Circularly polarized and total luminescence as probes of nucleation and growth in chiral nanocrystals

Abstract

Nucleation of crystals as well as their growth is difficult to study experimentally. We have recently demonstrated that chiral Eu³⁺‐doped terbium phosphate nanocrystals are an interesting system for studying nanocrystal formation mechanisms and chiral symmetry breaking, occurring during their formation, directed by chiral ligands, such as tartaric acid. In this paper, we show how simultaneous, in situ monitoring of both total emission intensity and circularly polarized luminescence magnitude and sign versus time during nanocrystal formation provides considerable information on the mechanisms of nanocrystal nucleation and growth. Specifically, we show that the presence of tartaric acid leads to the formation of chiral prenucleation clusters, which deterministically transform into nanocrystals of a specific handedness. Additionally, we demonstrate that both unseeded and seeded nanocrystal syntheses behave differently mechanistically and that the addition of seed nanocrystals catalyses both enantio‐specific (also called secondary nucleation) as well as nonspecific nucleation.

In situ circularly polarized and total luminescence were measured during the synthesis of chiral terbium phosphate nanocrystals and revealed many details about their nucleation and growth mechanisms.

1. INTRODUCTION

Nanocrystals (NCs) are crystalline solids of which at least one dimension is on the order of nanometers (1–100 nm). In the past 30 years, colloidal NCs have been the focus of intensive academic and industrial attention, both for their useful physicochemical properties (catalysis, sensing, light emission, spin control, and optoelectronics) and structural–chemical issues connected with their formation (crystal growth and NC shape control). ¹ , ² , ³ , ⁴ , ⁵ , ⁶ , ⁷ , ⁸ This is due to their unique properties compared with bulk crystals, mainly high surface area to volume ratio and size‐dependent properties arising from a variety of physical effects. Recent studies of the formation of colloidal NCs have proved highly useful in obtaining better insight into crystal nucleation and growth in general. ⁹ , ¹⁰ , ¹¹ It is now well accepted that the process through which NCs form in solution is a complex one, which may occur through many pathways, sometimes through several competing ones simultaneously. ¹²

NC formation is commonly separated into two distinct regimes, nucleation and growth. ¹³ , ¹⁴ Nucleation and the initial growth steps are challenging to study due to the very small size of atomic clusters that form the crystal (pre)nuclei, their inherent instability, and their complex dynamics in suspensions. So far, few experimental tools have been available for monitoring crystal nucleation phenomena. ¹⁵ , ¹⁶ The recent development of liquid cell transmission electron microscopy shows potential to visualize such events, ¹⁷ under the limitation of the high‐energy electron beam, which could strongly interfere with the observed processes. ¹⁸

We have recently demonstrated that combining luminescence and nuclear magnetic resonance studies in lanthanide‐based NCs is useful for elucidating the kinetics of NC nucleation and growth. ¹⁹ The system consists of luminescent Eu³⁺‐doped TbPO₄·H₂O single crystal nanorods (chiral space group P3₁21), where the terbium ions within the NCs are excited by near‐ultraviolet light (365 nm), transfer the excitation energy to the Eu³⁺ ions, which emit light in a series of characteristic emission lines corresponding to various ⁵D₀ → ⁷F_J (J = 0–6) transitions. As the free Eu³⁺ ions in solution barely absorb light at the excitation wavelength, emission is only possible when they are positioned in close proximity to the Tb³⁺ ions as the lattice begins to form, that is, the crystals nucleate. As each individual Eu³⁺ ion incorporated into the forming NCs independently contributes to the emission intensity, the growth of the NCs can be quantitatively monitored by following the increase in Eu³⁺ luminescence intensity from the NC dispersion in real time. In addition, tuning the synthesis temperature allows control of nucleation and growth kinetics and thus bringing them to a convenient timescale of minutes, which is easy to monitor, even with weak luminescence signals. We have recently discovered that the NC formation starts with an induction period, where no Eu³⁺ luminescence occurs, followed by a swift increase in emission reflecting a rapid NC nucleation and growth phase. ¹⁹ We concluded that the observed delayed nucleation occurs due to growth of disordered polymeric structures involving both phosphate and lanthanide ions, devoid of Eu³⁺ luminescence, which undergo a phase transformation into crystal nuclei at a certain critical size (accompanied by appearance of Eu³⁺ luminescence) and further grow by particle attachment.

