Nonsolvent induced reconfigurable bonding configurations of ligands in nanoparticle purification

Abstract

The geometry of the organic ligands on colloidal nanoparticles (NPs) is central for understanding the self-assembly behavior and many properties of NP-based soft matter. However, the current models of the NP–media interface give incomplete and often incorrect representations of its structure, especially for aqueous systems. Here, we show that NPs capped with 3-mercaptopropionic acid (MPA) serve as a convenient model for elucidation of surface ligand geometry and its effect on collective noncovalent forces of highly concentrated NPs. Contrary to expectations, MPA was found to form surface layers with different thicknesses varying from monolayer to multilayers depending on the ratio of isopropanol to the as-made NP dispersions in the purification. Solution-state nuclear magnetic resonance (NMR) and theoretical simulations uncovered that the monolayer includes strongly coordinated ligands and adsorbed ligands interacting with the NP surface non-covalently. Increased density of the non-covalently attached ligands results in a multilayered ligand shell. Consequentially, we demonstrate that the viscoelasticity of concentrated NP dispersions carrying a monolayer of MPA is higher than those of multilayered NPs. These observations enable a better understanding of the complexity of water-soluble NP dispersions and collective non-covalent forces of supramolecular NPs.

---

New concepts

Enumeration of forces between nanoparticles is a difficult but much needed task for the design of numerous soft matter and self-assembling systems. Current models developed for macroscale colloids are not applicable to nanoparticles due to them losing sight of the reconfigurability of the nanoparticle surface layer. In this work, we elucidate the surface ligand geometry and its effect on the collective noncovalent forces using a typical carboxy-thiol bifunctional ligand capped nanoparticle that serves as a convenient model. The conceptual novelty of this work includes several innovative aspects. (1) The introduction of low and high volume ratio of isopropanol to crude nanoparticle dispersions in the purification results in the different thickness of the surface ligand varying from monolayer to multilayer. (2) Nuclear magnetic resonance and theoretical simulations uncover that the monolayer includes bound ligands and strongly adsorbed ligands inserted, while the multilayer is composed of the bound ligand layer entangled with strongly and weakly adsorbed ligands. (3) The ligand can have both monolayered and multilayered configuration that represent a completely unexpected fact not accounted for in any current NP models. (4) The viscoelasticity of nanoparticle assemblies carrying a monolayer of thiol ligand is higher than those of multilayered nanoparticles. This work provides clear evidence for the ligand binding model in colloidal nanoparticles and lays the foundation of multiscale computational models adequately applying to many assembly systems and macroscale biomaterial designs.

Introduction

The spectrum of interactions between inorganic nanoparticles (NPs) coated with organic surface ligands includes van der Waals and electrostatic interactions, hydrogen bonds, coordination bridges, and other interactions.^1–3 This benefits from the nanoparticle ligands with chemical bonding and/or physical adsorption, which is different from the model of polymer adsorption mainly driven by the high conformational entropy.⁴ Their accurate enumeration and prediction of organic surface ligand-coated inorganic NPs are currently unattainable building on the obscure ligand structures, such as the organic shells on metal NPs and quantum dots.^5,6 However, current knowledge about the molecular configurations of surface ligands for the wide variety of NPs with short hydrophilic ligands is limited, let alone how the NPs self-assemble into superstructures via non-covalent forces.^2,7–9 Furthermore, derivations made from existing knowledge of microscale colloids^10,11 or other NPs lead to conclusions about their structure and the different forces between the NPs that are uncertain at best. Consequently, deciphering the connections between the molecular structure at Ångström and nanometer levels and macroscale properties represents one of the most pressing scientific problems today.

Highly concentrated NP dispersions as a convenient research model elucidate the structure of nanoscale interfaces including the complex case of aqueous media.^12,13 Here, we show that NPs capped with 3-mercaptopropionic acid (MPA) form hydrogels with dramatically different viscoelasticity when prepared using slight changes in the purification of their as-made NP dispersions. The properties of the viscoelastic gels translate the interparticle interactions from the nanoscale to macroscale implying intrinsic differences in the structure of the NP surface. We performed nuclear magnetic resonance (NMR) spectroscopy on the NP colloids and revealed different bonding configurations of the MPA ligands. Combined with simulations, it was also found that NP assembly may have a widely variable surface ligand configuration from monolayer to multilayers depending on the preparation procedure. The viscoelasticity of hydrogels assembled from monolayered NPs is higher than those of multilayered NPs. Our investigation uncovered structure–property relationships for NP assemblies and illustrated the possibility to engineer ligands with two or more functional groups to form superstructures with tailored properties originating from collective interactions in the surface ligands on NPs.^14–17

Results and discussion

Properties of highly concentrated MPA–CdTe dispersions

MPA-stabilized CdTe NPs (MPA–CdTe) were synthesized with an average diameter of ∼2.7 nm and ∼3.2 nm in aqueous solution (Fig. 1a–c and Fig. S1, S2, ESI†).^18–20 The charged NPs with an average diameter of ∼2.7 and ∼3.2  nm were synthesized by controlling the time of the reflux stage and remained stable for several days. The sizes of NPs differ from each other by one additional atomic layer of CdTe. The as-made dispersions of MPA-stabilized NPs usually contain many salts and free Cd–MPA complexes, which form dynamically balanced concentrations of each species through adsorption and desorption with NPs.²¹ These unbound complexes, formed through the non-covalent forces between NPs, keep the superficial configuration of NPs stable and thus enrich the physical and chemical properties of the NPs.²²

Fig. 1 Characterization of CdTe NPs and hydrogels. (a) HRTEM image of the 3.2 nm CdTe NPs. The size distributions of (b) 2.7 nm CdTe NPs and (c) 3.2 nm CdTe NPs, respectively. (d) The images of the as-made NP dispersions precipitated by 1 : 1 and 1 : 3 volume fraction of IPA. (e) UV-vis spectra and (f) normalized PL (λ_ex = 450 nm) of the original MPA–CdTe NP dispersions (green lines) and the MPA–CdTe NP dispersions purified by a volume ratio of IPA to NP dispersion with 1 : 1 (red lines) and 1 : 3 (blue lines). The dotted lines are the spectra of 2.7 nm CdTe NPs; the solid lines are the spectra of 3.2 nm CdTe NPs. (g) ¹H NMR spectra of Cd–MPA complexes and CdTe NP gels in deuterium oxide solution; the signals in yellow and gray background are the resonances of the bound ligands and free Cd–MPA complexes, respectively.

