1. Introduction

Various application fields, such as energy technology and information communication, require high-performance wavelength conversion to resolve fundamental issues related to the mismatch between the wavelengths emitted by light sources and the wavelengths at which detectors are most sensitive. One of the promising ideas for realizing high-yield wavelength conversion is to use fluorescent materials, such as quantum dots (QDs) and fluorescent dyes [1–5]. Because such materials exhibit high quantum yield in their absorption and emission processes, effective wavelength conversion is expected. Among various ideas for using fluorescent materials, there have been some proposals to mix and couple heterogeneous materials to realize novel emission mechanisms based on their combinations [1,6,7]. The most attractive advantage of such proposals is that absorption and emission wavebands can be readily tuned by controlling the mixing and coupling conditions of components to meet the requirements of various applications. However, the behavior of this mechanism fundamentally depends on the relative spatial positions of the components in the mixed structure. Therefore, it is technically difficult to ensure homogeneity and stability of their optical functions, and advanced fabrication techniques are required [8–12].

On the other hand, the field of nanophotonics has seen rapid progress in recent years, and various applications have been actively developed [13]. Nanophotonics exploits the local interactions between nanometer-scale particles via optical near fields induced by incident light. Several characteristics of optical near fields can be described by the behavior of a dressed photon (DP), which is a quasi-particle representing the coupled state of a photon and an electron in a nanometric space [14]. A DP excites a multi-mode coherent phonon in a nanometric material, and the DP state is coupled with the excited coherent phonon state [15–18]. Because the coupled state can be regarded as an intermediate state during the excitation and relaxation process of the material, multistep excitation and relaxation, and corresponding optical functions, are allowed. By utilizing such a phenomenon, called a phonon-assisted transition, some experimental demonstrations on high-yield emission of upper-converted optical energy by using organic dye grains [16,19] and high-intensity emission from indirect transition type semiconductors [17,18,20] have been reported.

Previously, we have proposed a novel method of coupling heterogeneous QDs in a photo-curable polymer solution by utilizing a phonon-assisted photo-curing process [21–23]. We call such coupled-QDs, when they are encapsulated by locally cured polymer, a nanophotonic droplet (ND). All we have to do to form NDs is to mix all components together and irradiate the mixture with light. Here, the irradiated energy must be lower than the curable energy of the polymer. The formation process is induced only when heterogeneous QDs encounter each other at the optimum distance to induce appropriate optical near field interactions between the two. Therefore, their relative spatial positions and corresponding optical functions are autonomously determined, which is a strongly desired property of wavelength conversion.

In this paper, we focus on the implementation of wavelength conversion utilizing the characteristics of NDs. First, we briefly review the principles of the phonon-assisted photo-curing process and our method of forming NDs. Here, we assume the use of a thermo-curable polymer that is transparent to the incident and emitted light, instead of a photo-curable polymer. Then, we describe experimental demonstrations on forming NDs and their quantitative evaluation as wavelength conversion elements. We embedded a large number of NDs into a transparent material and compared a figure-of-merit for wavelength conversion with a corresponding sample in which we embedded randomly dispersed QDs.

2. Principles

An ND is formed of coupled heterogeneous QDs encapsulated by locally cured polymer [21–23]. During the process of fabricating NDs, incident light having a lower photon energy than the curable energy of the polymer is radiated into a polymer solution that contains a number of nanometric components to be coupled. The incident light induces a phonon-assisted process [16–19], namely, multistep photo-curing, which locally cures the polymer with transitions via activated phonon levels. In particular, a thermo-curable polymers that is transparent over a wide range of wavelengths is used instead of a photo-curable polymer, so that light is losslessly input to and output from the sample. Moreover, in this case, the conventional photo-curing process due to the absorption of incident optical energy cannot be expected to occur. A detailed description of the formation process of NDs is given below.

