1864 Emer Duffy Dimitar P. Mitev Pavel N. Nesterenko Artaches A. Kazarian Brett Paull Australian Centre for Research on Separation Science, University of Tasmania, Hobart, Tasmania, Australia
Received October 7, 2013 Revised March 3, 2014 Accepted March 9, 2014
Electrophoresis 2014, 35, 1864–1872
Research Article
Separation and characterisation of detonation nanodiamond by capillary zone electrophoresis A new method for the characterisation of purified detonation nanodiamond (DND) using CZE has been developed. The influence of BGE conditions on electrophoretic mobility, peak shape and particle aggregation was investigated, with resultant observations supported by zeta potential approximations and particle size measurements. Sodium tetraborate (pH 9.3), Tris (pH 9.3) and sodium phosphate (pH 7) were used in studying the BGE concentration effect on a commercial source of chemically stabilised DND. The BGE concentration had a strong effect on the stability of DND in suspension. The formation of aggregates of various sizes was observed as BGE concentration increased. The effect of pH on the electromigration of DND was examined using sodium phosphate (pH 8 and 10). The CZE method was subsequently applied to four different DND samples, which had undergone different routes of purification following detonation synthesis. Each sample produced a unique electrophoretic peak or profile in sodium tetraborate buffer (pH 9.3), such that the actual separation of DND samples from different sources could be achieved. Keywords: Aggregation / Capillary electrophoresis / Detonation nanodiamond / Size distribution / Zeta potential DOI 10.1002/elps.201300488
Additional supporting information may be found in the online version of this article at the publisher’s web-site
1 Introduction 1.1 General Nanoscale diamond (sp3 carbon) is an increasingly applied member of the nanocarbon family, which exhibits unique properties, originating both from its core lattice structure and its complex surface chemistry [1]. Favourable optical and mechanical properties of nanodiamond (ND), together with their excellent thermal stability and conductivity, and biocompatibility [2] have generated considerable interest in these relatively uncharacterised and chemically variable nanomaterials. Detonation ND (DND) is produced by the detonation of explosives in closed containers, followed by the subsequent collection and thorough purification of the soot to obtain
Correspondence: Professor Brett Paull, Australian Centre for Research on Separation Science, University of Tasmania, Hobart, Tasmania 7001, Australia E-mail: [email protected] Fax: +61-3-6226-2858
Abbreviations: DLS, dynamic light scattering; DND, detonation nanodiamond; DS, detonation soot; ND, nanodiamond; SD-DND, single-digit detonation nanodiamond C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
graphite-free diamond [3]. In terms of its surface properties, DND is a diverse material, in both variety and concentration of functional groups, elemental impurities, zeta ( ) potentials and subsequent propensity to form aggregates, which are highly variable in size. The material is generally considered to be relatively chemical stable, despite the varied surface chemistry, which may include carboxyl, hydroxyl, ketones, ethers, and lactones, together with adsorbed inorganic impurities [1, 4, 5]. This complexity in the surface chemistry of DND is associated with a possible presence of sp2 carbon layers at the outer surface. The exact composition and nature of these surface properties is primarily determined by the methods applied during production and purification. Oxygen-containing groups are usually present on the surface of DND due to the postsynthesis cooling and the purification methods, which usually involve oxidation [6]. The typical diameter of DND produced by detonation synthesis is 4–5 nm [7], however, these particles have a very strong tendency to form aggregates up to several hundreds of nanometres in size. DND suspensions are generally stabilised through naturally occurring surface charge, or via adsorbed charged species. Such
Colour Online: See the article online to view Figs. 2 and 5 in colour.
www.electrophoresis-journal.com
CE and CEC
Electrophoresis 2014, 35, 1864–1872
electrostatic stabilisation allows for measurement of electrophoretic mobility and calculation of potentials. Potentials above +30 mV and below −30 mV are generally observed for such stable suspensions; however, DND can exhibit a very wide distribution of potentials ranging from as low as −100 to +70 mV, often dependent upon the purification procedures applied. A wide variety of applications for DND are currently being explored, including their use within analytical chemistry, specifically chromatography [2, 8–11], their incorporation in nanocomposites and coatings [12], their use in optoelectronics [13], and their emerging application within the field of biomedicine, such as their use within surgical implants and as vehicles for drug delivery [14–17]. However, to further these applications, it is quite clear that there are many challenges to be overcome in understanding the exact properties and behaviour of DND, and in developing new processes to ensure physical and chemical DND homogeneity. To meet both challenges, the development of reliable characterisation methods is essential. A number of techniques for the characterisation of DND have been reported to date. SEM and high-resolution transmission electron microscopy are often used for characterisation of nanoparticle size and morphology. However, these techniques are obviously expensive and time consuming, and in the case of DND, changes in particle aggregation as a result of desolvation present a problem. TOF-MS has also been reported for the analysis of DND samples, and has provided information on size distribution, aggregation and surface chemistry [18]. The ionisation of DND samples resulted in the formation of singly charged carbon clusters in different regions of the mass spectrum providing information on size distributions. Studying electrophoretic parameters has proven useful in the characterisation of DND. For example, dynamic light scattering (DLS) and laser Doppler electrophoresis offer a more rapid method of obtaining information on particle diameters, estimating electrophoretic mobilities and potentials. Petrova et al. [19] used such measurements to investigate the stability of DND in different salt solutions and calculated potentials based on the size of the different particles. Chiganova et al. also looked at the electrophoretic behaviour of DND in increasing concentrations of aluminium chloride, where concentration affected potential as the DND surface became modified with aluminium [20]. CZE offers a potentially rapid, automated and robust method for the characterisation of such properties of DND, and indeed nanoparticles in general. The technique is compatible with a large range of solvent and buffer systems, and has low sample volume requirements, and CZE should be capable of providing information on nanoparticle size distributions, surface charge, dispersion quality and stability. CZE has previously been employed as a method for the size and surface-charge characterisation of a variety of nanoparticles, including metal nanoparticles [7, 21, 22], carbon nanoparticles, such as graphene oxide and chemically converted graphene [23–25] carbon dots [26] and carbon nan C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
1865
otubes [27–33], and a number of reviews have been published on the topic [34–37]. Chang et al. monitored the growth of gold nanoparticles using CZE by simultaneously analysing the electrophoretic profile and the absorption spectra of the samples to obtain information on particle size and distribution [38]. Baker and Col´on used CZE to investigate the influence of buffer conditions on the mobility and fractionation of carbonaceous nanoparticles in a soot sample obtained from the flame of an oil lamp in order to better understand the complex sample [39]. However, to date few studies using CZE for actual DND separations have emerged. Therefore, herein the authors present the first report on the use of CZE for the separation and characterisation of commercial and non-commercial DND. The influence of BGE conditions on the electro-migration and the separation of DND was investigated and used to approximate potentials, and data were supported with particle size measurements using DLS. The effect of pH and increasing concentrations of a variety of common buffers on the electrophoretic profile and particle aggregation of DND samples was explored. Optimum CZE conditions were subsequently applied to a variety of commercial and non-commercial DND samples, which had undergone different routes of purification following detonation synthesis.
1.2 Theoretical background NDs exhibit a wide range of potentials and particle sizes and thus should be treated as a colloidal particle in this instance, rather than a small ion [39]. Therefore, a brief overview of the theory related to the electrophoretic behaviour of charged colloidal particles is presented here. A cloud of ions (referred to as the shear surface) surrounds a charged colloidal particle in a BGE, which has a net charge that determines the electrophoretic mobility of the particle. As the particle migrates in an electric field, it is accompanied by a specific quantity of the surrounding liquid and the charged species, which it contains. The potential is the value of electric potential at the surface of shear (can and usually does include some of the bound counterions) [40], and it is related to the electrophoretic mobility of the particle by the following equation: = (2ε /3) f (a),
(1)
where ε is the permittivity of the medium which has viscosity . This applies to particles with a low potential ( 25 mV at 25°C) where Henry’s function, f(a) ranges from 1 at a 1 (Smoluchowski limit) to 1.5 at a 1 (H¨uckel limit). The inverse Debye–H¨uckel parameter describes the electric double layer thickness and is dependent on the ionic strength (I) of the BGE as follows: √ (2) (nm−1 ) ≈ 3.288 I(mol/L), at 25°C in water [41, 42]. This equation applies to a symmetrical electrolyte; however, it may be used in cases of a nonsymmetrical electrolyte when the majority of counterions populating the diffuse www.electrophoresis-journal.com
1866
E. Duffy et al.
Figure 1. Plot showing the theoretical relationship of mobility as a function of a (product of inverse Debye length and the particle radius) for different values of potential (‘low’ , i.e. 25 mV, and ‘high’ , i.e. >25 mV) [40, 41]. Experimental results are superimposed for comparative purposes and to allow an approximation of potentials for single-digit nanodiamonds in Tris buffer ( ) at concentrations of 10, 50 and 140 mM (pH 9.3), and in sodium phosphate buffer () at concentrations of 10, 20 and 30 mM (pH 7, 8 and 10).
part of the electric double layer are monovalent [41]. Within Henry’s function, a is the hydrodynamic or stokes radius of the particles. The particle radius can actually be substituted for a in most cases, as the distance between solid and shear surfaces generally does not exceed 1 nm [42]. In reality, most nanoparticles lie in the range where a is near to one and they have much higher potentials. In order to account for these factors, it is necessary to apply the Overbeek–Booth theory [40] which accounts for electrophoretic retardation and the relaxation effect, and this was presented numerically for samples with a greater range of 25–150 mV [43, 44]. This is depicted in Fig. 1, where the theoretical dependence of reduced electrophoretic mobility on a (or the ratio of a particle’s hydrodynamic radius to the thickness of the double layer) for both ‘low’ ( = 25 mV) and ‘high’ ( ⬎ 25 mV) values of potential is shown.
