Abstract

A specific organization of optically active nanoscale objects can greatly affect the optical response of a system. Here, we report the controlled modification of the fluorescent emission by the assembly of water-soluble quantum rods (QRs). Our study combines optical, electron microcopy, and X-ray scattering characterizations to reveal a correlation between the self-assembly behavior of QRs into ordered 3D-arrays and the optical properties (luminescence) of formed assemblies, where the observed optical response is highly dependent on the QR aspect ratio. Specifically, shorter, 18 nm long QRs (QR18), exhibiting a well-defined smectic packing, demonstrate an enhancement of the emission intensity accompanied by a red shift and a lifetime reduction. In contrast, 40 nm long QRs (QR40), forming a columnar phase, does not show these optical properties.

1. Introduction

Assembling colloids, thanks to bottom-up approaches offers an effective way to create a wide range of new materials with specific properties that emerge from the organization of uniform objects into structures periodic at all scales [14]. Such realizations are nowadays possible because of the diversity and the versatility of the method for most of the preparation of the nanoparticles with a high level of control over their size and shape, as well as their assembly in well-ordered superstructures [410]. In particular, quantum dots (QDs) are spherical colloids made of semiconductor materials, 2 to 10 nm in size, which exhibit a robust, efficient, and spectrally well-defined fluorescence emission [47]. Taking advantage of these remarkable fluorescence properties and of the opportunities offered by superstructures, many applications in the field of electronics and photonics have emerged, from lasing [812] to light harvesting devices [ 13, 14], and new physical effects have been evidenced, from fluorescence intensity enhancement to superradiance [1518].

The optoelectronics properties of self-assembled QDs derive from a range of near-field interactions including energy transfer [19] and electromagnetic field enhancement. Thus, they are directly affected by structural and electrostatic parameters such as the strength of the interparticle Coulomb interactions, the dielectric constant of the material, the topology, and the morphology of the assembly [11, 12]. Among these parameters, the interparticle distance has tremendous effect on the optical coupling between neighboring nanoparticles [13]. In particular, due to the exciton coupling between proximal QDs, the fluorescence emission intensity can either increase or decrease, and the maximum wavelength of emission can also be shifted. [15, 19, 20] The coupling of QD and plasmonic nanoparticles can also alter their intensity and fluorescent lifetime as well as modulate their polarization properties [1, 22, 23].

Many strategies have been developed to organize QDs [1, 2426] into 2D and 3D higher-order architectures of well-controlled topology. One of the most efficient method to adjust the interdistance makes use of molecular recognition between biomolecular objects that are grafted on the nanoparticles, for instance DNA [27, 28] or proteins [21, 29, 30]. Another strategy is to exploit electrostatic interactions between protein filaments [31] and/or membranes [32] to generate 3D templates with monodisperse repeat distance in which the nanoparticles can fit. A strong increase in fluorescence intensity, as well as a bathochromic shift of the emission maximum, has already been observed in some of these assemblies [31].

Besides such biomimetic approaches, a simple strategy is to harness the tendency of anisotropic nanoparticles to self-assemble into 3D-arrays, herein called supracrystals. More specifically, as predicted by Onsager’s theory [23, 24], it has been demonstrated that anisotropic semiconductor nanocrystals, also named quantum rods (QR), exhibit a rich polymorphism, which includes nematic and smectic liquid crystal phases [10, 3335]. Previously, we have described a simple, yet powerful, strategy for shaping self-assemblies of water-soluble rod-like nanoparticles, either gold nanorods [36] or QRs [37]: an aqueous suspension is imprinted between a flat substrate and a topologically micropatterned mold, and, after water has evaporated, walls displaying a smectic B packing are obtained, organized on a regular lattice over several square millimeters. On the basis of this work, we here more deeply report on the influence of a variation of the QR aspect ratio on the structural properties of the obtained supracrystals. Furthermore, we characterized the fluorescence emission properties of the obtained assemblies and found that QR organization into 3D-ordered smectic crystal induced a red shift of the wavelength and an enhancement of the intensity of the fluorescence emission as well as a reduction of its lifetime.

