arXiv:1102.0364v1 [physics.optics] 2 Feb 2011
Conversion between electromagnetically induced transparency and
absorption in a lambda system
Sapam Ranjita Chanu, Kanhaiya Pandey,1and Vasant Natarajan∗
Department of Physics, Indian Institute of Science, Bangalore 560 012, India
1Presently at Centre for Quantum Technologies, National University of Singapore, Singapore 117543, Singapore
∗Corresponding author: email@example.com
Compiled February 3, 2011
We show that it is possible to change from a subnatural electromagnetically induced transparency (EIT)
feature to a subnatural electromagnetically induced absorption (EIA) feature in the same three-level Λ system.
The change is effected by turning on a second control beam counter-propagating with respect to the first beam.
We observe this change in the D2line of Rb in a room-temperature vapor cell. The observations are supported
by density-matrix analysis of the sublevel structure, but can also be understood qualitatively from the effects
of optical pumping.
c ? 2011 Optical Society of America
OCIS codes: 020.4180, 270.1670
Electromagnetically induced transparency (EIT) is a
well-understood phenomenon in three-level systems [1,2].
The effect arises when a strong control laser on one tran-
sition is used to modify the absorption properties of a
weak probe laser on a second transition. The control
laser creates new dressed states  of the combined atom-
photon system due to the AC Stark shift, and it is a com-
bination of this energy shift and the interference between
the dressed states that gives rise to the EIT phenomenon.
In addition, the control laser can cause population redis-
tribution due to optical pumping, which further affects
the absorption of the probe. The phenomenon of EIT has
wide-ranging applications, such as high-resolution spec-
troscopy [4,5], nonlinear optics , slowing of light ,
and quantum computation. EIT is particularly impor-
tant in lambda (Λ) systems, where the presence of two
ground levels allows the existence of a dark state .
Thus, probe absorption is exactly zero at line center, so
that the width of the EIT resonance is subnatural when
the control laser has a sufficiently low power. Even at
high powers, the width of the EIT resonance in hot va-
por shows a surprising narrowing due to Doppler aver-
The phenomenon of electromagnetically induced ab-
sorption (EIA) [10, 11], though related to EIT, is not
that well studied. In EIA, the probe laser shows en-
hanced absorption in the presence of a control laser. In
this work, we show that it is possible to convert from EIT
to EIA in the same Λ system by using two control beams,
counter-propagating with respect to each other. All the
other aspects of the phenomenon remain the same, just
the narrow resonance changes from a transparency peak
to an absorption dip. The observations are supported
by a density-matrix calculation of the complete sublevel
structure. It is also possible to understand this change
qualitatively as arising from optical pumping among the
various magnetic sublevels induced by the control beams.
The population is redistributed so that the sublevels
form a Λ-type system with one control beam and N-type
Fig. 1. (Color online) Partial energy-level diagram in Rb
showing the transitions of the D2line used to form the
systems  with two control beams.
We have performed these experiments in a room-
temperature Rb vapor cell using the 5S1/2→ 5P3/2tran-
sition, which is the well-known D2line at 780 nm. The
relevant energy levels forming the lambda system are
shown in Fig. 1. The two lower levels are the F = 1 and
F = 2 hyperfine levels of the ground state. The common
upper level is the F = 2 level of the excited state. The
control laser is on the F = 2 ↔ F′= 2 transition, while
the probe laser is on the F = 1 ↔ F′= 2 transition. The
powers in the two lasers are characterized by the respec-
tive Rabi frequencies Ω. The decay rate to the ground
levels is Γ, which is 2π × 6 MHz for these transitions.
The experimental schematic is shown in Fig. 2. The
probe and control beams are derived from two home-
built grating-stabilized diode laser systems . The in-
stantaneous linewidth of the lasers after stabilization is
about 1 MHz. The beams are elliptic and have a size
of 2 mm × 3 mm. The probe beam (P) and the co-
propagating control beam (CP) have orthogonal polar-
izations. This allows us to mix and separate them us-
ing polarizing beam splitter cubes (PBS’s), and detect
only the probe beam. The λ/4 retardation plate ensures
that the two beams have orthogonal circular polariza-
tions (σ−for P and σ+for CP) in the experimental cell.
The counter-propagating control beam (CO) also passes
through a λ/4 retardation plate, and therefore has the
Fig. 2. (Color online) Schematic of the experiment. Fig-
ure key: BS – beamplitter, λ/2 – halfwave retardation
plate, λ/4 – quarterwave retardation plate, M – mirror,
PBS – polarizing beamsplitter cube, PD – photodiode.
same circular polarization as the probe beam (σ−). The
experimental cell is a cylindrical vapor cell of dimensions
25-mm diameter × 50-mm length. The cell has a three
layer magnetic shield that reduces the stray fields to be-
low 0.1 mG.
