Determination of Handedness in a Single Chiral Nanocrystal via

Subscriber access provided by AUSTRALIAN NATIONAL UNIV

Article

Determination of Handedness in a Single Chiral Nanocrystal via Circularly Polarized Luminescence Eitam Vinegrad, Uri Hananel, Gil Markovich, and Ori Cheshnovsky ACS Nano, Just Accepted Manuscript • DOI: 10.1021/acsnano.8b07581 • Publication Date (Web): 06 Dec 2018 Downloaded from http://pubs.acs.org on December 6, 2018

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 25 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Nano

Determination of Handedness in a Single Chiral Nanocrystal via Circularly Polarized Luminescence Eitam Vinegrad1,§, Uri Hananel2,§, Gil Markovich2 and Ori Cheshnovsky2,* 1. School of Physics, Raymond and Beverly Faculty of Exact Sciences, Tel Aviv University, 69978 Tel Aviv, Israel. 2. School of Chemistry, Raymond and Beverly Faculty of Exact Sciences, Tel Aviv University, 69978 Tel Aviv, Israel.

ABSTRACT: The occurrence of biological homochirality is attributed to symmetry breaking mechanisms which are still debatable. Studies of symmetry breaking require tools for monitoring the population ratios of individual chiral nano-objects, such as molecules, polymers or nanocrystals. Moreover, mapping their spatial distributions may elucidate on their symmetry breaking mechanism. While luminescence is preferred for detecting single particle chirality due to its high signal to noise ratio, the typical low optical activity of chromophores limits its applicability. Here, we report on handedness determination of single chiral lanthanide based luminescent nanocrystals with a total photon count of 2×104.

Due to the large emission

dissymmetry we could determine the handedness of individual particles using only a single circular polarization component of the emission spectrum, without polarization modulation. A machine learning algorithm, trained to several spectral lineshape features, enabled us to determine and

ACS Paragon Plus Environment

1

ACS Nano 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 25

spatially map the handedness of individual nanocrystals with high accuracy and speed. This technique may become invaluable in studies of symmetry breaking in chiral materials.

KEYWORDS:

chiroptical

activity,

single-particle

spectroscopy,

circularly

polarized

luminescence, lanthanide nanocrystals Life has evolved in a way that virtually all proteins, sugars and nucleic acids are homochiral. While the origin of this biological homochirality is still under debate, some theories suggest an event of spontaneous symmetry breaking from initially racemic conditions.1 In order to understand the transition from a racemic to a homochiral state, tools that allow identification of handedness of single nano-objects are needed. Recently, single chiral nanostructures (single chiral assemblies) were probed by differential circularly polarized light scattering: Link and coworkers showed that individual chemically linked nano-dumbbell dimers show high levels of optical activity.2 Hentschel and coworkers showed that the chiroptical activity of single plasmonic antenna ensembles depends on minute structural differences.3 Boyd and coworkers showed that in nano-sphere trimers in which symmetry is broken by material selection a chiroptical response is induced.4 Narushima and Okamoto have used a sophisticated discrete modulation technique to exclude linear dichroic/birefringence contributions in scattering based circular dichroism (CD) microscopy.5 Genet and coworkers used optical trapping to rapidly identify the handedness of single chiral nano-structures in solution by measuring scattering CD.6 These studies were based on high scattering from different configurations of plasmonic particles (size range of 100-1000 nm) which yielded large dissymmetry factors (differential scattering normalized to total scattering) of the order of 10-1.


2

Page 3 of 25 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Nano

We have recently measured absorption CD of single inorganic nanocrystals (NCs)

7

with

dissymmetry factor as low as 10-2. While all these methods can be used over a wide range of samples, they are extremely sensitive to optical alignment and excitation polarization induced artifacts, and consequently are relatively slow and suffer from a limited confidence level. Light emission from nanoparticles and molecules enables their detection with better signal to noise ratios than absorption and scattering based methods, and since the light emission spectra are controlled by the emitting species structure and electronic properties, the NC's emission polarization only depends on its orientation and structure and not on the polarization of the excitation beam, thus the system is less prone to alignment artifacts (CD microscopy depends on both illumination and collection polarization artifacts). Moreover, the emission signal is free from interference effects which may occur in scattered and reflected laser light which complicate the analysis of the signal. Therefore, whenever applicable, emission based chiroptical measurements are preferable for single chiral nano-object characterization. The use of fluorescence intensity to detect absorption CD was studied on individual chiral molecules.8 However, these measurements were shown to be erroneous, as they must have contained a significant linear polarization anisotropy contribution which far exceeded the CD magnitude.9 Chirality can also be manifested through circularly polarized luminescence (CPL), the unequal emission of left- and right- circularly polarized light.10 The magnitude of this difference depends on the emitter properties, as well as on the degree of enantiopurity of the examined material. The measured quantity is the emission circular intensity differential (ECID), 𝐼𝐿 ― 𝐼𝑅, where IL and IR are the intensities of the emitted left- and right-handed circular polarization components, respectively. The degree of optical activity is given by the dimensionless dissymmetry factor 𝑔𝑒𝑚 𝐼𝐿 ― 𝐼𝑅

