Dynamic Control of Quantum Dot Localization in Nematic Liquid Crystal Matrix by Means of Photoinduced Phase Transition

Abstract

The stimulated assembly/disassembly of particles is a technique allowing for precise spatial and temporal control over the resulting structures to be realized. The application of a photosensitive liquid crystal (LC) allows the use of a photo-initiated order–disorder transition for the ordering and redistribution of dispersed nanoparticles. The semiconductor quantum dots (QDs) among them are useful for the imaging of such redistribution through simple luminescent microscopy with excitation by laser radiation at a wavelength of 532 nm. Doping the LC matrix with azo-chromophore molecules allowed us to localize the light-driven phase transition of the LC from the organized to the isotropic phase inside the spot, illuminated by ultraviolet (UV) light through a slit. The phase transition leads to a redistribution of the QDs within the matrix, followed by QD-rich region formation. After the termination of UV illumination, the QDs were found to form droplets in the region where UV illumination resulted in a homogeneous distribution of the QDs. The translation of the sample through the UV-illuminated spot resulted in QD accumulation inside the isotropic phase at the borders of the isotropic phase. The results obtained provide a good agreement with the model calculations of nanoparticle diffusion at the LC phase–isotropic liquid interface.

1. Introduction

Ensembles of semiconductor quantum dots (QDs), due to their unique properties, are extremely promising as a laser medium [1]. However, since for QD composite media the laser generation threshold strongly depends on the QD spatial distribution, tailoring their positions, e.g., the formation of QD-containing waveguides, gratings, or other structures, is of great importance. To achieve this goal, the QD can be arranged in a medium that allows a restructuring induced by external actions, e.g., in liquid crystal (LC). Among the widely studied materials, LCs have attracted much attention due to their unique physical properties such as orientational order, birefringence, high molecular mobility, etc. This aggregate state of matter, inherent in specific organic compounds of anisotropic form, is in a phase equilibrium between the crystalline state of a solid and isotropic liquids. The ability to orient and reorient LC molecules under external influence plays a key role in their use in various applications such as LC displays, optical lenses, infrared reflectors, and various photonic devices. Currently, the introduction of various inorganic particles into nematic liquid crystal is one of the most well-known and studied methodologies, since it improves the properties of nematic systems [2,3,4,5,6,7,8,9,10,11,12]. The LC matrix can drive the organization of introduced nanoparticles and provide broad opportunities to use such combined systems in quantum computing devices, photovoltaic devices, light-emitting diodes (LEDs), and many others, but various nanoparticles can also control the properties of the LC medium. There have been numerous attempts to improve the conductivity, anisotropy, and other characteristics of LCs by mixing them with inorganic QDs. However, few studies have examined the effect of changing the structure of QD surface ligands, which can significantly influence the coupling interaction with the LC environment and the final properties of the composite system [13,14], in particular, the photoinduced phase transition.

Since quantum dots themselves tend to aggregate, the choice of a QD stabilizing ligand plays a key role in the QD–medium interaction. For LC systems, the process of QD aggregation during the isotropic liquid–LC transition has been well studied and is associated with the thermodynamically driven tendency of the LC system to exclude sources of defects from its volume [15]. The introduction of mesogenic stabilizers into the QD shell is a common method for improving the compatibility of QDs with the LC matrix [13,14,15,16]. Such stabilizers also promote QD ordering during the isotropic liquid–LC transition. Mesogenic stabilizers in most cases contain aromatic rings, which are responsible for short-range π–π interactions between the shells of closely packed QDs, stabilizing the resulting structures [16].

The pivotal idea for controllable QD adjustment is the variation in the QD solubility, distribution, and concentration when the LC matrix undergoes a phase transition from the organized to the isotropic phase [10,11]. Controllable phase transitions allowing the desired structure to be formed can be performed with the help of photochromic additives in the LC molecule. The isotropization of the LC matrix caused by the photoinduced isomerization of the photochromic molecule, which takes place at a temperature below the isotropization temperature, causes a redistribution of the QD, which is a base for the production of a desired QD structure.

