RESULTS AND DISCUSSIONS

Selected pressure dependent fluorescence spectra are shown in Figure 1 and the corresponding high pressure values are marked in the figure. The main peaks are attributed to band edge emission of the QDs. The typical blue shift with different amount can be observed with increasing pressures below ~7 GPa, subsequently, the fluorescence disappeared abruptly above the pressure. Figure 2 is a plot of the fluorescence peaks as a function of applied hydrostatic pressures. Obviously, quadratic relationship can be observed, that is, the slope decreases with increasing pressure. It implied that the pressure coefficient is not a

2

constant, but depends on the pressures. The equation

2

( )

0

E P = E + α P + β P

can be used to fit the overall experimental data. Here,

α

and

β

are average pressure coefficients, and is the energy gap at ambient pressure. The fitting parameters for 、

E

0

E

0

α

^、

β

are 623 nm, 32 meV/GPa, and -0.001 meV/GPa², respectively. To compare with previous reports [4~6], we divided the overall data into two parts, relative low and high pressure regimes. The threshold pressure is assigned to 3 GPa due to dramatic change of the slop. This threshold is just close to transition pressure of CdSe bulk. The linear function

E P ( ) = E

₀

+ α P

can be used to fit two regimes. The pressure coefficients are 39 and 25 meV/GPa for low and high pressure regimes. The former is close to bulk value [10]. On the contrary, the latter resembles the value of nano-crystals [5].

According to above findings, actually, there is a nonlinear factor that influences the relationship between fluorescence peaks and applied pressures. This effect can be explicitly observed in relative high pressure regime. We thus consider Murnaghan equations which describes the relation between applied pressures and lattice constants,

3( )

bulk modulus, is lattice constant under atmospheric pressure, is the decrement of the lattice constant, which is usually small, induced by the applied pressures.

One can then obtain

a

0

the equation can be approximated to the form of

2

Δ

. If the external applied pressure is high enough, the

second order term cannot be neglected. Hence, the relationship between the fluorescence peaks and applied pressures exhibit quadratic behavior due to additional second order term. From the viewpoint of volume change under applied pressures, bulk modulus is not a constant and increases with increasing pressures. It implies that colloidal QDs are getting harder to be compressed in relative high pressure regime. Hence, the pressure coefficients are getting smaller.

One might argue that why we are able to observe relative high pressure regime compared to previous reports [4~6].

It may be ascribed to two reasons: First, the structures of the colloidal CdSe QDs with ZnS capped layer are more stable. Second, previous studies have shown that the transformation pressure of CdS nano-crystals can be increased by using de-ionized water instead of using the conventional methanol and ethanol solution [8] as the pressure medium. In the case of core/shell CdSe/ZnS QDs in water, fluorescence can be observed up to 6~7 GPa. This higher phase transformation pressure is possibly the main reason to exhibit the quadratic relationship.

For QDs under high pressures, the blue shift of PL is not only caused by the lattice contraction but also by the enhanced quantum confinement. For quantum confinement effect, the changes in electronic energy under applied

pressure are given by

2 2

where f is the volume compressive ratio due to pressures and

R

is the radius of the QDs. By employing the typical parameters of CdSe QDs with 4 nm in radius and set f to be about 0.8, the change of the electron energy induced by the increasing of quantum confinement due to the applied pressure up to the phase transformation is still much small. This implies the blue shift of fluorescence

3

neglected.

Raman scattering measurement is a very sensitive and powerful tool to probe the lattice property of crystalline structures. To study the lattice quadratic behavior thoroughly, Raman spectra were also recorded at the same time. Figure 3 shows the Raman spectrum of colloidal CdSe/ZnS QDs under the pressure of 1.43 GPa at room temperatures. From left to right, three peaks can be observed, and are assigned as longitudinal optical (LO) phonon of CdSe core, LO phonon of ZnS shell, and 2-LO phonon of CdSe, respectively. Figure 4 shows pressure dependent Raman spectra of colloidal CdSe/ZnS QDs under various high pressures. With the increasing of pressure, all phonon peak shift to higher frequency up to 7 GPa. Above this pressure, these CdSe related Raman peaks disappear abruptly. Combining PL with Raman scattering, it is appropriate to conclude that the structures transformation occurs at this pressure. Figure 5 (a), (b) show the Raman peak shifts of the LO phonon and 2 LO phonon of CdSe QDs as a function of applied pressures.

