonance’, Phys. Rev. Lett., 62, 1033.
[21] IMAMOGLU,A., and HARRIS, S. E., 1989, ‘Lasers without inversion: Interference of
dressed lifetime-broadened states’, Optics Lett., 14, 1344.
[22] FILL,E. E., SCULLY, M. O., and ZHU,S.-Y., 1990, ‘Lasing without inversion via the
Lambda quantum-beat laser in the collision dominated regime’, Optics Commun.,
77, 36.
[23] ZIBROV, A. S., LUKIN,M D., NIKONOV, D. E., HOLLBERG, L., SCULLY, M. O.,
VELICHANSKY, v. L., and ROBINSON, H. G., 1995, ‘Experimental observation of laser
oscillation without population inversion via quantum interference in Rb’, Phys. Rev.
Lett., 75, 1499.
[24] PADMABANDU, G. G., WELCH, G. R., SHUBIN, I. N., F R Y , E. S., NIKONOV,D. E., L U K I N ,
M. D., and SCULLY, M. O., 1996, ‘Laser oscillation without population inversion in
a sodium atomic beam’, Phys. Rev. Lett., 76, 2053.
[25] KOCHAROVSKAYA, O., 1992, ‘Amplification and lasing without inversion’, Phys. Rep.,
219, 175.
[26] MANDEL, P., 1993, ‘Lasing without inversion: A useful concept?’, Contemp. Phys., 34,
235.
[27] LEE, D. S., and MALLOY, K. J., 1994, ‘Analysis of reduced interband absorption
mechanisms in semiconductor quantum wells’, IEEE Jl quant. Electron., 30, 85.
[28] IMAMOGLU, A., and RAM, R. J., 1994, ‘Semiconductor lasers without population
inversion’, Optics Lett., 19, 1744.
[29] DALTON, B. J., and KNIGHT,P. L., 1983, ‘Laser Physics, Lecture notes in Physics’,
Laser Physics, Lecture Notes in Physics, Vol. 182, edited by J . D. Harvey and D. F.
Walls (Berlin: Springer).
[30] Yoo, H. I . , and EBERLY, J. H., 1985, ‘Dynamical theory of an atom with two or three
levels interacting with quantized cavity fields’, Phys. Rep., 118, 239.
[31] SHORE, B. W . , 1990, Theory of Coherent Atomic Excitation, (New York: Wiley).
[32] ARIMONDO, E., 1996, ‘Coherent population trapping in laser spectroscopy’, Prog.
Optics, 35, 257.
[33] KNIGHT, P. L., 1984, ‘Laser induced continuum structure’, Commun. at. molec. Phys.,
15, 193.
[34] KNIGHT, P. L., LAUDER, M. A., and DArroN, B. J., 1990, ‘Laser induced continuum
structure’, Phys. rep,, 190, 1.
[35] AI.ZETTA, G., GOZZINI, A., Moi, L., and OROLS, G., 1976, ‘An experimental method for
the observation of R.F. transitions and laser beam resonances in orientated Na
vapour’, Nuovo Cim. B, 36, 5.
[36] Wwri.m, R. M., and STROUD, C. R. JR, 1976, ‘Double optical resonance’, Phys. Rev.
A, 14, 1498.
[37] RmaionE, P. M., and KNimrr, P. L., 1982, ‘Population trapping and dispersion in a
three-level system’, J. Phys. B, 15, 561.
[38] OHEG,J., HIOE,F. T., and EBERLY, J . H., 1984, ‘Adiabatic following in multilevel
systems’, Phys. Rev. A, 29, 690.
[39] S t i o R E , B. W., 1995, ‘Counter intuitive physics’, Contemp. Phys., 36, 15.
[40] G . x m m , U., RUDECKI, P., SCHIEMAN, S., and BERGMAN, K., 1990, ‘Population transfer