As mentioned earlier, the NCs are chiral, and because they are luminescent, the enantiomeric excess (ee) of an entire NC dispersion can be probed by circularly polarized luminescence (CPL) spectroscopy. ²⁰ , ²¹ , ²² The normalized magnitude of the CPL signal is called the dissymmetry factor, defined as g _lum  = 2ΔI/I, where ΔI is the difference between the left‐ and right‐handed circular polarization components in the emitted light and I is the total emission intensity. g _lum is proportional to the ee of the emitting species. By introducing an organic chiral ligand into the NC synthesis solution, the handedness of the resulting inorganic NCs in the suspension can be controlled. ²⁰ It was shown that the chiral dicarboxylic acid, tartaric acid (TA), has the strongest chiral induction effect out of several chiral organic acids examined, resulting in 100% NC ee. ²¹ , ²²

In this communication, we add a new dimension to the kinetic studies of this system: the evolution of chirality (NC ee) during the formation of the NCs. By adding TA to the NC synthesis, we drive the system out of a racemic mixture of NCs and can thus follow the ee evolution through in situ measurements of CPL during the NCs' nucleation and growth.

2. MATERIALS AND METHODS

The NCs were synthesized according to a procedure based on our previous work (Schwartz et al. ¹⁹ ) with the addition of L/D‐TA to the lanthanide solution with 1:1 TA:lanthanide molar ratio. Briefly, a solution containing an acidic (pH ~ 2) solution of terbium and europium ions (Tb:Eu molar ratio of 95:5%) in D₂O was heated to 50°C, after which a Na₂HPO₄ solution in D₂O, also heated to the same temperature, was rapidly added to the solution while stirring. It should be noted that the syntheses were performed in a D₂O environment due to the partial quenching of Eu³⁺ emission by H₂O molecules' vibrational overtones, which increases the nonradiative decay rate and decreases the luminescence intensity. ²³ In the D₂O solution, the luminescence quenching is negligible. It should be noted that the chemical formula of the NCs is then actually TbPO₄·D₂O. In situ measurements were performed in a homebuilt photoelastic modulator‐based CPL spectrometer using a temperature‐controlled aluminium block similar to the one described in Schwartz et al. ¹⁹ Both luminescence intensity and CPL magnitude of the Eu³⁺ ‐doped TbPO₄·H₂O NCs were measured as they were being formed, while stirring the solution in the measurement cell, with a strong 365‐nm light‐emitting diode excitation source.

For convenience, we term L‐NCs those prepared with L‐TA, regardless of their real, unknown handedness. Seed NCs were initially synthesized with L‐TA; hence, they were purely L‐NCs. In experiments where seeding was performed, a small quantity (20 μL of NC solution out of the total 2 mL synthesized or 1% of the total synthesis quantity of NCs) of enantio‐pure L‐NCs, purified by centrifugation, was added to a synthesis solution containing either L‐ or D‐TA.

3. RESULTS AND DISCUSSION

Figure 1 displays a comparison between the time evolution of the total luminescence and CPL signals for an NC synthesis performed at 50°C. The initial period of ~25–30 min, where no Eu³⁺ luminescence is observed, was previously attributed to formation of prenucleation clusters/polymers, which transform into crystalline nuclei at the end of this period. ¹⁹ The excitation light used in the experiments (wavelength of 365 nm) excites only the Tb³⁺ ions, which transfer the excitation energy to the Eu³⁺ ions. The emission at the measured wavelength (704 nm, one of the ⁵D₀ → ⁷F_J, J = 0–6 transitions) is solely from Eu³⁺ ions and may only occur when they are pinned at close proximity to Tb³⁺ ions. The CPL signal appears to follow the total luminescence signal, especially at early growth times. Remarkably, as seen in Figure 1C, the g _lum (which is proportional to the ee) rises sharply at the end of the induction period, much faster than its two components, I and ΔI. This indicates that when the terbium phosphate NCs are formed in the presence of a high concentration of enantiomerically pure TA, g _lum (and consequently the NC's ee) increases sharply to its peak value roughly corresponding to ee~100% and remains high from the very early stages of NC formation. In other words, all NC nucleation events are of the same handedness. We currently do not fully understand the slight decline of g _lum at longer times. As similar reaction conditions were found to yield 100% ee at the end of the synthesis, ²² we are of the opinion that it is an artifact in the measurement of the signals at later stages of the synthesis when the sample is strongly scattering the emitted light. However, due to this uncertainty, we will present subsequent data in the form of CPL (ΔI) plots only, qualitatively representing the changes in ee.

