Accurate Measurement of the Molecular Thickness of Thin Organic

Bauer , C. A.; Stellacci , F.; Perry , J. W. Relationship between structure and solubility of thiol-protected silver nanoparticles and assemblies Top...
0 downloads 0 Views 4MB Size
Download PDF
Article pubs.acs.org/Langmuir

Accurate Measurement of the Molecular Thickness of Thin Organic Shells on Small Inorganic Cores Using Dynamic Light Scattering Matthew P. Shortell,* Joseph F. S. Fernando, Esa A. Jaatinen, and Eric R. Waclawik School of Chemistry, Physics and Mechanical Engineering, Science and Engineering Faculty, Queensland University of Technology, GPO Box 2434, Brisbane, Queensland 4001, Australia S Supporting Information *

ABSTRACT: Dynamic light scattering (DLS) has become a primary nanoparticle characterization technique with applications from material characterization to biological and environmental detection. With the expansion in DLS use from homogeneous spheres to more complicated nanostructures comes a decrease in accuracy. Much research has been performed to develop different diffusion models that account for the vastly different structures, but little attention has been given to the effect on the light scattering properties in relation to DLS. In this work, small (core size < 5 nm) core−shell nanoparticles were used as a case study to measure the capping thickness of a layer of dodecanethiol (DDT) on Au and ZnO nanoparticles by DLS. We find that the DDT shell has very little effect on the scattering properties of the inorganic core and, hence, can be ignored to a first approximation. However, this results in conventional DLS analysis overestimating the hydrodynamic size in the volume- and number-weighted distributions. With the introduction of a simple correction formula that more accurately yields hydrodynamic size distributions, a more precise determination of the molecular shell thickness is obtained. With this correction, the measured thickness of the DDT shell was found to be 7.3 ± 0.3 Å, much less than the extended chain length of 16 Å. This organic layer thickness suggests that, on small nanoparticles, the DDT monolayer adopts a compact disordered structure rather than an open ordered structure on both ZnO and Au nanoparticle surfaces. These observations are in agreement with published molecular dynamics results.

■

INTRODUCTION Small core−shell inorganic−organic nanoparticles are of importance in most areas of nanotechnology and fundamental studies in nanoscience.1 Their significance is even observed in everyday life, where they have been used to detect toxic metal ions to the parts per billion, improving drinking water safety,2 and in sunscreens to block harmful ultraviolet radiation.3 Metallic nanoparticles (MNPs) supporting localized surface plasmons (LSPs) and small semiconductor nanoparticles (quantum dots or QDs) are of particular interest because of their size-dependent properties. In the case of MNPs, size and shape determine plasmon resonance energy, while for QDs, size determines the bandgap through quantum confinement.4 The organic shell has several functions: it not only protects the core from particle growth and oxidation but can also be functionalized for selective attachment (for example, to cancer cells in photodynamic therapy).5 Measuring the size of the core and shell accurately and easily is therefore essential, because of the direct effect of the size on the optical, electrical, chemical, and transport properties of these nanoparticle types. Self-assembled nanoparticle films are also particularly interesting, because the optical6−8 and electronic9 properties of these films are highly dependent upon the distance between nanoparticles. Since the pioneering work by Heath et al.,10 many have inferred that the distance between core−shell nanoparticles in monolayer films is affected by interdigitation of © 2014 American Chemical Society

the capping molecules (see Figure 1a) because the distance between the core surfaces is much less than twice the extended

Figure 1. Different representations of core−shell nanoparticle films are shown. (a) Shells with an extended structure requiring interdigitation and (b) shells with a compact structure without interdigitation are shown.

length of the capping molecules.11−17 However, some simulations have shown that no interdigitation should occur;18,19 instead, the thickness of the capping layer is much smaller than the extended length of the capping molecule, accounting for the reduced distance between the cores (see Received: September 3, 2013 Revised: November 3, 2013 Published: January 3, 2014 470

dx.doi.org/10.1021/la403391t | Langmuir 2014, 30, 470−476

Langmuir

Article

We demonstrate that the scattering cross-section of a small core−shell inorganic−organic nanoparticle is best described by the core radius rather than their hydrodynamic radius, using a core−shell model and a core multi-shell model to incorporate solvent penetration. This simplification, without the need to include any material refractive indices, permits easy modification of the output size distributions from DLS software if either the core size or organic shell thickness is known. We use these results to demonstrate experimentally, quantitative measurement of the effective hydrodynamic thickness of a capping layer of dodecanethiol (DDT) on Au and ZnO core nanoparticles in hexane and chloroform, respectively. The hydrodynamic thickness obtained is in close agreement with published molecular dynamics results.