2. QRs Synthesis and Morphological Characterization

Starting from the same batch of quasispherical CdSe seeds, we synthesized hydrophobic QRs of different lengths using a slightly modified version [37, 38] of the protocol reported by Carbone et al. [39, 40] (see Supporting Information for details). These QRs are obtained solubilized in toluene and are stabilized by a mixture of trioctyl phosphine oxide (TOPO), octadecylphosphonic acid (ODPA), and hexylphosphonic acid (HPA) (Figure 1(a)). Two batches were characterized by transmission electron microscopy (TEM). Both particle mean diameters, , were found around 6 nm, and the length was 18 and 40 nm (Figure S1); thus, we called them and , respectively. For further solubilization in water, the hydrophobic QRs were first incubated in the presence of an aqueous solution of the TCCPAC peptide (Figure 1(b)) also containing tetramethylammonium hydroxide. Ligands (TOPO, ODPA, HPA) are exchanged for TCCPAC, and the QRs dissolve into the aqueous phase. The resulting hydrophilic QRs were finally purified by gel exclusion chromatography to remove the ligands in excess, and it yielded two samples that we, respectively, denoted and .

Upon transfer from the organic phase to the aqueous one, both batches of QRs exhibited a slight red shift of their maximum of fluorescence emission intensity, from 581 to 584 nm. A typical evolution of the fluorescence spectra exhibits a small decrease in their emission quantum yield, from 0.55 to 0.40 as shown in Fig.S2a for QR18 and similar fluorescence emission decay (Fig.S2b, Table S1A). The fluorescent emission red shift may be attributed to the ligand tricystein interaction with the surface Zn and S atoms [32, 41, 42] and the solvent change.

3. QR Supracrystals Synthesis and Morphological Characterization

We next focused on obtaining long-range order in 3D colloidal crystals having a size of at least several tens of microns. To do so, we used a previously developed approach in which we let dry an aqueous suspension of QRs in between a flat substrate, either a silicon wafer or a glass coverslip, and a PDMS stamp on which round pillars, 35 μm high and 100 μm in radius, are regularly spaced by 75 μm on a square lattice (Figure 2(a)) [36, 43]. More precisely, sample was prepared from at 2.5 μM supplemented by Tween 20 at 96.4 μM whereas sample was similarly obtained starting from a solution of (Table S2A). In both cases, after unmolding, one could observe fluorescent walls forming a maze-like pattern on the substrate. These tridimensional walls connect the positions previously occupied by the pillars and typically have a triangular section in the range of 10 to 20 μm (Figure 2(b)).

Studying the internal organization of such self-assembled supracrystals is essential for understanding the relationship between 3D structuration and emergent optical properties. First, the micropatterns made of QRs were observed by scanning electron microscopy (SEM). Sample cleavage cracked the walls, thereby opening the windows through which one could gain insight into colloidal packing. In some of the crevices obtained on both and , dots arranged on a hexagonal lattice were observed (Figures 2(c) and 2(d), respectively). In some other crevices, it was also possible to observe nanoparticles from the side, and here, the two structures differed markedly. Whereas in , rods are clearly longitudinally organized in lamellae (Figure 2(e)), in , one could only view surfaces covered with long and parallel linear motifs (Figure 2(f)).

Yet, SEM can only provide information on the inner structure of the supracrystals through cracks. It is also difficult to obtain precise lattice parameters from such measurements. Thus, small-angle X-ray scattering (SAXS) was used to overcome these limitations and investigate the bulk of QR assemblies. An aqueous suspension was poured into a glass capillary and slowly dried under vacuum, so as to yield the sample after one week (see Supporting Information for details). With a beam size smaller than any of the capillary dimensions, it was first possible to check for polymorphism within the sample: none was found since all the recorded diffraction patterns were totally superimposable (Fig. S4). The formed supracrystal could next be indexed as a smectic B (SmB) phase with an ABC packing (Figure 2(g)). The unit cell parameters are  Å and  Å, which correspond to a lateral hexagonal spacing  Å and to a lamellar repeat distance  Å (see Tab. S3 for full indexation). As a matter of fact, SAXS results fully support SEM observations although preparation conditions were slightly different. Indeed, Tween 20 was added in the starting solution used for the microstructured patterns. As a control, we recorded the SAXS intensity plot in the presence of detergent, and no change could be noted (Fig. S5), proving that this compound does not influence QR packing.