2 Materials and methods 2.1 Chemicals and reagents R All solutions were prepared using Milli-Q water (Millipore Gradient System). Commercial DND samples (single-digit detonation nanodiamonds (SD-DND) 20 mL aqueous 5% solution (50 mg/mL)) were purchased from PlasmaChem (Berlin, Germany) in a 50 mg/mL aqueous suspension of chemically stabilised negatively charged NDs, which was later diluted to 2 mg/mL with Milli-Q water for the preparation of samples for injection. Three DND samples, namely NSPA,
C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Electrophoresis 2014, 35, 1864–1872
NSHCl and NSFPA, were prepared in our laboratory from the detonation soot (DS) obtained from the Space and Solar Terrestrial Research Institute at the Bulgarian Academy of Sciences (SSTRI-BAS) (Sofia, Bulgaria) and purified by acid oxidation at the University of Tasmania (Hobart, Australia). Briefly, the NSPA sample was purified from DS using acidic oxidation of DS with a mixture of HNO3 , H2 SO4 and HClO4 acids. The NSHCl sample was treated similarly using a mixture of HNO3 , H2 SO4 and HCl acids, and the NSFPA sample using a mixture HNO3 ·H2 SO4 , HClO4 and HF. Subsequent washing was performed to reduce the digested impurities and residual acids. The Ru-Kr DND sample was obtained from the Institute of Biophysics at the Siberian Division of the Russian Academy of Sciences (Krasnoyarsk, Russia). All DND samples were negatively charged under the pH conditions used (pH 7, 8, 9.3 and 10). Acetone (99.5%) was sourced from Chem-Supply (Gillman, Australia). HCl was purchased from Merck (Darmstadt, Germany) and NaOH from Scharlau (Barcelona, Spain). The reagents for BGE preparation purchased from Sigma-Aldrich were sodium phosphate monobasic (99.0%; St. Louis, MO, USA), sodium phosphate dibasic (99.5%) (Seelze, Germany), sodium tetraborate (99.998%) and Tris (99.9%; Castle Hill, Australia). Amberlite resin IRA-93 (analytical grade) used in preparing DND samples for SEM imaging was purchased from BDH Chemicals (Poole, England). 2.2 Capillary zone electrophoresis Experiments were performed on an Agilent 7100 CE system with detection using a DAD (Agilent Technologies, Germany). Absorption was monitored at 195 and 210 nm at data acquisition rates 40 Hz. Data acquisition was performed using OpenLAB CDS ChemStation Edition software for Windows 7. Data were processed using Origin 9 (OriginLab) software. All reported results were obtained in triplicate. Untreated fused silica capillaries (Polymicro Technologies, Phoenix, AZ, USA) with a 75 m id, 365 m od were used in this study. Detection windows were burned with a butane torch at 8.5 cm from the capillary end with a total capillary length of 66 cm. New capillaries were pre-treated with 0.5 M NaOH for 10 min, followed by H2 O for 10 min, and finally BGE for 10 min. Capillaries were thermostated at 25°C in all experiments. All DND samples contained 0.04 wt% DND dispersed in BGE, and injections were performed by applying a pressure of 5 kPa for 5–20 s, followed by the injection of acetone at 5 kPa for 1 s. The analytes were separated under an applied potential of +15.0 kV. Prior to each run, the capillary was conditioned by flushing H2 O for 3 min, 0.5 M NaOH for 2 min, H2 O for 2 min and BGE for 2 min. Acetone was used as the neutral EOF marker. 2.3 BGE preparation Buffers were prepared by mixing and dilution of sodium phosphate monobasic and sodium phosphate dibasic, and www.electrophoresis-journal.com
Electrophoresis 2014, 35, 1864–1872
by diluting stock solutions of sodium tetraborate and Tris. The pH of buffer solutions was adjusted using dilute HCl and NaOH, and pH values were measured using a model 210 pH meter from Activon (Thornleigh, Australia) and a Mettler Toledo LE409 probe (Port Melbourne, Australia). Tetraborate and Tris buffers were maintained at pH 9.3 in this work. BGEs were degassed with a Soniclean ultrasonic bath and filtered through 0.45 mm disc filters (Activon). 2.4 Particle size measurements Particle size distributions of DND in water and solutions of sodium tetraborate, Tris chloride and sodium phosphate were measured using a ZetaSizer NanoZS (Malvern Instruments, UK). The reported results were obtained in triplicate, and performed at 25°C as in CZE. Measurements were performed using 0.01 wt% suspensions of ND. The concentration and the pH of tetraborate and Tris (pH 9.3) and sodium phosphate (pH 7, 8 and 10) used in this analysis were the same as per CZE experiments. Particle imaging was carried out using field emission high-resolution SEM using a Hitachi Ultra-High Resolution Analytical FE-SEM SU-70 instrument (Hitachi High Technologies America, USA). The DND sample was prepared on Amberlite IRA943 resin (BDH Chemicals) and sputtered with a thin layer of Pt before imaging. Briefly, sample preparation involved adding 0.1 g of Amberlite resin to 1 mL of 0.64% wt. suspension of DNDtype NSHCl, and diluting further with 20 mL Milli-Q water. The mixture was left for 15 h and subsequently washed with Milli-Q water and dried before imaging.
3 Results and discussion 3.1 Buffer chemistry and concentration effects The effect of buffer concentration within the BGE on the peak shapes, apparent particle aggregation points, and electrophoretic mobility of a commercial so-called SD-DND (PlasmaChem) was studied, using sodium phosphate, sodium tetraborate and Tris buffers. Observation of the aggregation properties of certain classes of carbon nanoparticles in various BGE buffers using CZE has been described previously, including for graphene oxide [24] and carbon nanotubes [27]. However, no such study involving DND has been reported to date. In the case of graphene oxide NPs, M¨uller et al. were able to obtain a single broad peak using CZE in BGE buffer concentrations of 2 mM (ammonium acetate) and below, whilst the onset of aggregation and the emergence of multiple random spikes were observed at buffer concentrations of 5 mM and above. Generally, potentials in aqueous colloidal suspensions are a function of the charge at the shear plane and the free ion concentration. Electrostatic potential decays exponentially with distance from the shear plane, and the inverse of this decay constant is called the Debye double layer thickness. Increasing the free ion concentration increases the decay and C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
CE and CEC
1867
so reduces this double layer thickness. At a certain ion concentration, the reduction of the double layer will reach a point whereby its electrostatic repulsion forces are less than underlying attractive forces, such as van der Waals forces, at which point suspension stability will be lost and spontaneous aggregation can occur. However, in the case of DND, a number of recent studies have shone new light upon the nature of particle aggregation within solution, dismissing van der Waals interactions and identifying strong electrostatic forces as being behind the spontaneous self-assembly of ND agglutinates (60 nm) and larger agglomerates (100–200 + nm) [43–47]. Barnard and co-workers classify ND aggregates according to these approximate size distributions, with the former, namely agglutinates, interacting through interfacial type alignment, based upon electrostatic attraction in an ordered configuration. The larger agglomerates are also interfacially aligned via electrostatic attractions, but configured in a more random orientation [47]. Herein, Fig. 2A shows the typical electropherograms obtained following injection of this particular SD-DND sample in tetraborate buffer-based BGEs of increasing concentration, ranging from 10 to 60 mM, each fixed at a pH of 9.3. The negatively charged NDs migrated after the acetone EOF marker, as a single well-defined broad peak for tetraborate buffer concentrations of between 10 and 32.5 mM. The absence of spikes within the broad Gaussian-shaped peak would suggest (10 mM tetraborate BGE) a stable suspension of DND ‘particles’ under such conditions. Under these low-concentration buffer conditions, it can be assumed that the DND is in the form of small agglutinates of a relatively small size distribution, rather than single-digit (⬍10 nm) particles. Such agglutinates are sufficiently small and surface charged to exhibit both stable suspension and migration within the capillary. Indeed, this assumption was supported by average particle size data obtained for the SDDND sample in 10 mM tetraborate buffer, obtained using DLS (see Table 1). Here, an average particle size of 33 ± 4 nm was recorded. Increasing the concentration of the tetraborate BGE to 35 mM and above resulted in the increased appearance of sharp peaks or spikes, and the simultaneous disappearance of the broad peak. This behaviour corresponds closely to that reported earlier for graphene oxide NPs [24], although herein the migration window for both the initial broad peak and the series of sharp spikes remained relatively constant. For example, at 60 mM tetraborate only a series of sharp spikes could be seen, covering a migration time window of 4 min, which was approximately equal to the width of the initial broad peak at 10 mM. The appearance of these spikes in higher concentrations of tetraborate can be attributed to the formation of larger randomly arranged agglomerates within the sample, formed as the repulsion forces between the ND agglutinates are sufficiently reduced to allow spontaneous and random aggregation, to sizes no longer stable in suspension. Figure 1C illustrates this visually, with the SD-DND sample diluted in water (stable suspension) and 60 mM tetraborate buffer solution (settled unstable agglomerates). The formation of www.electrophoresis-journal.com
1868
E. Duffy et al.
Electrophoresis 2014, 35, 1864–1872
Figure 2. Electropherograms generated by single-digit nanodiamonds (SD-DND) samples in different concentrations of (A) – tetraborate buffer (60, 40 and 10 mM) and (B) – Tris buffer (140, 50 and 10 mM) at a fixed pH of 9.3. Conditions: injection of acetone for 1 s and nanodiamond for 20 s with detection by DAD at 195 nm. (C) Image of a stable suspension of SD-DND in water (left) and an aggregated SD-DND sample in 60 mM tetraborate (right) where large aggregated particles have dropped out of suspension.