The spectra were obtained with the probe laser locked
to the F = 2 ↔ F′= 1 hyperfine transition, while the
control laser was scanned around the F = 2 ↔ F′= 2
transition. Using a locked probe beam and a scanning
control beam is a technique that we have developed to
overcome the first-order Doppler effect (this is different
from usual EIT experiments where the control beam is
fixed and the probe beam is scanned) [13,14]. In effect,
the locked probe beam addresses only the zero-velocity
atoms, and its absorption remains flat (Doppler free) un-
til it is modified by the control beam. The probe beam
was locked using fm modulation spectroscopy in a sepa-
rate saturated-absorption spectroscopy cell.
The observations are shown in Fig. 3. The probe-
absorption spectrum on the top is the usual EIT line-
shape obtained with just the co-propagating control
beam on. The powers are 0.9 mW for the control and
20 µW for the probe. Though one expects the spectrum
away from line center to be flat in our technique of scan-
ning only the control beam, there is a broad (30-MHz
wide) enhanced absorption peak away from line center.
This is due to optical pumping by the strong control
laser, which transfers population into the F = 1 level
and thereby increases probe absorption . The EIT dip
at line center is only 4.4 MHz wide (0.7Γ) and is subnat-
ural. When the counter-propagating control beam (with
a power of 1.5 mW) is also turned on, the EIT dip gets
inverted into an EIA peak. The width of the resonance
remains subnatural at ∼ 4.7 MHz.
In order to understand these line shapes theoretically,
we consider the complete sublevel structure shown in
Fig. 4(a). The probe beam and the counter-propagating
control beam are σ−polarized, and couple sublevels with
the selection rule ∆m = −1. The co-propagating control
beam is σ+polarized and couples sublevels with the se-
lection rule ∆m = +1. In the density-matrix approach,
Control detuning (MHz)
Probe absorption (arb. units)
EIT – CP (σ
+ ) 0.9 mW
− ) Off
EIA – CP (σ
+ ) 0.9 mW
− ) 1.5 mW
Fig. 3. (Color online) EIT to EIA conversion. Change of
lineshape from enhanced probe transmission (EIT) with
only the co-propagating control beam to enhanced probe
absorption (EIA) with both control beams. The two con-
trol beams have orthogonal circular polarizations.
probe absorption is proportional to Im(ρ12), where |1?
and |2? are the two levels coupled by the probe laser. The
calculated spectra, which take into account all the al-
lowed transitions (weighted by their respective Clebsch-
Gordan coefficients) and Doppler averaging in the room-
temperature vapor, are shown in Fig. 4(b). The time
evolution is stopped at 30 µs, by which time all the
transients have died down. The Rabi frequencies of the
two control beams are taken to be 12 MHz and 15.5
MHz, respectively, which corresponds to the experimen-
tal powers if we take into account the ∼ 10% loss at the
cell entrance window. There are no other adjustable pa-
rameters. The calculations reproduce the main features
of the observed spectra, namely EIT with one control
beam and EIA with both control beams. However, the
calculated linewidths for both the EIT and EIA features
are smaller than the observed values, probably because
of the 1-MHz linewidth of the probe laser (which is not
included in the calculation) and misalignment between
The conversion from EIT to EIA can be understood
qualitatively if we consider the effects of optical pump-
ing. Optical pumping implies that only a few sublevels
contribute to probe absorption. In the figure, these sub-
levels are labeled from |1? to |6?, and the relevant cou-
plings are shown using bold lines. When only P and CP
are on, we have a Λ-type system with the levels |1? ↔
|2? ↔ |5?. It is well known that a Λ system shows EIT .
When CO is also turned on, the bold lines show two N-
type systems, involving the levels |1? ↔ |2? ↔ |3? ↔ |4?
and the levels |1? ↔ |2? ↔ |5? ↔ |6?. It is again known
that N systems show EIA at line center .