Previously, we reported the synthesis of CdSe/ZnS QD-LC composites based on cyanobiphenyl LC with embedded azobenzene-containing molecules and reported their applicability and demonstrated ‘order–disorder’ phase transition in them [12]. However, the prospective employment of these composite media for the formation of the QD structure needs a detailed knowledge of the structural, electronic, and recombinational properties of the QDs inside the QD-LC composite. In particular, we should obtain knowledge of the distribution and concentration of the QDs in the composite and their variations caused by light action, the possible sizes of the occurring QD structures and their temporal characteristics, the possibility to accumulate the QDs, etc. To reach these aims, we employed a photoluminescent (PL) microscopy technique for the detection of the rearrangement of CdSe/ZnS QDs in a matrix of nematic LC with embedded azo-chromophore molecules induced by UV illumination.

The main idea of the work could be demonstrated by the diagram shown in the graphical abstract. QD-rich microphases, or droplets, are uniformly distributed in the photosensitive LC matrix. Localized UV irradiation leads to an LC-isotropic phase transition in the exposed area, followed by QD homogenization in the isotropic melt. Moving the sample relative to the UV-irradiated spot could lead to the accumulation of QDs inside the isotropic phase area.

2. Results and Discussion

In our study, we used an LC-QD composite based on nematic LC with embedded 10 wt.% azo-chromophore molecules undergoing a configurational (E-Z) transition stimulated by UV light and containing 0.1 wt.% CdSe/ZnS QDs. The composite is characterized by an isotropization temperature ($T_{NI}$) of 60 °C, as was previously found with the help of a differential scanning calorimetry technique and dielectric permittivity measurements [4].

2.1. PL Spectrum

The employed sample of the QD-LC composite demonstrates an easily seen PL under excitation from laser radiation at 532 nm (Figure 1). The PL spectrum of the sample includes two bands with maxima at 580 and 725 nm, which can be attributed as exciton and defect ones; the latter band is associated with defects on the surfaces of QDs, e.g., selenium atom vacancies acting as charge carrier traps [17]. The variation in the sample temperature does not change the form of the PL spectrum, although the PL intensity slightly decreases with the temperature increase. It is worth noting that a rise in the temperature above T_NI results in a noticeable fall of the exciton band signal, whereas the defect band remains stable.

2.2. Dynamics of the QD Distribution at a Fixed Position

All the results mentioned below on the dynamics of the QD redistribution in the LC matrix were obtained at a temperature of 53 °C, i.e., 7 °C below T_NI.

Details of the QD distribution in the composite QD-LC medium can be found from the PL microscopy images shown in Figure 2a. The initial (unilluminated) sample region demonstrates both a rather dim inhomogeneous background and much brighter droplets of about 10 μm in diameter (Figure 2b) enriched with QDs. As one can see, the inhomogeneity of the PL signal, and therefore the QD concentration, reaches up to 1:20 contrast. Converting the PL intensity to the QD concentration, we can conclude that the concentration of QDs in the QD-rich regions is 0.9 wt.% (for a mean concentration of 0.1 wt.% [3]), while in LC regions this value is 0.05 wt.%.

Figure 3a–f present PL microscopy images of the sample before, during, and after its UV illumination. The temporal dependences of the PL signals integrated in different regions of interest (ROIs), which are stripes of 150 µm in width (Figure 3d), are shown in Figure 3g. As one can see, initially, mean PL signals for different ROIs differ noticeably, although the difference is less than 12%; this effect is determined by variations in the initial QD concentration in the sample studied. Moreover, before the illumination starts, the PL signal varies significantly for different small regions, which is determined by variations in the initial spatial QD concentration in the studied sample caused by the redistribution of QDs between the LC phase and droplets enriched by QDs. This redistribution is induced by the transition from the isotropic to the LC phase of the matrix at a temperature lower than T_NI.