large agglomerates in higher concentrations of tetraborate buffer (60 mM) was visible to the naked eye, however, for lower concentrations where small agglomerates were beginning to form (20–40 mM), no visible changes to the suspension were noticeable. Once again to confirm the process occurring, particle size data were obtained (Table 1), in this instance for 40 and 60 mM tetraborate solutions, revealing average particle sizes of 868 and 1359 nm, respectively (the
average particle size of SD-DND in 60 mM tetraborate actually increased with each consecutive DLS measurement, as large aggregates formed rapidly in the sample). It appears that the critical (tetraborate, pH 9.3) buffer concentration for aggregation of these particular ND agglutinates into large unstable agglomerates is between 30 and 35 mM. Figure 2B shows the typical electropherograms obtained of the same SD-DND sample, this time separated with a
Table 1. Average particle size data and polydispersity index (PDI) obtained (by dynamic light scattering using number-based size distributions) for single-digit nanodiamond suspensions in H2 O and various concentrations of tetraborate and Tris buffers at pH 9.3 and sodium phosphate pH 7, 8 and 10, and for NSPA, NSHCl, Ru-Kr and NSFPA suspensions in H2 O and 10 mM tetraborate pH 9.3
Sample
BGE
Concentration (mM)
SD-DND
Tetraborate
SD-DND
Tris
SD-DND
Sodium phosphate
10 40 60 10 50 140 10
20
30
NSPA NSHCl Ru-Kr NSFPA
Tetraborate Tetraborate Tetraborate Tetraborate
10 10 10 10
C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
pH 9.3
9.3
7 8 10 7 8 10 7 8 10 9.3 9.3 9.3 9.3
Size (nm)
PDI
33 ± 4 868 ± 75 1359 ± 391 32 ± 5 31 ± 4 30 ± 7 31 ± 3 24 ± 5 18 ± 3 42 ± 4 38 ± 7 20 ± 4 48 ± 5 45 ± 5 200 ± 17 2292 ± 347 1613 ± 120 76 ± 2 Polydisperse
0.218 0.274 0.391 0.234 0.247 0.256 0.292 0.283 0.249 0.187 0.202 0.213 0.249 0.21 0.41 0.225 0.127 0.243
www.electrophoresis-journal.com
Electrophoresis 2014, 35, 1864–1872
CE and CEC
1869
deviations). Particle size measurements also showed stable agglutinates of 23–37 nm in size up to a concentration of 140 mM Tris. Figure 3A shows the electromigration of SD-DND in a sodium phosphate BGE at a fixed pH of 7, at concentrations of 10, 20 and 30 mM. Corresponding potential approximations are shown in Fig. 1 and particle size data are also listed within Table 1. Over these conditions, the SD-DND sample remained stable in solution, with no signs of aggregation. However, particle size data indicated a possible growth in the size of the stable ND agglutinates from 31, to 42, to 48 nm, respectively. This also corresponds to a decrease in effective mobility (see Fig. 3B), and for 30 mM phosphate, a noticeable increase in peak width. In phosphate BGEs above 30 mM, a reduction in sample stability was noted and agglomerate spikes began to appear in 40 mM sodium phosphate at pH 7.
3.2 Buffer pH on the electro-migration of DND
Figure 3. (A) Electropherograms generated by single-digit nanodiamonds (SD-DND) samples in different concentrations of sodium phosphate (10, 20 and 30 mM) at a fixed pH of 7. Conditions: injection of acetone for 1 s and nanodiamond for 20 s with detection by DAD at 210 nm. (B) Effective mobilities (CZE measurements) for SD-DND in different BGEs. Figure legend for the upper plot sodium phosphate pH 7, sodium phosphate pH 8, × sodium phosphate pH 10, tetraborate pH 9.3. Figure legend for the lower plot: Tris pH 9.3.
Tris BGE of increasing concentration. Separations were performed in concentrations ranging from 10 to 140 mM Tris, once more at a fixed pH of 9.3. Here, as with the tetraborate BGE above, the EOF decreased over the increasing concentration range due to the double layer compression. However, here the mobility of the sample peak was far less affected by the change in buffer concentration, and also showed no signs of particle aggregation, even at the highest (140 mM) BGE concentration. A similar characteristic broad peak was observed for the SD-DND sample, of 4-min width, with no spikes due to aggregation visible. The stability of the SDDND sample over the wider concentration range with the Tris buffer, as compared to that of the tetraborate BGE, was concluded to be an ionic strength effect rather than any surface adsorption effects. Figure 1 shows the approximate potentials for SD-DND under these BGE conditions, where potential is observed to decrease with increased ionic strength. As the ionic strength increased, the inverse Debye length and stokes radius also increased, corresponding to the observed reduction in mobility (see Fig. 3B – each data point is the mean of three measurements, error bars = standard C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Table 1 also shows data for SD-DND in sodium phosphate BGE at 10, 20 and 30 mM, but over a range of pH values. For each pH value, a similar trend for reduced mobility as a function of increased buffer concentration was seen (Fig. 3B). However, the effect of pH upon particle size appeared to change as buffer concentration increased. For 10 and 20 mM, a decrease in particle size was observed as buffer pH increased, for example, from 31, to 24, to 18 nm, at 10 mM phosphate BGE concentration for pH 7, 8 and 10, respectively. At 30 mM phosphate, pH 10, the system became unstable and large agglomerates began to form, with particle sizes of ⬎200 nm apparent. The reasons for these observed effects are unclear, although again it is possible that this is predominantly related to ionic strength. The approximate values for potential within Fig. 1 show no obvious trend to ascertain if changes in surface charge could also account for such size variation. The values remain between the trend lines for −51 and −76 mV, with pH 7 and 8 causing more of a change in potential of SD-DND compared to pH 10. Figure 4A shows the electropherograms obtained for the SD-DND sample in 10 mM sodium phosphate BGE and the effect of BGE pH (pH 8 and 10 shown) on the peak shape is evident with the SD-DND peak becoming narrower and peak intensity increasing as the pH increases. Peak widths of 4.5 min at 10 mM phosphate, pH 7, reduce to 3 min at pH 8 and 2.5 min at pH 10. Simultaneously, as mentioned above, average particle size decreases from 31, to 24, to 18 nm. Thus, at 10 mM BGE concentration, it would appear some evidence of peak width correlating with average particle size in the first instance. Similar, although less obvious, observations could be made with peaks shapes observed in 20 mM phosphate BGEs, over the three pH values, although as is clear within Fig. 4B, at pH 10 with this concentration of phosphate BGE, the critical aggregation point seems to have been reached, as the sudden appearance of agglomerate spikes can be seen. Corroborating the unusual particle size www.electrophoresis-journal.com
1870
E. Duffy et al.
Electrophoresis 2014, 35, 1864–1872
Figure 4. (A) Electropherograms generated by single-digit nanodiamonds (SD-DND) samples in 10 mM sodium phosphate at pH 8 and 10, (B) in 20 mM sodium phosphate at pH 8 and 10, (C) in 30 mM sodium phosphate at pH 8 and 10. Conditions: injection of acetone for 1 s and nanodiamond for 20 s with detection by DAD at 210 nm.
data listed within Table 1 for the 30 mM phosphate BGE at pH 10, is Fig. 4C that shows very dramatically the electropherograms for the ND agglomerates of 200 (±17) nm size.
3.3 Analysis of DND samples Based on the above observations on optimal conditions for the SD-DND sample, a variety of DND samples of differing origin were investigated using a 10 mM tetraborate BGE, at pH 9.3. Four DND samples were investigated, namely, NSPA, NSHCl, NSFPA and Ru-Kr. Each sample had undergone differing post-detonation purification, with the result that each had markedly differing surface chemistry, particle size and aggregation tendencies. Obtained under the above conditions, resulting electropherograms are shown in Fig. 5A. Figure 5A clearly shows the four samples could be classified as either containing stable agglutinates (Ru-Kr and NSFPA) or larger unstable agglomerates (NSPA and NSHCl), under such conditions. The absorbance spectra for each of the above four samples and that of the SD-DND sample can be seen within Supporting Information Fig. 1, which clearly shows the similarity between the absorbance spectra of the SD-DND, NSFPA and Ru-Kr samples, whilst the spectra for the aggregated samples are similarly shaped, whilst obviously rather irregular and noisy. Figure 5A(i) and (ii) shows expanded electropherograms of the latter two samples. The above classifications are supported by the average particle size data listed in Table 1 that confirms the very large agglomerated micron-sized particles, of size 1.4 to 1.5 m, for NSHCl and NSPA samples, whilst the Ru-Kr sample was found to be in the order of 76 ± 2 nm. To further investi C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
gate the aggregation of DND, the NSHCl was analysed using field emission high-resolution scanning electron microscopy and the resultant SEM image, taken at high magnification (×100K), is shown in Fig. 5D. Clear agglomerates of the order 500–1500 nm can be seen in this sample. Interestingly, the results from particle size measurements for the NSFPA sample were inconclusive and noted only as a polydisperse material. Upon close inspection of Fig. 5A, the reason for this finding becomes clear, as the emergence of spikes from ND aggregation can just be seen, suggesting the sample is at its critical aggregation point, and contains both agglutinates and agglomerates in solution, hence its polydispersity [38]. The NSPA sample formed large visible aggregates in water and 10 mM tetraborate, whilst the NSHCl sample formed slightly smaller aggregates in 10 mM tetraborate, but formed a stable suspension in water alone. Interestingly, the particle size ranges for these two samples proved to be larger (1945– 2639 nm for NSPA and 1493–1733 nm for NSHCl), which should be reflected in differences in average mobilities for the two samples. Indeed, Fig. 5A(i) and (ii) clearly shows these differences, and highlighted how the CZE method can in fact be used to completely separate two batches of different NDs. This is demonstrated in Fig. 5C wherein samples containing mixtures of two DND samples, either both present as larger agglomerates (e.g. NSPA and SD-DND samples at 60 mM tetraborate BGE, pH 9.3), or where one of which is present as small agglutinates and the other as larger agglomerates (e.g. NSPA and SD-DND, 10 mM tetraborate, pH 9.3), can be successfully separated. This is the first demonstration of CZE to physically separate DND samples obtained from different suppliers, or samples obtained via differing synthetic routes, and could prove a useful approach to help identify both source www.electrophoresis-journal.com
CE and CEC
Electrophoresis 2014, 35, 1864–1872
1871
Figure 5. (A) Electropherograms generated by four different nanodiamond samples: NSPA, NSHCl, Ru-Kr and NSFPA in 10 mM tetraborate (pH 9.3). (A(i)) Magnification of NSPA and (A(ii)) NSHCl showing sharp peaks characteristic of aggregated nanodiamond. (B) Image of the five different detonation nanodiamond samples under investigation (characteristic electrophoretic profile for single-digit nanodiamonds (SD-DND) in 10 mM tetraborate is shown in Fig. 1(A)). (C) Electropherograms showing the separation of single-digit and NSPA nanodiamonds in 60 and 10 mM tetraborate buffer (pH 9.3). Conditions: injection of acetone for 1 s and nanodiamonds for 20 s with detection by DAD at 195 nm. (D) HR-SEM image of DND-type NSHCl on Amberlite IRA943 resin. Magnification: ×100 000.
and nature of unknown DND suspensions, and/or provide information on composite suspensions.