Both curves are all in quadratic behavior. The solid line is a

quadratic fit by equation of

ω ( ) P = ω α

₀

+ P + β P

²

to all experimental data. The average pressure coefficient for Raman shift is 4.23 cm^-1/GPa, which is consistent with previous results of bare CdSe colloidal QDs [6]. This is a clear evidence for hydrostatic pressures applied to the colloidal QD core even over-coated with a ZnS thin layer.

We can observe also that there are two pressure regimes with different slope divided at the pressure of 3 GPa. From the pressure dependent Raman spectra, the Gruneisen parameter can be derived by the equations

of ⁰ equal to 53 GPa for CdSe bulk [13]. One then obtains

γ

=0.11, and the LO phonon frequency at ambient pressure is 216 cm^-1, which is larger than the bulk value of

compressive stress resulted from the over-coated ZnS [14].

Previous work has reported that QDs over-coated with ZnS are able to increase the Raman shift with a value of 2

cm

⁻¹.

IV. CONCLUSION

We have studied the electronic and vibrational states of colloidal core/shell CdSe/ZnS QDs at room temperatures with high pressure optical measurements. Pressure dependent quadratic lattice behavior can be observed explicitly from the PL and Raman spectras up to ~7 GPa.

This quadratic relationship is consistent with the theoretical prediction. The average pressure coefficients for PL and Raman measurements, as well as Gruneisen parameter are obtained to be 32 meV/GPa , 4.2 cm^-1 and 0.11, respectively.

ACKNOWLEDGEMENTS

This work was supported by MOE-ATU and the National Science Council under the grant numbers of NSC 94-2112-M-009 -024, NSC 95-2112-M-009 -047 and 94-2120-M-033-001.

REFERENCES

[1] M. Jr Bruchez, M. Moronne, P. Gin, S. Weiss, and A. P.

Alivisatos, Science 281, 2013 (1998).

[2] W. C. W. Chan and S. M. Nie, Science 281, 2016 (1998).

[3] X. Michalet, F. F. Pinaud, L. A. Bentolila, J. M. Tsay, S.

Doose, J. J. Li, G. Sundaresan, A. M. Wu, S. S. and Gambhir, S. Weiss, Science 307, 538 (2005).

[4] W. Shan, W. Walukiewicz, J. W. Ager, K. M. Yu, and J.

Wu, Appl. Phys. Lett. 84, 67 (2003).

[5] J. Li, G. H. Li, J. B. Xia, J. B. Zhang, Y. Lin and X. R.

Xiao, J. Phys. Condens. Matter 13, 2033 (2001).

[6] A. P. Alivisatos, T. D. Harris, L. E. Brus, and A.

Jayaraman, J. Chem. Phys. 89, 5979 (1988).

[7] S. H. Wei and A. Zunger, Phys. Rev. B 60, 5404 (1999).

4

Phys. Rev. B 43, 12580 (1991). FIGURE CAPTIONS [9] M. A. Hines and P. G. Sionnest, J. Phys. Chem. 100, 468,

(1996).

Figure 1: Pressure dependent fluorescence spectra of colloidal CdSe/ZnS QDs at room temperature.

[10] A. L. Edwards and H. G. Drickamer, Phys. Rev. 122, 1149 (1961).

Figure 2: Fluorescence peak energy as a function of applied pressures.

[11] S. H. Tolbert, A. B. Herhold, C. S. Johnson, and A. P.

Alivisatos, Phys. Rev. Lett. 76, 4384 (1996).

Figure 3: Raman spectrum of colloidal CdSe/ZnS QDs at 1.43 GPa at room temperature.

[12] M. R Silvestri and J. Schroeder, J. Phys. Condens.

Matter 7, 8519 (1995).

Figure 4: Pressure dependent Raman spectra of colloidal CdSe/ZnS QDs at room temperature.

[13] S. H. Wei, snf A. Zunger, Phys. Rev. B 60, 5404 (1999). Figure 5: Raman peak shifts of (a) LO and (b) 2 LO phonon as a function of applied pressure.

[14] R. W. Meulenberg, Phys. Rev. B 70, 235311 (2004).

5

5 0 0 5 5 0 6 0 0 6 5 0 7 0 0 8 . 3 2 G P a 6 . 4 1 G P a

6 G P a 5 . 7 3 G P a

4 . 6 5 G P a 2 . 5 G P a

Normalized intensity (a.u.)