= 2𝐼𝐿 + 𝐼𝑅, taking values between ± 2. ECID or 𝑔𝑒𝑚 can be used to monitor the handedness of a


3

ACS Nano 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 4 of 25

light emitting sample, and have been measured so far only in ensemble of chiral organic molecules and their metal ion complexes,11 and semiconductor NCs coated with chiral molecules.12,13 The strongest 𝑔𝑒𝑚 was measured in chiral lanthanide complexes.14,15 Recently, large 𝑔𝑒𝑚 was measured in emission lines of inorganic lanthanide phosphate NCs which crystallize in a chiral spacegroup.16 These experiments led us to strongly suspect that at certain synthesis conditions only one NC enantiomer is formed. However, until recently there has been no experimental method which could quantify the enantiomer population ratios in a chiral NC sample within a reasonable amount of time and confidence. Here we describe a methodology which allows us to perform highthroughput, high fidelity single particle quantitative assessment of the enantiomer population ratios of lanthanide based chiral NCs. RESULTS AND DISCUSSION Rod-shaped Eu3+ doped (5%) TbPO4D2O lanthanide NCs were synthesized by a simple aqueous procedure described in the Methods section.17 The lanthanide NC handedness was directed with either D- or L- tartaric acid present in the synthesis solution, and we label the resulting crystals as ‘D’ and ‘L’, accordingly. The NCs had typical dimensions of 1.4 µm  50 nm. Emission measurements of NCs dispersed in D2O exhibited sharp peaks typical of Eu3+ in the visible range, attributed to the 0𝐷5→7𝐹𝐽 transitions where 𝐽 = 1,..,4.18 In both ensemble and single particle measurements we excite the Tb3+ ions (at 365 nm and 488 nm, respectively). The Tb3+ ions transfer the excitation energy to the Eu3+ ions, which emit after a characteristic lifetime on the order of a millisecond.19,20 Eu3+ doping was performed mainly due to the record high CPL values that Eu3+ emission lines are known to have when complexed with chiral molecules (significantly higher than chiral Tb3+ CPL values). The Tb/Eu combination was chosen since Tb3+ absorbs more efficiently than Eu3+ at the chosen excitation wavelength, transferring the energy to the emitting


4

Page 5 of 25 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Nano

Eu3+ dopant. Lastly, Eu3+ has more emission lines in the visible, which would be useful for handedness determination as demonstrated later on. The NCs displayed ECID lines with varying relative intensities throughout the emission spectrum, with inverted polarity for the two opposite enantiomers (Figure 3d), as expected from a chiral emitter. High dissymmetry emission lines display the greatest relative line-shape difference between the two enantiomers, and can thus serve for identification of the handedness of single particles. The 596 nm and 650 nm emission lines displayed particularly high dissymmetry values (0.15 and ≥0.4, respectively)17. The majority of the emission at and above 590 nm stems from the Eu3+ dopant ions in the interior of the NC, as surface emission is quenched.19,21 Since the NCs are all single crystals of a particular handedness,16 all of the emitting ions within an individual NC are located in the same type of local chiral environment and thus produce the same ECID polarity. Taking into account the average diameter of the nanorods and the size of the diffraction-limited excitation spot we estimate the numbers of emitters illuminated by the excitation beam to be in the order of ~105. The low absorption cross section of these rods leads to ~200 photons/s at our EM-CCD detector for the 596 nm peak. Considering the overall optical transmission and the collection solid angle, this translates to a total of ~8000 photons/s emitted by this particle at the particular emission line, which is several orders of magnitude lower than a typical single fluorescent molecule like Rhodamine 6G. Single Particle CPL and ECID measurements. For single particle ECID measurements, the NCs were drop-cast on a Formvar film coated copper grid to enable both transmission electron microscopy (TEM) and luminescence microscopy for the same population of individual NCs. Single particle distribution was mapped on the grid by TEM. The grid was then placed on a low fluorescence fused-silica cover glass, and the same particles were measured by CPL microscopy, using the setup illustrated in Figure 1.