In the case of the sample, which was prepared in the same manner as sample but starting from , SAXS data show that the structure obtained after drying can be indexed as a hexagonal columnar phase with a unit cell parameter  Å (Figure 2(h), and Tab. S4 for full indexation). Once more, X-ray scattering measurements agree with SEM observations. Additionally, the shoulder named T gives an interesting detail about the columnar arrangement, which could not be evidenced by SEM. By analogy to what we have previously observed on gold nanorods columnar phases [36], it suggests that some QR columns are shifted by half of a nanoparticle length with respect to the adjacent columns.

Such smectic/columnar polymorphism of nanorods is usual and has been observed in the case of goethite inorganic materials [44], and of cylindrical viruses [45]. Both polydispersity and repulsion interaction features are invoked to impose the packing at the thermodynamical equilibrium. Lateral interactions are similar in both and . However, the relative polydispersity (Fig. S1) of is higher than . By analogy, goethite nanorods could be the origin of the observed smectic ordering.

To summarize our structural analyses, we can say that the surface chemistry of the functionalized, water-soluble QRs and the use of a microstructured stamp make possible to form supracrystals with a well-defined positional and orientational smectic order at a large scale in the case of QR18 and a columnar type order for QR40. In addition, the good quality of the long-range ordering could not be achieved with QR18. Importantly, whereas the surface-to-surface distances are in the same range, ≈3-4 nm for both QR18 and QR40, the center-to-center distance is much smaller and uniform for QR18, owing to the nature of the smectic B arrangement with respect to the one of a columnar hexagonal one. Indeed, for QR40, the center-to-center distance is on average longer (due to the higher aspect ratio and displacement in columnar state) and therefore varies a lot due random vertical shift. In the following, the QR40 will consequently be used as a control.

4. QR Supracrystals Fluorescence Emission Properties

In order to correlate the organization of the emitters with the optical properties of their assemblies, we recorded the fluorescence emission spectra and the fluorescence emission decay curves of the two hydrophilic QR batches in two distinct condensed, dried states. The first one, , corresponds to the microstructured supracrystals described above and measurements with a confocal setup have been performed in the upper part of the walls presented in Figure 2(b). The second type of condensed state, , corresponds to aggregates resulting from an undirected drying process without the PDMS stamp. In such aggregates, nanoparticles locally display a crystal liquid order, the phasing identical to the one observed in the capillary and in the walls. Finally, since for both QR lengths, sample preparation proceeded from nanoparticles dispersed at the same concentration in water (2.5 μM), we can assume that the number density of emitters in a given condensed state is more or less the same, within experimental errors and within a factor that would account for the difference in volume between QR18 and QR40.

We first analyzed the fluorescence spectra of the condensed dried states and compare them to the spectra obtained from noninteracting QRWater in an aqueous suspension. We noted that the ratio of emission intensities at between and is equal to 2.5 (Figures 3(a) and 3(c), green curves), which can be interpreted as a consequence of a roughly twice higher density of emitters for the sample made of roughly twice shorter nanoparticles. In contrast, the ratio of emission intensities at between and is equal to 190 (Figures 3(a) and 3(c), gray curves). Furthermore, with a 9 nm bathochromic shift, the sample is the only one that displays a maximum intensity wavelength different from the 583–584 nm value (Table 1).

Now considering the fluorescence emission decay curves, it can be observed that nanoparticles in dried states display shorter average lifetimes ( in comparison to the noninteracting in water (. In the condensed dried states, the contribution of the larger lifetime strongly decreases, whereas the ones of the two smaller lifetimes increase (see and in Table 1 and Figures 3(b) and 3(d)). To be more precise, in aggregates, the values of the preexponential factors are in the range of , and for both QR lengths resulting in an average lifetime . The microstructured supracrystals are characterized by similar figures (, and ) with an average lifetime , but for , the decrease in lifetime is less marked with preexponential factors values , and .