Figure 1b). An accurate measurement of the molecular thickness of thin organic shells on small inorganic cores is therefore very important and should be measured accurately to understand the final film structure. There are several techniques that can be used to measure the thickness of organic shells. The simplest techniques, such as Taylor dispersion and sedimentation techniques,20 are useful but are overly simplistic for more complicated colloids. Smallangle scattering21 (SAS) and two-dimensional (2D) diffusionordered nuclear magnetic resonance (NMR) spectroscopy22 (DOSY) are some of the more popular high-resolution techniques. SAS uses a collimated beam of X-rays or neutrons and measures the deflected intensity as a function of the angle. It is useful because it can acquire both the core and organic shell size, but the instrumentation required is relatively sophisticated and, therefore, costly and not readily available. Two-dimensional DOSY uses magnetic gradients to extrapolate the diffusion coefficient of the organic content in a deuterated solvent. Because the shell is organic, it effectively gives a surface-area-weighted size distribution. This technique is growing in popularity but requires strict cleanliness (any byproducts, excess ligands, or residual non-deuterated solvent may overpower the nanoparticle signal) and is currently not available as benchtop equipment. Dynamic light scattering (DLS) is a fast and inexpensive technique that can be applied to any nanoparticles that are undergoing Brownian motion.23 The technique measures the diffusion coefficient of nanoparticles and relates it back to their hydrodynamic radius. Currently, this is the only well-developed benchtop technique that can probe the physical thickness of the capping layer of an inorganic−organic core−shell nanoparticle by comparing to a known core size. In particular, DLS has been widely used to measure the thickness of adsorbed layers of polymers on large nanoparticles.24,25 The modernization of DLS has created many application opportunities. It has been used to detect many different organic molecules using gold nanoparticles as probes and measuring the change in hydrodynamic size. For example, viruses,26 trinitrotoluene (TNT),27 proteins,28,29 and metal ions30−32 have all been detected using DLS methods. The particle size distributions from DLS are intensityweighted because the acquired signal is the light scattered from the nanoparticles. For meaningful comparisons to other size measurement techniques, this intensity-weighted distribution, PI(R), must be converted into number, PN(R), or volume, PV(R), weighted distributions, using eqs 1 and 2, respectively PN(R ) =

PI(R ) A I (R )

PV(R ) = R3PN(R )B

■

THEORY Modeling the Scattering of Small Core−Shell Inorganic−Organic Nanoparticles. The electrostatic approximation was used to theoretically model the interaction between the core−shell nanoparticle and electromagnetic field because the particle size was much smaller than the 633 nm laser wavelength used for DLS. Four different nanoparticle radial models were considered, resulting in four distinct dielectric functions. In the first, we consider the nanoparticle as homogeneous, with a refractive index of the inorganic crystal and a radius given by the core radius (core model). Assuming the same composition but with a radius equal to the hydrodynamic radius yields the hydrodynamic model used by DLS software, which can only directly measure the hydrodynamic size. We also use a more sophisticated model to include the impact of the organic shell (core−shell model). Finally, to allow for the solvent penetration into the shell to be modeled, we use a multiple shell model (multi-shell). A schematic depiction of these models for the ZnO−DDT core− shell nanoparticle in chloroform with a core radius of r = 1 nm and capping thickness of L = R − r = 1 nm is given in Figure 2.

(1)

Figure 2. Four different refractive index models used in scattering derivations are shown: (a) core model, (b) hydrodynamic model, (c) core−shell model, and (d) multi-shell model. Orange, blue, and aqua represent the core, DDT, and chloroform refractive indices, respectively.

(2)

where I(R) is the scattering intensity cross-section of a single particle of hydrodynamic radius R, at the laser wavelength λ, and A and B normalize the distributions. Most modern DLS instruments automatically evaluate these distributions using the Mie theory by assuming that the particles are homogeneous spheres. While DLS has been proven effective under these conditions, less is known about the accuracy of the technique when particles are not spherical or not homogeneous. In this paper, we show that small (R ≪ λ) core−shell inorganic− organic nanoparticles must be treated differently because of the relatively low refractive index contrast between the organic shell and the solvent.