The principle of our method is schematically shown in Fig. 1. Here we assume a mixture containing two types of QDs, namely, QD_A and QD_B, and a thermo-curable polymer. The mixture is irradiated with incident light, what we call assisting light, having photon energy hv_assist. The excitation energies of QD_A and QD_B are E_A:bg and E_B:bg, respectively, and the curable energy of the thermo-curable polymer is E_poly:act. In the case where these energies satisfy the condition E_A:bg<hv_assist<E_poly:act<E_B:bg, the following process can be induced by radiating the assisting light. If the numbers of QDs, or in other words, their volume densities, in the mixture are not sufficiently high and they rarely encounter each other, only QD_A spontaneously emits light by absorbing the optical energy of the assisting light due to the energy condition E_A:bg<hv_assist<E_B:bg. In this case, no subsequent physical or chemical reaction occurs. On the other hand, if the density is sufficiently high that the QDs can frequently encounter each other, multistep photo-curing occurs due to induction of DPs and corresponding optical near-field interactions between neighboring QD_A and QD_B. Because the activated phonon-levels are real states, the excitation rate is expected to be much higher than any other multi-step excitation, such as conventional second harmonic generation [24] and phosphorescence using a multi-step transition [25]. As a result, the thermo-curable polymer is locally cured, and the spatial alignment of the encountered QDs is physically fixed by the cured polymer. The electronic transitions via such a coupled state induced by the assisting light have been described in a previous report by the authors [21].

 

Fig. 1 Schematic diagram of process of forming a thermo-curable polymer-based ND via the phonon-assisted photo-curing process.

Download Full Size | PDF

Since the spatial distribution of the DPs generated on the surface of the QDs is expressed by a Yukawa function [26], the separation between the two QDs in each ND is theoretically defined also by the Yukawa function. Moreover, because the above process occurs only when heterogeneous components encounter each other, the cured polymer necessarily contains coupled heterogeneous components. Such accuracy and homogeneity of their alignment and combinations of QDs ensure efficient induction of optical near-field interactions and the resulting wavelength conversion process, where QD_A and QD_B act as emitters and absorbers of light in each ND, respectively. A schematic diagram of wavelength conversion from a higher optical energy hv_IN to a lower optical energy hv_OUT-A is shown in Fig. 2.

 

Fig. 2 (a) Schematic diagram of enhanced emission from a CdSe-QD via optical energy transfer from a CdS-QD to the CdSe-QD, and (b) individual emissions from the CdSe-QD and the CdS-QD.

Download Full Size | PDF

As shown in Fig. 2(a), in the case of coupled CdSe-QDs, because the rate of an optically forbidden transition between neighboring QDs induced by optical near-field interactions is much higher than the relaxation rate in each QD, the optical energy absorbed by QD_B is preferably transferred to the neighboring QD_A. As a result, enhanced spontaneous emission is observed from QD_A. In this case, the emission intensity from QD_B is necessarily decreased. On the other hand, in the case of emission from isolated QDs, as shown in Fig. 2(b), each QD emits individually, and enhanced emission cannot occur. Therefore, though similar wavelength conversion occurs in both cases, the yield in the case of Fig. 2(b) is much lower than that in the case of Fig. 2(a). Besides, the time of such a transition between coupled QDs, which is a few hundred femtoseconds, has been theoretically and experimentally verified to be at least ten times faster than the relaxation rate of each QD, which is a few nanoseconds, in the case where specific QDs are assumed [27]. Furthermore, the processing time of the wavelength conversion is fundamentally limited by the transition rate between coupled QDs.