between molecular vibrational levels by stimulated Raman-scattering with partially
overlapping laser fields’, J . chem. Phys., 92, 5363.
[41] DAN,N., OUDEJANS, L., and REUSS,J., 1990, ‘Relaxation rates of ethylene obtained
from their effect on coherent transients’, Chem. Phys., 140, 217.
[42] LOLISIS, B., and COHEN-TANNOUDJI, C., 1992, ‘Coherent population trapping and Fano
profiles’, J. Phys. 11,Paris, 2, 579.
[43] HARRIS, S. E., and MACKLIN, J. J., 1989, ‘Lasers without inversion: Single-atom
transient response’, Phys. Rev. A, 40, 4135.
[44] CotiEN, J. L., and BERMAN, P. R., 1997, ‘Amplification without inversion:
Understanding probability amplitudes, quantum interferences and Feynman rules
in a strongly driven system’, Phys. rev. A, 55, 3900.
[45] m i < HAAH, D., 1961, ‘Theory and application of the density matrix’, Rep. Prog. Phys.,
24, 304.
[46] S.ARC;ENT, M. 111, SCULLY, M. O., and LAMB, W. E. JR, 1974, ‘Laser Physics’, Laser
Physics (Reading, Massachusetts: Addison-Wesley).
[47] W i ~ c o x ,L. R., and LAMB,W. E. J R , 1960, ‘Fine structure of short-lived states of
hydrogen by a microwave-optical method. II’, Phys. Rev., 119, 1915.
[48] SC.HLOSSBERG, H. R., and J A V A N , A., 1966, ‘Saturation behaviour of a Doppler-
broadened transition involving levels with closely spaced structure’, Phys. Rev., 150,
267.
t i , T., and TOSCHEK, P., 1970, ‘Theory of three-level gas laser amplifier’, Z.
s., 236, 213.
[SO] CoiiE~-TaNNoLiDIj, C., ZAMBON, B., and ARIMONDO, E., 1993, ‘Quantum-jump
approach to dissipative processes: application to amplification without inversion’,
J. Opt. SOC.A m . B, 10, 2107.
t i , J. C., KEITEL, C. H., KNiwrr, P. L., and MARANGOS, J. P., 1996, ‘Role of
electromagnetically induced transparency in resonant four-wave mixing schemes’,
Phys. Rev. A, 53, 543.
[52] Li, Y.-Q., and XIAO,M., 1995, ‘Transient properties of an electromagnetically induced
transparency in three-level atoms’, Optics Lett., 20, 1489.
[53] HARRIS, S. E., 1993, ‘Electromagnetically induced transparency with matched pulses’,
[54] HAImis, S. E., 1994, ‘Normal-modes for electromagnetically induced transparency’,
[55] ERERLY, J . H., PONS,M. L., and HAQ,H. R., 1994, ‘Dressed-field pulses in an
absorbing medium’, Phys. Rev. Lett., 72, 56.
[56] H.mRIs, S. E., and LUO,2.-F., 1995, ‘Preparation energy for electromagnetically
induced transparency’, Phys. Rev. A, 52, R928.
[57] DOWLIXG, J. P., and BOWDEN, C. M., 1993, ‘Near dipole-dipole effects in lasing without
inversion: An enhancement of gain and absorptionless index of refraction’, Phys. rev.
Lett., 70, 1421.
[58] MOSELEY, R. R., S I N C L A I R , B. D., and DUNN, M. H., 1994, ‘Local field effects in the
three-level atom’, Optics Commun., 108, 247.
500 Electromagnetically induced transparency
[59] FIELD, J. E., 1993, ‘Vacuum-Rabi-splitting-inducedtransparency’, Phys. Rev. A, 47,