These experiments show three specific photophysical effects that are observed in the case of : (i) a decrease in lifetime accompanied by (ii) an enhancement of the emission intensity with (iii) a 9 nm bathochromic shift. It is worthwhile to discuss if the particular behavior of could be directly related to the smectic order present in the supracrystals, for which both the in-plane and the normal organization of QRs might favor an efficient long range optical coupling.

The observed fluorescence exaltation is an increase of the emitted photons per time-unit which can find origin either in a nonradiative process vanishing or in a photon emission turnover increase. Indeed, an excited molecule or nanostructure relaxes to a lower energy state through radiative emission of a photon and nonradiative processes can depopulate the excited state, with respective decay rates and . The emission flux is proportional to the quantum yield and can increase through an increase of or a decrease of . The mean fluorescence lifetime is . Observing decrease concomitantly with an increase of quantum yield shows that the core mechanism is a reduction of the radiative lifetime. These observations are in lines with the ones reported in the literature with other types of nanoparticles superlattices [29, 46, 47]. This kind of photophysical behavior is rare and mostly related to collective effects. This can be enlightening by noting that similar results are observed for strongly coupled such as gas or dyes [4851] and are attributed to superradiance or superfluorescence. Since no coherence has been observed, we should invoke an energy transfer that might be either by tunnel effect (such as FRET) or by dipolar electromagnetic interactions (such as superradiance). The smallest interdistance between QR surfaces () is large (here around 3.7 nm in the hexagonal lattice). For such distance, QR excited states do hardly interact via the tunnel effect [50]. In the present case, dipolar interaction must generate a cascade of coherent fluorescence called superradiance [4852]. Such radiation reaction is accompanied by a significant energy loss, e.g., a slight red shift [53]. It has to be noticed that fluorescence intensity varies quadratically as , being the number of coherent emission dipoles. In the present case, efficient absorber was required with in order to not saturate the PM. Noteworthy, the superradiance is very sensitive to the positional order of the emiters due to the factor of interference between emission waves (Equation (1)).

This pair-interference term can be compared to the interference term for X-ray scattering. In our case, they are identical. This means that the pair-interference terms are exalted when absorption of dipoles are located in a 3D crystal. As mentioned, superradiance [48] does not correspond to any tunnel transfer of fluorescence such as in FRET. The excited states relax via dipolar interactions more efficiently in presence of an electromagnetic wave oscillating at a frequency close the emission energy difference, like in NMR. The emitted photon has energy close to the initial but is not coherent, neither in energy nor in direction. Such radiation reaction is accompanied by a slight bathochromic shift [31] which could be in agreement with the spectra we have recorded on . In the case of , we do not observe any enhancement of the emission intensity regardless of the lengths of the nanoparticles, which we attribute to the smaller scale over which they are organized. For , the nanoparticles are ordered on a larger scale but their respective lateral positions are less controlled than in the smectic phase characterizing . We evidenced the sensitivity of collective optical properties to the 3D positional order by the comparison between and . The former sample is well organized in a 3D crystal and exhibits a fluorescence increase with a bathochromic shift linked to a lifetime reduction, while the latter is organized only in 2D and exhibits fluorescence properties similar to the individual QRs.

5. Conclusion

Our bottom-up strategy enables to self-organize in hierarchical organization: nanoparticles obey a 3D positional and orientational order over hundreds of nanometers, they are part of supracrystal having the shape of walls with a triangular cross-section a few tens of microns large, and these walls are patterned on a substrate at the millimeter scale. More specifically, a smectic B (ABC) phase is obtained with 18 nm long QRs and a hexagonal columnar phase is obtained with 40 nm long QRs, which reveal that the nanoparticle length can be used to tune the structural properties of the assemblies, and consequently their optical ones. Indeed, confocal measurements of the fluorescence emission spectra and decay curves have revealed that, although in the two supracrystals emission is faster than for noninteracting shorter species, and the most organized, it is the only one that displays a strong enhancement of the emitted intensity. Thus, our bottom-up approach could be of interest in the field of photonics, in order to simply realize optical systems with strong fluorescence quantum yield.

