1. Introduction

Quantum dot (QD) nanocrystals are non-organic nanocrystals whose strong luminescence and fluorescence are widely used in material science and life science applications [1–7]. Especially in life science, surface-coated QDs have been widely used in cellular microstructural probes [3–5,8]. Their high brightness, photostability, and quantum yield make them suitable for imaging techniques. With the advancements in nanotechnology, it has become possible to synthesize rod-like QDs (quantum rods: QRs) [9]. QDs/QRs have been used for studying three-dimensional molecular orientations by polarized single molecule microscopy. Moreover, QDs/QRs have also been used for environmental sensing [10,11]. In the case of pH-sensitive QDs, the pH change is evaluated by detecting the decrease or increase in the fluorescence intensity. However, the pH change is also thought to cause other photophysical effects on QDs.

Blinking of QDs/QRs is a photophysical property [12,13] which is likely to vary with the pH of their solution. It is important for the understanding and classification of QDs/QRs. Blinking has been quantified using single-particle detection [12] and fluorescence correlation spectroscopy (FCS) [13,14]. By single particle detection, the blinking event of QDs/QRs can be easily observed by statistically analyzing their fluorescent on- and off-times in the emitted photon trace. The relaxation time of blinking ranges from several ms to 100 s. Blinking relaxation time of several μs has also been reported using FCS. The main reason for this is the difference in the properties of the investigated QDs/QRs and the time-resolution of the methods used for measuring their blinking relaxation times. The single particle method can used to observe only relatively slow events because it uses a charge coupled device (CCD) camera. In the case of FCS, the time resolution is high. This property makes suitable for analyzing rapidly changing phenomena.

Antibunching is a blinking-like phenomenon [15]. A single fluorophore emits only one photon during its lifetime after which it can emit another photon. This short-time blinking is called antibunching. It can be theoretically analyzed by autocorrelation analysis. We used two avalanche photodiodes (APDs) system to analyze the blinking property of QDs/QRs in this study. Previously, the properties of QDs were investigated by FCS and fluorescence cross-correlation spectroscopy (FCCS) [3,7,13,14,16–24]. FCS is a fluorescence fluctuation analyzing method which utilizes confocal microscopy. FCS was developed in the 1970s [25–27]. With the development of confocal fluorescence microscopy, the FCS technique was also improved gradually. FCCS is expanded technique of FCS. It needs more than two fluorophores in the same sample to estimate the molecular interaction and diffusion coefficient, simultaneously. FCS can be used to estimate the molecular hydrodynamic radius from the translational diffusion coefficient using the Stokes-Einstein equation. Thus, FCS allows the characterization of QDs. The measurement of the hydrodynamic radius of water-soluble bare CdTe QDs by FCS was reported in 2005 [16]. In this study, the QD radius was estimated by FCS. The estimated value was consistent with that obtained by electron microscopy. Liedl et al. reported a molecular crowding relationship between the translational diffusion and hydrodynamic radius of QDs [17]. Moreover, QDs have broad excitation spectra. This property is exploited for double-excitation FCCS using a single laser source [28]. FCS has also been used to analyze QRs. The rotational diffusion of QRs is studied by polarization-dependent FCS [29,30].

A number of pH indicators have been reported to be suitable for living cell and in vivo imaging. Genetically encoded pH indicators such as SypHer2 are used for measuring the pH in living cells and mice [31]. A great advantage of ratiometric pH indicators is that they allow the measurement of pH under physiological conditions (pH = 6.7–8.0). However, fluorescence proteins suffer from photobleaching and require specially selected excitation optics because the fluorescent molecules in them have their characteristic excitation peak wavelengths. Organic pH indicators are also widely used [32]. BCDEF is an intensity-based organic pH indicator. Such organic molecules also have poor photobleaching resistance. On the other hand, QDs/QRs show noticeably better fluorescence. The photobleaching resistance, wide excitation spectra, and narrow emission spectra of QDs/QRs make them suitable for use in pH probes for solutions or living cells.