Here we used a strategy based on judicious control of attractive interactions between NPs. The typical method to purify the as-made NP solution is to remove excessive unbound molecules by nonsolvent (i.e., methanol, ethanol, and IPA) to vary solution polarity.^3,23–25 Considering the ligand configuration and similar size for the core and ligand, the nonsolvent dosage plays a dominant role to ligands for the surface ligand adsorption. Here, we added different volume ratios of IPA (or other nonsolvents, see Fig. S3 and S4, ESI†) to the as-made NP dispersions to partially remove free complexes and subsequently control the amount of ligand adsorption on the NP surface (i.e. v_NP : v_IPA = 1 : 1 and ν_NP : ν_IPA = 1 : 3, here v_NP is the volume of NP dispersion and v_IPA is the volume of IPA). We noticed that phase separation occurred and the two groups of IPA-enriched NP dispersion exhibited different viscosities in the fluid phase. The NP enriched phase was separated by the centrifugal speed at 5000 rpm for 5 minutes and then dried for 24 hours under vacuum. The dried NP solid was hydrated by deionized water into gel-like concentrated NP dispersions with a high-quantum yield of about ∼31%. Optical images show that there is an obvious difference between the two hydrogels: the morphology of the one formed with 1 : 1 NP : IPA is dry and brittle, and that of 1 : 3 is dense and sticky (Fig. 1d). High resolution transmission electron microscopy (HRTEM) images show structural similarities for both NP hydrogels in that they are composed of densely packed CdTe NPs. Scanning electron microscopy (SEM) images show slight differences; the morphology of the 1 : 1 sample is more like a brittle solid compared to the NP gels from the 1 : 3 sample (Fig. S1, ESI†). The SAXS measurements for both hydrogels show zero-characteristic peaks, meaning that nanoparticles in hydrogels have similarly random structures (Fig. S5, ESI†). In this paper, to differentiate the as-made MPA–CdTe dispersions that have not been purified by isopropanol (IPA), we shall refer to the concentrated MPA–CdTe dispersions as: gel-1 : x-NP size (i.e., gel-1 : 1-2.7, gel-1: 3-2.7, gel-1 : 1–3.2, and gel-1 : 3-3.2), where x is the volume ratio of the IPA.

Ultraviolet and visible (UV-vis) spectra and photoluminescence (PL) spectra provide a preliminary evaluation of the IPA–NP ratio effect in solution, for example, the relative intensity of the absorption spectra below 300 nm for gel-1 : 1 is much lower than that of gel-1 : 3 under the same NP concentrations,²⁶ showing that a higher content of free Cd–MPA complexes was retained in gel-1 : 3 (Fig. 1e and Fig. S3, ESI†). Meanwhile, the aqueous dispersion of negatively charged NPs with sizes 2.7 and 3.2 nm displayed PL peaks at 519 and 575 nm, respectively, whereas PL peaks of the corresponding hydrogels were positioned at 544 and 590 nm, respectively. The quantitative intensity of PL spectra of gel-1 : 1 is lower than that of gel-1 : 3 (Fig. 1f), meaning that the low IPA ratio is more efficient at removing Cd–MPA complexes for purification.

The effect of IPA on the ligands and morphology

The addition of nonsolvent alcohol to crude NP dispersion not only removes extra free complexes but also potentially strips adsorbed ligands coating the NP core.²³ Surface ligands enable control of NP colloidal stability and self-assembly. IPA was used as the purification nonsolvent because it prevents chemical damage of surface ligands and etching of the CdTe surface. When exposed to deionized water, MPA–NPs retain uniform dispersity. Adding IPA reduces the electrostatic repulsion of NPs in solution and removes free Cd–MPA complexes from NP dispersion by changing the polarity. Thus, the MPA–CdTe coated surface ligands and even part of the free complexes can be precipitated, and most of the extra free salts and complexes were effectively removed.²¹ Herein, we tentatively compare the number of ligands by thermogravimetric analysis (TGA). The ligand ratio of gel-1 : 1 with sizes of 2.7 nm or 3.2 nm is about 10% w/w, which is significantly lower than that of gel-1 : 3 (∼26% w/w), indicating that more complexes were retained in the gel-1 : 3 NP dispersion (Fig. S6, ESI†).

Bound and unbound surface ligands

¹H NMR spectroscopy is able to quantify and differentiate the bound and unbound ligands in the NP dispersions.^5,27,28 The bound ligands on NPs are usually broadened due to reduced tumbling rates^29,30 compared with the free Cd–MPA complexes (Fig. S7, ESI†). The methylene signals of the MPA–CdTe dispersion appear at two different chemical shifts (Fig. 1g). The broadened peaks with larger chemical shifts belong to bound ligands (designated as (MPA)S–NP) with positions at δ_H ∼ 2.5 ppm and ∼2.9 ppm (with yellow background). The resonances positioned at δ_H ∼ 2.4 ppm and ∼2.8 ppm (with gray background) belong to unbound Cd–MPA complexes which are mostly interacting with NPs. Notably, in gel-1 : 3, the integral area of Cd–MPA complexes is ∼6 times that of the bound ligands; while in gel-1 : 1, the adsorbed Cd–MPA complex is only ∼0.6 times that of the bound ligands. More Cd–MPA complexes are retained in gel-1 : 3 than in gel-1 : 1, which is in line with the UV-vis and TGA results. In addition, gel-1 : 3 contains ∼5% IPA (the residual nonsolvent) resonating at 1.5 and 4.0 ppm (Fig. S8, ESI†). And gel-1 : 1 exhibits small triplets at 2.5 and 2.8 ppm (Fig. S8, ESI†) corresponding to ∼2% dithiodipropionic acid (DTDPA), the by-product produced in the synthetic process of Cd–MPA complexes.³¹