5064.
[60] FIELD,J . E., and IMAMOGLU, A., 1993, ‘Spontaneous emission into an electro-
magnetically induced transparency’, Phys. Rev. A, 48, 2486.
[61] FLEISCHHAUER, M., 1994, ‘Correlation of high-frequency phase fluctuations in electro-
magnetically induced transparency’, Phys. Rev. Lett., 72, 989.
[62] HARRIS, S. E., FIELD, J. E., and KASAPI, A., 1992, ‘Dispersive properties of electro-
magnetically induced transparency’, Phys. Rev. A, 46, R29.
[63] SCULLY, M. O., and ZHU,S.-Y., 1992, ‘Ultra-large index of refraction via quantum
interference’, Optics Commun., 87, 134.
[64] FLEISCHHAUER, M., KEITEI,,C. H., SCULLY, M. O., and Su, C., 1992, ‘Lasing without
inversion and enhancement of the index of refraction via interference of incoherent
pump processes’, Optics Commun.,87, 109.
[65] WILSON-GORDON, A. D., and FRIEDMAN, H., 1992, ‘Enhanced index of refraction: a
comparison between two- and three-level systems’, Optics Commun.,94, 238.
[66] FLEISCHHAUER, M., KEITEL,C. H., SCULLY, M. O., Su, C., ULRICH,B. T., and ZHU,
S.-Y., 1992, ‘Resonantly enhanced refractive index without absorption via atomic
coherence’, Phys. Rev. A, 46, 1468.
[67] RATHE,u., FLEISCHHAUER, M., ZHU,s.-Y., HANSCH,T. w., and SCLILLY, M. o., 1993,
“on-linear theory of index enhancement via quantum coherence and interference’,
Phys. Rev. A, 47, 4994.
[68] MANKA, A. S., DOWLING, J. P., BOWDEN, C. M., and FLEISCHHAUER, M., 1994, ‘Piezo-
photonic switching due to local-field effects in a coherently prepared medium of 3-
level atoms’, Phys. Rev. Lett., 73, 1789.
[69] HARRIS, S. E., 1994, ‘Refractive-index control with strong fields’, Optics Lett., 19,
2018.
[70] FLEISCHHAUER, M., and SCULLY, M. O., 1994, ‘Quantum sensitivity limits of an optical
magnetometer based on atomic phase coherence’, Phys. rev. A, 49, 1973.
[71] CARDIMONA, D. A,, RAYMER, M. G., and STROUD, C. R. JR, 1982, ‘Steady-state quantum
interference in resonance fluorescence’, J. Phys. B, 15, 55.
[72] ZHU,S.-Y., and SCULLY, M. O., 1996, ‘Spectral line elimination and spontaneous
emission cancellation via quantum interference’, Phys. Rev. Lett., 76, 388.
[73] ZHOU,P., and SWAIN, S., 1996, ‘Ultra-narrow spectral lines via quantum interference’,
[74] XIA, H.-R., YE, C.-Y., and ZHU, S.-Y., 1996, ‘Experimental observation of
spontaneous emission cancellation’, Phys. Rev. Lett., 77, 1032.
[75] TEWARI, S. P., and AGARWAL, G. S., 1986, ‘Control of phase-matching and nonlinear
generation in dense media by resonant fields’, Phys. Rev. Lett., 56, 1811.
[76] MOSELEY, R. R., SINCLAIR, B. D., and DUNN,M. H., 1993, ‘Enhanced conversion
efficiency in three-wave mixing via electromagnetically reduced fundamental wave
absorption’, Optics Commun.,101, 139.
[77] AGARWAL, G. S., and TEWARI, S. P., 1993, ‘Large enhancements in non-linear
generation by external electromagnetic fields’, Phys. Rev. Lett., 70, 1417.
[78] LAKSHMI, P. A., and SWAIN, S., 1991, ‘Effects of laser linewidths on nonlinear optical
processes exploiting electromagnetically induced transparency’, J . mod. Optics, 38,
2031.
[79] H A R S H A M , P. J., BLUM,o., GUSTAFSON, T. K., and KELLEY, P. L., 1993, ‘Applications
of radiative renormalisation to electromagnetically induced transparency’, Optics
Lett., 18, 1706.
[80] CERBONESCHI, E., and ARIMONDO, E., 1996, ‘Matched pulses and electromagnetically
induced transparency for the interaction of laser pulse pairs with a double-vee
system’, Optics Commun.,127, 55.
[81] FULTON, D. J., MOSELEY, R. R., SHEPHERD, S., SINCLAIR, B. D., and DUNN,M. H.,
1995, ‘Effects of Zeeman splitting on electromagnetically induced transparency’,
Optics Commun.,116, 231.
[82] ZHANC, G. Z., KASURACAWA, M., HAKUTA, K., THOMPSON, R. I., and STOICHEFF, B. P.,
1995, ‘Sum-frequency generation using strong-field coupling and induced
transparency in atomic hydrogen’, Phys. Rev. A, 52, 1584.