The optical properties of small nanoparticles can be found from the polarizability (α) of the nanoparticle in the solvent. The scattering cross-section is given by33 Csca =

k4 2 |α | 6π

(3)

where k is the wavevector of the laser light (k = 2π/λ). The polarizabilities of the nanoparticles (NPs) for the core, hydrodynamic, and core−shell models are given by33 471

dx.doi.org/10.1021/la403391t | Langmuir 2014, 30, 470−476

Langmuir

Article

αcore = 4πr 3

εc − εsol εc + 2εsol

αhydrodynamic = 4πR3

εc − εsol εc + 2εsol

radius is more accurate than measuring the minor or major axis or the mean of the two (see the Supporting Information). It also makes measurement more repeatable because the measurer only has to determine the edge of the nanoparticle against the carbon grid and not the locations of the semi-minor and major axes. The TEM data were placed into 10 equally sized bins in MATLAB, and the mode was found. DLS Measurements. DLS measurements were performed on a Malvern Zetasizer Nano ZS in a 173° backscatter configuration. For ZnO nanoparticles, high concentrations in the range of 1−5 mg/mL were used because the low refractive index of ZnO provided a weak or low signal. Measurements were made over a range of concentrations to ensure that concentration effects were negligible.23 Because the refractive index contrast between Au and the solvent is large, the signal from the Au nanoparticles was much greater than that obtained from ZnO particles. As a result, much lower concentrations of Au were used (Au < 1 mM), making it unnecessary to undertake concentrationdependent measurements for those particles.

(4)

(5)

αcore−shell = 4πR3((R3(εsh − εsol)(εc + 2εsh) + r 3(εc − εsh)(εsol + 2εsh)) /(R3(εsh + 2εsol)(εc + 2εsol) + r 3(2εsh − 2εsol) (εc − εsh)))

(6)

where εc, εsh, and εsol are the dielectric constants of the core, shell molecule, and solvent, respectively. The multi-shell model assumes that the dielectric constant in each shell is a volumeweighted average of the solvent molecule and the capping molecule. This is calculated by assuming the capping molecules are cylindrical and have the maximum packing density (packing fraction = 0.91) on the surface of the cores. Because of the large number of equations involved, this must be solved numerically. For more details, see the Supporting Information.

■

■

RESULTS AND DISCUSSION Scattering Models. The results for the scattering models are plotted in Figure 3. The scattering cross-section of the

EXPERIMENTAL SECTION

Synthesis. ZnO QDs were synthesized following a modified procedure of Spanhel and Anderson.34 First, 4 mmol of zinc acetate dihydrate [Zn(CH3CO2)2·2H2O, Aldrich, analytical reagent (AR) grade] was dissolved in 100 mL of boiling ethanol (Aldrich, AR grade) and cooled to 0 °C in an ice bath. Then, 6.4 mmol of NaOH (Aldrich, AR grade) in 100 mL of ethanol solution was added quickly to the zinc acetate [Zn(OAc)2] solution under vigorous stirring, and ZnO QDs were formed. The solution was then removed from the ice bath and allowed to return to room temperature under mild stirring for a period of growth time. When the QDs had grown to the required size, capping of ZnO QDs by DDT (Aldrich) was performed by adding 2 mmol of DDT to the ZnO QD solution under vigorous stirring. Because of the surface modification of the QDs, they were no longer stable in ethanol and precipitated. The supernatant was removed, and the precipitate was washed 3 times in ethanol. The nanoparticle precipitates were then removed and air-dried on filter paper. Solutions in chloroform were prepared by dispersing nanoparticle powder in chloroform with the aid of ultrasonic agitation and filtered using a 220 nm syringe filter. Au MNPs capped with DDT dispersed in hexane were prepared using a phase-transfer technique by Martin et al.35 using AR-grade chemicals. ZnO Core Measurement. The core size of ZnO QDs was measured by optical absorption spectroscopy using the method derived by Pesika et al.36 Briefly, the absorption spectrum of a single QD is modeled as a step function, with the step location equal to the band gap of the QD. In an ensemble of QDs, the step function is spread out and the inflection point gives the volume-weighted mode of the QD distribution. This is then converted into a particle radius using the empirical formula derived by Viswanatha et al.37 Although this model does not represent the true discrete nature of QD electronic transitions, it does give good agreement with experimental results, as long as we restrict our samples to QDs with band gap shifts over 0.2 eV.37 Au Core Measurement. Au core size distributions were measured using transmission electron microscopy (TEM). Immediately following DLS measurements, a few drops of Au−DDT colloid were drop-cast onto a TEM grid on filter paper and allowed to air dry. TEM measurements were performed on a JEOL 1400 transmission electron microscope at an accelerating voltage of 80 kV. The cross-sectional area of at least 250 nanoparticles per sample was measured using ImageJ and converted into a circle equivalent radius. The area was chosen because this was predicted to give the closest correlation to DLS results, namely, the diffusion coefficient. If we consider that any nanoparticle could be approximated as a spheroid, the area-equivalent