For unilluminated regions, neither the images nor the PL signals integrated over the ROIs significantly vary with time. In contrast, illuminated region (ROI3) is characterized by a discernible redistribution of the emitting QDs. At the start of the UV illumination, the PL signal integrated over ROI3 increases up to 10%, which is a result of additional PLs excited by UV light. As one can see from Figure 3b,g, the very first variation in the PL microscopy image was found after UV illumination for 1 min, which causes an additional 10% PL signal increase. Illumination over a longer time results in the homogenization of the QD distribution inside the stripe of about 200 μm in thickness, which slightly exceeds the UV illuminated region, caused by the isotropization of the LC matrix inside the stripe. After UV illumination termination, QD-enriched droplets with a higher diameter than the ones in the initial LC-QD composite are formed (Figure 2b). They are surrounded by regions depleted in QDs; initial microphases with a moderate QD content could be hardly found after the UV exposure. The contrast of the PL signals increased in comparison with the initial one and reached up to 1:40, with a QD concentration less than 0.02 wt.% outside the droplets (Figure 2b).

It is worth noting that the standard deviation of the PL intensity calculated in the frames of the whole ROI can serve as a very useful quantitative characteristic of QD concentration equalization during the isotropization of the composite matrix caused by UV illumination. In the UV-illuminated region (ROI3), the standard deviation value starts to fall drastically after 1 min of illumination, which is in precise agreement with the variation in the PL microscopy image (Figure 3b and enlarged fragment of Figure 3g). A complete homogeneity of the QD distribution, which corresponds to a decrease in the standard deviation more than three times due to the homogeneous QD distribution in the isotropic phase, is achieved within 10 min of UV irradiation (Figure 3d,g). Note, however, that during this period mean value of the PL signal has no significant variations. The slow fall of the mean PL signal taking place after UV illumination for 15 min is possibly connected with the QD degradation under continuous UV illumination. The formation of the QD-enriched droplets after UV illumination termination results in a huge rise in the PL signal standard deviation (Figure 3g for time exceeded 2900 s). Thus, the formation of QD-enriched droplets upon switching on and off UV irradiation is reversible at temperatures slightly lower than $T_{NI}$.

The maximum speed of movement of the region exposed to UV radiation will be determined by the time it takes for the composite to achieve a uniform concentration of QDs. After the composite matrix becomes isotropic under UV radiation, a leveling diffusion of quantum dots from the regions enriched with them must occur.

Fick’s second law predicts how diffusion causes concentration to change with respect to time. It is a partial differential equation which in one dimension reads:

$${\displaystyle \frac{\partial C}{\partial t}}=D{\displaystyle \frac{{\partial }^{2}C}{\partial x^2}},$$

(1)

where $C$ is the QD concentration, $t$ is time, $D$ is the diffusion coefficient, and $x$ is the position. The solution to Equation (1) on an infinite interval is the convolution of the initial concentration data and the kernel:

$$C\left(x,t\right)={\displaystyle \frac{1}{\sqrt{4\pi Dt}}} \exp \left(-{\displaystyle \frac{x^2}{4Dt}}\right).$$

(2)

The diffusion coefficient for a single QD may be estimated using the Stokes–Einstein–Sutherland equation for the diffusion of spherical particles through a liquid with a low Reynolds number:

$$D={\displaystyle \frac{k_{B}T}{6\pi \eta r_{QD}}},$$

(3)

where $k_B$ is the Boltzmann constant, $T=326 K$ is the absolute temperature, $\eta$ is the dynamic viscosity of the matrix in the isotropic phase, and $r_{QD}$ is the Stokes radius (effective solvated radius in solution) [18] of the spherical QD.