5

ACS Nano 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 6 of 25

Figure 1. Schematic of the experimental setup. The broadband supercontinuum source is filtered by an Acusto-optical tunable filter (AOTF) to select the 488 nm excitation line. The excitation beam wavelength is further cleaned via 488 nm laser line (LL) and 500 nm low pass (LP) filters. A /4 waveplate is used to convert the excitation polarization into circular, to remove potential dependence of the NCs orientation on excitation polarization. Then, the excitation beam is passed through 50:50 non-polarizing beamsplitter, separating the light to a reference beam and a probe beam. The probe beam is focused on the sample, which is scanned by a piezoelectric motor. This enables the detection and localization of single NCs. The transmitted light is collected by another objective and is filtered by a 550 nm high pass filter (HP) to transmit only the NCs luminescence. The emitted light is then passed through a super-achromatic /4 waveplate, followed by a linear polarizer at 45° to the fast axis of the /4 waveplate to a spectrograph equipped with an EM-CCD camera. This enables measurement of circularly polarized emission of a specific handedness. ECID is measured by subtracting this spectrum from another in which the polarizer is rotated by


6

Page 7 of 25 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Nano

90 . The reflected light is used to locate the NCs using extinction measurement (PD), while the camera is used to correlate the locations of the particles to the TEM micrographs.

These nanorods exhibit a large variety of emission peak intensities. This phenomenon is depicted in Figure 2a, showing several unpolarized emission spectra from an aqueous dispersion ensemble, an aggregate of many NCs as well as two single NCs. While the general line-shapes and peak wavelengths of these spectra are similar, they exhibit distinct variations in the relative peak intensities, which most likely stem from different defects and doping efficiency in each NC.22,23 Since the ECID signal is usually a small fraction of the luminescence signal, it is desirable to employ fast modulation between the right (r-CPL) and left (l-CPL) handed circular polarization luminescence, combined with lock-in detection. The difficulty in such approach is that polarization modulation of the emitted light (e.g. using liquid crystal variable retarder or photoelastic modulator) requires wavelength dependent tuning and temperature sensitive retardation calibration. Hence, the emitted light is collected sequentially by scanning over the relevant wavelength range (in this case 580 nm - 720 nm). The large dissymmetry values of several emission lines enabled us to use two faster schemes for identifying the handedness of the individually measured particles: I. Measuring ECID by subtracting two accumulated spectra of rCPL and l-CPL. II. Analyzing the spectral line-shape of only r-CPL or l-CPL to deduce the handedness.


7

ACS Nano 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 8 of 25

Figure 2. Unpolarized and polarized emission spectra of the NCs. (a) Unpolarized luminescence spectra from an aqueous dispersion ensemble (black), an aggregate of many NCs (blue) as well as a typical 'L' (orange) and 'D' (purple) lanthanide NCs. The 'J' labels correspond to the different J levels of the 0𝐷5→7𝐹𝐽 transition. (b) Right- (red) and left- (blue) handed circularly polarized luminescence (r-CPL and l-CPL, respectively) from a single 'D' lanthanide NC. Shaded areas mark the wavelength ranges that exhibit significant (>10%) ECID signal. All l-CPL spectra were


8

Page 9 of 25 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Nano

normalized by scaling with respect to the same particles' r-CPL spectra as described in the Methods section. (c) Effect of imperfect CPL detection of the emission lines observed through NC orientation dependence: r-CPL spectra averaged from 7 NCs at each of two perpendicular NC orientations (a ±10° margin was allowed for each orientation), collected for both NC enantiomers. The same normalization used for panel b was applied here. The two orientations were chosen so that the maximal difference was obtained. We also examined a third orientation (not shown here), at 45° to each of the two perpendicular orientations. The emission lineshape of this orientation was, as expected, an average of the two perpendicular orientations.

To measure the entire r-CPL or l-CPL spectrum in a single measurement cycle we used a superachromatic 𝜆/4 waveplate combined with a linear polarizer, switching the polarizer between +45 and 45, with respect to the fast axis of the waveplate. Figure 2b shows the r-CPL and l-CPL from a single 'D'-NC. These spectra show areas with significant emission dissymmetry (>10-2, shaded in grey). Ideally, our emission analysis detects exclusively circularly polarized emission. However, due to imperfections in the system residual linear polarization effects in the emission are detected. To estimate these effects, we have averaged the emission spectra of two subgroups of nanocrystals with perpendicular orientation. Fig. 2c shows that for some emission peaks (shaded grey in Fig. 2b) the difference in circular polarization emission between two enantiomers dramatically exceeds the effects of orientation, related to linear polarization effects. Other peaks such as the 610-620 and 696 nm peaks showed a significant amount of orientation dependence exceeding the circularly polarized emission difference between the two chiral enantiomers and thus were not used for handedness determination.