3. Experimental demonstration

In order to experimentally verify the proposed process of forming NDs with a thermo-curable polymer, as well as the mechanism of wavelength conversion using the formed NDs, as QD_A and QD_B in Fig. 1, we used commercially available CdSe-QDs (Sigma-Aldrich, Lumidots) and CdS-QDs (NN-Labs, Nanocrystals), respectively, which emit visible light with emission wavelengths of 560 nm and 420 nm, in toluene solutions. In this case, wavelength conversion from ultraviolet (UV) light to visible light is expected. The QD solutions were dispersed in a thermo-curable polymer (Dow Corning Toray, Sylgard 184), which consists of polydimethylsiloxane (PDMS). To form the NDs, the mixture was irradiated with assisting light emitted from a 200 mW laser diode with a wavelength of 457 nm for 30 minutes. These experimental conditions surely fulfilled the previously described energy conditions for inducing the sequential process of the phonon-assisted photo-curing method, as schematically shown in Fig. 1. The total amount of the mixture was limited to 1.0 mL to maintain spatially uniform illumination in our experimental setup. This volume contains about 10¹⁸ of each QD. Under these experimental conditions, the QDs can be assumed to encounter each other at a sufficiently high frequency to induce the photo-curing process. After irradiation with the assisting light, the mixture was separated into cured and uncured materials by centrifugation at 10,000 rpm for 5 min. The extracted cured material was assumed to contain a large number of NDs. Before demonstrating wavelength conversion using these NDs, they were dispersed in a toluene solution and uniformly dispersed on a Si substrate to observe their appearances with a fluorescence microscope. Figure 3 shows a fluorescence image of the formed NDs under UV light irradiation.

 

Fig. 3 Microscope fluorescence image of densely formed NDs under UV light irradiation and (inset) a magnified view.

Download Full Size | PDF

As shown, a number of NDs with similar sizes and emission intensities were successfully obtained. In contrast to our previous NDs using a photo-curable polymer [21–23], which had diameters of a few micrometers, the diameters of the NDs in Fig. 3 are only a few hundred nanometers. This is due to suppression of the encapsulation process during formation of the NDs as a result of absorption of spontaneously emitted light from coupled QDs. Because the surrounding thermo-curable polymer does not absorb such emitted light, the NDs cannot grow as much as in the case where a photo-curable polymer is used. On the other hand, the distance between and the distribution of QDs in each ND are expected to be quite similar to those of our previous NDs because the QD coupling process during formation of NDs fundamentally depend on size of each QD and not on the types of polymers. Moreover, as schematically illustrated in the inset of Fig. 3, each ND is assumed to contain physically coupled QDs, which consist of CdSe-QDs as emitters and CdS-QDs as absorbers. By such alignment of the QDs, the optical energy absorbed by the CdS-QDs is preferentially transferred to the CdSe-QD before being emitting from it. Therefore, the emission from the CdSe-QD is expected to be enhanced. Related research that experimentally demonstrated optical energy transfer between QDs and corresponding emission enhancement has been previously reported [28].

Next, the NDs were extracted and embedded in another pure PDMS solution. Then, the solution was cured for 2 hours by heating it at 150 °C after degasing to remove air bubbles in the mixture. We call this sample a w/-NDs (with NDs) sample in this paper. In order to discuss the effectiveness of embedding NDs, a corresponding sample without NDs, called a w/o-NDs sample, was prepared by embedding randomly dispersed QDs. The numbers of QDs in the w/-NDs and w/o-NDs samples were set to be equivalent. Figures 4(a) and 4(b) show the appearances of both samples under room light and their absorption spectra, respectively. As shown, it is quite difficult to discern a difference between the two. However, as shown in Fig. 4(c), their appearances under UV light seem quite different in terms of their color tones: the w/-NDs sample showed a more monochromatic green appearance than that of the w/o-NDs sample. Such appearances support the mechanism of the emission processes in both samples shown in Fig. 3, where the w/-NDs sample is expected to show enhanced light emission from the CdSe-QDs, whereas the w/o-NDs sample is expected to show individual emissions from both CdSe-QDs and the CdS-QDs.

 

Fig. 4 Appearance of w/-NDs sample (left) and w/o-NDs sample (right) under (a) room light and (b) the absorption spectra of w/-NDs sample (red line) and w/o-NDs sample (blue line). (c) Appearance of w/-NDs sample (left) and w/o-NDs sample (right) under UV light.