R2 = 0.9999 and an LOD of 0.0125 mg/mL. The corresponding calibration curve and electropherograms can be viewed within Supporting Information Fig. 3.
3.4 Quantitation of DND solutions using CZE
4 Concluding remarks
The increasing interest in medical applications of DND-type materials also requires quantitative assays of DND suspensions and their stability. Therefore, herein, to demonstrate the potential use of CZE to quantify the concentration of DND within solution (as stable agglutinates), both injection and detection linearity were briefly explored. The SD-DND sample was injected in triplicate for 2, 5, 10, 15 and 20 s injections in 10 mM tetraborate, pH 9.3. A plot of injection time versus corrected peak area resulted in a linear correlation of R2 = 0.9998. The resultant calibration curve and corresponding electropherograms can be viewed within Supporting Information Fig. 2. A serial dilution of the SD-DND sample (starting at 0.4 mg/mL) was then carried out to determine both sample concentration linearity and approximate DND detection limits, with a resultant linear correlation of
As shown herein, CZE has been found to offer some unique capabilities in this area. CZE can be a quantitative tool for the characterisation of DND, providing information on sample charge, size, stability and tendency to agglomerate. Coupled with particle size measurements, it can provide a better understanding of the surface properties and dispersion quality of NDs in different buffer systems. CZE offers the possibility of separating different NDs, which could prove highly useful for sample fractionation in the future.
C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
The Authors would like to acknowledge the Australian Research Council for financial support (ARC Discovery Grant DP110102046). The authors have declared no conflict of interest. www.electrophoresis-journal.com
1872
E. Duffy et al.
5 References [1] Krueger, A., J. Mater. Chem. 2008, 18, 1485–1492. [2] Nesterenko, P. N., Haddad, P. R., Anal. Bioanal. Chem. 2010, 396, 205–211.
Electrophoresis 2014, 35, 1864–1872
[25] Takayanagi, T., Morimoto, M., Yabutani, T., Anal. Sci. 2013, 29, 769–771. [26] Hu, Q., Paau, M. C., Zhang, Y., Chan, W., Gong, X. J., Zhang, L., Choi, M. M. F., J. Chromatogr. A 2013, 1304, 234–240.
[4] Paci, J. T., Man, H. B., Saha, B., Ho, D, Schatz, G. C., J. Phys. Chem. C 2013, 117, 17256–17267.
[27] Doorn, S. K., Strano, M. S., O’Connell, M. J., Haroz, E. H., Rialon, K. L., Hauge, R. H., Smalley, R. E., J. Phys. Chem. B 2003, 107, 6063–6069.
[5] Mitev, D. P., Townsend, A. T., Paull, B., Nesterenko, P. N., Carbon 2013, 60, 326–334.
[28] Suarez, B., Simonet, B. M., Cardenas, S., Valcarcel, M., J. Chromatogr. A 2006, 1128, 282–289.
[6] Krueger, A., Lang, D., Adv. Funct. Mater. 2012, 22, 890–906.
[29] Yamamoto, T., Murayama, S., Kato, M., J. Sep. Sci. 2011, 34, 2866–2871.
[7] Mochalin, V. N., Shenderova, O., Ho, D., Gogotsi, Y., Nat. Nanotechnol. 2012, 7, 11–23.
[30] Yamamoto, T., Murakami, Y., Motoyanagi, J., Fukushima, T., Maruyama, S., Kato, M., Anal. Chem. 2009, 81, 7336–7341.
[3] Krueger, A., Adv. Mater. 2008, 20, 2445–2449.
[8] Saini, G., Jensen, D. S., Wiest, L. A., Vail, M. A., Dadson, A., Lee, M. L., Shutthanandan, V., Linford, M. R., Anal. Chem. 2010, 82, 4448–4456. [9] Nesterenko, P. N., Fedyanina, O. N., J. Chromatogr. A 2010, 1217, 498–505. [10] Fedyanina, O. N., Nesterenko, P. N., Russ. J. Phys. Chem. 2011, 85, 1773–1777. [11] Wiest, L. A., Jensen, D. S., Hung, C.-H., Olsen, R. E., Davis, R. C., Vail, M. A., Dadson, A. E., Nesterenko, P. N., Linford, M. R., Anal. Chem. 2011, 83, 5488–5501.
[31] Xiao, H., Zou, H. F., Pan, C. S., Jiang, X. G., Le, X. C., Yang, L., Anal. Chim. Acta 2006, 580, 194–199. [32] Lopez-Pastor, M., Dominguez-Vidal, A., Ayora-Canada, M. J., Simonet, B. M., Lendl, B., Valcarcel, M., Anal. Chem. 2008, 80, 2672–2679. [33] Doorn, S. K., Fields, R. E., Hu, M., Hamon, M. A., Haddon, R. C., Selegue, J. P., Majidi, V., J. Amer. Chem. Soc., 2002, 124, 3169–3174. [34] Surugau, N., Urban, P. L., J. Sep. Sci. 2009, 32, 1889–1906.
[12] Behler, K. D., Stravato, A., Mochalin, V., Korneva, G., Yushin, G., Gogotsi, Y., ACS Nano 2009, 3, 363–369.
[35] Pyell, U., Electrophoresis 2010, 31, 814–831.
[13] Williams, O. A., Semicond. Sci. Technol. 2006, 21, R49–R56.
[36] Radko, S. P., Chrambach, A., Electrophoresis 2002, 23, 1957–1972.
[14] Huang, H., Pierstorff, E., Osawa, E., Ho, D., Nano Lett. 2007, 7, 3305–3314.
[37] Rodriguez, M. A., Armstrong, D. W., J. Chromatogr. B 2004, 800, 7–25.
[15] Schrand, A., Hens, S. A. C., Shenderova, O., Crit. Rev. Solid State Mater. Sci. 2009, 34, 18–74.
[38] Chang, T.-H., Liu, F.-K., Chang, Y.-C., Chu, T.-C., Chromatographia 2008, 67, 723–730.
[16] Xing, Y., Dai, L., Nanomedicine 2009, 4, 207–218.
´ L. A., J. Chromatogr. A 2009, 1216, [39] Baker, J. S., Colon, 9048–9054.
[17] Ho, D., ACS Nano 2009, 3, 3825–3829. ˇ ˜ ´ [18] Houska, J., Panyala, N. R., Pena-M endez, E. M., Havel, J., Rapid Comm. Mass Spectrom. 2009, 23, 1125–1131.
[40] Overbeek, J. T. G., Wiersema, P. H., in: Bier, M. (Ed.), Electrophoresis: Theory, Methods and Applications, Vol. II, Academic Press, New York 1967, pp. 1–52.
[19] Petrova, N., Zhukov, A., Gareeva, F., Koscheev, A., Petrov, I., Shenderova, O., Diam. Relat. Mater. 2012, 30, 62–69.
[41] Radko, S. P., Stastna, M., Chrambach, A., Electrophoresis 2000, 21, 3583–3592.
[20] Chiganova, G. A., Bondar, V. A., Chiganov, A. S., Colloid J. Russ. Acad. 1993, 55, 774–775.
´ F., Varenne, A., Gareil, P., Electrophoresis 2008, [42] d’Orlye, 29, 3768–3778.
[21] Liu, F.-K., Ko, F.-H., Huang, P.-W., Wu, C.-H., Chu, T.-C., J. Chromatogr. A 2005, 1062, 139–145.
[43] Wiersema, P. H., Loeb, A. L., Overbeek J. T. G., J. Colloid Interf. Sci. 1966, 22, 78–99.
[22] Liu, F.-K., Lin, Y.-Y., Wu, C.-H., Anal. Chim. Acta 2005, 528, 249–254.
[44] O’Brien R. W., White L. R., J. Chem. Soc., Faraday Trans. 1978, 74, 1607–1626.
[23] Zhao, J., Chen, G., Zhang, W., Li, P., Wang, L., Yue, Q., Wang, H., Dong, R., Yan, X., Liu, J., Anal. Chem. 2011, 83, 9100–9106.
[45] Huang, H., Dai, L., Wang, D. H., Tan, L-S., Osawa, E., J. Mater Chem. 2008, 18, 1347–1352.
¨ [24] Muller, M. B., Quirino, J. P., Nesterenko, P. N., Haddad, P. R., Gambhir, S., Li, D., Wallace, G. G., J. Chromatogr. A 2010, 1217, 7593–7597.
[47] Chang, L-Y., Osawa, E., Barnard, A. S., Nanoscale 2011, 3, 958–962.