9

ACS Nano 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 10 of 25

Figure 3a,b shows the ECID spectra (𝐼𝐿 ― 𝐼𝑅) of several single NCs of both handedness. These spectra show that some of the ECID peaks are stable between different NCs, e.g. the 596 nm and 650 nm peaks, while other peaks, like the 612 nm peak, vary significantly between individual particles and display a 'random' sign independent of NC handedness. This most likely stems from the different orientations of the single nanocrystals, in the emission peaks where the circular polarization dissymmetry (how 'chiral' the transition is) are small (1% or less) compared to the orientation-dependent linear polarization effects. These spectra, which clearly define the handedness of the individual particle, were obtained with a total photon count of 2104 photons (collected during 100 s) originating from about 105 Eu3+ ions located within the diffraction limited illuminated volume of an individual NC. Figure 3c shows the average ECID over seven particles for each of the two enantiomers. This ECID average compares quite well in peak location and sign to the aqueous dispersion spectrum depicted in Figure 3d, since the orientation dependence is averaged out. However, the relative amplitudes still differ. This probably originates from the difference in the c-axis alignment in the two measurement configurations. In the aqueous dispersion the long axis (c-axis) of the NCs is randomly aligned, while in the CPL microscopy sample it is confined to the substrate plane. Since the spatial distribution of emission intensities is known to vary between different transitions in such NCs,24 variations in spatial alignment can be translated to different relative intensities. The measurement of ECID over a large distribution of individual particles more than doubles the total measurement time compared to only measuring r- or l-CPL, and increases the experiment's sensitivity to sample drifts, limiting the usefulness of this method. Looking back at the raw circularly polarized emission spectra in Figure 2b, we note that the high dissymmetry of some of


10

Page 11 of 25 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Nano

the emission peaks leads to slight yet discernible differences in the luminescence spectral lineshape between the two orthogonal circular polarizations.

Figure 3: ECID spectra of individual NCs and their ensembles. (a) ECID spectra of seven different single 'D' NCs. (b) ECID spectra of seven different single 'L' NCs. The L- and r-CPL spectra in (a,b) were separately normalized by their 688 nm peak intensity, and corrected for the polarization collection efficiency before subtraction of 𝐼𝐿 ― 𝐼𝑅 . (c) Average ECID spectra for the seven lanthanide 'D' NCs (red line), and seven 'L' NCs (blue line). Shaded gray areas mark spectral regions for which the ECID of the single NCs has the lowest variance between the measured NCs.


11

ACS Nano 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 12 of 25

These spectral regions were then used for the input features for the machine learning classifier. (d) Ensemble ECID spectra of 'L' (blue), 'D' (red) and the racemic (black) aqueous dispersion samples. The spectra in (d) was scaled by a constant to be of the same range as in (c).

These line-shape differences, which are inverted between the two NC enantiomers, allowed us to determine the NC handedness without measuring ECID spectra, i.e., by measuring only a single circular polarization. This single measurement also minimizes the offset of ECID values originating from residual asymmetry of the experimental setup response to right- and left-handed circularly polarized light. Handedness classification. In order to establish the use of line-shape analysis of single r- or lCPL spectra to determine the NC handedness, we measured many single particles' r-CPL spectra over two different samples which, according to ensemble aqueous dispersion ECID measurements, were expected to be pure 'L'- and pure 'D'-NCs.16 For handedness determination based on spectral features we employ a machine learning Support Vector Machine (SVM) classifier.25 For the SVM training data we measured 64 individual NCs from the ‘D’ sample and 60 individual NCs from the ‘L’ sample. The SVM algorithm input consisted of four distinct features of the luminescence line-shape, and the corresponding NC handedness. We chose the peaks which display the highest dissymmetry in dispersion measurements, and which show the least dependence on NC orientation. Since the luminescence intensity is not a quantitative attribute, as it is dependent on NC size, individual doping efficiency and illumination intensity, we normalized each peak of interest to a neighboring peak which exhibits a lower dissymmetry factor in ensemble measurements. Figure 4a shows a scatter plot of the training data using three of the four selected spectral features. The scatter plot


12

Page 13 of 25 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Nano

displays a clear separation surface between the two NC enantiomers. Figure 4b outlines the four normalized spectral features, and shows their individual scatter plots and the classification accuracy of the SVM classifier used with only that single feature. Using the four spectral features, the SVM algorithm classified the training data with an error rate of less than one percent (0.8%). Only one NC from the ‘L’ sample was classified as a ‘D’ NC. A close examination of this NC lineshape in comparison to the line-shape of other ‘L’ and ‘D’ particles lead us to the conclusion that this specific particle is indeed a ‘D’ handed NC, essentially bringing the classification accuracy to 100%. The discovery of a 'D' NC impurity in the supposedly enantiopure ‘L’ NCs sample further demonstrates the importance of the single particles handedness study, and the robustness of the SVM classifier.