Composition and content

The one-dimensional NMR spectra can only monitor substances whose chemical shifts are distinguishable, while ¹H–¹³C HSQC (heteronuclear single quantum coherence) experiments are used to distinguish signals with superimposed ¹H chemical shifts. The use of HSQC experiments is necessary because either the thiol or the carboxylate group of MPA can bond to metals, which, in principle, could lead to a variety of coordination structures. For free MPA in solution, the ¹³C chemical shift of the alpha-methylene carbon (position no. 2) is ∼23 ppm, and that of the beta-methylene carbon (position no. 3) is ∼42 ppm. These resonances shift slightly downfield when MPA is bound to the Cd²⁺ or CdTe NPs (Fig. S8, ESI†). We first tested the reaction precursor (Cd–MPA complexes) by HSQC (Fig. S9, ESI†). The result shows that the Cd–MPA solution contains both (MPA)COO–Cd in which Cd²⁺ binds to the carboxylate group and (MPA)S–Cd complexes in which Cd²⁺ binds to the thiol group. The (MPA)S–Cd complexes are the dominant component in the mixture because the bonding strength of the thiol group with cadmium is greater than that of a carboxylate group. The gel-1 : 1-3.2 dispersion contains three components, including 63% (MPA)S–NP (MPA binds to NP surface), 37% (MPA)S–Cd complexes, and 2% free DTDPA (Fig. 2a). The gel-1 : 3-3.2 dispersion contains 14% (MPA)S-NP, 65% (MPA)S–Cd complexes, 16% (MPA)COO–Cd complexes and 5% free IPA (Fig. 2b and Fig. S10, ESI†). Later, we will show that the (MPA)S–Cd complexes are not free in the NP gel dispersion and they are closely associated with NPs forming a multicomponent structure as a whole. In addition, we found that the thiolate alpha-methylene protons appear at δ_H ∼ 2.9 ppm, indicating that the sulfhydryl group of the bound ligands forms a terminal bond to single cadmium ions (S–Cd) in the NP hydrogels.³¹

Fig. 2 Solution state NMR analysis of CdTe NP dispersions. ¹H–¹³C HSQC spectra of (a) gel-1 : 1-3.2 and (b) gel-1 : 3-3.2. The alpha- and beta-methylene carbon atoms of sulfhydryl correspond to numbers 2 and 3, respectively. The subscript b and e represent the bound ligands and adsorbed complexes, respectively. The numbers and dotted line with the same color belong to the same molecule. ¹H–¹H NOESY spectra of (c) gel-1 : 1-3.2 and (d) gel-1 : 3-3.2. Blue and red contours denote negative and positive signal intensity, respectively. Negative NOE (and transfer NOE) effects have positive correlation peaks which have the same sign as the peaks on the diagonal. DOSY spectra of the (e) gel-1 : 1-3.2 and (f) gel-1 : 3-3.2 in deuterium oxide solution.

Using the elemental analysis results (Table S2, ESI†) and the quantification results by NMR as described in the previous paragraphs (Table S3, ESI†), we calculated the mass ratios of bound ligands with respect to the inorganic core of CdTe NPs (Table S4, ESI†) and finally obtained the average number of bound ligands and adsorbed ligands per nanoparticle as shown in Table 1.

Table 1 The number of different organic ligands per CdTe nanocrystal

NP gels	Bound ligands (MPA)S-NP	Adsorbed ligands (MPA)S–Cd
gel-1 : 1-3.2	160	95
gel-1 : 3-3.2	240	1120

---

NP–surface ligand interactions

NOESY (nuclear Overhauser effect spectroscopy) experiments can be used to understand NP–ligand and ligand–ligand interactions in the liquid state. The bound ligands and the inorganic cores of NPs move together with slowed tumbling rates like macromolecules. As a result, their ¹H signals have negative cross relaxation rate constants and generate NOE cross-peaks with the same sign as their diagonal peaks (Fig. 2c and d, blue contour). In contrast, the cross peaks of free Cd–MPA complexes have a red contour (with positive cross relaxation rate constants, Fig. S11a, ESI†).³² The interaction and exchange between bound and free ligands will also generate transfer NOE (trNOE) signals which have the same sign as their diagonal peaks. The trNOE signals appeared at frequencies of the free ligands because those NOE signals were generated in the bound state and transferred to the free state.³² Molecules that had no interaction with the NP surface generated zero trNOE cross-peaks. The strong NOE cross-peaks with the same sign (in blue contours) at 2_b2_e, 2_e2_b, 3_b3_e, and 3_e3_b in Fig. 2c and d demonstrate that the (MPA)S–Cd complexes are adsorbed and desorbed from the NP surface dynamically by exchanging between the two states. In the gel-1 : 3-3.2 spectrum (Fig. 2d), the cross peaks with the opposite sign (red contour) at 2_e3_e and 3_e2_e actually arise from the (MPA)COO–Cd complexes, which tumble freely without interacting with the NP surfaces. Therefore, (MPA)S–Cd complexes interact strongly with NPs probably due to the hydrogen bonds between the carboxylate group themselves and those of bound ligands on the NP surface, while (MPA)COO–Cd complexes behave more like free molecules because the thiol groups have no significant hydrogen bond. The gel-1 : 1-3.2 and gel-1 : 3-3.2 contain different amounts of these species and therefore lead to different macroscopic properties. Although trNOE spectroscopy is the more robust technique for detecting chemical exchange signals, they can also be verified by rotating-frame nuclear Overhauser effect correlation spectroscopy (ROESY).³³ The ROESY experiment also shows that 2_b2_e, 2_e2_b, 3_b3_e, and 3_e3_b signals originate mainly from the exchange of adsorbed and desorbed (MPA)S–Cd complexes (Fig. S12a and b, ESI†).