FIGURE 1.

(A) Total luminescence intensity (I), (B) circularly polarized luminescence (CPL) (ΔI) spectrum, and (C) dissymmetry factor (g _lum ) evolution with time. In a colloidal synthesis of Eu³⁺‐doped TbPO₄·D₂O NCs with L‐TA at 50°C in D₂O, measured at 704 nm (excitation wavelength = 365 nm). The 704‐nm emission line is attributed to the ⁵D₀ → ⁷F₄ transitions of the Eu³⁺ ion. The curves were normalized to their peak magnitude during the measurement period. The negative CPL is obtained due to the use of L‐TA and would be positive if D‐TA is used.

A new set of results was obtained from seeding experiments, where a small volume of fully grown L‐NCs was added to the TA‐containing precursor solution. Figures 2 and 3 show the results of these experiments, where the seeds were injected either at the beginning of the synthesis (t = 0 in Figure 2) or toward the end of the induction period (t = 26 min in Figure 3). Several observations are reported here:

1. As seen in Figure 2A, the luminescence of the sample containing L‐TA increases faster than that of D‐TA, after an initial overlap of the two curves during the first ~5 min. This indicates that about 5 min after injection of L‐NC seeds, the rate of increase in total NC volume becomes faster in the presence of L‐TA compared with D‐TA. The absence of the induction period in the seeded NC formation must imply that the seed particles catalyze the formation of the NCs, regardless of handedness of seed particles.

2. Figure 2B displays the evolution of the CPL signal upon injection of the L‐seed particles. In the simpler case of injection of L‐seed particles into a synthesis with L‐TA, we observe fast increase of the CPL to high negative values, closely following the total luminescence intensity, hence again, indicating a large L‐NC ee (~100%) sustained throughout the NC formation. This is as expected, because even without the synergetic effect of seeding the synthesis produces enantiomerically pure NCs, as previously described ²⁰ , ²² and as observed in Figure 1B. However, when the L‐seeds were injected into the D‐TA based synthesis, we observed, first, an increase toward negative CPL, which indicates the increase in L‐NCs volume. However, after about 30 min, the influence of the L‐seeds seems to be offset by the nucleation and growth of D‐NCs, turning the direction of CPL evolution toward the positive side.

3. The last interesting piece of information was obtained when the same seeding protocol was performed at the end of the induction period (see Figure 3). This time, on injecting L‐seeds into D‐TA‐based synthesis, the NCs evolve towards the positive CPL side, reaching almost complete enantiopurity (ee = 92%) of D‐NCS, dictated almost entirely by the presence of the D‐TA.

FIGURE 2.

(A) Luminescence intensity and (B) circularly polarized luminescence (CPL) versus time measured at 50°C in D₂O, following the addition of L‐NC seed particles (20 μl of seed solution) in the presence of D‐ or L‐TA. L‐seeds with L‐TA synthesis yields 100% ee of the NCs, whereas L‐seeds with D‐TA yields a lower ee (~20%). CPL curves were normalized to the value of the enantiomerically pure NC sample (Figure 1B). The inset in (A) is a magnification of the first minutes of luminescence increase, showing that the two curves overlap at the first ~5 min.

FIGURE 3.

(A) Luminescence intensity versus time of both seeded (20‐μl L‐NCs added after 26 min) and unseeded experiments, where the synthesis solution contained D‐TA. (B) Circularly polarized luminescence (CPL) versus time of the seeded experiment measured at 50°C, yielding a very high ee (92%). The CPL curve was normalized to the value obtained for enantiomerically pure sample as obtained in Figure 1B.

We will now attempt to integrate all the above observations into a coherent description of the nucleation and growth mechanisms, as we currently understand them.

3.1. Formation of prenucleation clusters

Similar to our previous work, on the same NCs without TA, ¹⁹ we believe that nucleation starts from nonluminescent, hence noncrystalline species (clusters/polymers). These form during the induction period and transform into crystallites as they grow or aggregate beyond some critical size. However, when TA is present in the synthesis, it is incorporated in the prenucleation clusters; hence, they are chiral. The chirality is propagated from TA to clusters to NCs in a deterministic manner, where prenucleation clusters evolve into crystallites of a well‐defined handedness determined by the handedness of the TA enantiomer. This conclusion is supported by the results shown in Figure 3. These indicate that at the end of the induction period, most of the precursors are in the form of D‐TA containing clusters/polymers and will nucleate exclusively as D‐crystallites, despite the presence of seed NCs of the opposite handedness.