Since we have no precise information on either the dynamic viscosity of our matrix or the Stokes radius of the QD, the product of the dynamic viscosity and the Stokes radius of the QD $\eta r_{QD}$ was evaluated to estimate the diffusion coefficient. The radius of the QD determined from the TEM results is 2.5 nm. For nematic LCs not far from $T_{NI}$, the mean dynamic viscosity may be estimated by an order of magnitude as 0.05 Pa·s [19]. Thus, the product $\eta r_{QD}$ can be estimated as 1.25 × 10⁻¹⁰ N·s/m. On the basis of the obtained estimate, we simulated the leveling diffusion from the QD-enriched droplet and estimated the typical time it takes. For this purpose, a single droplet enriched with the QDs 7 µm in diameter was selected, with the initial profile shown in Figure 4a. Figure 4b presents the experimentally found dependence of its PL profile on the time elapsed after the start of UV irradiation, which demonstrates the significant fall of the QD concentration in the time interval of 35 s. Variations in both the size of the QD-enriched droplets and the QD concentration in it were estimated through the direct solution of Equation (1) for the cases of the initial QD concentration profile approximated by the Gauss function, with a maximum of 0.9 wt.% (Figure 2b), and two different values of the product $\eta r_{QD}$: 1.25 × 10⁻¹⁰ N s/m (Figure 4c) and 1.25 × 10⁻⁹ N s/m (Figure 4d). As one can see, for the former value of the product $\eta r_{QD}$ and the higher diffusion coefficient, the PL signal decreases almost 4 times, whereas the size of the droplet increases more than twice for 35 s, i.e., the QD-enriched droplet is re-dispersed in less than 1 min (Figure 4c). Estimations for the higher value of the product $\eta r_{QD}$ and lower diffusion coefficient demonstrate a significantly slower dissolving of the droplet, exceeding the experimentally found one (Figure 4d).

Thus, comparing experimental and simulation results, we can find a good qualitative agreement of the simulation for viscosities of 0.05 Pa·s (Figure 4c) and the data from PL microscopy (Figure 4b). The simulation gives us the estimation of the QD diffusion rate, which is about 40 µm/min by the value order.

2.3. Dynamic Accumulation of the QDs with the Help of Moving UV Light Spot

The movement of the UV-illuminated spot over the sample allows the QDs to be accumulated at the edges of the spot. The important parameter is the speed of the light spot movement, which should not exceed a certain limit for a steady rise in the PL signal with the light spot movement. According to previous results, the speed should be 10 µm/min in order of magnitude, whereas a higher speed cannot result in a complete QD redistribution. Figure 5 presents PL microscopy images and, averaged over 500 µm in the direction across the travel line cross-sections, the QD concentrations for various illumination times and positions of the illuminated field at the sample surface for the sample moving right to left with a speed of 26 µm/min. For 5 min of illumination, based on our previous observation of the disappearance of the QD droplets in the isotropic phase of LC, we can attribute the obviously distinguished region in the center of the images as an isotropic phase surrounded with an organized phase containing the QD-rich droplets. The width of the isotropic phase region is about 250 μm, which exceeds the mask slit width (170 μm) and slightly increases with time. For longer illuminations, returning the LC from the isotropic to the nematic phase results in the QDs’ exclusion from the amorphous region, as was predicted in Ref. [10]. Thus, the depletion in the QDs takes place in the previously illuminated region (left to the isotropic phase spot); the QD concentration in this region falls up to two times, combined with an increase in the QD concentration more than three times in comparison with the unilluminated region at the back (left) edge of the isotropic spot. However, for the UV illumination of 1 h, a small increase in the QD concentration in the depleted region is found (cf. PL signal cross-sections for 30 and 60 min).