[83] GEA-BXNACLOCHE, J., LI, Y.-Q., J I N , S.-Z., and XIAO,M., 1995, ‘Electromagnetically
induced transparency in ladder-type inhomogeneously broadened media-theory and
experiment’, Phys. Rev. A, 51, 576.
[84] SIII.:PHERD, S., FUL‘I’ON, D. J., and DUNN,M. H., 1996, ‘Wavelength dependence of
coherently induced transparency in a Doppler-broadened cascade medium’, Phys.
Rev. A, 54, 5394.
[85] Xiao, M., L I , Y.-Q., and GEA-BANACLOCHE, J., 1995, ‘Measurements of dispersive
properties of electromagnetically induced transparency in Rubidium atoms’, Phys.
Rev. Lett., 74, 666.
[86] MOSELEY, R. R., SHEPHERD, S., FULTON, D. J., SINCLAIR, B. D., and DUNN,M. H.,
1995, ‘Spatial consequences of electromagnetically induced transparency:
Observation of electromagnetically induced focusing’, Phys. Rev. Lett., 74, 670.
[87] MISLINGA, M., MUKAI,T., WATANABE, K., and MUKAI,T., 1996, ‘Dressed-atom
spectroscopy of cold Cs atoms’, J . Opt. SOC.B, 13, 2696.
[88] HOPKINS, S. A., USADI, E., CHEN,H. x., and DURRANT, A. V., 1997, ‘Electro-
magnetically induced transparency of laser-cooled rubidium atoms in three-level A
type systems’, Optics Commun., 134, 185.
[89] AC;ARWAL, G. S., 1976, ‘Exact solution for the influence of laser temporal fluctuations
on resonance fluorescence’, Phys. Rev. Lett., 37, 1383.
[90] S\VAIS, S., 1980, ‘Theory of atomic processes in strong resonant electromagnetic
fields’, Adv. at. molec. Phys., 16, 159.
[91] DALTON, B. J., and KNIGHT, P. L., 1982, ‘The effect of laser field fluctuations on
coherent population trapping’, J . Phys. B, 15, 3997.
[92] ZOLLER, P., and LAMBROPOULOS, P., 1979, “on-Lorentzian laser lineshapes in intense
field-atom interaction’, J . Phys. B, 12, L547.
[93] DIXIT,S. N., ZOLLER,P., and LAMBROPOULOS, P., 1980, “on-Lorentzian laser
lineshapes and the reversed peak asymmetry in double optical renounces’, Phys.
Rev. A, 21, 1289.
[94] Ktws, A,, COULSTON, G. W., HE, G . Z., BERGMANN, K., and WARREN, W. S., 1992,
‘Population transfer by stimulated Raman-scattering with delayed pulses using
spectrally broad light’, J. chem. Phys., 96, 4215.
[95] BELL,W. E., and BLOOM, H. L., 1961, ‘Optically driven spin precession’, Phys. Rev.
Lett., 6 , 280.
[96] AI:.I.LER, S. H., and TOWNES, C. H., 1955, ‘Stark effect in rapidly varying fields’, Phys.
Rev., 100, 703.
[97] ADELLA,I. D., KURNIT, N. A,, and HARTMANN, S. R., 1966, ‘Photon echoes’, Phys.
Rev., 141, 391.
[98] M o i ~ o w B. , R., 1969, ‘Power spectrum of light scattered by 2-level systems’, Phys.
Rev., 188, 1969.
[99] MOLLOW, B. R., 1972, ‘Stimulated emission and absorption near resonance for driven
systems’, Phys. Rev. A, 5, 2217.
[loo] KNic;wr, P. L., and MILONI, P. W., 1980, ‘The Rabi frequency in optical spectra’,
Phys. Rep., 66, 21.
[ l o l l ALLEN,L., and EBERLY, J. H., 1982, Optical Resonance and 2-Level Atoms (New York:
Dover Publications).
[lo21 ARIMONDO, E., and ORRIOLS, G., 1976, Lett. Nuovo Cim., 17, 333.
[lo31 A L Z F ~ I ’G.,A , Moi, L., and ORRIOLS, G., 1979, ‘Nonabsorption hyperfine resonances
in a sodium vapour irradiated with a multimode dye laser’, Nuovo Cim. B, 52, 209.
[lo41 GRAY, H. R., WHITLEY, R. M., and STROUD, C. R., 1978, ‘Coherent trapping of atomic
populations’, Optics Lett., 3, 218.
[lo51 ASPECT, A., ARIMONDO, E., KAISER, R., VANSTEENKISTE, N., and COHEN-TANNOUDJI, C.,
1988, ‘Laser cooling below the one-photon recoil energy by velocity-selective
coherent population trapping’, Phys. Rev. Lett., 61, 826.
[lo61 BARDOIJ, F., SAUBAMEA, B., LAWALL, J., SHIMIZU, K., EMILE,O., WESTBROOK, C.,
ASPECT,A., and COHEN-TANNOUDJI, C., 1994, ‘Sub-recoil laser cooling with pre-
cooled atoms’, C.r. hebd. Seanc Acad. Sci. Paris, 11, 318, 877.
[lo71 DOERY, M. R, WIDMER, M. T., BELLANCA, M. J., BUELL,W. F., BERGEMANN, T. H.,
METCALF, H., and VRENBREGT, E. J. D. 1995, ‘Population accumulation in dark states