Figure 3. Scattering cross-sections of (a) Au−DDT and (b) ZnO− DDT as a function of the core radius. Dot, dashed, and solid lines are hydrodynamic, core−shell, and multi-shell models, respectively. Gray and black lines refer to a capping thickness of 8 and 16 Å, respectively. The scattering cross-section is normalized by the core model.

hydrodynamic, core−shell, and multi-shell models, normalized by the core model, for both Au nanoparticles (εAu = −11.6 + 1.2i) in hexane (εhexane = 1.37272) and ZnO nanoparticles (εZnO = 2.00412) in chloroform (εchloroform = 1.492) are shown. Two different capping thicknesses of DDT (εDDT = 1.4592) are shown for each nanoparticle type. The results for the core−shell and multi-shell models demonstrate that the capping layer has very little effect on the scattering intensity. Furthermore, the hydrodynamic model will greatly overestimate the effect of the shell on the scattering intensity. An interesting result of the shell models is that a resonance effect can occur, whereby at a certain size, the polarizability goes to 0 and the nanoparticles become “invisible”.38 For the core−shell model, this occurs when the numerator in eq 6 reduces to 0. Under experimental conditions, when the polarizability approaches 0, problems arise when a 472

dx.doi.org/10.1021/la403391t | Langmuir 2014, 30, 470−476

Langmuir

Article

the distribution weighting required. Samples with PDIs less than 5% are usually considered monodisperse and are highly desired. Volume-weighted distributions are also preferential because they are more accurately obtained than numberweighted distributions in DLS. When both of these characteristics are fulfilled, the error in using DLS software directly (i.e., hydrodynamic model) to find the hydrodynamic radius should be minimal. Only in this ideal case could the molecular thickness be calculated directly as the difference between the DLS software result for the hydrodynamic radius and the corresponding core radius measurement. In the other extreme (when a number-weighted distribution is required and the sample is polydisperse), the error in experimental calculation of the molecular thickness would be very large (>100% overestimation). Both the ZnO and Au samples used in this research fall between these two limiting cases. Volumeweighted distributions are required for the ZnO samples, but these samples are polydisperse. Conversely, the Au samples are considered monodisperse, but number-weighted distributions are required. As a result, the core model should be used to determine particle size distributions and/or capping thicknesses for both the ZnO and Au samples. Core Size Results. The absorption spectra for the ZnO− DDT QDs investigated are given in Figure 5. The horizontal

shell model is used for converting DLS intensity-weighted distributions to volume- or number-weighted distributions, because essentially the calculation approaches a singularity. Therefore, even a small error in the measured DLS intensityweighted distribution near the resonance could dominate the calculated volume- or number-weighted distribution. Furthermore, material constants are required, and a shell structure (homogeneous shell or multiple shells) would need to be assumed. Because the shell has little effect on the scattering, we suggest that the core model would be simpler and just as effective. This simplification has the benefit that no material refractive index constants need to be determined. To demonstrate this, we calculate the theoretical error in measuring the thickness of the shell using DLS, assuming that the multi-shell model best represents the actual scattering intensity of the nanoparticles. We relate this investigation directly to our nanoparticles of interest, Au−DDT core−shell and ZnO−DDT with polydispersity indices (PDIs)39 of 0.05 and 0.25, respectively. We assume that the actual size distribution is a log-normal distribution in number- and volume-weighted distributions for Au and ZnO, respectively. The choice of different distributions comes from our experimental methods of determining the core size of the nanoparticles. Because we are most concerned with measuring the shell thickness with DLS, we compute the relative error (ΔL/L) as a function of the core radius for capping thicknesses of 8 and 16 Å. The hydrodynamic model yields the capping thickness that one would obtain if the core radius (from TEM or absorption spectroscopy) was subtracted from the hydrodynamic radius, as determined directly from DLS software as a usual practice. The results of the analysis are shown in Figure 4. The results obtained with the core model are significantly better, with