W a v e l e n g t h ( n m )

C d S e / Z n S Q D S a t r o o m t e m p e r a t u r e

0 . 9 G P a

Figure 2

0 1 2 3 4 5 6 7

2.00 2.02 2.04 2.06 2.08 2.10 2.12 2.14 2.16 2.18

α=25 meV/GPa

PL peak energy (eV)

Applied pressure (GPa) E(P)=1.99+0.032P-0.001P2

α=39 meV/GPa

6

Figure 3

Figure 4

150 200 250 300 350 400 450 500 550

0.5 0.6 0.7 0.8 0.9 1.0

ZnS LO phonon

CdSe 2LO phonon CdSe LO phonon

Normalized intensity (a.u.)

Raman shift (cm-1 )

1.43 GPa

150 200 250 300 350 400 450 500 550 600

8.32 GPa CdSe/ZnS QDs

295 k

6.0 GPa

6.41 GPa 5.73 GPa 4.65 GPa 2.5 GPa

Normalized intensity (a.u.)

Raman shift (cm

-1

₎

0.9 GPa

7

0 1 2 3 4 5 6 7

215 220 225 230 235 240

α=3.18 cm-1/GPa

Ramman shift (cm-1)

Applied pressure (GPa) R(P)=216+4.23P-0.11P2

α=4.35 cm-1/GPa (a) CdSe/ZnS LO phonon

0 1 2 3 4 5 6 7

420 430 440 450 460 470

α=6.29 cm-1/GPa

Raman shift (cm-1)

Applied pressure (GPa) α=9.39 cm-1/GPa

(b) CdSe/ZnS 2LO phonons

8

of single molecule detection technique

C. T. Yuan¹, R.W.Chou^!，Y. N. Chen^1,2, W. C. Chou¹ and D. S. Chuu¹

1Department of Electrophysics, National Chiao Tung University, HsinChu 30010, Taiwan

2National Center for Theoretical Sciences, National Cheng Kung University, Tainan 701, Taiwan

ABSTRACT

Fluorescence enhancement for colloidal CdSe quantum dots (QDs) modified by capped layers has been studied by means of single molecule detection technique. It is found that modification of the ZnS capped layer does not increase the on-time fraction but can enhance the quantum yields (QYs) of on-time duration. With the attachment of additional hexanediamine (HDA) surface ligands, both on-time fraction and QYs can be enhanced up to 2 and 13 fold for colloidal CdSe/ZnS QDs.

In this case, the fluorescence decay profile exhibits close to single exponential behavior with longer lifetimes.

1 Electronic address: [email protected]

Manuscript submitted for publishing

From the macroscopic point of view, ensemble colloidal CdSe quantum dots (QDs) are attractive fluorophores for potential applications in biological labels due to excellent fluorescence properties, such as brightness, higher photo-stability, and tunable emission spectra [1,2]. Consequently, single colloidal QDs are expected to be one of the suitable candidates for single molecular probes [3]. However, ensemble measurements merely reflect the average properties of an inhomogeneous sample. Deep studies of fluorescence properties at single QD level are necessary. With the development of single-molecule detection technique, some interesting phenomena, which are missing in the ensemble-averaged experiments, of single QDs have been discovered, such as fluorescence intermittency (blinking) and lifetime fluctuation [4,5]. This blinking behavior (fluorescence switches between on and off states under continuous excitation) is a fatal problem for practical applications. It causes fluorescence to become complex and not directly correlated with ensemble measurements. For example, QYs are not consistent between ensemble and individual QDs. In Ref [6], they proposed QY of individual QD is the same for all bright QDs, and ensemble QYs can be determined by the ratio of bright QDs to total ones.

The surfaces of colloidal QDs play an important role in determining fluorescence properties due to large surface to volume ratio. It has been demonstrated that CdSe fluorescence properties are sensitive to surface of QDs, even overcoated with a thick ZnS shell [7,8]. Attaching organic or inorganic capped layers is a popular method to enhance the ensemble fluorescence [9,10]. However, the mechanism is not explicit known, especially, from single QD viewpoint, how much of the enhancement is due to the increasing of on-time duration and how much is due to enhancement of QYs within the on-time or both. Therefore, studies on fluorescence enhancement from single QD viewpoint can facilitate to directly understand and improve the fluorescence properties of colloidal QDs.

In this work, fluorescence enhancement for colloidal CdSe QDs modified by organic and inorganic capped layer are studied by means of single molecule detection technique. It is found that the conventional ZnS capped layer does not increase the on-time fraction , however, it can enhance QYs of on-time duration. Attaching additional HDA surface ligands, both on-time fraction and QYs can be enhanced up to 2 and 13 fold for colloidal CdSe/ZnS QDs. In this case, the fluorescence decay profiles exhibit close to

single exponential behavior with longer lifetimes.