Figure 4. Machine learning handedness classification procedure. (a) Scatter plot of the classification training data from presumably 'L' (blue) and 'D' (red) enantiopure dispersions. Accordingly, blue and red spheres are the correctly classified NCs out of the training set. The blue cube is a NC classified in contradiction to its expected handedness. Data points were rescaled to the range [0, 1] for each axis independently. The scatter plot uses three out of the four features that led to the highest classification accuracy. Feature 1 is the ratio of the 596 nm peak to that of the


13

ACS Nano 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 14 of 25

592 nm shoulder. Feature 2 is the area ratio between the 650 nm peak and the 653 nm shoulder. Feature 4 is the ratio of the 596 nm peak to that of the 650 nm peak height. (b) Scatter plots for each of the individual data features used for data training and classification. The percentage over each feature represent the SVM training accuracy reached when using only that feature for training.

Once the SVM classifier was trained, we moved on to a racemic NC sample; an equal mixture of the ‘D’ and ‘L’ NCs, which displayed zero ECID in ensemble measurements. We were able to predict the handedness of individual NCs in this sample. These predictions were randomly checked over 7 particles of each handedness, with 100% success (shown in Figure 3a,b). For the 162 NCs measured out of the racemic mixture we found that the D to L particle ratio is 90:72, so that the probability of finding a 'D' enantiomer is 𝑃(𝐷) = 0.555 ± 0.064, which is within the statistical error margin from the expected 50% ratio. Figure 5 shows TEM micrographs of the racemic NCs sample with individual NCs artificially colored according to the SVM classifier handedness analysis. Unexpectedly, the spatial distribution of the two opposite NC enantiomers in the racemic sample seems to be heterogeneous: NCs of the same handedness are more likely to be found in proximity to their own kind. This effect, which has been observed for other salts,26 may be related to symmetry breaking and chiral amplification recently observed in the synthesis of these NCs,16 and will be further studied in the future using the technique presented here.


14

Page 15 of 25 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Nano

Figure. 5. TEM micrographs of NCs from a racemic dispersion. The sample was prepared by dropcasting a NC racemate dispersion on a Formvar film. The measured NCs were color coded according to their classification results, blue for the 'L' enantiomer and red for the 'D' one. Scale bars are 10 µm. The uncolored NCs were not measured. Insets are magnifications of the areas surrounded by the dashed rectangles, showing the high spatial resolution of the method. CONCLUSIONS In summary, we have measured CPL spectra and used their analysis to determine the handedness of individual NCs. We have demonstrated a methodology for handedness sorting of single nanostructures based on CPL. This has been achieved for rod-shaped Eu3+ doped (5%) TbPO4D2O NCs with 105 emitters illuminated with a diffraction limited spot. The low photon count needed for CPL measurement of single particles suggests that this method could be extended to single chiral molecules or molecular aggregates, provided that the emission dissymmetry is larger than 10 ―2, while emitting 107 photons before bleaching, and care is given to omit linear effects in the CPL analysis. Considering the recent progress in the field of chiral organic chromophores this challenge seems surmountable.27 This system could also be beneficial for bio-labeling using lanthanide-based luminescent markers, providing an additional degree of freedom besides color separation. Finally, the ability to identify and count handedness over a large number of particles


15

ACS Nano 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 16 of 25

together with the two dimensional spatial information forms a valuable tool to study symmetry breaking mechanisms and their kinetics.

METHODS Chiral lanthanide nanorod synthesis. Terbium chloride hexahydrate (99.9%) and europium chloride hexahydrate (99.9%) (Fisher Scientific). Sodium dibasic phosphate (≥99.0%), D-tartaric acid (99%) and L-tartaric acid (≥99.5%) (Sigma). Hydrogen chloride (32%) (BioLab). Deuterated water (D2O) (ARMAR Chemicals). All chemicals were used without further purification. All precursor chemicals were diluted in D2O to 100 mM stock, except the hydrogen chloride solution which was diluted in D2O to 1M. The synthesis consisted of two steps: 1) Synthesizing spherical 'seed' nanoparticles of ~5 nm diameter, and 2) Seeded growth of these seeds into larger (~1 uM on 50 nm) nanorods. All of the syntheses were performed in D2O, which enhanced the lanthanide NC luminescence, as water is a known Eu3+ quencher.17,19,21 1. Seed synthesis: Solution A preparation: in a 20 ml scintillation vial, to 10 mL D2O were added under constant stirring, at the order as listed: 285 μL of TbCl36H2O, 15 μL of EuCl36H2O, 150 μL of HCl 1M and 300 μL of L- or D-TA, depending on which NC enantiomer is desired.. Solution B preparation: in a glass vial, under stirring, to 1 mL of D2O we added 0.6 mL of Na2HPO4 and 150 μL of HCl 1M. The solutions were allowed 30 minutes to equilibrate at the desired temperature, following which solution B was rapidly added to solution A. The addition of the phosphate solution to the Ln3+ solution prompted the immediate precipitation of Eu3+ doped TbPO4D2O NC seeds. The aqueous dispersion remained clear. The seeds were purified by adding 1:1 volume of acetone and