Figure 5. Absorption spectra for the ZnO−DDT samples prepared. Growth time and size increase from dark to light lines.

axis is the change in bandgap from bulk ZnO (3.32 eV). The results are consistent with other publications,40,41 with a blueshifted absorption edge (from bulk ZnO) because of quantum confinement and an additional edge because of higher order electronic transitions. The data are smoothed, and the inflection point is found by the crossover point of the second derivative. The inflection point gives the mode bandgap shift for the volume-weighted distributions. This is then used with the formula by Viswanatha et al.37 to find the core radius for comparison to DLS results. Au TEM measurements are shown in Figure 6. The particles are roughly spherical and agglomerate around the edges of the carbon grid holes. Samples were relatively monodisperse (PDI ∼ 0.05) and appeared more log-normally distributed than Gaussian. Some samples still contained a large amount of organic material and were difficult to image; therefore, these were subsequently discarded. DLS Results. Implementing the core model for DLS results is relatively straightforward. The volume- or number-weighted distributions from the DLS software (hydrodynamic model) require the simple correction

Figure 4. Error in calculating the organic shell thickness for (a) Au− DDT and (b) ZnO−DDT as a function of the core radius. Dashed and solid lines are hydrodynamic and core models, respectively. Gray and black lines refer to a capping thickness of 8 and 16 Å, respectively.

errors in L that are relatively independent of the core size and typically less than 2.5 and 10% for Au and ZnO, respectively. This is not overly surprising because the standard hydrodynamic model will always overestimate the shell thickness and depends upon the core size for a fixed PDI. Typically, there are two main characteristics that determine the quality of a DLS measurement: the PDI of the sample and

PN,corrected(r ) = 473

R6 PN,software(r )A r6

(7)

dx.doi.org/10.1021/la403391t | Langmuir 2014, 30, 470−476

Langmuir

Article

Viswanatha et al. to convert the change in bandgap to the ZnO core radius. The data points are the DLS results for the mode core radius of the volume-weighted distribution assuming various capping thicknesses (from 0 to 8 Å from top to bottom). The results show convincingly that the capping thickness of DDT on ZnO QDs is approximately 7 Å and independent of the core nanoparticle size (an inherent assumption in this work). The results for Au−DDT nanoparticles are similar to ZnO− DDT with an average shell thickness of 7.6 Å. These results, along with the ZnO−DDT results, are summarized in Figure 8

Figure 8. Measured capping thickness as a function of the measured core radius. Filled and open data points are calculated using the hydrodynamic and core models, respectively. Squares and circles refer to ZnO−DDT and Au−DDT samples, respectively.

and show the importance of using the core scattering model. Using the hydrodynamic model gives a larger capping thickness, much closer to the extended chain length of DDT (∼16 Å). It also suggests that the capping thickness decreases with increasing size. This is likely due to the decrease in PDI with increasing size; when the PDI decreases, the error also decreases, resulting in a smaller measured capping thickness. Using the more accurate core model gives a capping thickness less than half of the extended chain length and independent of the core radius for the range that we have considered. One possible reason for the small measured thickness of the DDT capping layer could be the low packing fraction of thiols on the core surface, leading to a tilted or stacked-like structure of the thiols. This has been observed for 2D self-assembled monolayers on small grain Au films, where the final monolayer had a thickness of about 7 Å.42 To investigate this, an excess of DDT was added to the ZnO−DDT colloids to increase the packing fraction of DDT on the core surface. However, this led to digestive ripening of the nanoparticles observed in both DLS and absorption spectroscopy. We also note that there is a significant excess of DDT in the synthesis of the Au−DDT colloid. This leads us to believe that the small capping thickness observed was not due to a low packing fraction. A more likely reason for the small thickness is the large excess cone volume available to the DDT molecules, resulting in significant bending of the C12 chains back toward the core surface. Molecular dynamics simulations by Tay and Bresme18,19 on small Au140 dodecanethiol-passivated nanoparticles supports this conclusion. They found that, despite having over 70% trans dihedral bonds, the dodecanethiol molecules were not found in extended conformations; instead, the shell thickness was ∼7 Å. They also observed that the core structure had very little effect on the overall surfactant layer structure, which is very similar to results found here for both ZnO and Au nanoparticles.