Data Availability

Data in supplementary information files (included).

Disclosure

Elsa Mazari-Arrighi’s present address is CYTO/PCV/BIG/DRF/CEA, Unité de Thérapie Cellulaire, Institut Universitaire d’Hématologie, Hôpital Saint Louis, 1 avenue Claude Vellefaux, 75010 Paris, France. Christophe Dupuis’ present address is Centre de Nanosciences et de Nanotechnologies, UMR 9001 CNRS/Université Paris-Saclay, 10 boulevard Thomas Gobert, 91120 Palaiseau, France. Charlie Gosse’s present address is Institut de Biologie de l’Ecole Normale Supérieure, ENS, CNRS, INSERM, PSL Research University, 46 rue d’Ulm, 75005 Paris, France.

Conflicts of Interest

The authors declare that they have no conflicts of interest.

Authors’ Contributions

Thomas Bizien is responsible for synthesis and structural and optical measurements. Marie Postic is responsible for surface deposition of nanoparticles. Pascale Even-Hernandez is responsible for supervision of the QR colloidal characterization. Pascal Panizza is responsible for process of surface deposition. Cristelle Mériadec is responsible for SAXS experiments. Florian Meneau is responsible for SAXS experiments. Elsa Mazari-Arrighi is responsible for SEM analysis. Christophe Dupuis is responsible for SEM measurements. Charlie Gosse is responsible for supervision of SEM analysis and manuscript writing. Mircea Cotlet is responsible for supervision of the optical properties and lifetime particle measurements. Oleg Gang is responsible for structural analysis. Valérie Marchi is responsible for manuscript writing and planning of the QR synthesis and characterization. Franck Artzner is responsible for supervision of the project, the structural analysis and of the optical properties.

Acknowledgments

The Agence Nationale de la Recherche (ANR contract No. 16-CE09-0027-02 HYPNAP and contract No. 14-CE08-0004 ARTEMIS) funded this work. We are indebted to the Region Bretagne, the French Direction Generale de l’Armement, and the ANR for the PhD financial supports to T. Bizien, M. Postic, and E. Mazari-Arrighi. We also acknowledge Université de Rennes for a travel grant to T. Bizien, so as to enable him to perform measurements at the Center for Functional Nanomaterials of the Brookhaven National Laboratory and Rennes Métropoles for equipment support. Oleg Gang was supported by the US Department of Defense, Army Research Office, grant W911NF-19-1-0395. This research used resources of the Center for Functional Nanomaterials, supported by U.S. DOE Office of Science Facilities at Brookhaven National Laboratory under contract No. DE-SC0012704.

Supplementary Materials

Experimental details are given about synthesis of the CdSe/CdS hydrophobic nanorods, transmission electron microscopy, UV-Vis absorption spectroscopy, quantum yield determination preparation of the “QR” supracrystals, scanning electron microscopy, small angle X-ray scattering, and confocal fluorescence spectromicroscopy. Figure S1: diameters and lengths of the hydrophobic s measured by transmission electron microscopy (TEM). (a) . (b) . Figure S2: Characterization of the fluorescence emission properties of the s dispersed in toluene and in water at 2.5 μM. Figure S3: deconvolution of the fluorescence emission spectra of hydrophilics QRs dissolved at 2.5 μM in a drop of water placed over a coverslip. Figure S4: Detailed SAXS data acquired on . Figure S5: SAXS intensity plot acquired when drying proceeds. Table S1: influence of the ligand and solvent on the maximum intensity wavelength () and on the decay of fluorescence emission for 18 nm long QRs. Table S2: influence of the length of hydrophilic QRs on their fluorescence emission spectra. Table S3: full indexation for the SAXS intensity plot measured on sample (Figure 2(f)). Table S4: Full indexation for the SAXS intensity plot measured on sample (Figure 2(g)). (Supplementary Materials)