A characteristic feature of the QD concentration distribution is a strong inhomogeneity near the interface between the isotropic and LC phases. Near this interface, the LC phase is depleted in QDs, while the isotropic phase is enriched in them, with the QD concentration near the interface being higher than in the isotropic phase located far from the interface. This feature is explained by the finite lifetime of the interface between the isotropic and LC phases. To confirm this statement, a simulation of the dependence of the QD concentration near the interface was performed. The product $\eta r_{QD}$ was equal to 1.25 × 10⁻¹⁰ N s/m, and the equilibrium concentrations of QDs were 0.1 wt.% for the initial LC phase and 1.05 wt.% for maximum in the isotropic phase. The simulation results are shown in Figure 6. These results demonstrate a noticeable inhomogeneity of the QD concentration and are consistent with the experimental results. Indeed, the simulation results for the right (front) edge of the UV illuminated region, where the isotropic liquid meets the organized phase, qualitatively describe abrupt jumps of the QD concentration as well as an increase in the thickness of both QD-enriched and QD-depleted zones with illumination time. Generally, the asymmetry of the QD distributions, i.e., the accumulation of a greater amount of the QDs at the left (back) edge of the illuminated region, is a result of the sample movement from right to left. The QDs moving with the sample are unable to penetrate the phase interface. However, this consideration is valid for the low speed of the sample, which does not exceed the diffusion speed.

In contrast, the higher movement speed (87 µm/min), significantly exceeding the diffusion speed, results in a totally different QD distribution. The UV-illuminated region also demonstrates a depletion in the QD concentration for an illumination for 13 min; however, further illumination results in the detachment of rather large droplets (up to 100 µm) enriched with the QDs (Figure 7). It is noticeable that, in contrast to the case of lower speed for the sample movement (Figure 5), the QDs are partially accumulated at the left (back) edge of the amorphous spot and partially overcome the barrier caused by the QD diffusion and are deposited as QD-enriched droplets in the QD-depleted region behind the left (back) edge, with the step-like increase in QD concentrations at the front edge of the isotropic spot being not formed. These facts indicate that the sample speed of 87 µm/min is higher than the speed of the QD diffusion. Estimating UV exposure times yields values of 350 s and 100 s for the velocities 26 µm/min and 87 µm/min, respectively. A comparison of the obtained UV exposure values with estimates based on static measurements demonstrates their consistency.

3. Materials and Methods

We employed a eutectic mixture of cianobiphenyl-based low-molecular-weight mesogenic molecules as a nematic LC matrix; the compound forms only a nematic phase in the temperature range of −30 to +64 °C [12]. To ensure the control of transitions between anisotropic and isotropic phases, the photochromic azo-compound AX-8 was introduced into the LC mixture (10 wt.%). Spherical CdSe/ZnS QDs of an average diameter of about 5 nm, stabilized with mesogenic carboxyl acid, were used in the mixture, with a fraction of 0.1 wt.%. The QD surface stabilizer was used to improve QD solubility. The isotropization temperature variation after the QD introduction is less than 1 °C, which indicates no significant effect of the QDs at the used concentration on LC ordering. Detailed information on the synthesis of the compound is contained in Appendix A.

The obtained mixtures were placed between two glass slides divided by a spacer of 25 µm thickness. The glass slides were rinsed with ethanol and wiped dry before use. The chosen cell depth is optimal to provide a highly detectable PL signal and to avoid redistribution of the QDs in depth. The mixture demonstrated properties of homeotropic LC for temperatures below $T_{NI}$ = 60 °C. Polarizing optical microscopy (POM) in crossed polarizers demonstrated a black field for temperatures below 59.6 °C (Figure 8a), which is typical for homeotropic media, a schlieren texture at 59.6 °C (Figure 8b), which accompanies the reorientation of mesogens during the process of transition from anisotropic to isotropic phases, and a black field again for the isotropic phase of the LC at a temperature of 60 °C (Figure 8c); cooling of the sample results in the occurrence of a more pronounced schlieren texture at 56 °C (Figure 8d), which is also a manifestation of the phase transition into the LC phase, and a black field for the homeotropic LC phase at a lower temperature (Figure 8e).