and subrecoil laser cooling’, Phys. Rev. A, 52, 2295.
[lo81 BOLLER, K. J., IMAMOCLU, A., and HARRIS, S. E., 1991, ‘Observation of electro-
magnetically induced transparency’, Phys. Rev. Lett., 66, 2593.
[lo91 FIELD, J. E., HAHN,K. H., and HARRIS, S. E., 1991, ‘Observation of electro-
magnetically induced transparency in collisionally broadened lead vapor’, Phys. Rev.
Lett., 67, 3062.
[ 1101 KASAPI, A., 1996, ‘Enhanced isotope discrimination using electromagnetically
induced transparency’, Phys. Rev. Lett., 77, 1035.
[ l l l ] L I , Y.-Q., and XIAO,M., 1995, ‘Electromagnetically induced transparency in a three-
level A system in rubidium atoms’, Phys. Rev. A, 51, R2703.
[112] LI, Y.-Q., and XIAO,M., 1995, ‘Observation of quantum interference between the
dressed states in an electromagnetically induced transparency’, Phys. Rev. A, 51,
4959.
[113] MOSELEY, R. R., SHEPHERD, S., FULTON, D. J., SINCLAIR, B. D., and DUNN,M. H.,
1996, ‘Electromagnetically induced focusing’, Phys. Rev. A, 53, 408.
[114] ZIBROV, A. S., LUKIN,M. D., HOLLBERG, L., NIKONOV, D. E., SCULLY, M. O.,
ROBINSON, H. G., and VELICHANSKY, V. L., 1996, ‘Experimental demonstration of
enhanced index of refraction via quantum coherence in Rb’, Phys. Rev. Lett., 76,
3935.
[115] DURRANT, A. V., CHEN,H. X., HOPKINS, S. A., and VACCARO, J. A., 1997
‘Experimental Observation of atomic coherence between Zeeman states in laser
cooled rubidium atoms’ (to be published).
[116] VAN DER VELDT,T., ROCH,J.-F., GRELU, P., and GRANGIER, P., 1997, ‘Nonlinear
absorption and dispersion of cold Rb-87 atoms’, Optics Commun., 137, 420.
[117] SCANDOLO, S., and BASSANI, F., 1992, “on-linear sum rules: T h e three-level and
anharmonic-oscillator modes’, Phys. Rev. B, 45, 13 257.
[118] CATALIOTTI, F. S., FORT,C., HANSCH, T. W., INCUSCO, M., and PREVEDELLI, M., 1997,
‘Electromagnetically induced transparency in cold free atoms: Test of a sum rule for
non-linear optics’, Phys. Rev. A.
[119] GAO,J., GUO,C., GUO,X., JIN, G., WANG,P., ZHAO,J., ZHANG, H., JIANG, Y.,
WANG,D., and JIANG,D., 1992, ‘Observation of light amplification without
population inversion’, Optics Commun., 93, 323.
[120] NARDUCCI, L. M., Doss, H. M., Ru, P., SCULLY, M. O., ZHU,S.-Y., and KEIWL,
C. H., 1991, ‘A simple model of a laser without inversion’, Optics Commun., 81, 379.
[121] GAO,J., ZHANG, H. Z., CUI,H. F., Guo, X. Z., JIANG, Y., WANG,Q. W., JIN,G . X.,
and LI, J. S., 1994, ‘Inversionless amplification in sodium’, Optics Commun., 110,
590.
[122] KLEINFELD, J. A., and STREATER, A. D., 1994, ‘Observation of gain due to coherence
effects in potassium-helium mixture’, Phys. Rev. A, 49, R4301.
[123] FRY, E. S.,LI, X., NIKONOV, D., PADMABANDU, G. G., SCULLY, M. O., SMITH, A. V.,
Ti-r‘rEL, F. K., WANG,C., WILKINSON, S. R., and ZHU, S.-Y., 1993, ‘Atomic
coherence effects within the sodium D1 line: Lasing without inversion via
population trapping’, Phys. Rev. Lett., 70, 3235.
[124] NOTTELMAN, A,, PETERS, C., and LANGE, W., 1993, ‘Inversionless amplification of
picosecond pulses due to Zeeman coherence’, Phys. Rev. Lett., 70, 1783.
[125] VAN DER VEER,W. E., VAN DIENST, R. J. J., D ~ N S E L M AA.,
N Nand
, VAN LINDEN V A N DEN
HUEVELL, H. B, 1993, ‘Experimental demonstration of light amplification without
population’, Phys. Rev. Lett., 70, 3243.
[126] PETERS, C., and LANG, W., 1996, ‘Laser action below threshold inversion due to
coherent population trapping’, Appl. Phys. B, 62, 221.
[127] KASAPI, A., JAIN,M., YIN, G. Y., and HARRIS, S. E., 1995, ‘Electromagnetically
induced transparency: Propagation dynamics’, Phys. Rev. Lett., 74, 2447.
[128] JAIN,M., MERRIAM, A. J., KASAPI,A., YIN, G. Y., and HARRIS,S. E., 1995,
‘Elimination of optical self-focusing by population trapping’, Phys. Rev. Lett., 75,
4385.
[129] KAS;\PI,A., YIN, G . Y . , J A I N , M., and HARRIS, S. E., 1996, ‘Measurement of