16

Page 17 of 25 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Nano

centrifugation at 1070 RCF for 5 minutes, re-dissolution of the precipitate in D2O, vigorous vortex and sonication for at least 20 minutes. 2. Nanorod synthesis: Solution A was prepared in the same manner as the seeded synthesis, only heated on a hot plate set so the solution temperature was 75°C. Solution B was prepared in the same manner, only 10ml of D2O was used instead of 1 ml, and it was also heated on the same hot plate as solution A. The solutions were allowed 30 minutes to equilibrate at the desired temperature, following which 90 µl of the cleaned seed solution was added to solution A and immediately after solution B was rapidly added to solution A. The addition of the phosphate solution to the Ln3+ solution prompted the precipitation of Eu3+ doped TbPO4D2O NC rods, which caused the aqueous dispersion to grow turbid. The aqueous dispersion was left under heating and stirring for ~5 hours. Sample preparation: The nanorods were purified in the same manner the 'seed' nanocrystals were purified. Several µl of the cleaned NC dispersion was drop-cast on a Formvar/Carbon 300mesh Cu TEM grids, which was held in place on a fused-silica glass. For the racemic sample, TEM grids were made hydrophilic by O2 plasma in a Diener Pico plasma oven set to 60 W for 1 minute prior to the sample deposition. The racemic NC dispersion was prepared by titrating one NC dispersion with the other NC enantiomer until the ensemble CPL of the dispersion was below noise level for these systems (glum ~ 10-5). Single particle measurements. The output of a supercontinuum laser (SC400 4W, Fianium) was filtered by an acousto-optic tunable filter (Gooch & Housego, AOTF 2885-02) to match the 488 nm excitation wavelength of the NCs. The output was then further spectrally cleaned by a 500 nm low pass filter (FESH0500, Thorlabs) and a 488 nm laser line (FF01-488/10-25, Semrock). A quarter waveplate (AQWP05M-600, Thorlabs) converts the linearly polarized light to circularly


17

ACS Nano 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 18 of 25

polarized light to avoid orientation sensitivity in the excitation of the needle-like NCs. The beam is then split by a 50:50 non polarizing beamsplitter into an excitation and reference beams. The excitation beam was then focused on the sample using a 0.85NA polarization objective (Plan NEOFLUAR 0.85 POL, Zeiss) and collected using another 0.8NA polarization objective (TU plan 0.8 POL, Nikon). The reflected light was focused onto a confocal pinhole and collected by photodiode (Two Color Sandwich Detectors, OSI Optoelectronics). In the transmitted light path a 550 nm high pass emission filter (FELH0550, Thorlabs) blocked the excitation beam and allowed the measurement of the NC's luminescence. The emitted light is then passed through a super achromatic λ/4 waveplate (APSAW-7, Astropribor) followed by a broadband wire grid linear polarizer (WP12L-UB, Thorlabs) at 45° to the fast axis of the λ/4 waveplate. This enables measurement of circular polarized emission of a specific handedness at the full emission bandwidth. The emerging linearly polarized luminescence was then focused on a 400 µm fiber and fed into a spectrograph (Shamrock 303i, Andor) equipped with an EM-CCD camera (Newton 971, Andor) cooled to -80ºC. Each NC was measured using EM gain set to max, in full vertical binning mode. The signal was accumulated for 100-300 exposures for a total of 35-105 seconds. The TEM grid containing the NC sample was placed on a low fluorescence highly pure fused silica 170 µm cover glass (Schott HPFS 7980, Mark Optics). The cover slide was cleaned by sonication in acetone for 15 minutes, followed by sonication for 15 minutes in isopropanol, and lastly heat treated in a furnace at 1000º C for 5 hours. The final step reduced the glass residual internal fluorescence by 50%. For comparisons of l-CPL and r-CPL spectra (as done in Figure 2) intensity normalization was required. This was done by scaling a single NC's l-CPL spectrum with respect to the same particles' r-CPL spectrum to compensate for the polarization dependent light collection efficiency of the two


18

Page 19 of 25 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Nano

orthogonal circular polarizations, which had also a small wavelength dependence. Scaling factors were extracted from 8 points which displayed the lowest dissymmetry in ensemble measurements at the wavelengths: 592.5, 613.2, 621.6, 653.5, 688.5, 692, 700, 707 nm. A linear fit was done on the scaling factors as a function of wavelength over the entire spectral range. Linear SVM classification. four spectral features (see Figure 4) were selected as input to Matlab's SVM machine learning algorithm. Each input feature was rescaled to the range of [0, 1] 𝐹𝑖 ― min (𝐹𝑖𝑡𝑜𝑡𝑎𝑙 )

using 𝐹𝑖𝑠𝑐𝑎𝑙𝑒𝑑 = max (𝐹𝑖

𝑡𝑜𝑡𝑎𝑙

𝑖

) ― min (𝐹𝑖𝑡𝑜𝑡𝑎𝑙 ) , where 𝐹𝑡𝑜𝑡𝑎𝑙 are all instances of the i'th feature. The training