C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
[46] Barnard, A. S., J. Mater Chem. 2008, 18, 4038–4041.
www.electrophoresis-journal.com
Received October 7, 2013 Revised March 3, 2014 Accepted March 9, 2014
Electrophoresis 2014, 35, 1864–1872
Research Article
Separation and characterisation of detonation nanodiamond by capillary zone electrophoresis A new method for the characterisation of purified detonation nanodiamond (DND) using CZE has been developed. The influence of BGE conditions on electrophoretic mobility, peak shape and particle aggregation was investigated, with resultant observations supported by zeta potential approximations and particle size measurements. Sodium tetraborate (pH 9.3), Tris (pH 9.3) and sodium phosphate (pH 7) were used in studying the BGE concentration effect on a commercial source of chemically stabilised DND. The BGE concentration had a strong effect on the stability of DND in suspension. The formation of aggregates of various sizes was observed as BGE concentration increased. The effect of pH on the electromigration of DND was examined using sodium phosphate (pH 8 and 10). The CZE method was subsequently applied to four different DND samples, which had undergone different routes of purification following detonation synthesis. Each sample produced a unique electrophoretic peak or profile in sodium tetraborate buffer (pH 9.3), such that the actual separation of DND samples from different sources could be achieved. Keywords: Aggregation / Capillary electrophoresis / Detonation nanodiamond / Size distribution / Zeta potential DOI 10.1002/elps.201300488
Additional supporting information may be found in the online version of this article at the publisher’s web-site
1 Introduction 1.1 General Nanoscale diamond (sp3 carbon) is an increasingly applied member of the nanocarbon family, which exhibits unique properties, originating both from its core lattice structure and its complex surface chemistry [1]. Favourable optical and mechanical properties of nanodiamond (ND), together with their excellent thermal stability and conductivity, and biocompatibility [2] have generated considerable interest in these relatively uncharacterised and chemically variable nanomaterials. Detonation ND (DND) is produced by the detonation of explosives in closed containers, followed by the subsequent collection and thorough purification of the soot to obtain
Correspondence: Professor Brett Paull, Australian Centre for Research on Separation Science, University of Tasmania, Hobart, Tasmania 7001, Australia E-mail: [email protected] Fax: +61-3-6226-2858
Abbreviations: DLS, dynamic light scattering; DND, detonation nanodiamond; DS, detonation soot; ND, nanodiamond; SD-DND, single-digit detonation nanodiamond C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
graphite-free diamond [3]. In terms of its surface properties, DND is a diverse material, in both variety and concentration of functional groups, elemental impurities, zeta ( ) potentials and subsequent propensity to form aggregates, which are highly variable in size. The material is generally considered to be relatively chemical stable, despite the varied surface chemistry, which may include carboxyl, hydroxyl, ketones, ethers, and lactones, together with adsorbed inorganic impurities [1, 4, 5]. This complexity in the surface chemistry of DND is associated with a possible presence of sp2 carbon layers at the outer surface. The exact composition and nature of these surface properties is primarily determined by the methods applied during production and purification. Oxygen-containing groups are usually present on the surface of DND due to the postsynthesis cooling and the purification methods, which usually involve oxidation [6]. The typical diameter of DND produced by detonation synthesis is 4–5 nm [7], however, these particles have a very strong tendency to form aggregates up to several hundreds of nanometres in size. DND suspensions are generally stabilised through naturally occurring surface charge, or via adsorbed charged species. Such
Colour Online: See the article online to view Figs. 2 and 5 in colour.
www.electrophoresis-journal.com
CE and CEC
Electrophoresis 2014, 35, 1864–1872
electrostatic stabilisation allows for measurement of electrophoretic mobility and calculation of potentials. Potentials above +30 mV and below −30 mV are generally observed for such stable suspensions; however, DND can exhibit a very wide distribution of potentials ranging from as low as −100 to +70 mV, often dependent upon the purification procedures applied. A wide variety of applications for DND are currently being explored, including their use within analytical chemistry, specifically chromatography [2, 8–11], their incorporation in nanocomposites and coatings [12], their use in optoelectronics [13], and their emerging application within the field of biomedicine, such as their use within surgical implants and as vehicles for drug delivery [14–17]. However, to further these applications, it is quite clear that there are many challenges to be overcome in understanding the exact properties and behaviour of DND, and in developing new processes to ensure physical and chemical DND homogeneity. To meet both challenges, the development of reliable characterisation methods is essential. A number of techniques for the characterisation of DND have been reported to date. SEM and high-resolution transmission electron microscopy are often used for characterisation of nanoparticle size and morphology. However, these techniques are obviously expensive and time consuming, and in the case of DND, changes in particle aggregation as a result of desolvation present a problem. TOF-MS has also been reported for the analysis of DND samples, and has provided information on size distribution, aggregation and surface chemistry [18]. The ionisation of DND samples resulted in the formation of singly charged carbon clusters in different regions of the mass spectrum providing information on size distributions. Studying electrophoretic parameters has proven useful in the characterisation of DND. For example, dynamic light scattering (DLS) and laser Doppler electrophoresis offer a more rapid method of obtaining information on particle diameters, estimating electrophoretic mobilities and potentials. Petrova et al. [19] used such measurements to investigate the stability of DND in different salt solutions and calculated potentials based on the size of the different particles. Chiganova et al. also looked at the electrophoretic behaviour of DND in increasing concentrations of aluminium chloride, where concentration affected potential as the DND surface became modified with aluminium [20]. CZE offers a potentially rapid, automated and robust method for the characterisation of such properties of DND, and indeed nanoparticles in general. The technique is compatible with a large range of solvent and buffer systems, and has low sample volume requirements, and CZE should be capable of providing information on nanoparticle size distributions, surface charge, dispersion quality and stability. CZE has previously been employed as a method for the size and surface-charge characterisation of a variety of nanoparticles, including metal nanoparticles [7, 21, 22], carbon nanoparticles, such as graphene oxide and chemically converted graphene [23–25] carbon dots [26] and carbon nan C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
1865
otubes [27–33], and a number of reviews have been published on the topic [34–37]. Chang et al. monitored the growth of gold nanoparticles using CZE by simultaneously analysing the electrophoretic profile and the absorption spectra of the samples to obtain information on particle size and distribution [38]. Baker and Col´on used CZE to investigate the influence of buffer conditions on the mobility and fractionation of carbonaceous nanoparticles in a soot sample obtained from the flame of an oil lamp in order to better understand the complex sample [39]. However, to date few studies using CZE for actual DND separations have emerged. Therefore, herein the authors present the first report on the use of CZE for the separation and characterisation of commercial and non-commercial DND. The influence of BGE conditions on the electro-migration and the separation of DND was investigated and used to approximate potentials, and data were supported with particle size measurements using DLS. The effect of pH and increasing concentrations of a variety of common buffers on the electrophoretic profile and particle aggregation of DND samples was explored. Optimum CZE conditions were subsequently applied to a variety of commercial and non-commercial DND samples, which had undergone different routes of purification following detonation synthesis.
1.2 Theoretical background NDs exhibit a wide range of potentials and particle sizes and thus should be treated as a colloidal particle in this instance, rather than a small ion [39]. Therefore, a brief overview of the theory related to the electrophoretic behaviour of charged colloidal particles is presented here. A cloud of ions (referred to as the shear surface) surrounds a charged colloidal particle in a BGE, which has a net charge that determines the electrophoretic mobility of the particle. As the particle migrates in an electric field, it is accompanied by a specific quantity of the surrounding liquid and the charged species, which it contains. The potential is the value of electric potential at the surface of shear (can and usually does include some of the bound counterions) [40], and it is related to the electrophoretic mobility of the particle by the following equation: = (2ε /3) f (a),
(1)
where ε is the permittivity of the medium which has viscosity . This applies to particles with a low potential ( 25 mV at 25°C) where Henry’s function, f(a) ranges from 1 at a 1 (Smoluchowski limit) to 1.5 at a 1 (H¨uckel limit). The inverse Debye–H¨uckel parameter describes the electric double layer thickness and is dependent on the ionic strength (I) of the BGE as follows: √ (2) (nm−1 ) ≈ 3.288 I(mol/L), at 25°C in water [41, 42]. This equation applies to a symmetrical electrolyte; however, it may be used in cases of a nonsymmetrical electrolyte when the majority of counterions populating the diffuse www.electrophoresis-journal.com
1866
E. Duffy et al.
Figure 1. Plot showing the theoretical relationship of mobility as a function of a (product of inverse Debye length and the particle radius) for different values of potential (‘low’ , i.e. 25 mV, and ‘high’ , i.e. >25 mV) [40, 41]. Experimental results are superimposed for comparative purposes and to allow an approximation of potentials for single-digit nanodiamonds in Tris buffer ( ) at concentrations of 10, 50 and 140 mM (pH 9.3), and in sodium phosphate buffer () at concentrations of 10, 20 and 30 mM (pH 7, 8 and 10).
part of the electric double layer are monovalent [41]. Within Henry’s function, a is the hydrodynamic or stokes radius of the particles. The particle radius can actually be substituted for a in most cases, as the distance between solid and shear surfaces generally does not exceed 1 nm [42]. In reality, most nanoparticles lie in the range where a is near to one and they have much higher potentials. In order to account for these factors, it is necessary to apply the Overbeek–Booth theory [40] which accounts for electrophoretic retardation and the relaxation effect, and this was presented numerically for samples with a greater range of 25–150 mV [43, 44]. This is depicted in Fig. 1, where the theoretical dependence of reduced electrophoretic mobility on a (or the ratio of a particle’s hydrodynamic radius to the thickness of the double layer) for both ‘low’ ( = 25 mV) and ‘high’ ( ⬎ 25 mV) values of potential is shown.