Figure 6. Representative TEM images and histograms for prepared Au−DDT samples. Histograms have 10 equally spaced bins, and the scale bars are 100 nm.

PV,corrected(r ) =

R3 PV,software(r )B r3

(8)

where A and B are again normalization constants and r = R − L. Here, L is the unknown thickness that we are seeking. In practice, this is found by recalculating the measured distributions for a series of proposed L values and matching the results to the core radius found by other methods (TEM or absorption spectroscopy in this work). The experimental DLS results for ZnO−DDT QDs are shown in Figure 7. The solid black line shows the equation by

Figure 7. DLS results for ZnO−DDT samples. The circles are the DLS core size assuming different capping thicknesses from 0 to 8 Å from top to bottom. The solid line is the formula by Viswanatha et al. 474

dx.doi.org/10.1021/la403391t | Langmuir 2014, 30, 470−476

Langmuir

■

Article

nanoparticles for quantitative imaging of transdermal transport. Biomed. Opt. Express 2011, 2, 3321−3333. (4) Eustis, S.; El-Sayed, M. A. Why gold nanoparticles are more precious than pretty gold: Noble metal surface plasmon resonance and its enhancement of the radiative and nonradiative. Chem. Soc. Rev. 2006, 35, 209−217. (5) Ghosh Chaudhuri, R.; Paria, S. Core/shell nanoparticles: Classes, properties, synthesis mechanisms, characterization, and applications. Chem. Rev. 2012, 112, 2373−2433. (6) Ozel, T.; Hernandez Martinez, P. L.; Mutlugun, E.; Akin, O.; Nizamoglu, S.; Ozel, I. O.; Zhang, Q.; Xiong, Q.; Demir, H. V. Observation of selective plasmon-exciton coupling in nonradiative energy transfer: Donor-selective vs. acceptor-selective plexcitons. Nano Lett. 2013, 13, 3065−3072. (7) Komarala, V. K.; Bradley, A. L.; Rakovich, Y. P.; Byrne, S. J.; Gun’ko, Y. K.; Rogach, A. L. Surface plasmon enhanced Förster resonance energy transfer between the CdTe quantum dots. Appl. Phys. Lett. 2008, 93, 123102. (8) Akin, O.; Ozel, T.; Demir, H. V. Cascading plasmonic and nonradiative energy transfer interactions by plasmon-coupling only donor or only acceptor quantum dots of the energy transfer pairs. Proceedings of the 2010 IEEE Photinic Society’s 23rd Annual Meeting; Denver, CO, Nov 7−11, 2010; Vol. 1, pp 582−583. (9) Markovich, G.; Collier, C.; Heath, J. Reversible metal−insulator transition in ordered metal nanocrystal monolayers observed by impedance spectroscopy. Phys. Rev. Lett. 1998, 80, 3807−3810. (10) Heath, J. R.; Knobler, C. M.; Leff, D. V Pressure/temperature phase diagrams and superlattices of organically functionalized metal nanocrystal monolayers: The influence of particle size, size distribution, and surface passivant. J. Phys. Chem. B 1997, 101, 189− 197. (11) Wang, Z. . Self-assembled superlattices of size- and shapeselected nanocrystals. Mater. Charact. 1999, 42, 101−109. (12) Chen, S. Langmuir monolayers of gold nanoparticles: From ohmic to rectifying charge transfer. Anal. Chim. Acta 2003, 496, 29− 37. (13) Bauer, C. A.; Stellacci, F.; Perry, J. W. Relationship between structure and solubility of thiol-protected silver nanoparticles and assemblies. Top. Catal. 2008, 47, 32−41. (14) Motte, L.; Pileni, M. P. Influence of length of alkyl chains used to passivate silver sulfide nanoparticles on two- and three-dimensional self-organization. J. Phys. Chem. B 1998, 102, 4104−4109. (15) Pileni, M. P. Nanocrystal self-assemblies: Fabrication and collective properties. J. Phys. Chem. B 2001, 105, 3358−3371. (16) Sastry, M.; Gole, A.; Patil, V. Lamellar Langmuir−Blodgett films of hydrophobized colloidal gold nanoparticles by organization at the air−water interface. Thin Solid Films 2001, 384, 125−131. (17) Kulkarni, G. U.; Thomas, P. J.; Rao, C. N. R. Mesoscale organization of metal nanocrystals. Pure Appl. Chem. 2002, 74, 1581− 1591. (18) Tay, K.; Bresme, F. Computer simulations of two dimensional gold nanoparticle arrays: The influence of core geometry. Mol. Simul. 2005, 31, 515−526. (19) Tay, K. A.; Bresme, F. Wetting properties of passivated metal nanocrystals at liquid−vapor interfaces: A computer simulation study. J. Am. Chem. Soc. 2006, 128, 14166−14175. (20) Wuelfing, W. P.; Templeton, A. C.; Hicks, J. F.; Murray, R. W. Taylor dispersion measurements of monolayer protected clusters: A physicochemical determination of nanoparticle size. Anal. Chem. 1999, 71, 4069−4074. (21) Korgel, B. A.; Fullam, S.; Connolly, S.; Fitzmaurice, D. Assembly and self-organization of silver nanocrystal superlattices: Ordered “soft spheres”. J. Phys. Chem. B 1998, 102, 8379−8388. (22) Canzi, G.; Mrse, A. A.; Kubiak, C. P. Diffusion-ordered NMR spectroscopy as a reliable alternative to TEM for determining the size of gold nanoparticles in organic solutions. J. Phys. Chem. C 2011, 115, 7972−7978. (23) Pecora, R. Dynamic light scattering measurement of nanometer particles in liquids. J. Nanopart. Res. 2000, 2, 123−131.