data set was divided into ten groups so that each will be used separately to train the classifier while the rest are used to as test data, thus reducing the likelihood of over-fitting the model. The training and test data produced an error rate of ~0.8% based on our initial ansatz that the training samples were enantiopure. The single 'wrongly' classified NC spectrum was then rechecked visually, leading to the conclusion that the misclassified 'L' NC was in fact a 'D' handed particle. The exclusion of the 'wrong' particle brings the accuracy of the SVM classifier to 100%. Following the training, data of unknown NCs from the racemic mixture was fed into the SVM to predict their handedness. Dispersion (bulk) measurements. Measurements were performed in a home-built CPL system, using a photoelastic modulator and lock-in detection, built according to the general setup described in ref 15. Samples were excited by a high-power near-UV Hamamatsu LED light source (7 W, 365 nm) on the sample. Light emitted from the samples was collected at 90° to the excitation, with both excitation and emission edgepass filters (FF01-424/SP-25, Semrock and FEL0450, Thorlabs) to minimize light leakage from the excitation source, passed through the monochromator and detected by a photomultiplier.


19

ACS Nano 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 20 of 25

AUTHOR INFORMATION Corresponding Author * Corresponding author: [email protected]

AUTHOR CONTRIBUTIONS E.V. and O.C. conceived and designed the experiment. E.V. built the experimental setup, wrote the control software, conducted the experiments, analyzed the data and wrote the paper. U.H. and G.M conceived and synthesized the chiral lanthanide nanocrystals. U.H. conducted ensemble experiments and performed the TEM microscopy and wrote the paper. All authors contributed to the discussion and interpretation of the results and edited the manuscript. § These authors contributed equally to this work.

ACKNOWLEDGMENT We are grateful to the staff of the Tel-Aviv university chemistry department mechanical workshop for their tremendous help with custom mechanical parts that made our measurement system possible. This research was supported in part by The Israel Science Foundation grants no. 507/14 and 1716/13. We thank the reviewers for their insightful comments.

REFERENCES (1)

Blackmond, D. G. The Origin of Biological Homochirality, Cold Spring Harb. Perspect Biol., 2010, 2 , a002147.


20

Page 21 of 25 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Nano

(2)

Wang, L.-Y. Y.; Smith, K. W.; Dominguez-Medina, S.; Moody, N.; Olson, J. M.; Zhang, H.; Chang, W.-S. S.; Kotov, N.; Link, S. Circular Differential Scattering of Single Chiral Self-Assembled Gold Nanorod Dimers, ACS Photonics, 2015, 2 , 1602–1610.

(3)

Karst, J.; Strohfeldt, N.; Schäferling, M.; Giessen, H.; Hentschel, M. Single Plasmonic Oligomer Chiral Spectroscopy, Adv Opt. Mater., 2018, 6 , 1800087.

(4)

Banzer, P.; Woźniak, P.; Mick, U.; De Leon, I.; Boyd, R. W. Chiral Optical Response of Planar and Symmetric Nanotrimers Enabled by Heteromaterial Aelection, Nat. Commun., 2016, 7, 13117.

(5)

Narushima, T.; Okamoto, H. Circular Dichroism Microscopy Free from Commingling Linear Dichroism via Discretely Modulated Circular Polarization, Sci. Rep., 2016, 6, 35731.

(6)

Schnoering, G.; Poulikakos, L. V.; Rosales-Cabara, Y.; Canaguier-Durand, A.; Norris, D. J.; Genet, C. Three-Dimensional Enantiomeric Recognition of Optically Trapped Single Chiral Nanoparticles, Phys. Rev. Lett., 2018, 121 , 023902.

(7)

Vinegrad, E.; Vestler, D.; Ben-Moshe, A.; Barnea, A. R.; Markovich, G.; Cheshnovsky, O. Circular Dichroism of Single Particles, ACS Photonics, 2018, 5 , 2151–2159.

(8)

Hassey, R.; Swain, E. J.; Hammer, N. I.; Venkataraman, D.; Barnes, M. D. Probing the Chiroptical Response of a Single Molecule, Science, 2006, 314 , 1437–1439.

(9)

Tang, Y.; Cook, T. A.; Cohen, A. E. Limits on Fluorescence Detected Circular Dichroism of Single Helicene Molecules, J. Phys. Chem. A, 2009, 113 , 6213–6216.

(10)

Riehl, J. P.; Richardson, F. S. Circularly Polarized Luminescence Spectroscopy, Chem.


21

ACS Nano 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 22 of 25

Rev., 1986, 86, 1–16. (11)

Longhi, G.; Castiglioni, E.; Koshoubu, J.; Mazzeo, G.; Abbate, S. Circularly Polarized Luminescence: A Review of Experimental and Theoretical Aspects, Chirality, 2016, 28, 696–707.