2 Materials and methods 2.1 Chemicals and reagents R All solutions were prepared using Milli-Q water (Millipore Gradient System). Commercial DND samples (single-digit detonation nanodiamonds (SD-DND) 20 mL aqueous 5% solution (50 mg/mL)) were purchased from PlasmaChem (Berlin, Germany) in a 50 mg/mL aqueous suspension of chemically stabilised negatively charged NDs, which was later diluted to 2 mg/mL with Milli-Q water for the preparation of samples for injection. Three DND samples, namely NSPA,
C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Electrophoresis 2014, 35, 1864–1872
NSHCl and NSFPA, were prepared in our laboratory from the detonation soot (DS) obtained from the Space and Solar Terrestrial Research Institute at the Bulgarian Academy of Sciences (SSTRI-BAS) (Sofia, Bulgaria) and purified by acid oxidation at the University of Tasmania (Hobart, Australia). Briefly, the NSPA sample was purified from DS using acidic oxidation of DS with a mixture of HNO3 , H2 SO4 and HClO4 acids. The NSHCl sample was treated similarly using a mixture of HNO3 , H2 SO4 and HCl acids, and the NSFPA sample using a mixture HNO3 ·H2 SO4 , HClO4 and HF. Subsequent washing was performed to reduce the digested impurities and residual acids. The Ru-Kr DND sample was obtained from the Institute of Biophysics at the Siberian Division of the Russian Academy of Sciences (Krasnoyarsk, Russia). All DND samples were negatively charged under the pH conditions used (pH 7, 8, 9.3 and 10). Acetone (99.5%) was sourced from Chem-Supply (Gillman, Australia). HCl was purchased from Merck (Darmstadt, Germany) and NaOH from Scharlau (Barcelona, Spain). The reagents for BGE preparation purchased from Sigma-Aldrich were sodium phosphate monobasic (99.0%; St. Louis, MO, USA), sodium phosphate dibasic (99.5%) (Seelze, Germany), sodium tetraborate (99.998%) and Tris (99.9%; Castle Hill, Australia). Amberlite resin IRA-93 (analytical grade) used in preparing DND samples for SEM imaging was purchased from BDH Chemicals (Poole, England). 2.2 Capillary zone electrophoresis Experiments were performed on an Agilent 7100 CE system with detection using a DAD (Agilent Technologies, Germany). Absorption was monitored at 195 and 210 nm at data acquisition rates 40 Hz. Data acquisition was performed using OpenLAB CDS ChemStation Edition software for Windows 7. Data were processed using Origin 9 (OriginLab) software. All reported results were obtained in triplicate. Untreated fused silica capillaries (Polymicro Technologies, Phoenix, AZ, USA) with a 75 m id, 365 m od were used in this study. Detection windows were burned with a butane torch at 8.5 cm from the capillary end with a total capillary length of 66 cm. New capillaries were pre-treated with 0.5 M NaOH for 10 min, followed by H2 O for 10 min, and finally BGE for 10 min. Capillaries were thermostated at 25°C in all experiments. All DND samples contained 0.04 wt% DND dispersed in BGE, and injections were performed by applying a pressure of 5 kPa for 5–20 s, followed by the injection of acetone at 5 kPa for 1 s. The analytes were separated under an applied potential of +15.0 kV. Prior to each run, the capillary was conditioned by flushing H2 O for 3 min, 0.5 M NaOH for 2 min, H2 O for 2 min and BGE for 2 min. Acetone was used as the neutral EOF marker. 2.3 BGE preparation Buffers were prepared by mixing and dilution of sodium phosphate monobasic and sodium phosphate dibasic, and www.electrophoresis-journal.com
Electrophoresis 2014, 35, 1864–1872
by diluting stock solutions of sodium tetraborate and Tris. The pH of buffer solutions was adjusted using dilute HCl and NaOH, and pH values were measured using a model 210 pH meter from Activon (Thornleigh, Australia) and a Mettler Toledo LE409 probe (Port Melbourne, Australia). Tetraborate and Tris buffers were maintained at pH 9.3 in this work. BGEs were degassed with a Soniclean ultrasonic bath and filtered through 0.45 mm disc filters (Activon). 2.4 Particle size measurements Particle size distributions of DND in water and solutions of sodium tetraborate, Tris chloride and sodium phosphate were measured using a ZetaSizer NanoZS (Malvern Instruments, UK). The reported results were obtained in triplicate, and performed at 25°C as in CZE. Measurements were performed using 0.01 wt% suspensions of ND. The concentration and the pH of tetraborate and Tris (pH 9.3) and sodium phosphate (pH 7, 8 and 10) used in this analysis were the same as per CZE experiments. Particle imaging was carried out using field emission high-resolution SEM using a Hitachi Ultra-High Resolution Analytical FE-SEM SU-70 instrument (Hitachi High Technologies America, USA). The DND sample was prepared on Amberlite IRA943 resin (BDH Chemicals) and sputtered with a thin layer of Pt before imaging. Briefly, sample preparation involved adding 0.1 g of Amberlite resin to 1 mL of 0.64% wt. suspension of DNDtype NSHCl, and diluting further with 20 mL Milli-Q water. The mixture was left for 15 h and subsequently washed with Milli-Q water and dried before imaging.
3 Results and discussion 3.1 Buffer chemistry and concentration effects The effect of buffer concentration within the BGE on the peak shapes, apparent particle aggregation points, and electrophoretic mobility of a commercial so-called SD-DND (PlasmaChem) was studied, using sodium phosphate, sodium tetraborate and Tris buffers. Observation of the aggregation properties of certain classes of carbon nanoparticles in various BGE buffers using CZE has been described previously, including for graphene oxide [24] and carbon nanotubes [27]. However, no such study involving DND has been reported to date. In the case of graphene oxide NPs, M¨uller et al. were able to obtain a single broad peak using CZE in BGE buffer concentrations of 2 mM (ammonium acetate) and below, whilst the onset of aggregation and the emergence of multiple random spikes were observed at buffer concentrations of 5 mM and above. Generally, potentials in aqueous colloidal suspensions are a function of the charge at the shear plane and the free ion concentration. Electrostatic potential decays exponentially with distance from the shear plane, and the inverse of this decay constant is called the Debye double layer thickness. Increasing the free ion concentration increases the decay and C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
CE and CEC
1867
so reduces this double layer thickness. At a certain ion concentration, the reduction of the double layer will reach a point whereby its electrostatic repulsion forces are less than underlying attractive forces, such as van der Waals forces, at which point suspension stability will be lost and spontaneous aggregation can occur. However, in the case of DND, a number of recent studies have shone new light upon the nature of particle aggregation within solution, dismissing van der Waals interactions and identifying strong electrostatic forces as being behind the spontaneous self-assembly of ND agglutinates (60 nm) and larger agglomerates (100–200 + nm) [43–47]. Barnard and co-workers classify ND aggregates according to these approximate size distributions, with the former, namely agglutinates, interacting through interfacial type alignment, based upon electrostatic attraction in an ordered configuration. The larger agglomerates are also interfacially aligned via electrostatic attractions, but configured in a more random orientation [47]. Herein, Fig. 2A shows the typical electropherograms obtained following injection of this particular SD-DND sample in tetraborate buffer-based BGEs of increasing concentration, ranging from 10 to 60 mM, each fixed at a pH of 9.3. The negatively charged NDs migrated after the acetone EOF marker, as a single well-defined broad peak for tetraborate buffer concentrations of between 10 and 32.5 mM. The absence of spikes within the broad Gaussian-shaped peak would suggest (10 mM tetraborate BGE) a stable suspension of DND ‘particles’ under such conditions. Under these low-concentration buffer conditions, it can be assumed that the DND is in the form of small agglutinates of a relatively small size distribution, rather than single-digit (⬍10 nm) particles. Such agglutinates are sufficiently small and surface charged to exhibit both stable suspension and migration within the capillary. Indeed, this assumption was supported by average particle size data obtained for the SDDND sample in 10 mM tetraborate buffer, obtained using DLS (see Table 1). Here, an average particle size of 33 ± 4 nm was recorded. Increasing the concentration of the tetraborate BGE to 35 mM and above resulted in the increased appearance of sharp peaks or spikes, and the simultaneous disappearance of the broad peak. This behaviour corresponds closely to that reported earlier for graphene oxide NPs [24], although herein the migration window for both the initial broad peak and the series of sharp spikes remained relatively constant. For example, at 60 mM tetraborate only a series of sharp spikes could be seen, covering a migration time window of 4 min, which was approximately equal to the width of the initial broad peak at 10 mM. The appearance of these spikes in higher concentrations of tetraborate can be attributed to the formation of larger randomly arranged agglomerates within the sample, formed as the repulsion forces between the ND agglutinates are sufficiently reduced to allow spontaneous and random aggregation, to sizes no longer stable in suspension. Figure 1C illustrates this visually, with the SD-DND sample diluted in water (stable suspension) and 60 mM tetraborate buffer solution (settled unstable agglomerates). The formation of www.electrophoresis-journal.com
1868
E. Duffy et al.
Electrophoresis 2014, 35, 1864–1872
Figure 2. Electropherograms generated by single-digit nanodiamonds (SD-DND) samples in different concentrations of (A) – tetraborate buffer (60, 40 and 10 mM) and (B) – Tris buffer (140, 50 and 10 mM) at a fixed pH of 9.3. Conditions: injection of acetone for 1 s and nanodiamond for 20 s with detection by DAD at 195 nm. (C) Image of a stable suspension of SD-DND in water (left) and an aggregated SD-DND sample in 60 mM tetraborate (right) where large aggregated particles have dropped out of suspension.