Diffusion speed

DOSY (diffusion ordered NMR spectroscopy) can differentiate species based on their diffusivity and reveal interactions and dispersity in solution. As shown in Fig. 2e and f, DTDPA (2.1 × 10⁻¹⁰ m² s⁻¹) and IPA (7.6 × 10⁻¹⁰ m² s⁻¹) have a much smaller diffusion coefficient than that of the nanoparticles (∼1.4 × 10⁻¹⁰ m² s⁻¹) which indicates that these species move freely and interact little with the NPs. Their small molecule-like movement can be verified by COSY (Fig. S11, ESI†). In gel-1 : 1-3.2 and gel-1 : 3-3.2, water has the same diffusion coefficient as 1.7 × 10⁻⁹ m² s⁻¹, which is consistent with the diffusion coefficient of deionized water (Fig. S13, ESI†). However, the (MPA)S–Cd complexes have the same diffusion coefficient as the bound ligands (MPA)S–NP, suggesting the formation of a larger complex. The hydrodynamic diameter, d, of the complex can be derived from the diffusion coefficient, D, by the Stokes–Einstein equation (eqn (1)).³⁴Here, k is the Boltzmann constant, T is the absolute temperature, and η is the solvent viscosity. The hydrodynamic diameter includes the core of the NP, its ligand layer and the surface associated species in solution. The NP complex in gel-1 : 1-3.2 has a hydrodynamic diameter of 3.2 ± 0.1 nm based on the diffusion coefficients of (1.5 ± 0.1) × 10⁻¹⁰ m² s⁻¹, and that in NP-1 : 3-3.2 has a hydrodynamic diameter of 3.6 ± 0.1 nm based on the diffusion coefficient of (1.3 ± 0.1) × 10⁻¹⁰ m² s⁻¹. For gel-1 : 1-3.2, the hydrodynamic diameter is almost the same as that of the core size of NPs and we expect that the bound ligands and adsorbed (MPA)S–Cd form a monolayered structure. In gel-1 : 3-3.2, most of the (MPA)S–Cd also interact strongly with the NPs in the aqueous solution and form a multilayered structure with a considerably larger hydrodynamic diameter.

Computational modeling of the binding configurations

To complement experimental NMR observations, density functional theory (DFT) simulation was performed to analyze the binding structure of MPA surface ligands on the CdTe NP surface. In the simulation, the models were constructed with one MPA anion and different CdTe NP surfaces, including CdTe(100) and CdTe(111), the two main facets of the CdTe NP (see the ESI† for the detailed description of the simulation methods). From a series of simulations, we concluded that when the MPA anion is placed on the CdTe(100) facet, the most stable geometry of optimized MPA is that each thiol–sulfur bridges with two Cd atoms (E_bind = 1.30 eV, Fig. 3a). The top binding structure (i.e. each thiol–sulfur interacting with one Cd atoms) is also a stable geometry, although the binding energy (E_bind = 0.92 eV, Fig. 3b) is smaller than the bridge binding structure. When the MPA anion is at the CdTe(111) facet, the only stable configuration is the top binding (E_bind = 0.70 eV, Fig. 3c). This result points out that there exist two binding structures of MPA on the CdTe(100) facet including the bridge and the top binding structures, and one binding structure on CdTe(111). Moreover, the stronger binding energy of MPA on CdTe(100) reflects that the (100) facet is the dominant component of the CdTe NP surface, which is in general accordance with the earlier study of CdSe NPs.³⁵

Fig. 3 Simulation results of MPA on CdTe NPs. DFT-optimized MPA geometries on the CdTe(100) surface with MPA bonded in (a) the bridge fashion or (b) the top fashion, and on CdTe(111) surfaces with MPA bonded in (c) the top fashion. Please see ESI† for the calculation method and binding energies. Schematic diagrams of (d) MPA bound to the edge of the NPs, (e) MPA coated on the facet of the NPs, and (f) the interaction of free and bound ligands. Schematic representations of a CdTe NP in (g) gel-1 : 1-3.2 with a monolayer of ligands and (h) gel-1 : 3-3.2 with multiple layers of ligands.

We also simulated the interaction between free MPA–Cd and the MPA coated on CdTe(100) or (111) facets (Fig. 3f and Fig. S14, ESI†). The results showed that when free MPA–Cd is close to the bound MPA on CdTe, hydrogen bonding exists between the carboxylate groups of each MPA molecule. The energy of hydrogen bonding ranges from 0.18 eV to 0.38 eV. These results confirm that strong interactions exist between the CdTe NPs and free MPA–Cd complexes, which agrees with our experimental results.

Furthermore, some free ligands may adsorb onto the NPs by van der Waals forces. In gel-1 : 1-3.2, the nanoparticles retain a monolayer structure, indicating that the bound ligand layer contains some free ligands (Fig. 3g). But in gel-1 : 3-3.2, the surface is a multilayered structure which includes a bound ligand layer entangled with adsorbed ligands and an additional layer of adsorbed ligands. As shown in Fig. 3h, the adsorbed ligands could be further divided into strongly-associated ligands and loosely-associated ligands (via van der Waals forces).

Collective non-covalent forces of NP dispersions

When the NPs are formed into hydrogels, their surface ligand layer affects their collective forces in terms of their viscoelastic properties. Typically, the noncovalent bond in accordance with loss modulus results revealed that the stress increases linearly with the strain amplitude applied to the gels in the initial stage; then the stress deviates from linearity at high strains (Fig. 4a–d). Moreover, the less free MPA–Cd complexes in gel-1 : 1 would always induce high shear modulus for all strains (Fig. 4e and g), and in particular, small particles easily contribute to the formation of that behaviour. When more free MPA-Cd complexes are retained in gel-1 : 3, the test result is the opposite (Fig. 4f and h). Frequency responses of gel-1 : 3-2.7 (Fig. 4j) are remarkably different from the other three gels, and the latter three gels (Fig. 4i, k and l) revealed G′ > G′′ over the entire range of frequencies, while the former showed that G′ < G′′. This result indicated that more free MPA–Cd complexes may transform the properties of the gel made by small nanoparticles from elastic to viscous behaviour. As mentioned above, the low volume fraction of IPA (ν_NP : ν_IPA = 1 : 1) with less free complexes has higher purification efficiency for the MPA–CdTe and facilitates the formation of a gel of high stiffness. However, MPA–CdTe purified by the high volume fraction of IPA (ν_NP : ν_IPA = 1 : 3) contains more free complexes, which leads to viscous behaviour.