Figure 1 shows the ensemble fluorescence spectra for colloidal CdSe QDs with original TOPO ligands, high band gap materials (ZnS), and ZnS/HDA surface from the same batch. The red shift of the peak position can be observed for both ZnS and ZnS/HDA coated QDs due to exciton wavefunction extended into ZnS layers [13]. Upon attaching additional surface layers, ensemble fluorescence can be increased. The enhanced factors are 3.13, and 8.3 EXPERIMENTAL DETAILS

For single-isolated QD measurements, dilute colloidal solution of nano-molar concentration was dispersed onto a clean cover slide by spin coating.[11]

In this case, the mean separation between two individual QDs is larger than 1 micron and can be detected by far field optical microscopy.

Fluorescence measurements were performed based on a laser scanning confocal microscope equipped with a piezo-scanner with nanometer spatial resolution. The excitation pulses at the wavelength of 405 nm, 10 MHz repetition rate were focused to a nearly diffraction limited spot by an oil-immersion objective with 1.4 N.A. Fluorescence was collected by the same objective and guided to a single photon avalanche photon diodes after passing a 50 micron confocal pinhole. Then, fluorescence and relative synchronized laser reference fed into very fast electronics system (Time harp 200, PicoQuant) to process time-correlated analysis.

Time-tagged, time-resolved (TTTR) measurements by using Timeharp 200 (Picoquant GmBH) are performed on each single QD. TTTR measurements are distinct from conventional time-correlated single photon counting techniques.

For TTTR acquisition modes, both start-stop time (time between laser pulses and single photon emission) and absolute arrival time (from the experimental start to single photon emission) can be recorded simultaneously [12]. The fluorescence decay profiles can be constructed by histograming start-stop time for many cycles by time-resolved modes. The transient fluorescence trace can be formed by integrating all photons in a given bin time (1 milliseconds) by time-tagged modes. From this time trace, we can construct on-off time and burst sizes histogram to statistical analysis.

As for the ensemble fluorescence measurements, the concentrated solution sample was directly excited by a tungsten lamp filtered by a monochromator.

Fluorescence was dispersed by a spectrometer and guided to a photo-multiplied tube.

RESULTS AND DISCUSSIONS

ensemble enhancement by introducing capped layer is generally assigned to passivate surface states of QDs [10,14].

To clarify the fluorescence enhancement, we utilized single molecule detection technique to monitor the fluorescence properties of single-isolated QDs from the same batch. Figure 2 shows a

3 3 m × μ

²

fluorescence image for colloidal CdSe/ZnS QDs obtained by laser scanning confocal microscope.

Some streaky patterns with diffraction limited spots of ~300 nm can be observed and are attributed to fluorescence intermittency [4]. This is a criterion to identify the single-isolated colloidal QDs. In order to monitor individual QDs, the laser spot can be moved to specific QD position to record transient fluorescence. Figure 3 displays the fluorescence time traces with time window of 10 seconds for original TOPO (a), ZnS (b), and ZnS/HDA (c) capped QDs.

For single QDs measurements, to avoid dots to dots heterogeneity, more than 10 individual QDs for the same species can be measured and the bin time for taking every data is 1 millisecond. From fluorescence time traces, blinking phenomena can be observed clearly, especially for ZnS and ZnS/HDA mixture capped QD. This is a hallmark and evidence of the detection of a single-isolated QD [4]. The reasons for the dark periods are usually attributed to the formation of the charged QDs due to Auger ionization [15]. When a charged QD absorbs a photon and generates an exciton, it becomes a three particles system. In this case, the energy transfer from exciton to third particle is faster (~ps) than radiative recombination process (~ns) [16]. For this duration, the QD does not fluoresce, even absorbs the excitation photons. Only after the neutralization of the charged QD, it starts to emit photon. To obtain quantitative and meaningful parameters, the general approach is to define an intensity threshold from time trace to distinguish between on and off states.[17] Then, we can define on (off) duration time, which all photons are above (below) the threshold, and construct an on (off) time distribution histogram from overall experimental data.