Herein, we demonstrate that the blinking, antibunching, and aggregation of QDs are pH-dependent. A high-speed FCS technique was used to analyze the variations in the blinking property of QDs in high-pH solutions. Moreover, the correlation between their aggregation and blinking property is reported for the first time.

2. Materials and methods

2.1 Experimental setup

The FCS experiment was performed using an FCS setup developed in-house, as described previously [33]. A commercial upright microscope (Axioplan2, Carl Zeiss, Germany) was used to expand and construct the system, as described previously. An Ar + laser (IMA101020BOS, Melles Griot, USA) with a wavelength of 488 nm was used as the excitation source for QDs/QRs. The guided excitation light was reflected by a dichroic mirror (FT510, Carl Zeiss, Germany) towards the samples and was focused by an objective lens (C-Apochromat 40x, numerical aperture = 1.2, water immersion, Carl Zeiss, Germany). The emitted fluorescent light was collected using the same objective lens and was made to pass through an emission filter (BLP01-488R-25, Semrock, USA). To exclude the undesired noise signal from the detectors, the collected fluorescence was divided into two ways using a polarization-independent 50:50 beam splitter (Thorlabs, USA) and was then coupled to a multimodal fiber (M31L02, Thorlabs, USA). The core diameter of the multimodal fiber was 62.5 μm. The confocal volume of our system was calculated to be 2.8 fL. This volume is sufficient for the analysis of nanocrystals. The light guided by the fibers was detected by two APDs (APD1 and APD2, SPCM-CD3017, PerkinElmer, USA). The photon-counting data was sent to a digital hardware correlator (Flex02-01D, Correlator.com, USA) to obtain the correlation function at a rate faster than that obtained with the previous setup. The time resolution of Flex02-01D was 1.56 ns, which is eight times faster than the time resolution used in the previous setup. The intensity cross-correlation function (CCF) was calculated by Flex02-01D using the photon-counting data.

2.2 QDs and Quantum rods (QRs)

CdSe-ZnS QDs, the diameter is 15-20 nm, were purchased from Thermo Fisher Scientific (USA). Two types of QDs were used. The maximum emission wavelength of the QDs was 525 (QD525) and 545 nm (QD545). The surface of the QDs was coated by polymers and –COOH to make them water-soluble. The QRs were synthesized according to a previously reported method [34] and were coated with glutathione (GSH) to make them water-soluble. These QDs/QRs were stable in neutral solutions. The QRs had a dimension of 19.5 nm x 4.4 nm, as determined from the transmission electron microscopy (TEM) images [Figs. 1(a)–1(c)] reported previously [29]. Thus, in solutions, their hydrodynamic radii did not change significantly.

 

Fig. 1 QR quantification. (a) TEM image of QRs. Yellow region of interests (ROIs) indicate the QRs selected for quantification. (b) Length distribution. (c) Width distribution.

Download Full Size | PDF

2.3 Preparation of QDs/QRs solution

Each QDs/QR was diluted in a PBS buffer containing 170 mM NaCl, 27 mM KCl, 81 mM Na₂HPO₄, and 14.7 mM KH₂PO₄. The pH-controlled buffer was prepared by titrating H₂NaPO₄, HNa₂PO₄, and NaOH. Each QDs/QR was diluted in a pH solution. The final concentration of QDs/QRs is 5 nM.

2.4 Nonlinear curve fitting

The CCF was calculated by Flex02-01D. The raw correlation-data was fitted by a model equation (Eq. (1) using Origin®2016 (MathWorks, USA).

Here, $G_{1,2}(\tau )$ is a normalized CCF calculated by using the photon-counting data. The photon data sets were detected by APD1 and APD2. $\tau$ is the delay time. $〈I_i(t)〉$ is the ensemble average of the photon intensity detected by a detector $i$. $G_D(\tau )$ is a correlation function of translational diffusion defined as:

where $N$ is the averaged particle number in the observation volume. ${\tau }_D$ is the relaxation time of the translational diffusion. $s$ is the structural parameter which is defined as the aspect ratio of the observation volume. $G_B(\tau )$ is the correlation function of blinking and is defined as:

Here, $f_B$ is the fraction of blinking against the translational diffusion. ${\tau }_B$ is the relaxation time of blinking. $G_{AB}(\tau )$ is a correlation function of antibunching and is defined as:

Here, ${\tau }_{AB}$ is the relaxation time of antibunching.