large agglomerates in higher concentrations of tetraborate buffer (60 mM) was visible to the naked eye, however, for lower concentrations where small agglomerates were beginning to form (20–40 mM), no visible changes to the suspension were noticeable. Once again to confirm the process occurring, particle size data were obtained (Table 1), in this instance for 40 and 60 mM tetraborate solutions, revealing average particle sizes of 868 and 1359 nm, respectively (the
average particle size of SD-DND in 60 mM tetraborate actually increased with each consecutive DLS measurement, as large aggregates formed rapidly in the sample). It appears that the critical (tetraborate, pH 9.3) buffer concentration for aggregation of these particular ND agglutinates into large unstable agglomerates is between 30 and 35 mM. Figure 2B shows the typical electropherograms obtained of the same SD-DND sample, this time separated with a
Table 1. Average particle size data and polydispersity index (PDI) obtained (by dynamic light scattering using number-based size distributions) for single-digit nanodiamond suspensions in H2 O and various concentrations of tetraborate and Tris buffers at pH 9.3 and sodium phosphate pH 7, 8 and 10, and for NSPA, NSHCl, Ru-Kr and NSFPA suspensions in H2 O and 10 mM tetraborate pH 9.3
Sample
BGE
Concentration (mM)
SD-DND
Tetraborate
SD-DND
Tris
SD-DND
Sodium phosphate
10 40 60 10 50 140 10
20
30
NSPA NSHCl Ru-Kr NSFPA
Tetraborate Tetraborate Tetraborate Tetraborate
10 10 10 10
C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
pH 9.3
9.3
7 8 10 7 8 10 7 8 10 9.3 9.3 9.3 9.3
Size (nm)
PDI
33 ± 4 868 ± 75 1359 ± 391 32 ± 5 31 ± 4 30 ± 7 31 ± 3 24 ± 5 18 ± 3 42 ± 4 38 ± 7 20 ± 4 48 ± 5 45 ± 5 200 ± 17 2292 ± 347 1613 ± 120 76 ± 2 Polydisperse
0.218 0.274 0.391 0.234 0.247 0.256 0.292 0.283 0.249 0.187 0.202 0.213 0.249 0.21 0.41 0.225 0.127 0.243
www.electrophoresis-journal.com
Electrophoresis 2014, 35, 1864–1872
CE and CEC
1869
deviations). Particle size measurements also showed stable agglutinates of 23–37 nm in size up to a concentration of 140 mM Tris. Figure 3A shows the electromigration of SD-DND in a sodium phosphate BGE at a fixed pH of 7, at concentrations of 10, 20 and 30 mM. Corresponding potential approximations are shown in Fig. 1 and particle size data are also listed within Table 1. Over these conditions, the SD-DND sample remained stable in solution, with no signs of aggregation. However, particle size data indicated a possible growth in the size of the stable ND agglutinates from 31, to 42, to 48 nm, respectively. This also corresponds to a decrease in effective mobility (see Fig. 3B), and for 30 mM phosphate, a noticeable increase in peak width. In phosphate BGEs above 30 mM, a reduction in sample stability was noted and agglomerate spikes began to appear in 40 mM sodium phosphate at pH 7.
3.2 Buffer pH on the electro-migration of DND
Figure 3. (A) Electropherograms generated by single-digit nanodiamonds (SD-DND) samples in different concentrations of sodium phosphate (10, 20 and 30 mM) at a fixed pH of 7. Conditions: injection of acetone for 1 s and nanodiamond for 20 s with detection by DAD at 210 nm. (B) Effective mobilities (CZE measurements) for SD-DND in different BGEs. Figure legend for the upper plot sodium phosphate pH 7, sodium phosphate pH 8, × sodium phosphate pH 10, tetraborate pH 9.3. Figure legend for the lower plot: Tris pH 9.3.
Tris BGE of increasing concentration. Separations were performed in concentrations ranging from 10 to 140 mM Tris, once more at a fixed pH of 9.3. Here, as with the tetraborate BGE above, the EOF decreased over the increasing concentration range due to the double layer compression. However, here the mobility of the sample peak was far less affected by the change in buffer concentration, and also showed no signs of particle aggregation, even at the highest (140 mM) BGE concentration. A similar characteristic broad peak was observed for the SD-DND sample, of 4-min width, with no spikes due to aggregation visible. The stability of the SDDND sample over the wider concentration range with the Tris buffer, as compared to that of the tetraborate BGE, was concluded to be an ionic strength effect rather than any surface adsorption effects. Figure 1 shows the approximate potentials for SD-DND under these BGE conditions, where potential is observed to decrease with increased ionic strength. As the ionic strength increased, the inverse Debye length and stokes radius also increased, corresponding to the observed reduction in mobility (see Fig. 3B – each data point is the mean of three measurements, error bars = standard C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Table 1 also shows data for SD-DND in sodium phosphate BGE at 10, 20 and 30 mM, but over a range of pH values. For each pH value, a similar trend for reduced mobility as a function of increased buffer concentration was seen (Fig. 3B). However, the effect of pH upon particle size appeared to change as buffer concentration increased. For 10 and 20 mM, a decrease in particle size was observed as buffer pH increased, for example, from 31, to 24, to 18 nm, at 10 mM phosphate BGE concentration for pH 7, 8 and 10, respectively. At 30 mM phosphate, pH 10, the system became unstable and large agglomerates began to form, with particle sizes of ⬎200 nm apparent. The reasons for these observed effects are unclear, although again it is possible that this is predominantly related to ionic strength. The approximate values for potential within Fig. 1 show no obvious trend to ascertain if changes in surface charge could also account for such size variation. The values remain between the trend lines for −51 and −76 mV, with pH 7 and 8 causing more of a change in potential of SD-DND compared to pH 10. Figure 4A shows the electropherograms obtained for the SD-DND sample in 10 mM sodium phosphate BGE and the effect of BGE pH (pH 8 and 10 shown) on the peak shape is evident with the SD-DND peak becoming narrower and peak intensity increasing as the pH increases. Peak widths of 4.5 min at 10 mM phosphate, pH 7, reduce to 3 min at pH 8 and 2.5 min at pH 10. Simultaneously, as mentioned above, average particle size decreases from 31, to 24, to 18 nm. Thus, at 10 mM BGE concentration, it would appear some evidence of peak width correlating with average particle size in the first instance. Similar, although less obvious, observations could be made with peaks shapes observed in 20 mM phosphate BGEs, over the three pH values, although as is clear within Fig. 4B, at pH 10 with this concentration of phosphate BGE, the critical aggregation point seems to have been reached, as the sudden appearance of agglomerate spikes can be seen. Corroborating the unusual particle size www.electrophoresis-journal.com
1870
E. Duffy et al.
Electrophoresis 2014, 35, 1864–1872
Figure 4. (A) Electropherograms generated by single-digit nanodiamonds (SD-DND) samples in 10 mM sodium phosphate at pH 8 and 10, (B) in 20 mM sodium phosphate at pH 8 and 10, (C) in 30 mM sodium phosphate at pH 8 and 10. Conditions: injection of acetone for 1 s and nanodiamond for 20 s with detection by DAD at 210 nm.
data listed within Table 1 for the 30 mM phosphate BGE at pH 10, is Fig. 4C that shows very dramatically the electropherograms for the ND agglomerates of 200 (±17) nm size.
3.3 Analysis of DND samples Based on the above observations on optimal conditions for the SD-DND sample, a variety of DND samples of differing origin were investigated using a 10 mM tetraborate BGE, at pH 9.3. Four DND samples were investigated, namely, NSPA, NSHCl, NSFPA and Ru-Kr. Each sample had undergone differing post-detonation purification, with the result that each had markedly differing surface chemistry, particle size and aggregation tendencies. Obtained under the above conditions, resulting electropherograms are shown in Fig. 5A. Figure 5A clearly shows the four samples could be classified as either containing stable agglutinates (Ru-Kr and NSFPA) or larger unstable agglomerates (NSPA and NSHCl), under such conditions. The absorbance spectra for each of the above four samples and that of the SD-DND sample can be seen within Supporting Information Fig. 1, which clearly shows the similarity between the absorbance spectra of the SD-DND, NSFPA and Ru-Kr samples, whilst the spectra for the aggregated samples are similarly shaped, whilst obviously rather irregular and noisy. Figure 5A(i) and (ii) shows expanded electropherograms of the latter two samples. The above classifications are supported by the average particle size data listed in Table 1 that confirms the very large agglomerated micron-sized particles, of size 1.4 to 1.5 m, for NSHCl and NSPA samples, whilst the Ru-Kr sample was found to be in the order of 76 ± 2 nm. To further investi C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
gate the aggregation of DND, the NSHCl was analysed using field emission high-resolution scanning electron microscopy and the resultant SEM image, taken at high magnification (×100K), is shown in Fig. 5D. Clear agglomerates of the order 500–1500 nm can be seen in this sample. Interestingly, the results from particle size measurements for the NSFPA sample were inconclusive and noted only as a polydisperse material. Upon close inspection of Fig. 5A, the reason for this finding becomes clear, as the emergence of spikes from ND aggregation can just be seen, suggesting the sample is at its critical aggregation point, and contains both agglutinates and agglomerates in solution, hence its polydispersity [38]. The NSPA sample formed large visible aggregates in water and 10 mM tetraborate, whilst the NSHCl sample formed slightly smaller aggregates in 10 mM tetraborate, but formed a stable suspension in water alone. Interestingly, the particle size ranges for these two samples proved to be larger (1945– 2639 nm for NSPA and 1493–1733 nm for NSHCl), which should be reflected in differences in average mobilities for the two samples. Indeed, Fig. 5A(i) and (ii) clearly shows these differences, and highlighted how the CZE method can in fact be used to completely separate two batches of different NDs. This is demonstrated in Fig. 5C wherein samples containing mixtures of two DND samples, either both present as larger agglomerates (e.g. NSPA and SD-DND samples at 60 mM tetraborate BGE, pH 9.3), or where one of which is present as small agglutinates and the other as larger agglomerates (e.g. NSPA and SD-DND, 10 mM tetraborate, pH 9.3), can be successfully separated. This is the first demonstration of CZE to physically separate DND samples obtained from different suppliers, or samples obtained via differing synthetic routes, and could prove a useful approach to help identify both source www.electrophoresis-journal.com
CE and CEC
Electrophoresis 2014, 35, 1864–1872
1871
Figure 5. (A) Electropherograms generated by four different nanodiamond samples: NSPA, NSHCl, Ru-Kr and NSFPA in 10 mM tetraborate (pH 9.3). (A(i)) Magnification of NSPA and (A(ii)) NSHCl showing sharp peaks characteristic of aggregated nanodiamond. (B) Image of the five different detonation nanodiamond samples under investigation (characteristic electrophoretic profile for single-digit nanodiamonds (SD-DND) in 10 mM tetraborate is shown in Fig. 1(A)). (C) Electropherograms showing the separation of single-digit and NSPA nanodiamonds in 60 and 10 mM tetraborate buffer (pH 9.3). Conditions: injection of acetone for 1 s and nanodiamonds for 20 s with detection by DAD at 195 nm. (D) HR-SEM image of DND-type NSHCl on Amberlite IRA943 resin. Magnification: ×100 000.
and nature of unknown DND suspensions, and/or provide information on composite suspensions.