Lorentzian linewidths by pulse-propagation delay’, Phys. Rev. A, 53, 4547.
[130] H A ~ ~ I TK., L . , and SwicHEFF, B. P., 1991, ‘Electric-field-induced second-
A , MARRIE’I’,
harmonic generation with reduced absorption’, Phys. Rev. Lett., 66, 596.
[131] H ~ t i t v .K
~ ., , MmhiET, L., and STOICHEFF, B. P., 1992, ‘Nonlinear optical generation
with reduced absorption using electric-field coupling in atomic hydrogen’, Phys.
Ree. A, 45, 5152.
[132] Zii,Asc;,G . Z., H.Atiti,rA, K., and STOICHEFF, B. P., 1993, ‘Nonlinear optical generation
using electromagnetically induced transparency in hydrogen’, Phys. Rea. Lett., 71,
3099.
[133] K.ATS~IRGAWA,
M., ZHAIVG,
G. Z., and HAKLTA,
K., ‘Sum-frequency generation using
strong-field coupling in atomic hydrogen: Effect of co-existing molecular hydrogen’,
Optics Commun., 129, 212.
[134] TiiohIPsov, R. I . , STOICHEFF, B. P., ZHANG,G . Z., and HAKIJTA, K . , 1995, ‘Time-
dependent effects in non-linear generation of VUV radiation with induced
transparency’, Appl. Phys. B, 60, S129.
[135] S i ~ o a m ~ c R.
; , S. D., KATSURAGAWA, M., ZHANG,G. Z., and HAKUTA, K . , 1996,
‘Quantum interference in resonant multi-photon ionisation processes for strongly
coupled atomic systems’, Phys. Rev. A, 54, 1551.
[136] DEN(;,L., G;\ruwr, W. R., PAYNE, M. G . , and LEE,D. Z., 1996, ‘Observation of a
critical concentration in laser-induced transparency and multi-photon excitation and
ionisation in rubidium’, Optics Lett., 21, 928.
[137] J A I S , M . , YIS, G. Y., FIELD, J . E . , and HARRIS, S. E., 1993, ‘Observation of
electromagnetically induced phase matching’, Optics Lett., 18, 998.
[138] L I , Y.-Q., and XIAO,M., 1996, ‘Enhancement of non-degenerate four-wave-mixing
based on electromagnetically induced transparency in rubidium atoms’, Optics Lett.,
21, 1064.
[139] HEMhlER, P. R., KATZ,D. P., DONOGHUE, J . , CRONIN-GOLOMB,
M., SHAHRIAR, M S.,
and KLIhlAH, P., 1995, ‘Efficient low-intensity optical phase conjugation based on
coherent population trapping in sodium’, Optics Lett., 20,982.
[140] C h o v E , T . T., SIIAHRIAR,
M. S., HEMMER, P. R., KUMAR, P., SUDARSHANAM, V. S., and
CHONIS-GOLOZIB, M., 1997, ‘Distortion-free gain and noise correlation in sodium
vapor with four-wave mixing and coherent population trapping’, Optics Lett., 22,
769.
[141] B A H I NS., , HINZE, U., TIERIANN, E., and WELLEGEHAUSEN, B., 1996, ‘Continuous resonant
four-wave-mixing in double-A level configurations of Na2’, Optics Lett., 21, 1186.
[142] AmiAoNstiii,A., BALISCHEV, S., HINZE,U., TIEMANN, E., and WELLEGEHAUSEN, B.,
1992, ‘Continuous frequency up-conversion in a double-A scheme in Naz’, Appl.
Phys. B, 64, 435.
[143] J A I N , M., X I A H , . , Y I N ,G. Y., MERRIAM, A. J., and HARRIS, S. E., 1996, ‘Efficient non-
linear frequency conversion with maximal atomic coherence’, Phys. Rev. Lett., 77,
4326.
[144] H ‘ A R R R I s , S. E., and J A N , M., 1997, ‘Optical parametric oscillators pumped by
population-trapped atoms’, Optics Lett., 22, 636.
[145] H.ARRIS, s. E., YIN,G . Y., X I A ,H., and M E R R I A M , A. J., 1997, ‘Nonlinear optics at
maximum coherence’, Phil. trans. R . SOC. A (to be published).
[146] H.AKI:TA, K., S M L J KM I ,. , KATSURAGAWA, M., and LI, J. Z., 1997, ‘Self-induced phase
matching in parametric anti-Stokes stimulated Raman scattering’, Phys. Rev. Lett.,
79, 209.
[147] Sctlniini., H., and I~iAhmc;i.c, A., 1996, ‘Giant Kerr nonlinearities obtained by
electromagnetically induced transparency’, Optics Lett., 21, 1936.
[148] HARRIS, S. E., and SOKOLOV, A. V., 1997, ‘Broadband spectral generation with
refractive index control’, Phys. Rev. A, 55, R4019.
[149] ZHAO,Y., WU, C., HAM,B.-Y., K I M , M. K., and AWAD,E., 1997, ‘Microwave induced
transparency in Ruby’, Phys. Rev. Lett., 79, 641.