CONCLUSION We have demonstrated that DLS is capable of accurately measuring the capping thickness of small inorganic−organic core−shell nanoparticles if the appropriate scattering crosssection is used and the core size is known. Here, we present a simple correction formula to be used for extracting corrected particle size distributions from standard DLS software. This technique can be used to find the capping thickness for different ligands, or if capping thickness is known, this technique can be used to find the correct particle size distributions using DLS alone. This method is also applicable to bare nanoparticles in polar solvents, where there will be an effective capping layer of solvent molecules. For the size range studied, we found that that the effective capping thickness of DDT on Au and ZnO cores is 7.3 ± 0.3 Å, much less than the extended chain length of 16 Å. This is likely a result of the large excess cone volume available, resulting in a compact disordered shell structure. Future work will focus on determining if larger nanocrystal cores behave in the same way because, in the large core radius limit, it is expected that the monolayer structure should resemble that of 2D self-assembled monolayers with a highly ordered geometry.

■

ASSOCIATED CONTENT

S Supporting Information *

Theoretical consideration of the diffusion coefficient of ellipsoids for measuring the TEM size and derivation of the scattering cross-section of core−shell nanoparticles. This material is available free of charge via the Internet at http:// pubs.acs.org.

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This material is based on research sponsored by the Air Force Research Laboratory, under agreement number FA2386-13114016. The views and conclusions contained herein are those of the authors and should not be interpreted as necessarily representing the official policies or endorsements, either expressed or implied, of the Air Force Research Laboratory or the U.S. Government. TEM and DLS were carried out at the Central Analytical Research Facility (CARF), Institute for Future Environments (IFE). Matthew P. Shortell gratefully acknowledges Monique Tourell for her assistance in manuscript preparation.

■

REFERENCES

(1) Sperling, R. A.; Parak, W. J. Surface modification, functionalization and bioconjugation of colloidal inorganic nanoparticles. Philos. Trans. R. Soc., A 2010, 368, 1333−1383. (2) Kalluri, J. R.; Arbneshi, T.; Afrin Khan, S.; Neely, A.; Candice, P.; Varisli, B.; Washington, M.; McAfee, S.; Robinson, B.; Banerjee, S.; Singh, A. K.; Senapati, D.; Ray, P. C. Use of gold nanoparticles in a simple colorimetric and ultrasensitive dynamic light scattering assay: Selective detection of arsenic in groundwater. Angew. Chem. 2009, 121, 9848−9851. (3) Song, Z.; Kelf, T. A.; Sanchez, W. H.; Roberts, M. S.; Rička, J.; Frenz, M.; Zvyagin, A. V. Characterization of optical properties of ZnO 475