R2 = 0.9999 and an LOD of 0.0125 mg/mL. The corresponding calibration curve and electropherograms can be viewed within Supporting Information Fig. 3.
3.4 Quantitation of DND solutions using CZE
4 Concluding remarks
The increasing interest in medical applications of DND-type materials also requires quantitative assays of DND suspensions and their stability. Therefore, herein, to demonstrate the potential use of CZE to quantify the concentration of DND within solution (as stable agglutinates), both injection and detection linearity were briefly explored. The SD-DND sample was injected in triplicate for 2, 5, 10, 15 and 20 s injections in 10 mM tetraborate, pH 9.3. A plot of injection time versus corrected peak area resulted in a linear correlation of R2 = 0.9998. The resultant calibration curve and corresponding electropherograms can be viewed within Supporting Information Fig. 2. A serial dilution of the SD-DND sample (starting at 0.4 mg/mL) was then carried out to determine both sample concentration linearity and approximate DND detection limits, with a resultant linear correlation of
As shown herein, CZE has been found to offer some unique capabilities in this area. CZE can be a quantitative tool for the characterisation of DND, providing information on sample charge, size, stability and tendency to agglomerate. Coupled with particle size measurements, it can provide a better understanding of the surface properties and dispersion quality of NDs in different buffer systems. CZE offers the possibility of separating different NDs, which could prove highly useful for sample fractionation in the future.
C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
The Authors would like to acknowledge the Australian Research Council for financial support (ARC Discovery Grant DP110102046). The authors have declared no conflict of interest. www.electrophoresis-journal.com
1872
E. Duffy et al.
5 References [1] Krueger, A., J. Mater. Chem. 2008, 18, 1485–1492. [2] Nesterenko, P. N., Haddad, P. R., Anal. Bioanal. Chem. 2010, 396, 205–211.
Electrophoresis 2014, 35, 1864–1872
[25] Takayanagi, T., Morimoto, M., Yabutani, T., Anal. Sci. 2013, 29, 769–771. [26] Hu, Q., Paau, M. C., Zhang, Y., Chan, W., Gong, X. J., Zhang, L., Choi, M. M. F., J. Chromatogr. A 2013, 1304, 234–240.
[4] Paci, J. T., Man, H. B., Saha, B., Ho, D, Schatz, G. C., J. Phys. Chem. C 2013, 117, 17256–17267.
[27] Doorn, S. K., Strano, M. S., O’Connell, M. J., Haroz, E. H., Rialon, K. L., Hauge, R. H., Smalley, R. E., J. Phys. Chem. B 2003, 107, 6063–6069.
[5] Mitev, D. P., Townsend, A. T., Paull, B., Nesterenko, P. N., Carbon 2013, 60, 326–334.
[28] Suarez, B., Simonet, B. M., Cardenas, S., Valcarcel, M., J. Chromatogr. A 2006, 1128, 282–289.
[6] Krueger, A., Lang, D., Adv. Funct. Mater. 2012, 22, 890–906.
[29] Yamamoto, T., Murayama, S., Kato, M., J. Sep. Sci. 2011, 34, 2866–2871.
[7] Mochalin, V. N., Shenderova, O., Ho, D., Gogotsi, Y., Nat. Nanotechnol. 2012, 7, 11–23.
[30] Yamamoto, T., Murakami, Y., Motoyanagi, J., Fukushima, T., Maruyama, S., Kato, M., Anal. Chem. 2009, 81, 7336–7341.
[3] Krueger, A., Adv. Mater. 2008, 20, 2445–2449.
[8] Saini, G., Jensen, D. S., Wiest, L. A., Vail, M. A., Dadson, A., Lee, M. L., Shutthanandan, V., Linford, M. R., Anal. Chem. 2010, 82, 4448–4456. [9] Nesterenko, P. N., Fedyanina, O. N., J. Chromatogr. A 2010, 1217, 498–505. [10] Fedyanina, O. N., Nesterenko, P. N., Russ. J. Phys. Chem. 2011, 85, 1773–1777. [11] Wiest, L. A., Jensen, D. S., Hung, C.-H., Olsen, R. E., Davis, R. C., Vail, M. A., Dadson, A. E., Nesterenko, P. N., Linford, M. R., Anal. Chem. 2011, 83, 5488–5501.
[31] Xiao, H., Zou, H. F., Pan, C. S., Jiang, X. G., Le, X. C., Yang, L., Anal. Chim. Acta 2006, 580, 194–199. [32] Lopez-Pastor, M., Dominguez-Vidal, A., Ayora-Canada, M. J., Simonet, B. M., Lendl, B., Valcarcel, M., Anal. Chem. 2008, 80, 2672–2679. [33] Doorn, S. K., Fields, R. E., Hu, M., Hamon, M. A., Haddon, R. C., Selegue, J. P., Majidi, V., J. Amer. Chem. Soc., 2002, 124, 3169–3174. [34] Surugau, N., Urban, P. L., J. Sep. Sci. 2009, 32, 1889–1906.
[12] Behler, K. D., Stravato, A., Mochalin, V., Korneva, G., Yushin, G., Gogotsi, Y., ACS Nano 2009, 3, 363–369.
[35] Pyell, U., Electrophoresis 2010, 31, 814–831.
[13] Williams, O. A., Semicond. Sci. Technol. 2006, 21, R49–R56.
[36] Radko, S. P., Chrambach, A., Electrophoresis 2002, 23, 1957–1972.
[14] Huang, H., Pierstorff, E., Osawa, E., Ho, D., Nano Lett. 2007, 7, 3305–3314.
[37] Rodriguez, M. A., Armstrong, D. W., J. Chromatogr. B 2004, 800, 7–25.
[15] Schrand, A., Hens, S. A. C., Shenderova, O., Crit. Rev. Solid State Mater. Sci. 2009, 34, 18–74.
[38] Chang, T.-H., Liu, F.-K., Chang, Y.-C., Chu, T.-C., Chromatographia 2008, 67, 723–730.
[16] Xing, Y., Dai, L., Nanomedicine 2009, 4, 207–218.
´ L. A., J. Chromatogr. A 2009, 1216, [39] Baker, J. S., Colon, 9048–9054.
[17] Ho, D., ACS Nano 2009, 3, 3825–3829. ˇ ˜ ´ [18] Houska, J., Panyala, N. R., Pena-M endez, E. M., Havel, J., Rapid Comm. Mass Spectrom. 2009, 23, 1125–1131.
[40] Overbeek, J. T. G., Wiersema, P. H., in: Bier, M. (Ed.), Electrophoresis: Theory, Methods and Applications, Vol. II, Academic Press, New York 1967, pp. 1–52.
[19] Petrova, N., Zhukov, A., Gareeva, F., Koscheev, A., Petrov, I., Shenderova, O., Diam. Relat. Mater. 2012, 30, 62–69.
[41] Radko, S. P., Stastna, M., Chrambach, A., Electrophoresis 2000, 21, 3583–3592.
[20] Chiganova, G. A., Bondar, V. A., Chiganov, A. S., Colloid J. Russ. Acad. 1993, 55, 774–775.
´ F., Varenne, A., Gareil, P., Electrophoresis 2008, [42] d’Orlye, 29, 3768–3778.
[21] Liu, F.-K., Ko, F.-H., Huang, P.-W., Wu, C.-H., Chu, T.-C., J. Chromatogr. A 2005, 1062, 139–145.
[43] Wiersema, P. H., Loeb, A. L., Overbeek J. T. G., J. Colloid Interf. Sci. 1966, 22, 78–99.
[22] Liu, F.-K., Lin, Y.-Y., Wu, C.-H., Anal. Chim. Acta 2005, 528, 249–254.
[44] O’Brien R. W., White L. R., J. Chem. Soc., Faraday Trans. 1978, 74, 1607–1626.
[23] Zhao, J., Chen, G., Zhang, W., Li, P., Wang, L., Yue, Q., Wang, H., Dong, R., Yan, X., Liu, J., Anal. Chem. 2011, 83, 9100–9106.
[45] Huang, H., Dai, L., Wang, D. H., Tan, L-S., Osawa, E., J. Mater Chem. 2008, 18, 1347–1352.
¨ [24] Muller, M. B., Quirino, J. P., Nesterenko, P. N., Haddad, P. R., Gambhir, S., Li, D., Wallace, G. G., J. Chromatogr. A 2010, 1217, 7593–7597.
[47] Chang, L-Y., Osawa, E., Barnard, A. S., Nanoscale 2011, 3, 958–962.
C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
[46] Barnard, A. S., J. Mater Chem. 2008, 18, 4038–4041.
www.electrophoresis-journal.com