While optical pumping is most easily seen when the
+2 +10 −1−2
-20 -100 10 20
Control detuning (MHz)
Fig. 4. (Color online) (a) Complete sublevel structure
showing the levels coupled by P (σ−), CP (σ+), and CO
(σ−). The bold lines show coupling between the rele-
vant sublevels when the effects of optical pumping are
taken into account. With only P and CP, we get a Λ-
type system – |1? ↔ |2? ↔ |5?. When CO is also on, we
get two N-type systems – |1? ↔ |2? ↔ |3? ↔ |4? and
|1? ↔ |2? ↔ |5? ↔ |6?. (b) Calculated probe absorption
plotted as Im(ρ12) for the two cases. The respective Rabi
frequencies (in MHz) are listed.
control beams are circularly polarized, it also happens
when the control beams are linearly polarized. Exper-
imentally, this configuration is easy to realize, we just
have to remove the λ/4 waveplates on either side of the
cell (see Fig. 2). The two control beams then have the
same linear polarization, which is orthogonal to the lin-
ear polarization of the probe beam. The observations
in this case are shown in Fig. 5. As before, we get
the usual EIT lineshape when only one control beam
is present, and it changes to an EIA lineshape when the
second control beam is also turned on. The spectra were
recorded with powers of 100 µW in the probe beam and
3.4 mW in the two control beams. The higher probe
power gives more signal-to-noise ratio and more promi-
In conclusion, we have shown that it is possible to
convert from enhanced probe transmission to enhanced
probe absorption in the same Λ system. We get the well-
known subnatural EIT feature when one co-propagating
control beam is on, which transforms to EIA when a sec-
ond counter-propagating control beam is turned on. The
observations are explained by density-matrix analysis of
the complete sublevel structure. However, it can also be
Control detuning (MHz)
CP(lin) 3.4 mW
CP(lin) 3.4 mW
CO(lin) 3.4 mW
Fig. 5. (Color online) EIT to EIA conversion with lin-
early polarized beams. The two control beams are lin-
early polarized in the same direction, which is orthogonal
to the probe polarization.
understood qualitatively if one takes into account the
effects of optical pumping among the different magnetic
sublevels, so that the relevant sublevels form a Λ system
with one control beam and N systems with two control
beams. The ability to not just reduce probe absorption
(normal EIT) but also enhance it in a straightforward
way should vastly increase the applications of this phe-
This work was supported by the Department of Sci-
ence and Technology, Government of India. One of us
(K.P.) acknowledges financial support from the Council
of Scientific and Industrial Research, India.
1. S. P. Tewari and G. S. Agarwal, “Control of phase
matching and nonlinear generation in dense media by
resonant fields,” Phys. Rev. Lett. 56, 1811–1814 (1986).
2. K.-J. Boller, A. Imamo˘ glu, and S. E. Harris, “Observa-
tion of electromagnetically induced transparency,” Phys.
Rev. Lett. 66, 2593–2596 (1991).
3. C. Cohen-Tannoudji and S. Reynaud, “Modification of
resonance Raman scattering in very intense laser fields,”
J. Phys. B. 10, 365–383 (1977).
4. U. D. Rapol and V. Natarajan, “Precise measurement
of hyperfine intervals using avoided crossing of dressed
states,” Europhys. Lett. 60, 195–200 (2002).
5. A. Krishna, K. Pandey, A. Wasan, and V. Natara-
jan, “High-resolution hyperfine spectroscopy of excited
states using electromagnetically induced transparency,”
Europhys. Lett. 72, 221–227 (2005).
6. S. E. Harris, J. E. Field, and A. Imamo˘ glu, “Nonlin-
ear optical processes using electromagnetically induced
transparency,” Phys. Rev. Lett. 64, 1107–1110 (1990).
7. L. V. Hau, S. E. Harris, Z. Dutton, and C. H. Behroozi,
“Light speed reduction to 17 metres per second in an
ultracold atomic gas,” Nature (London) 397, 594–598
8. M. Fleischhauer, A. Imamoglu, and J. P. Marangos,
“Electromagnetically induced transparency: Optics in
coherent media,” Rev. Mod. Phys. 77, 633 (2005).
9. S. M. Iftiquar, G. R. Karve, and V. Natarajan,
“Subnatural linewidth for probe absorption in an
electromagnetically-induced-transparency medium due Download full-text
to Doppler averaging,” Phys. Rev. A 77, 063807 (2008).
10. A. Lezama, S. Barreiro, and A. M. Akulshin, “Elec-
tromagnetically induced absorption,” Phys. Rev. A 59,
11. C. Goren, A. D. Wilson-Gordon, M. Rosenbluh, and
H. Friedmann, “Atomic four-level n systems,” Phys.
Rev. A 69, 053818 (2004).
12. A. Banerjee, U. D. Rapol, A. Wasan, and V. Natarajan,
“High-accuracy wavemeter based on a stabilized diode
laser,” Appl. Phys. Lett. 79, 2139–2141 (2001).
13. D. Das and V. Natarajan, “Hyperfine spectroscopy on
the 6P3/2state of133Cs using coherent control,” Euro-
phys. Lett. 72, 740–746 (2005).
14. S. M. Iftiquar and V. Natarajan, “Line narrowing of
electromagnetically induced transparency in Rb with a
longitudinal magnetic field,” Phys. Rev. A 79, 013808