Download Full Size | PDF

For more quantitative evaluation of their optical properties, we measured the excitation spectra, which are shown in Figs. 5(a) and 5(b). Insets in each figure show schematic diagrams of the assumed QD arrangement in each sample. The emission and excitation wavelengths were scanned from 300 nm to 800 nm and from 300 nm to 600 nm, respectively.

 

Fig. 5 Comparison of excitation spectra of (a) w/-NDs sample and (b) w/o-NDs sample.

Download Full Size | PDF

Peaks at 420 nm and 560 nm in the spectra correspond to emission from the CdS-QDs and CdSe-QDs, respectively. As shown, a clear difference was revealed between the two samples. Emission from the CdSe-QDs in the w/-NDs sample was enhanced, whereas the w/o-NDs sample revealed independent emissions from the CdSe-QDs and the CdS-QDs. Such spectra are the specific results of this research, and they indicate that NDs in the w/-NDs sample contained coupled CdS-QDs and CdSe-QDs, which exhibit optical energy transfer between them. As a result, the light emission from the CdSe-QDs was enhanced and that from the CdS-QDs was decreased, as we showed in Fig. 2(a). A quantitative difference on the yield of the wavelength conversion between the two samples was evaluated from the results of emission spectra with irradiation of 325 nm light. A result of comparison is shown in Fig. 6(a).

 

Fig. 6 (a) Comparison of emission spectra with 325 nm light irradiation, measured with w/-NDs sample (red line) and w/o-NDs sample (blue line), respectively. (b) Description of differential optical amount, D, calculated from the results in (a).

Download Full Size | PDF

As shown, the w/-NDs sample showed a 3.7-times higher intensity of emission from the CdSe-QDs, whereas the intensity of emission from the CdS-QDs was decreased, which is evidence of the successful existence of NDs in the w/-NDs sample. Moreover, a sideband-like spectral feature from 500 nm to 650 nm, which is much more clearly recognized in Fig. 6 than in Fig. 5, was also decreased due to the existence of NDs. This spectral feature can be attributed to emission from defect levels of the CdS-QDs.

As we can confirm by the existence of emission spectra from the CdS-QDs in the w/-NDs sample in Fig. 6(a), not all QDs in the w/-NDs sample are coupled with each other, and the sample also includes isolated QDs. In order to directly discuss improvement of the energy effectiveness due to the existence of NDs and to compensate effect of the emission from isolated QDs in the w/-NDs sample, we defined a metric called the differential optical amount, D = (I_w/-I_w/o)/hν, where I_w/ and I_w/o represent the normalized optical intensities in Fig. 6(a), and hν represents the photon energy of light at each wavelength. The calculated D is explained in Fig. 6(b). Regions filled with red and blue correspond to gain and loss caused by the existence of the NDs, respectively. As shown, the total gain due to the existence of the NDs was 4.2-times larger than that of the loss. The result is due to effective use of the incident optical energy by the induced energy transfer from the CdS-QDs to the CdSe-QDs. That is to say, while individual QDs necessarily reveal quenching of optical energy during their emission process, in the case of coupled QDs, optical energy transfer preferentially occurs before the quenching process. Therefore, the incident optical energy is utilized for wavelength conversion in NDs much more effectively than when using randomly dispersed QDs.

4. Summary

In this paper, we have reported the experimental demonstration of effective wavelength conversion based on novel optical functions of NDs due to the particular structural characteristics of their constituent elements. In order to verify the mechanism of wavelength conversion from UV to visible light, to form NDs, we used a thermo-curable polymer that is transparent to input and output light, instead of a photo-curable polymer as in our previous experiments. As a result, NDs were successfully obtained, and a sample with NDs showed much more effective wavelength conversion than a corresponding sample without NDs, which contained randomly dispersed QDs. This effectiveness is due to the autonomously realized accuracy and homogeneity of the structured components in each ND and the resulting induced optical energy transfer between each component. Such optimized structures can also be realized by utilizing other components instead of CdSe-QDs and CdS-QDs and are therefore expected to achieve other optical functions with higher-efficiency than existing techniques.