3. Results and discussions

3.1 Antibunching and blinking of QDs/QRs measured by FCS using CW laser

First, the FCS experiment was performed in the PBS buffer to observe the property of each QD/QR sample in aqueous conditions and to evaluate the sensitivity of this system. Figure 2 shows results of this experiment. Because of the improvement in the time-resolution, the antibunching of the QDs was clearly observed. The lifetime of QDs in neutral solutions has been reported to be 10–20 ns using a pulsed laser and TCSPC. The FCS result obtained in this study was consistent with the previous results. The calculated correlation function showed a precipitous rise from the left edge to approximately 50 ns. Thus, this is the antibunching fraction.

 

Fig. 2 Normalized CCFs of QDs/QRs in PBS. Antibunching, blinking, and translational diffusion were simultaneously observed in the CCFs. Measurement time: 120 s.

Download Full Size | PDF

In the case of the QRs, the initial rising was completed in 10–20 ns (as QDs curves were shifted to left), indicating that the fluorescence lifetime of the QRs was shorter than that of the QDs. On the other hand, the ridge of the correlation function was not flat but slanting. It can be defined as the blinking fraction. It is difficult to observe such a rapidly changing phenomenon using commercial FCS systems with a CW laser. Thus, the tool would be useful for observing nanosecond phenomena in freely diffusing solutions. The CCF of the QRs showed a large shift to the right [Fig. 2].

A right shift in the correlation function of a fluorescent particle indicates that the particle diffusion is slower than the non-shifted correlation function. Thus, in the case of the QRs, initially an oligomer or aggregation was formed in PBS because the size of single QRs are thorough to be as approximately same to that of QDs from Fig. 1.

3.2 Oligomerization and variation in blinking in pH-controlled solutions

Figure 3(a) shows the normalized CCFs of the QDs/QRs under different pH conditions. Both QD525 and QD545 showed a pH-dependent increase in the CCF amplitude at higher pH values (pH = 11–13). The CCF change at high pH values is believed to be caused by fluorescent blinking of the QDs because the time region of this change in the CCFs was approximately 10⁻⁷–10⁻⁴ s. This time region was well separated from that of the translational diffusion or antibunching [Fig. 2]. However, no pH-dependent changes were observed in the case of QR595. The differences in the properties of QDs and QRs were clearly observed by FCS.

 

Fig. 3 CCFs of QDs/QRs in a pH-controlled solution. (a) Normalized CCF in a pH-controlled solution. Measurements were performed just after the dissolution of the QDs/QRs in the solvent. (b)-(d) Blinking fraction, blinking relaxation time, and antibunching relaxation time of the QDs/QRs just after the dissolution. Fitting analysis was performed using the CCFs in Fig. 3(a). (e) Normalized CCF in a pH-controlled solution. The measurements were performed 1 h after the dissolution. (f)-(h) Blinking fraction, blinking relaxation time, and antibunching relaxation time of the QDs/QRs 1 h after the dissolution. Fitting analysis was performed using the CCFs in Fig. 3(e). Measurement time: 120 s

Download Full Size | PDF

From the fitting analysis, we obtained several parameters. The antibunching- and blinking-related parameters are shown in Figs. 3(b)–3(d). Figure 3(b) shows the blinking fraction as a function of the solution pH. A high blinking fraction was observed for the QD solutions with high pH. However, it was not the case with the QRs. Interestingly, with an increase in the pH, the blinking relaxation time of the QDs decreased gradually, as shown in Fig. 3(c). Thus, it can be stated that there exists a negative correlation between the blinking fraction and blinking relaxation time of QDs. The high blinking fraction of the QDs at high pH values is related to their blinking possibility, which is reported for the first time in this study using FCS.