(12)

Tohgha, U.; Deol, K. K.; Porter, A. G.; Bartko, S. G.; Choi, J. K.; Leonard, B. M.; Varga, K.; Kubelka, J.; Muller, G.; Balaz, M. Ligand Induced Circular Dichroism and Circularly Polarized Luminescence in CdSe Quantum Dots, ACS Nano, 2013, 7, 11094–11102.

(13)

Naito, M.; Iwahori, K.; Miura, A.; Yamane, M.; Yamashita, I. Circularly Polarized Luminescent CDs Quantum Dots Prepared in a Protein Nanocage, Angew. Chem. Int. Ed., 2010, 49, 7006–7009.

(14)

Zinna, F.; Di Bari, L. Lanthanide Circularly Polarized Luminescence: Bases and Applications, Chirality, 2015, 27, 1–13.

(15)

Carr, R.; Evans, N. H.; Parker, D. Lanthanide Complexes as Chiral Probes Exploiting Circularly Polarized Luminescence, Chem. Soc. Rev., 2012, 41, 7673.

(16)

Hananel, U.; Ben-Moshe, A.; Diamant, H.; Markovich, G. Spontaneous and Directed Symmetry Breaking in Chiral Nanocrystals, arXiv:1808.05406, 2018.

(17)

Luwang, M. N.; Ningthoujam, R. S.; Srivastava, S. K. Effects of Ce3+ Codoping and Annealing on Phase Transformation and Luminescence of Eu3+ Doped YPO4 Nanorods: D2O Solvent Effect, J. Phys. Chem. Soc., 2009, 56, 2979–2982.

(18)

Binnemans, K. Interpretation of europium(III) spectra, Coord. Chem. Rev., 2015, 295, 1– 45.


22

Page 23 of 25 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Nano

(19)

Bunzli, J. G.; Eliseeva, S. V. Basics of Lanthanide Photophysics; 2011.

(20)

Di, W.; Li, J.; Shirahata, N.; Sakka, Y.; Willinger, M.; Pinna, N. Photoluminescence, Cytotoxicity and in vitro Imaging of Hexagonal Terbium Phosphate Nanoparticles Doped with Europium, Nanoscale, 2011, 3, 1263–1269.

(21)

Beeby, A.; Clarkson, I. M.; Dickins, R. S.; Faulkner, S.; Parker, D.; Royle, L.; de Sousa, A. S.; Williams, J. a. G.; Woods, M. Non-radiative Deactivation of the Excited States of Europium, Terbium and Ytterbium Complexes by Proximate Energy-matched OH, NH and CH Oscillators: An Improved Luminescence Method for Establishing Solution Hydration States, J. Chem. Soc., Perkin Trans. 2, 1999, 493–504.

(22)

Gargas, D. J.; Chan, E. M.; Ostrowski, A. D.; Aloni, S.; Altoe, M. V. P.; Barnard, E. S.; Sanii, B.; Urban, J. J.; Milliron, D. J.; Cohen, B. E.; Schuck, P. J. Engineering Bright Sub10-nm Upconverting Nanocrystals for Single-molecule Imaging, Nat. Nanotech., 2014, 9, 300–305.

(23)

Gupta, S. K.; Reghukumar, C.; Kadam, R. M. Eu3+ Local Site Analysis and Emission Characteristics of Novel Nd2Zr2O7:Eu Phosphor: Insight Into the Effect of Europium Concentration on its Photoluminescence Properties, RSC Adv., 2016, 6, 53614–53624.

(24)

Kim, J.; Michelin, S.; Hilbers, M.; Martinelli, L.; Chaudan, E.; Amselem, G.; Fradet, E.; Boilot, J.-P.; Brouwer, A. M.; Baroud, C. N.; Peretti, J.; Gacoin, T. Monitoring the Orientation of Rare-earth-doped Nanorods for Flow Shear Tomography, Nat. Nanotech., 2017, 12, 914–919.

(25)

Steinwart, I.; Christmann, A. Support Vector Machines; Springer: New York, 2008.


23

ACS Nano 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

(26)

Page 24 of 25

Viedma, C.; McBride, J. M.; Kahr, B.; Cintas, P. Enantiomer-Specific Oriented Attachment: Formation of Macroscopic Homochiral Crystal Aggregates from a Racemic System, Angew. Chem. Int. Ed., 2013, 52, 10545–10548.

(27)

Kumar, J.; Nakashima, T.; Kawai, T. Circularly Polarized Luminescence in Chiral Molecules and Supramolecular Assemblies, J. Phys. Chem. Lett., 2015, 6, 3445–3452.


24

Page 25 of 25 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Nano

82x44mm (300 x 300 DPI)


Determination of Handedness in a Single Chiral Nanocrystal via

Recommend Documents