Fig. 4 Tests of mechanical properties of CdTe NP gels. Measurements of stress/strain (a–d) and continuous step moduli/strain (e–h). Each measurement was performed at least three times from the same sample. Rheological dynamic oscillatory frequency sweep tests (i–l). The triangles correspond to the storage moduli G′, which is associated with gel elasticity (stiffness). The diamonds correspond to the loss moduli G′′, which describes energy dissipation in the gel. Shear dynamics for each NP gel was measured from low to high strain starting at 0.01–25% at a frequency of 6.28 rad s⁻¹. The rheological dynamic oscillatory frequency sweep measurements were performed with a parallel fixed plate (diameter 25 mm) at a strain value of 0.05%. The gap was 1 mm.

Similar to van der Waals forces driving the polymer coated microspheres into a highly flocculated state,¹¹ the rheology observations of CdTe NPs are essentially caused by the difference in composition and thickness of surface ligands. In gel-1 : 1-3.2, bound ligands and strongly adsorbed ligands form the monolayer ligand structure, which strengthens the dipole–dipole interaction between the NPs.³⁶ The collective properties of gel-1 : 1-3.2 are close to the rigidity of the inorganic “hard” core due to the thin ligand layer. However, in gel-1 : 3-3.2, the bound ligand and its adsorbed ligand entangled with both strongly and weakly form the multilayered ligand structure, which weakens the dipole–dipole interactions because of the large separation between NPs. When these NPs are made into hydrogels, due to the thick ligand layer, the collective property of gel-1 : 3-3.2 is close to the viscosity and softness of organic ligands.

Conclusions

This work demonstrates that the binding patterns of bifunctional ligands on the NP surface have a strong influence on NP–NP interactions, making it possible to better understand non-additivity of various nanoscale non-covalent interactions and create multiscale computational models adequately describing their macroscale behavior.³⁷ The understanding of binding models between ligands and NPs leads to inorganic–organic hydrogels with high viscoelasticity and many applications such as biosensors.³⁸ Similar relationships between surface layer structures can also be generalized to the hierarchical assembly of NPs^1,9,13 and other nanoscale components such as clay, graphene, nanocellulose and other nanoscale components using different self-assembly techniques.^39–41 This study also enables the establishment of a parallel with the design of the nanoscale components for hydrogels in biology. The knowledge about the surface chemistry of inorganic NPs makes possible not only replication of the best viscoelastic materials found in nature,^12,42,43 but actually improving upon them by exploiting the collective supramolecular bonds between surface ligands on their surface. This study also enlightens the pathway to quantitative biomimetics when the design of such materials may not only use the structural blueprint from biological analogs but also the computational algorithms for predictions of their functionalities.

Methods

Experimental methods

Chemicals. Deuterium oxide and mercaptopropionic acid (MPA) were purchased from Sigma-Aldrich. Cadmium perchlorate hexahydrate was obtained from Alfa-Aesar. Aluminum telluride powder was purchased from KYD materials. Dithiodipropionic acid (DTDPA) was obtained from Energy Chemicals. All organic solvents (methanol, ethanol, and isopropanol) were obtained from Alfa-Aesar. These chemicals were used as received.

Preparation of CdTe NPs. Briefly, 0.019 M of Cd(ClO₄)₂·6H₂O (3.84 g) and 1.96 ml MPA were dissolved in 480 ml deionized water. The pH of the mixed solution was adjusted to 11 by adding 2 M NaOH under stirring. Then the solution was transferred to a three-necked flask and bubbled with nitrogen for about 60 min. Subsequently, H₂Te generated by a mixture of 0.01 M Al₂Te₃ and 10 ml 0.5 M sulfuric acid was carefully injected into the solution. The solution turned from colorless to red. Finally, CdTe NPs with different sizes were obtained at 100 °C by controlling reaction time.

Purification of CdTe NPs. IPA and the as-made MPA–CdTe dispersions (i.e. hardcore size with 2.7 nm and 3.2 nm) were mixed in a 1 : 1 (IPA : NP) or 3 : 1 volume ratio, and centrifuged at 5000 rpm/5 min. After the supernatant was removed quickly, the viscous transparent NP drops were directly obtained. Residual IPA was removed in a vacuum desiccator for 24 h at room temperature.

Preparation of MPA–CdTe hydrogels. MPA–CdTe hydrogels were obtained by adding 22 w/w% deionized water into the dried samples. All property tests were conducted after a maturation time of 8 h. The hydrogel samples were frozen by liquid nitrogen for 15 min and subsequently transferred to a cooling condenser at −80 °C until all samples were totally dry.

NMR analysis. CdTe NPs (50 mg) and Cd–MPA complexes (25 mg) were dispersed in 500 μl deuteroxide (99% D atom) and then transferred into Wilmad-glass NMR tubes. Solution NMR experiments, including ¹H spectrum, ¹³C spectrum, ¹H–¹³C HSQC (heteronuclear single quantum coherence), 2D COSY (homonuclear correlation spectroscopy), 2D NOESY (nuclear Overhauser effect spectroscopy), 2D ROESY (rotating-frame nuclear Overhauser effect spectroscopy), and DOSY (diffusion-ordered spectroscopy) were performed on a Bruker Advance DRX 500 spectrometer with the ¹H frequency at 500.13 MHz and temperature controlled at 298 K. All the ¹H and ¹³C signals were distinguished by chemical shifts (δ) and integral areas. The NOESY measurements of CdTe NPs were performed with a mixing time of 100 ms, and that of Cd–MPA complexes was 900 ms. The DOSY experiments were performed with the ledbpgp2s pulse sequence. A gradient pulse duration of 4.0 ms and a diffusion delay of 27 ms were used. The viscosity coefficient of the water at 298 K was used to calculate the diameter of the nanoparticles.