Figure 3(e) shows the CCFs of the QDs/QRs after incubation for 1 h at room temperature. In the case of the solutions with the lowest and highest pH values, the CCFs shifted to the right unlike the case shown in Fig. 3(a). It indicates that at the lowest and highest pH values, the QDs aggregated gradually within this time duration. The samples that showed high blinking fraction in Fig. 3(a) also showed a right shift in Fig. 3(e). This strongly indicates the relationship between blinking and aggregation. From Fig. 3(f), it is obvious that the blinking fraction of QD545 was pH-dependent. This is attributed to the formation of big aggregations of QD545 [Fig. 3(e)]. On the other hand, the blinking fraction did not show any pH-dependent change in Fig. 3(g).

The antibunching relaxation time of the QDs [Fig. 3(h)] also showed a pH-dependent change and is also correlated to their aggregated particle size [Fig. 3(e)]. Thus, it can be argued that the aggregation of QDs leads to a decrease in their antibunching time. On the other hand, QR595 did not exhibit any pH-dependent property throughout the experiment. The FCS results suggest that this technique can be used to investigate the photophysical properties of QDs/QRs under diffusing conditions.

3.3 Time-dependent oligomerization of QDs

The other interesting question is when does the aggregation occur in these solutions. Because of the high temporal resolution of FCS, it can be used to analyze the time-dependent changes. Figure 4 shows the data from these experiments. QD525 and QD545 were dissolved in high-pH solutions (pH = 12.0). Then, the FCS measurements were performed every 1 min up to 20 min after the dissolution. Figures 4(a) and 4(b) show the normalized CCFs obtained after the FCS measurements. The data indicates that the CCFs showed a right-shift in the case of the samples that were analyzed after a long time after the dissolution. In this case, blinking was quantified to be the same as in Fig. 3. In the short-time range, blinking was successfully obtained.

 

Fig. 4 Time-dependent oligomerization and change in the photophysical parameters of the QDs. (a) Time-dependent CCF change for QD525, (b) Time-dependent CCF change for QD545. (c)-(e). Blinking relaxation time, blinking fraction, and antibunching time obtained from the time-course data ((a) and (b)). Time point indicates the time at which (a) and (b) were measured. 1 denotes the shortest time, while 5 denotes the longest time after the dissolution. Measurement time: 60 s.

Download Full Size | PDF

A large antibunching fraction was evident. However, the fitted blinking relaxation time [Fig. 4(c)] was dramatically long (more than 1 ms), indicating that the blinking fraction was not correctly fitted. In the aggregated condition, there are many QDs with diverse sizes. It might be main reason for the blinking fraction miss-fitting. However, no blinking fraction was observed in the case of the QD545 sample incubated for a long time [Fig. 4(b)]. Thus, it is better to state that there was no blinking in the case of the aggregated QD545 sample. Antibunching was also quantified by fitting [Fig. 4(e)]. The antibunching relaxation time gradually decreased with an increase in the incubation time.

4. Conclusions

The FCS data shows the advantages of this technique for analyzing the properties of QDs/QRs. FCS allows the measurement of the variations in the photophysical properties of QDs/QRs from the ns–ms time region under diffusing conditions.

In the antibunching and blinking analysis, we investigated the correlation between the photophysical properties of QDs/QRs and the pH of the solutions in which they were dissolved by FCS using a CW laser for the first time. The investigation of the origin of the photophysical properties of QDs/QRs by FCS is challenging. Its fast and reliable measurement method can be used to perform the screening of new application of QDs/QRs under aqueous conditions.

The currently available QDs/QRs cannot be used as pH indicators in living cells because they exhibit photophysical properties at a pH of 10. The physiological pH in living cells lies within the range of 6.0–7.0. However, our findings showed that FCS can be used to measure the photophysical properties of QDs/QRs with high temporal resolution and reliability. Thus, this technique can be used to discover new environmental indicators.