PL microscopy is the main experimental technique employed to study spatial redistributions of the QDs. The measurements were carried out with the help of the optical microscope POLAM L-213 (LOMO, St. Petersburg, Russia), equipped with a thermal stage FP82HT (Mettler Toledo, Greifensee, Switzerland). The temperature of the tested samples was 53 °C, which is 7 °C below the isotropization temperature [4], and maintained with an accuracy of 0.1 °C. The standard halogen light source of POLAM L-213 has been replaced by a high power 365 nm LED CUN6GF1A (Seoul Viosys, Ansan, Republic of Korea) and a Wood optical filter. For the PL excitation radiation, a solid-state laser (OXlasers, Shanghai, China) at wavelength of 532 nm was used, which does not cause any configuration change in the azo-chromofore. The illumination of the sample is homogeneous (Supplementary Materials, Figure S1). The light from the laser was delivered to the thermal stage by an optical fiber equipped with a pair of collimators. Light intensities at the sample surface were up to 0. 01 W/cm² for UV radiation and 0.025 W/cm² for radiation at a wavelength of 532 nm, correspondingly. A notch filter and absorption optical filter separated the PL signal before reaching the camera, based on a CMOS sensor IMX269 (Sony, Tokyo, Japan). PL spectra of the studied QD-LC composites were obtained through excitation with pulse radiation from a laser PL2143A (Ekspla, Vilnius, Lithuania) at wavelengths of 532 nm (25 ps, 10 Hz, fluence 2 mJ/cm²) and registered with the help of a spectrometer Spectra Pro 2500i (Princeton Instruments, Acton, MA, USA). The obtained images and videos were processed using the software ImageJ version 1.54 g (National Institutes of Health, Bethesda, MD, USA). To minimize the impact of UV radiation and 532 nm laser radiation on the processed images, the camera’s red channel was used for measurements. The mean PL value and its standard deviation (StdDev) were chosen as characteristics describing the evolution of QD PL upon exposure of the composite to UV radiation. The regions of interest (ROIs) used for measuring are 150 µm by 1200 µm in dimension. Each ROI, consisting of approximately 450,000 pixels, was used to calculate mean PL and StdDev values. StdDev in this case characterized the presence of heterogeneity in the QD distribution and significantly decreased under UV irradiation. The contribution of the noise signal to StdDev was significantly smaller than the contribution from the spatial heterogeneity of the QD distribution.

A steel diaphragm with a cut 170 μm wide and a 1.8 mm long slit was placed onto the heating stage to shape the UV spot for localized illumination. For the static experiments, the UV light was passed through the diaphragm, forming a stripe crossing the entire observable sample area with homogeneous illumination. The slit width was found to be optimal for the dynamic experiment, since it allows us to detect separated QD-enriched regions at both borders of the light-illuminated region. In dynamic experiments, the sample slide moved at constant speeds of 26 or 87 μm/min, provided by a 1 rpm electric motor attached to the positioner wheel of the thermal stage through a gearbox, and was continuously UV-illuminated through the immobile slit.

4. Conclusions

Thus, we have demonstrated the redistribution of CdSe/ZnS QDs under the influence of UV radiation in composites based on the nematic LC containing the azo-chromophore. UV irradiation leads to the transition of the LC from the nematic to the isotropic phase, followed by the homogenization of the QD distribution. This homogenization is accompanied by the gradual disappearance of QD-enriched droplets, with diameters ranging from a few microns to 10 μm. The reverse process, with the appearance of QD-enriched droplets, is observed upon cessation of UV irradiation. It is shown that the characteristic times for the homogenization of the QD distribution in a matrix subjected to UV irradiation are 10–15 min for regions with characteristic sizes on the order of 100 μm. The concentration of the QDs can be significantly increased by moving the UV-illuminated spot relative to the sample at a velocity not exceeding the quantum dot diffusion rate. The translation of the UV-illuminated spot through the sample results in QD accumulation inside the isotropic phase induced by UV irradiation.