• Keine Ergebnisse gefunden

CHAPTER 5. ULTRAFAST DYNAMICS OF ZNO

5.5 I NTENSITY DEPENDENCE OF DYNAMICS

Ultrafast Dynamics of ZnO

64

Ultrafast Dynamics of ZnO

65 The rise time of stimulated emission clearly show a monotonous decrease with increasing excitation energy. This must be due to the enhancement in carrier-generation rate [26] with the increasing photon-flux. Assuming that absorption of each photon directly causes the generation of electron-hole pairs, the generation rate G is can be determined by the following equation.

(5.4) where α is the absorption coefficient, d is the distance into the material at which the generation-rate is measured and N0 is the incident photon-flux. For a given thickness and absorption coefficient, the generation rate is found to be directly proportional to the energy. The exciton levels will be thus be populated faster at higher pump-fluences resulting in the occurrence of the exciton-interactions earlier in the time sequence. A similar trend is observed for early time evolution of the contribution from exciton bleaching. It should be mentioned here that the anomalous rise time of bleaching for the low pump energies must be due to competition between the absorption increase from BGR and the weaker bleaching peak at such low intensities.

Figure 5.19 : Maximum area of stimulated emission, as a function of pump fluence It has already been shown in Figure 5.1 that amplitude of exciton bleaching increases non-linearly with increasing intensity. This should be due to the fact that two-photon absorption being a non-linear process will become more efficient at relatively higher intensities. The maximum amplitude of stimulated emission as a function of the pump-energy is plotted in Figure 5.19. The contribution shows a non-linear increase for fluences upto 48 mJ/cm2; but further increase in intensity results in saturation of the emission.

The bi-exponential decay behaviour of both the contributions was maintained at all excitation energies. The decay constants for exciton bleaching and stimulated emission corresponding to the excitation energies from 48 mJ/cm2 to 79 mJ/cm2 is tabulated in Table 5.1 and Table 5.2 respectively. The input intensity does not seem to have a significant influence on the rate of decay of exciton bleaching except for the amplitudes of the decay components. The ratio of amplitudes of the fast component to the slower component increases with the increasing excitation intensity. This is an indication that the trapping process becomes more efficient with the increasing availability of carriers.

Ultrafast Dynamics of ZnO

66 Fluence Decay constants (ps) Amplitude

Ratio

1b2b A1b/ A2b

48 mJ/cm2 2 28.5 1.62

64 mJ/cm2 2.1 29.8 2.56

79 mJ/cm2 2.3 29 3.62

Table 5.1: Decay-constants (relative error ~ 8 %) for bleaching contribution corresponding to different excitation energies

Fluence Decay constants (ps) Amplitude Ratio

1s2s A1s/ A2s

48 mJ/cm2 7.9 75.8 2.49

64 mJ/cm2 8.2 70.5 2.01

79 mJ/cm2 11.9 53.7 3.21

Table 5.2: Decay-constants (relative error ~ 8 %) for stimulated emission corresponding to different excitation energies

On the other hand, for the case of stimulated emission it is observed that the first decay component becomes slower with increasing input intensity while the slow component becomes faster. The increase in the density of trapped population at higher fluencies due to the increase efficiency of trapping must lead to an increase in their rate of recombination. This explains the trend observed for the slow time component. However a precise interpretation for the trend observed for the smaller of the decay constants could not be ascertained. Nonetheless it can be speculated that two or more decay channels contribute to this time constant at low intensities. As the intensity increases one or more of them might be blocked due to saturation or other effects which leads to an increase of the decay constant at higher fluences.