Rheological measurements. The rheometer (TA DHR-2) is equipped with parallel fixed plates (diameter 25 mm) and was used with 1 mm gap at 25 °C. The gels were directly attached on the plate. The top plate was carefully lowered in order to protect the hydrogel structure. The bottom plate was allowed to stand until the force practically dissipated to zero. The samples were covered with silicon oil to prevent the water evaporation during the testing process. Frequency dependencies for oscillatory shear deformations of the NP hydrogels were characterized with a fixed strain of 0.05%. The viscoelasticity and shear dynamics of NP gels were measured and characterized with an oscillatory strain sweep, which was carefully performed from low to high strain with a range of 0.01–25% at a frequency of 1 Hz or 6.28 rad s⁻¹ with the NP hydrogel loaded between the plates.

Other characterization. SEM images were tested on Hitachi SU-8010 scanning electron microscope. HRTEM images were taken on an FEI-T20F transmission electron microscope at 200 kV. Ultraviolet and visible spectroscopy (UV-vis) was carried out on the PE Lambda 25 UV-vis spectrophotometer, PerkinElmer. Fluorescence (PL) spectra were collected on a Horiba fluoromax-4. Thermogravimetric analysis (TGA) was measured on a PerkinElmer TGA8000. Freeze-drying of the NPs was performed on an FD-1A-50 or FD-2. Fourier transform infrared spectroscopy (FT-IR) spectra were acquired on a Nicolet 380 spectrophotometer. Elemental analysis (EA) was performed on an Elementary Vario MICRO instrument. Zeta potential was measured on the Zetasizer Nano ZS produced by Malvern Panalytical. The Small Angle X-ray Scattering (SAXS) was performed on a XEUSS SAXS/WAXS SYSTEM. The original gels (gel-1 : 1 and gel-1 : 3) were diluted in equal proportions to fulfill the solution test, including solution-state NMR, UV-vis, and PL measurement.

Calculation of the “relative” quantum yields of CdTe NPs. The “relative” quantum yields of NPs were calculated by taking rhodamine 6G (Rho.) as a reference with a known quantum yield ∼95% with 480 nm excitation in ethanol as follows:⁴⁴where F is the integrated intensity, and A is the optical density.

Calculation of the size of CdTe NPs. The calculation of the size of CdTe NPs was based on eqn (3) in ref. 45.

ΔE_g = ΔE_g,NP − ΔE_g,0 = a₁e^−d/b₁ + a₂e^−d/b₂ (3)

where E_g,NP is the band gap of NPs obtained from the wavelength of the first excitonic absorption peak, and E_g,0 is the bulk band gap. The first excitonic absorption peaks for all hydrogel NPs were calculated by the absorption of highly diluted hydrogel solutions in order to avoid the redshift caused by NP aggregation. In the MPA–NP hydrogels, the first excitonic absorption peak is 477 nm and 535 nm. For CdTe NPs, E_g,0 = 1.61 eV, d is the diameter of the CdTe NPs and a₁ = 5.77, b₁ = 8.45, a₂ = 1.33, b₂ = 43.73.

Calculation methods

In this work, the first principle density functional theory (DFT) calculations were performed using the Vienna ab initio simulation package (VASP).^46–48 The PAW method⁴⁹ implemented by Kresse et al. was used to calculate the electronic and atomic structures of all the systems studied here. Exchange–correlation effects were treated in generalized gradient approximation (GGA) with the Perdew–Burke–Ernzerhof (PBE) potential.⁵⁰ The kinetic energy cutoff was chosen to be 400 eV. Brillouin zone integration was sampled with 3 × 3 × 1 Monkhorst–Pack mesh K-points for CdTe(100) and CdTe(111) surface system relaxation, and 5 × 5 × 5 K-points for the relaxation of the bulk CdTe. The Gaussian smearing method⁵¹ with a width of 0.05 eV was employed to determine electron occupancies. The energy and force convergence criteria are set to be 10⁻⁵ eV and 0.03 eV Å⁻¹, respectively.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

The authors acknowledge financial support from the Natural Science Foundation of China (no. 21773172, no. 21603166 and 21573162 to Y. L. Z.), WIBEZD 2014001-02 to Y. L. Z., the Natural Science Foundation of Zhejiang Province (LQ17B030003) and S20170014 to F. L. T., J. Z. and X. K. acknowledge the National Key Research and Development Program of China 2016YFA0203600, Zhejiang Provincial Natural Science Foundation R19B050003, Zhejiang University K. P. Chao's High Technology Development Foundation.