3.2. NC growth mode

Beyond the induction period, NC growth may proceed by particle attachment (either of clusters or small crystallites, also known as “oriented attachment”) or by monomer addition (classical growth). Evidence for both types of growth is found in Figure 2A. The identical initial growth rate of L‐seeds injected to both D‐ and L‐TA (inset of Figure 2A) probably shows that the first minutes of seeded growth occur by monomer addition to the seed NCs, regardless of which TA enantiomer is present in solution. From the sigmoidal shape of the luminescence curves, we learn that only later, as TA‐containing prenucleation clusters form, the growth switches to particle attachment, which is expected to produce a different growth profile than the classical, monomer addition mode. NC growth through monomer addition only should have resulted in an increasing form of the type of exponential decay curve ( $1-e^{-t}$), which is not the case. An important observation is that the type of TA enantiomer also affects the growth rate of the seeded synthesis, at a later stage (after ~5 min): When seeds are of the same handedness as the TA, the total luminescence increases faster than when the seeds are of the opposite chirality. We interpret this result as evidence for handedness‐sensitive particle attachment, where D‐TA containing particles do not readily attach to L‐NCs, but preferably form separate D‐NCs, while L‐TA containing particles attach more readily to L‐NCs.

The largest peak growth rate (highest slope of luminescence vs. time) of the various experiments was obtained in the case of unseeded NC formation, as can be seen in Figure 3A, where the unseeded curve has a larger maximal slope relative to the seeded one. This can be explained by the coincident growth of the NCs occurring by particle attachment, as the induction period ends. When seeds are present, they grow mostly via monomer addition during the first minutes; this reduces the supersaturation level (the driving force for nucleation) and the total precursor concentration (which correlates to the growth kinetics). Hence, in the seeded growth case, the delayed bulk nucleation from prenucleation clusters occurs at lower precursor concentration, and therefore, the rates of nucleation and growth should be somewhat diminished.

3.3. Heterogeneous (seeded) nucleation

Besides spontaneous nucleation, seed particles may induce nucleation in two manners: secondary nucleation, which in the sense of chiral crystals would be chirality‐preserving nucleation on top of a seed particle, as also suggested for the chiral NaClO₃ system, ²⁴ , ²⁵ or more generally, heterogeneous nucleation, which would be a non‐chirality‐preserving, surface‐catalyzed effect. Evidence for the existence of chirality‐preserving secondary nucleation is found by comparing seeding L‐NC seeds into D‐TA containing synthesis at the beginning and end of the induction period (Figures 2 and 3). When seeds were added at the end of the induction period, their effect was minimal, inducing a few percent of L‐crystallites ee (Figure 3), which might also be simply the result of further growth of the seed particles. However, when seeding is performed at t = 0 (Figure 2), the effect of the seeds on the total ee is larger, diminishing the D‐oriented ee to a much greater degree (20% vs. 92%). This large effect cannot be solely attributed to growth of seed particles and is largely due the formation of new L‐NCs of the same handedness as the seeds and opposite of the TA in solution.

4. CONCLUSION

We have presented kinetic studies of the evolution with time of total luminescence and CPL in the seeded and unseeded preparation of chiral Eu³⁺‐doped terbium phosphate NCs, in the presence of L‐ or D‐TA. We show that many interesting mechanistic details about the nucleation, growth, and enantiomeric selection of the NCs can be obtained. It is shown that although unseeded growth probably occurs through particle attachment mode, as previously concluded on the same synthesis without TA, seeded growth involves surface catalytic effects which partially also cause secondary nucleation, that is, newly formed crystallites obtain the same handedness as the seed crystals.

The present work demonstrates that this type of NCs is an interesting model system for studying NC formation, as their luminescence allows for easy quantitative monitoring of the nucleation and growth and their chirality introduces another useful parameter for gaining further insight into these phenomena. For future work, we shall explore the origin of the chiral symmetry breaking observed for this system and attempt to gain detailed knowledge on the nature of the secondary nucleation mechanism.

Schwartz G, Hananel U, Markovich G. Circularly polarized and total luminescence as probes of nucleation and growth in chiral nanocrystals. Chirality. 2023;35(2):104‐109. doi: 10.1002/chir.23523

DATA AVAILABILITY STATEMENT

Data available from the authors upon request.