dx.doi.org/10.1021/la403391t | Langmuir 2014, 30, 470−476

Langmuir

Article

(24) Hierrezuelo, J.; Szilagyi, I.; Vaccaro, A.; Borkovec, M. Probing nanometer-thick polyelectrolyte layers adsorbed on oppositely charged particles by dynamic light scattering. Macromolecules 2010, 43, 9108− 9116. (25) Seyrek, E.; Hierrezuelo, J.; Sadeghpour, A.; Szilagyi, I.; Borkovec, M. Molecular mass dependence of adsorbed amount and hydrodynamic thickness of polyelectrolyte layers. Phys. Chem. Chem. Phys. 2011, 13, 12716−12719. (26) Driskell, J. D.; Jones, C. A.; Tompkins, S. M.; Tripp, R. A. Onestep assay for detecting influenza virus using dynamic light scattering and gold nanoparticles. Analyst 2011, 136, 3083−3090. (27) Dasary, S. S. R.; Senapati, D.; Singh, A. K.; Anjaneyulu, Y.; Yu, H.; Ray, P. C. Highly sensitive and selective dynamic light-scattering assay for TNT detection using p-ATP attached gold nanoparticle. ACS Appl. Mater. Interfaces 2010, 2, 3455−3460. (28) Watanabe, S.; Yamamoto, S.; Yoshida, K.; Shinkawa, K.; Kumagawa, D.; Seguchi, H. Surface plasmon resonance scattering and absorption sensing of Concanavalin A using glycoconjugated gold nanoparticles. Supramol. Chem. 2011, 23, 297−303. (29) Jans, H.; Liu, X.; Austin, L.; Maes, G.; Huo, Q. Dynamic light scattering as a powerful tool for gold nanoparticle bioconjugation and biomolecular binding studies. Anal. Chem. 2009, 81, 9425−9432. (30) Miao, X.; Ling, L.; Cheng, D.; Shuai, X. A highly sensitive sensor for Cu2+ with unmodified gold nanoparticles and DNAzyme by using the dynamic light scattering technique. Analyst 2012, 137, 3064−3069. (31) Xiong, C.; Ling, L. Label-free, sensitive detection of Hg(II) with gold nanoparticles by using dynamic light scattering technique. Talanta 2012, 89, 317−321. (32) Miao, X.; Ling, L.; Shuai, X. Ultrasensitive detection of lead(II) with DNAzyme and gold nanoparticles probes by using a dynamic light scattering technique. Chem. Commun. 2011, 47, 4192−4194. (33) Bohren, C. F.; Huffman, D. R. Absorption and Scattering of Light by Small Particles; John Wiley and Sons, Inc.: Hoboken, NJ, 1983; Vol. 1, pp 136−150. (34) Spanhel, L.; Anderson, M. A. Semiconductor clusters in the sol− gel process: Quantized aggregation, gelation, and crystal growth in concentrated zinc oxide colloids. J. Am. Chem. Soc. 1991, 113, 2826− 2833. (35) Martin, M. N.; Basham, J. I.; Chando, P.; Eah, S.-K. Charged gold nanoparticles in non-polar solvents: 10-min synthesis and 2D selfassembly. Langmuir 2010, 26, 7410−7417. (36) Pesika, N. S.; Stebe, K. J.; Searson, P. C. Relationship between absorbance spectra and particle size distributions for quantum-sized nanocrystals. J. Phys. Chem. B 2003, 107, 10412−10415. (37) Viswanatha, R.; Sapra, S.; Satpati, B.; Satyam, P. V.; Dev, B. N.; Sarma, D. D. Understanding the quantum size effects in ZnO nanocrystals. J. Mater. Chem. 2004, 14, 661−668. (38) Sihvola, A. H. Peculiarities in the dielectric response of negativepermittivity scatterers. Prog. Electromagn. Res. 2006, 66, 191−198. (39) The polydispersity index is given by PDI = (σ/μ)2, where σ and μ are the standard deviation and mean of the particle size distributions, respectively. (40) Wong, E. M.; Bonevich, J. E.; Searson, P. C. Growth kinetics of nanocrystalline ZnO particles from colloidal suspensions. J. Phys. Chem. B 1998, 102, 7770−7775. (41) Hu, Z.; Oskam, G.; Searson, P. C. Influence of solvent on the growth of ZnO nanoparticles. J. Colloid Interface Sci. 2003, 263, 454− 460. (42) Godin, M.; Williams, P. J.; Tabard-Cossa, V.; Laroche, O.; Beaulieu, L. Y.; Lennox, R. B.; Grütter, P. Surface stress, kinetics, and structure of alkanethiol self-assembled monolayers. Langmuir 2004, 20, 7090−7096.

476

dx.doi.org/10.1021/la403391t | Langmuir 2014, 30, 470−476