For single QD measurements, the fluorescence enhancement factors are 2.76, and 6.8 for ZnS and ZnS/HDA mixture capped QDs compared with bare QDs by summing up the total number of emitted photons above the predefined threshold. The values are near to the results of ensemble measurements. It indicated that ensemble fluorescence enhancement by introducing surface modification is mainly due to the increasing fluorescence of individual QDs instead of total numbers of bright ones. Figure 4 shows a histogram of on-time duration plotted by log-log scale for TOPO (a), ZnS (b), ZnS/HDA (c) capped QDs. The distribution of on-time duration

mean on-time duration calculated by averaging arithmetically are 1.69, 3.81, and 8.37 ms, respectively. We discovered that the mean on-time duration can be increased 2.3, 5 fold for ZnS and ZnS/HDA coated QDs. However, the corresponding off-time also increased (not shown). This result can be explained by the ionization blinking model [15].

ZnS capped layer can block electron ejected (returned) between QDs and surrounding matrix.

Hence, the enhanced on-time accompanied with the increasing of the off-time by introducing the ZnS high band gap materials. For comparison, the ratio of the on-time to total on-off time needs to be deduced.

The on-time fractions are 0.16 and 0.15 for bare and ZnS capped QDs. It implies that the on-time fraction can not be increased by ZnS capped layer. Therefore, the fluorescence enhancement of individual QDs with ZnS layer is not originated from increasing on-time fraction. Interestingly, the on-time fraction can be enhanced up to 2.1 fold by adding HDA surface ligands.

To study this enhancement explicitly, we need to concern not only the on-time fraction, but also the emitted photon number within the on-time duration.

From fluorescence time traces of Figure 1, the intensity fluctuated within on-time for individual QDs and varied with QDs capped by different surface modifications. Figure 5 shows the histogram of the burst sizes for TOPO (a), ZnS (b), and ZnS/HDA (c) coated QDs. The burst sizes are defined as the total photon number above the threshold for a given on-time duration. From burst sizes histogram, we can obtain total emitted photon number from a specific QD by integrating column area and get mean burst sizes by arithmetic average.

It can tell us how many emitted photons within mean on-time duration. The mean burst sizes are 5.6, 38.6, and 523.5 counts, respectively. It indicates that ZnS capped layer can enhance the burst sizes (fluorescence intensity) within on-time duration instead of increasing on-time fraction. However, for ZnS/HDA capped QDs, both QYs and burst sizes can be increased. As mentioned before, this enhancement is usually assigned to passivate trap states by introducing surface modification from ensemble measurements.

To confirm the above statements, we also performed the time-resolved fluorescence measurements to monitor the photoexcited carrier dynamics of single QDs at the same time. Figure 6 shows the fluorescence decay curves for bare TOPO (a), ZnS (b), ZnS/HDA (c) capped QDs. The decay lifetime is correlated to the fluorescence intensity, for which strong intensity possesses longer lifetime. The decay curve, which is more close to single exponential behavior with longest lifetime, can be

nonradiative decay rates. In general, radiative processes are not sensitive to surface modification and only nonradiative decay can be modified. In this case, nonradiative processes can be suppressed due to passivate trap states by introducing surface modification and increase the observed lifetime.

Here, we used the stretched exponential function

( )

0

exp( ( ) ) t

I t I

^β

= − τ

to fit our experimental data.

Where

τ

is fluorescence lifetime of single QDs,

0

[1] M. J. Bruchez, M. Moronne, P. Gin, S. Weiss, and A. P. Alivisatos, Science 281, 2013 (1998).

[2] W. C. W. Chan and S. M. Nie, Science 281, 2016 (1998).

[3] C. Y. Zhang, H. C. Yeh, M. T. Kuroki, and T. H.

Wang, Nature, 4, 826 (2005).

[4] M. Nirmal, B. O. Dabbousi, M. G. Bawendi, J. J.

Macklin, J. K. Trautman, T. D. Harris, L. E. Brus, Nature, 383, 802 (1996).

β 1

< ≤

, representing the distribution of decay rates. For

β

=1, fluorescence decay curve exhibits perfect single exponential decay behavior [19], which means radiative processes dominate. For ZnS/HDA capped QDs, it possesses the longest lifetime and largest

β

due to a better degree of passivation compared with bare and only ZnS capped QDs. It implies that the QYs for individual QDs can

< ≤

, representing the distribution of decay rates. For

β

=1, fluorescence decay curve exhibits perfect single exponential decay behavior [19], which means radiative processes dominate. For ZnS/HDA capped QDs, it possesses the longest lifetime and largest

β

due to a better degree of passivation compared with bare and only ZnS capped QDs. It implies that the QYs for individual QDs can

在文檔中 量子糾纏態之研究 (頁 39-54)