110
P. Vasa
To investigate the incoherent dynamics, the differential reflectivity R/R 0
(where, R is the difference in reflectivity in presence and absence of pump pulse
and R 0 is the sample reflectivity in absence of the pump pulse) measurements are
performed using a pump-probe set-up based on an optical parametric amplifier system operating at a repetition rate of 1 kHz [8, 9, 23]. 80-fs pump pulses centered
at 620 nm are used for excitation and a white-light continuum generated in a sapphire plate provides the time-delayed probe pulses covering the range from 630 to
750 nm (schematically depicted in Fig. 5.3a). The pump wavelength is chosen such
that it is non-resonant with the J-aggregate as well as the SPP mode of the grating
for the chosen range of incidence angles. This leads to incoherent excitation of the
J-aggregates via higher energy levels [28]. The optical pumping here mainly results
in altering the exciton population via incoherent population transfer and relaxation,
effect of which is monitored by recording the differential reflectivity spectra as a
function of pump-probe delay. Both pump and probe pulses are p-polarized and
nearly collinearly focused onto the sample with a beam diameter of ∼ 100 µm at an
incidence angle θ with respect to the sample normal. Angle-resolved R/R 0 spectra
from 30
◦ to 50
◦ are recorded at different pump-probe delays t under vacuum at a
temperature of T = 77 K to minimize J-aggregate photobleaching. Control experiments are performed on a planar dye-coated gold sample without nanoslit arrays [8,
23].
Angle-resolved R/R 0 spectra recorded at pump-probe delay of 150 fs and pump
fluence of 9 µJ/cm
2 are shown in Fig. 5.3b exhibit pronounced nonlinearity near the
polariton resonance energies even at angles and wavelength considerably away from
the exciton-SPP crossing. Under the experimental conditions used in these experiments, SPP modes essentially behave like a linear oscillator and do not contribute to
the nonlinear response [8, 14]. The remarkable observation of polariton nonlinearity
clearly demonstrates that the strong radiative interaction between excitons and SPPs
induces nonlinearity in the coupled mode even in case of a ensemble. Interestingly,
instead of the saturation of absorption of J-aggregate, with enhanced reflectivity at
polariton resonance, the polariton nonlinearity R exhibits a dispersive line shape
at all pump fluences and incidence angles. As shown in Fig. 5.3b, there is a positive nonlinear signal, (enhanced probe reflectivity), at energies slightly above (UP
branch) or slightly below (LP branch) the polariton energies. The enhancement is
accompanied by reduction in reflectivity at the low-energy (UP) or high-energy (LP)
side of ω U P,L P . Apart from the dispersive nonlinear lineshape, the fluence dependence also exhibits reduction in the normal mode splitting. These results support the
conclusion that most of the strong polariton nonlinearities seen in Fig. 5.3a, b result
from a saturation of NMS [8, 14, 23, 25–27]. A qualitative lineshape analysis along
with the expression for polariton frequencies, suggest that the origin of the nonlinear
response is the transient reduction in NMS (Fig. 5.3a).
The observed nonlinear response is attributed to the effect of the incoherent optical
pumping on NMS . The pump-generated exciton density saturates the exciton number
which, in turn transiently reduces the NMS . Thus, the pump-induced exciton creation
results in a time-dependent normal mode splitting
NMS (t) = NMS
√
n 0 (t) − n 1 (t),
where NMS is the normal mode splitting observed in the weak excitation limit and
P. Vasa
To investigate the incoherent dynamics, the differential reflectivity R/R 0
(where, R is the difference in reflectivity in presence and absence of pump pulse
and R 0 is the sample reflectivity in absence of the pump pulse) measurements are
performed using a pump-probe set-up based on an optical parametric amplifier system operating at a repetition rate of 1 kHz [8, 9, 23]. 80-fs pump pulses centered
at 620 nm are used for excitation and a white-light continuum generated in a sapphire plate provides the time-delayed probe pulses covering the range from 630 to
750 nm (schematically depicted in Fig. 5.3a). The pump wavelength is chosen such
that it is non-resonant with the J-aggregate as well as the SPP mode of the grating
for the chosen range of incidence angles. This leads to incoherent excitation of the
J-aggregates via higher energy levels [28]. The optical pumping here mainly results
in altering the exciton population via incoherent population transfer and relaxation,
effect of which is monitored by recording the differential reflectivity spectra as a
function of pump-probe delay. Both pump and probe pulses are p-polarized and
nearly collinearly focused onto the sample with a beam diameter of ∼ 100 µm at an
incidence angle θ with respect to the sample normal. Angle-resolved R/R 0 spectra
from 30
◦ to 50
◦ are recorded at different pump-probe delays t under vacuum at a
temperature of T = 77 K to minimize J-aggregate photobleaching. Control experiments are performed on a planar dye-coated gold sample without nanoslit arrays [8,
23].
Angle-resolved R/R 0 spectra recorded at pump-probe delay of 150 fs and pump
fluence of 9 µJ/cm
2 are shown in Fig. 5.3b exhibit pronounced nonlinearity near the
polariton resonance energies even at angles and wavelength considerably away from
the exciton-SPP crossing. Under the experimental conditions used in these experiments, SPP modes essentially behave like a linear oscillator and do not contribute to
the nonlinear response [8, 14]. The remarkable observation of polariton nonlinearity
clearly demonstrates that the strong radiative interaction between excitons and SPPs
induces nonlinearity in the coupled mode even in case of a ensemble. Interestingly,
instead of the saturation of absorption of J-aggregate, with enhanced reflectivity at
polariton resonance, the polariton nonlinearity R exhibits a dispersive line shape
at all pump fluences and incidence angles. As shown in Fig. 5.3b, there is a positive nonlinear signal, (enhanced probe reflectivity), at energies slightly above (UP
branch) or slightly below (LP branch) the polariton energies. The enhancement is
accompanied by reduction in reflectivity at the low-energy (UP) or high-energy (LP)
side of ω U P,L P . Apart from the dispersive nonlinear lineshape, the fluence dependence also exhibits reduction in the normal mode splitting. These results support the
conclusion that most of the strong polariton nonlinearities seen in Fig. 5.3a, b result
from a saturation of NMS [8, 14, 23, 25–27]. A qualitative lineshape analysis along
with the expression for polariton frequencies, suggest that the origin of the nonlinear
response is the transient reduction in NMS (Fig. 5.3a).
The observed nonlinear response is attributed to the effect of the incoherent optical
pumping on NMS . The pump-generated exciton density saturates the exciton number
which, in turn transiently reduces the NMS . Thus, the pump-induced exciton creation
results in a time-dependent normal mode splitting
NMS (t) = NMS
√
n 0 (t) − n 1 (t),
where NMS is the normal mode splitting observed in the weak excitation limit and