Notes and references

1. M. Boles, M. Engel and D. V. Talapin, Self-Assembly of Colloidal Nanocrystals: From Intricate Structures to Functional Materials, Chem. Rev., 2016, 116, 11220.
2. K. J. M. Bishop, C. E. Wilmer, S. Soh and B. A. Grzybowski, Nanoscale Forces and Their Uses in Self-Assembly, Small, 2009, 55, 1600–1630.
3. Z. Tang, N. A. Kotov and M. Giersig, Spontaneous Organization of Single CdTe Nanoparticles into Luminescent Nanowires, Science, 2002, 297, 237–240.
4. Y. J. Harry and G. Steve, Absorbed Polymer Multilayers in A Bad Solvent, Macromolecules, 1991, 24, 3023–3026.
5. H. Al-Johani, E. Abou-Hamad, A. Jedidi, C. M. Widdifield, J. Viger-Gravel, S. S. Sangaru, D. Gajan, D. H. Anjum, S. Ould-Chikh and M. N. Hedhili, et al., The Structure and Binding Mode of Citrate in the Stabilization of Gold Nanoparticles, Nat. Chem., 2017, 9, 890–895.
6. W. R. Algar, T. Jeen, M. Massey, W. J. Peveler and J. Asselin, Small Surface, Big Effects, and Big Challenges: Toward Understanding Enzymatic Activity at the Inorganic Nanoparticle–Substrate Interface, Langmuir, 2019, 35, 7067–7091.
7. J. C. Love, L. A. Estroff, J. K. Kriebel, R. G. Nuzzo and G. M. Whitesides, Self-Assembled Monolayers of Thiolates on Metals As A Form of Nanotechnology, Chem. Rev., 2005, 105, 1103–1169.
8. C. T. Yip, T. W. Lo, S.-C. Zhu, G. Y. Jia, H. Sun, C.-H. Lam and D. Lei, Tight-Binding Modeling of Excitonic Response In Van der Waals Stacked 2D Semiconductors Nanoscale Horiz., 2019, 4, 969–974.
9. N. A. Kotov, Nanoparticle Assemblies and Superstructures, Boca Raton, New York, London, Tokyo, CRC Press, 2005, p. 648.
10. S. Buzzaccaro, R. Rusconi and R. Piazza, “Sticky” hard spheres: Equation of state, phase diagram, and metastable gels, Phy. Rev. Lett., 2007, 99, 098301.
11. J. E. Smay, G. M. Gratson, R. F. Shepherd, J. Cesarano III and J. A. Lewis, Directed Colloidal Assembly of 3D Periodic Structures, Adv. Mater., 2002, 14, 1279–1283.
12. M. Yang, H. Chan, G. Zhao, J. H. Bahng, P. Zhang, P. Král and N. A. Kotov, Self-Assembly of Nanoparticles into Biomimetic Capsid-like Nanoshells, Nat. Chem., 2017, 9, 287–294.
13. S. Bhattacharya and S. K. Samanta, Soft-Nanocomposites of Nanoparticles and Nanocarbons with Supramolecular and Polymer Gels and Their Applications, Chem. Rev., 2016, 116, 11967–12028.
14. K. Skakuj, S. Wang, L. Qin, A. Lee, B. Zhang and C. A. Mirkin, Conjugation Chemistry-Dependent T-Cell Activation with Spherical Nucleic Acids, J. Am. Chem. Soc., 2018, 140, 1227–1230.
15. N. A. Kotov, Inorganic Nanoparticles as Protein Mimics, Science, 2010, 330, 188–189.
16. R. Klajn, K. J. M. Bishop and B. A. Grzybowski, Light-Controlled Self-Assembly of Reversible and Irreversible Nanoparticle Suprastructures, Proc. Natl. Acad. Sci. U. S. A., 2007, 104, 10305–10309.
17. F. Rechberger and M. Niederberger, Synthesis of aerogels: from molecular routes to 3-dimensional nanoparticle assembly, Nanoscale Horiz., 2017, 2, 6–30.
18. A. L. Rogach, T. Franzl, T. A. Klar, J. Feldmann, N. Gaponik, V. Lesnyak, A. Shavel, A. Eychmüller, Y. P. Rakovich and J. F. Donegan, Aqueous Synthesis of Thiol-Capped CdTe Nanocrystals:  State-of-the-Art, J. Phys. Chem. C, 2007, 111, 14628–14637.
19. A. L. Rogach, L. Katsikas, A. Kornowski, D. Su, A. Eychmüller and H. Weller, Synthesis and Characterization of Thiol-Stabilized CdTe Nanocrystals, Ber. Bunsen-Ges., 1996, 100, 1772–1778.
20. Y. Li, L. Jing, R. Qiao and M. Gao, Aqueous synthesis of CdTe nanocrystals: progresses and perspectives, Chem. Commun., 2011, 47, 9293–9311.
21. A. Shavel, N. Gaponik and A. Eychmüller, Factors Governing the Quality of Aqueous CdTe Nanocrystals:  Calculations and Experiment, J. Phys. Chem. B, 2006, 110, 19280–19284.
22. Y. Min, M. Akbulut, K. Kristiansen, Y. Golan and J. Israelachvili, The Role of Interparticle and External Forces in Nanoparticle Assembly, Nat. Mater., 2008, 7, 527–538.
23. E. Scalise, V. Srivastava, E. Janke, D. Talapin, G. Galli and S. Wippermann, Surface Chemistry and Buried Interfaces in All-Inorganic Nanocrystalline Solids, Nat. Nanotechnol., 2018, 13, 841–848.
24. Z. Chu, Y. Han, P. Král and R. Klajn, “Precipitation on Nanoparticles”: Attractive Intermolecular Interactions Stabilize Specific Ligand Ratios on the Surfaces of Nanoparticles, Angew. Chem., Int. Ed., 2018, 57, 7023–7027.
25. C. Wang, M. Fang, J. Han, H. Zhang, Y. Cui and B. Yang, Two Opposite Effects of Alcohols in the Precipitation of Aqueous Nanocrystals, J. Phys. Chem. C, 2009, 113, 19445–19451.
26. W. W. Yu, L. Qu, W. Guo and X. Peng, Experimental Determination of the Extinction Coefficient of CdTe, CdSe, and CdS Nanocrystals, Chem. Mater., 2003, 15, 2854–2860.
27. A. Ceccon, V. Tugarinov and G. M. Clore, TiO₂ Nanoparticles Catalyze Oxidation of Huntingtin Exon 1-Derived Peptides Impeding Aggregation: A Quantitative NMR Study of Binding and Kinetics, J. Am. Chem. Soc., 2018, 141, 94–97.