5.5.2 Intensity dependence of band-gap renormalization:

The absorption increase at ~ 382 nm attributed to band-gap renormalization is observed at all pump-intensities (Figure 5.20(a)). However, the magnitude of the absorption increase shows an interesting trend with increasing input intensity. An enhancement of the peak is observed when the pump-fluence is increased from 16 mJ/cm2 to 32 mJ/cm2, but on further increase of the input energy, the effect of renormalization seems to be diminished. The excited carrier density increases non-linearly with the increasing pump-fluence (clear from last subsection) and in consistent to the known theory of band-gap renormalization the band-gap energy should decreases monotonously with increasing intensity. In this context the only physically plausible explanation for the experimentally observed behavior could be

Ultrafast Dynamics of ZnO

67 that the increase of the BGR peak is masked by the non-linearly increasing peak from optical gain at higher intensities.

Figure 5.20: (a) Spectral variation at 0.3 ps and (b) Temporal evolution of area as a function of excitation energies for the band-gap renormalization effect

Both width and amplitude of the Gaussian curve added to describe the contribution was varied while the spectral peak position was kept constant at 382 nm. The area under the curve for the contribution as a function of time for different pump-intensities is plotted in Figure 5.20 (b). It is observed that the time scale till which the absorption increase persists before being overtaken by the stronger components, decrease with the increasing input intensity. This can be explained as an effect of the faster rise of the contribution from stimulated emission and exciton bleaching with increasing intensity. The increase in the rise time of the relatively weaker stimulated emission and exciton bleaching peak at lower intensities allows for the renormalization peak to be observed for a comparatively longer duration at lower fluences. This consistency of the temporal behaviour of the BGR and stimulated emission (and exciton bleaching) contribution among each other further reaffirms that our interpretation is indeed correct.

5.5.3 Intensity dependence of refractive index change and rise of lattice temperature

The amplitude of the modulations in the range of 400 nm to 600 nm is observed to increase with the increase in the pump-fluence (Figure 5.21 (a)).

Figure 5.21 Spectral variation of (a) oscillations at 0.4 ps and (b) absorption increase corresponding to thermal effects at 500 ps, for different excitation energies

(b)

(a) (b)

(a)

Ultrafas This wo value is in the respecti input-in graphs t significa

Figure We hav transmi increase increase tempera behavio Increase signatur observe thermal the high observa modulat (a)

t Dynamics o ould mean s induced by

model for ive paramet ntensities is that the con ant only wh

5.22 Temp ve already d

ssion spectr e of pump-f e in the ph ature betwe our of the pe

e in amplitu res of high ed that the l

effects) als her increase ations we

tions and th of ZnO

that a grea y the higher r simulating ters over th s shown in ntribution of

hen the phot

poral evoluti discussed th

ra is an eff fluence, mo hoton flux.

en the samp eak at ~ 37 ude as wel her lattice

ong time-re so increases e in temper

may safely he peak at ~

ater deviatio r intensity p g the obse he measured Figure 5.2 f Aca is almo ton-flux inc

ion of (a) 1 hat the long fect of incre

ore energy This in t ple before 7 nm is in l ll as red-sh temperatur esidual com s with the i rature for h y conclude

~ 377 nm is

on of the re pump-pulse erved modu d time sequ 22 (a) and

ost negligib creases.

1off and (b) g-delay abs ease in temp

will be abs turn will re and after ex line with ou hift of the p res, is cle mponent of t

increase in higher excita

e that attri indeed true

(b)

efractive in es. Both ε1o

ulations. T uence corre 5.22 (b). It ble at low in

Aca for diff sorption inc perature of sorbed by t esult in a xcitation at ur expectati peak positio

arly visible the refractiv pump-inten ation fluenc ibution ma

.

ndex from i

off and Aca w The evoluti

esponding t t is evident ntensities an

ferent pump crease in th f the sample the sample greater dif t higher flue ions (Figure on both of e. Moreove ve index (at nsity further ces. In ligh de to the

68 its original were varied

on of the o different t from the nd becomes

p-energies he transient

e. With the due to the fference of ences. The e 5.21 (b)).

which are er, it was ttributed to r reflecting ht of above long-time

Ultrafast Dynamics of ZnO

69

5.6 Effect of layer thickness on the femtosecond response of