Ref 1

Uploaded by

Copyright:

Available Formats

Ref 1

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Ref 1

Uploaded by

Copyright:

Available Formats

Journal of Modern Optics

ISSN: 0950-0340 (Print) 1362-3044 (Online) Journal homepage: http://www.tandfonline.com/loi/tmop20

Electromagnetically induced transparency

To cite this article: J. P. Marangos (1998) Electromagnetically induced transparency, Journal of

To link to this article: https://doi.org/10.1080/09500349808231909

Published online: 03 Jul 2009.

Submit your article to this journal

Article views: 1942

Citing articles: 401 View citing articles

Full Terms & Conditions of access and use can be found at

(Received 11 August 1997)

Abstract. The subject of electromagnetically induced transparency (EIT) is

0950-0340/98 $12.00 0 1998 Taylor & Francis Ltd.

optical properties. Recently classical field interference of an analogous type has

T h e transition amplitude at the (undressed) resonant frequency (E2 - E l ) & from

calculation will be comparable with those of the steady-state method, at least in so

2.2. Physical consequences of EIT electromagnetically induced transparency

tion profile at a detuning &/2. Instead there is a destructive interference at this

there is a fundamental difference. Paying attention to the centre of the profile at

2.3. Physical constraints to electromagnetically induced transparency

3.2. Electromagnetically induced transparency experiments with pulsed lasers

3.3. Continuous-wave, electromagnetically induced transparency experiments

3.4. Investigation of gain and lasing without inversion

Padmabandu et al. [24] demonstrating lasing without inversion in the A scheme in

3.5, Propagation effects and matched pulses

3.6. Electromagnetically induced transparency in nonlinear frequency mixing

T h e preparation of coherent population trapped states leading to the formation

4. Discussion and conclusion

development of strong coherence unless a large coupling, sufficiently large that it

Recommended books and reviews

[59] FIELD, J. E., 1993, ‘Vacuum-Rabi-splitting-inducedtransparency’, Phys. Rev. A, 47,

[129] KAS;\PI,A., YIN, G . Y . , J A I N , M., and HARRIS, S. E., 1996, ‘Measurement of

You might also like

Pfad - The Proxy pFad of © 2024 Garber Painting. All rights reserved.