28. L. Piveteau, T.-C. Ong, B. J. Walder, D. N. Dirin, D. Moscheni, B. Schneider, J. Bär, L. Protesescu, N. Masciocchi and A. Guagliardi, et al., Resolving the Core and the Surface of CdSe Quantum Dots and Nanoplatelets Using Dynamic Nuclear Polarization Enhanced PASS–PIETA NMR Spectroscopy, ACS Cent. Sci., 2018, 4, 1113–1125.
29. D. K. Nagesha, X. Liang, A. A. Mamedov, G. Gainer, M. A. Eastman, M. Giersig, J.-J. Song, T. Ni and N. A. Kotov, In₂S₃ Nanocolloids with Excitonic Emission: In₂S₃ vs CdS Comparative Study of Optical and Structural Characteristics, J. Phys. Chem. B, 2001, 105, 7490–7498.
30. D. Diaz, M. Rivera, T. Ni, J.-C. Rodriguez, S.-E. Castillo-Blum, D. Nagesha, J. Robles, O.-J. Alvarez-Fregoso and N. A. Kotov, Conformation of Ethylhexanoate Stabilizer on the Surface of Cds Nanoparticles, J. Phys. Chem. B, 1999, 103, 9854–9858.
31. S. A. Majetich, A. C. Carter, J. Belot and R. D. McCullough, ^lH NMR Characterization of the CdSe Nanocrystallite Surface, J. Phys. Chem., 1994, 98, 13705–13710.
32. B. Fritzinger, I. Moreels, P. Lommens, R. Koole, Z. Hens and J. C. Martins, In Situ Observation of Rapid Ligand Exchange in Colloidal Nanocrystal Suspensions Using Transfer NOE Nuclear Magnetic Resonance Spectroscopy, J. Am. Chem. Soc., 2009, 131, 3024–3032.
33. Z. Hens and J. C. Martins, A Solution NMR Toolbox for Characterizing the Surface Chemistry of Colloidal Nanocrystals, Chem. Mater., 2013, 25, 1211–1221.
34. B. Fritzinger, R. K. Capek, K. Lambert, J. C. Martins and Z. Hens, Utilizing Self-Exchange To Address the Binding of Carboxylic Acid Ligands to CdSe Quantum Dots, J. Am. Chem. Soc., 2010, 132(29), 10195–10201.
35. J. Zhou, C. Pu, T. Jiao, X. Hou and X. Peng, A Two-Step Synthetic Strategy toward Monodisperse Colloidal CdSe and CdSe/CdS Core/Shell Nanocrystals, J. Am. Chem. Soc., 2016, 138, 6475–6483.
36. Z. Zhang, Z. Tang, N. A. Kotov and S. C. Glotzer, Simulations and Analysis of Self-Assembly of CdTe Nanoparticles into Wires and Sheets, Nano Lett., 2007, 7(6), 1670–1675.
37. C. Batista-Silvera, R. Larson and N. A. Kotov, Non-Additivity of Nanoparticle Interactions, Science, 2015, 350(6257), 1242477.
38. T. Hu, J. Xu, Y. Ye, Y. Han, X. Li, Z. Wang, D. K. Sun, Y. L. Zhou and Z. H. Ni, Visual Detection of Mixed Organophosphorous Pesticide Using QD-AChE Aerogel Based Microfluidic Arrays Sensor, Biosens. Bioelectron., 2019, 136, 112–117.
39. L.-B. Mao, H.-L. Gao, H.-B. Yao, L. Liu, H. Cölfen, G. Liu, S.-M. Chen, S.-K. Li, Y.-X. Yan, Y.-Y. Liu and S.-H. Yu, Synthetic Nacre by Predesigned Matrix-Directed Mineralization, Science, 2016, 354, 107–110.
40. Y. Zhang, F. Zhang, Z. Yan, Q. Ma, X. Li, Y. Huang and J. A. Rogers, Printing, Folding and Assembly Methods for Forming 3D Mesostructures in Advanced Materials, Nat. Rev. Mater., 2017, 2, 17019.
41. S. E. Wolf, J. Leiterer, M. Kappl, F. Emmerling and W. Tremel, Early Homogenous Amorphous Precursor Stages of Calcium Carbonate and Subsequent Crystal Growth in Levitated Droplets, J. Am. Chem. Soc., 2008, 130(37), 12342–12347.
42. Y. Zhou, P. F. Damasceno, B. S. Somashekar, M. Engel, F. Tian, J. Zhu, R. Huang, K. Johnson, C. McIntyre, K. Sun, M. Yang, P. F. Green, A. Ramamoorthy, S. C. Glotzer and N. A. Kotov, Unusual Multiscale Mechanics of Biomimetic Nanoparticle Hydrogels, Nature Commun., 2018, 9, 181.
43. Z. Nie, A. Petukhova and E. Kumacheva, Properties and Emerging Applications of Self-Assembled Structures Made from Inorganic Nanoparticles, Nat. Nanotechnol., 2010, 5, 15–25.
44. Z. Yuan, A. Zhang, Y. Cao, J. Yang, Y. Zhu and P. Yang, Effect of Mercaptocarboxylic Acids on Luminescent Properties of CdTe Quantum Dots, J. Fluoresc., 2012, 22, 121–127.
45. S. Sapra and D. D. Sarma, Evolution of the Electronic Structure with Size in II–VI Semiconductor Nanocrystals, Phys. Rev. B: Condens. Matter Mater. Phys., 2004, 69, 125304.
46. G. Kresse and J. Hafner, Ab initio Molecular Dynamics for Liquid Metals, Phys. Rev. B: Condens. Matter Mater. Phys., 1993, 47, 558.
47. G. Kresse and J. Furthmüller, Efficiency of ab-initio Total Energy Calculations for Metals and Semiconductors Using a Plane-Wave Basis Set, Comput. Mater. Sci., 1996, 6, 15.
48. G. Kresse and J. Hafner, Ab initio Molecular-Dynamics Simulation of The Liquid-Metal–Amorphous-Semiconductor Transition in Germanium, Phys. Rev. B: Condens. Matter Mater. Phys., 1994, 49, 14251.
49. G. Kresse and D. Joubert, From Ultrasoft Pseudopotentials to the Projector Augmented-Wave Method, Phys. Rev. B: Condens. Matter Mater. Phys., 1999, 59, 1758.
50. J. P. Perdew, K. Burke and M. Ernzerhof, Generalized Gradient Approximation Made Simple, Phys. Rev. Lett., 1996, 77, 3865.
51. M. Methfessel and A. T. Paxton, High-Precision Sampling for Brillouin-Zone Integration in Metals, Phys. Rev. B: Condens. Matter Mater. Phys., 1989, 40, 3616.

---

Footnotes

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c9nh00397e

‡ These authors contributed equally to this work.

---