The detection and quantification of nanoparticles is a complex issue due to the need to combine “classical” identification and quantification of the constituent material, with the accurate determination of the size of submicrometer objects, usually well below the optical diffraction limit. In this work, the authors show that one of the most used analytical methods for silver nanoparticles, asymmetric flow field-flow fractionation, can be strongly influenced by the presence of dissolved organic matter (such as alginate) and lead to potentially misleading results. The authors explain the anomalies in the separation process and show a very general way forward based on the combination of size separation and size measurement techniques. This combination of techniques results in more robust AF4-based methods for the sizing of silver nanoparticles in environmental conditions and could be generally applied to the sizing of nanoparticles in complex matrices.

Nanotechnology-based products are increasingly being made available to consumers in different application areas, ranging from paint and inks to cosmetics and medicines. In particular, silver nanoparticles (AgNP) are used in consumer products due to their antimicrobial properties.1 Products containing silver nanoparticles have been shown to release AgNP during their normal use2,3 and ultimately end up in the environment. Nanoparticles can be difficult to characterize due to the need to both identify (and quantify) the constituent material(s) and to measure their size which is usually well below the optical diffraction limit. Due to the increased relevance of nanotechnology-based products on the market, the European Commission has proposed a definition of what constitutes a “nanomaterial.” This definition is based on the determination of the number-based particle size distribution (PSD).4 The determination of the particle size distribution of nanometer sized objects is even more difficult when measuring engineered nanoparticles in complex matrices5 and especially nanoparticles suspended in environmental matrices that may contain several other ingredients such as dissolved organic matter.

Asymmetric flow field flow fractionation (AF4) is a separation technique which can fractionate liquid dispersed particles as a function of their size. The technique is widely used to in the separation and characterization of nanoparticles,6 especially in complex matrices such as environmental and biological matrices. In ideal cases, it is possible to obtain the particle size distribution of unknown samples by converting the AF4 retention time to size by either using the AF4 theory or by performing a size calibration with nanoparticle standards of known size7,8 or to determine the PSD of AgNP commercial products in simple matrices.9 

Alginic acid, also called algin or alginate, is an industrially relevant anionic polysaccharide which is commonly found in the cell walls of brown algae and in biofilms produced by some bacteria. This nonbranched molecule can be used as a model of dissolved environmental organic matter which is known to play a critical role in the behavior and environmental fate of AgNP.10 

As shown before,11 the behavior of AgNP in water is modified in the presence of dissolved organic matter such alginic acid and humic acid, another common component of dissolved organic matter in water. It was found that these two materials, although chemically distinct, produced similar changes in particle behavior.

In this work, we show that the presence of alginic acid can strongly influence the AF4 separation of AgNP and lead to errors when attempting to size AgNP based on AF4 retention time. We address the origin of the phenomenon and show a very general solution based on the combination of the AF4 separation with the online size measurement.

High purity sucrose and dodecane were used in centrifugal liquid sedimentation (CLS) analysis and ammonium carbonate and NaOH for preparation of the AF4 eluent. All these reagents were sourced from Sigma-Aldrich and were of analytical grade purity or better. Low viscosity alginate sodium salt (Sigma-Aldrich) was used as a model for dissolved organic matter. The alginate stock solutions used were freshly prepared daily in MilliQ water (Millipore Advantage System, Merck Millipore) without further purification. Citrate stabilized AgNPs with nominal size of 60 nm and concentration of 20 μg/ml were purchased from Sigma-Aldrich and stored away from light at 4 °C in airtight glass vials.

Aliquots of the AgNP stock solution were diluted in MilliQ water from the nominal concentration of 20 μg/ml to a final concentration of either 0.5 μg/ml for AF4 analysis or 5 μg/ml, for CLS measurements. For AgNP-alginate complexed samples analysis, alginate was added to a solution containing 0.5 or 5 μg/ml AgNP, for CLS and AF4 analyses, respectively, at final concentrations ranging from 2 to 4 μg/ml. Alginate stock solution was prepared in MilliQ water at a concentration of 10 μg/ml and immediately mixed with the silver nanoparticles at concentrations ranging from 2 to 4 μg/ml. Mixtures were equilibrated at RT, with constant agitation for 1 h before analysis.

The size of the as-supplied AgNPs and their complexes with alginate were measured with dynamic light scattering (DLS) and CLS instruments.

A Zetasizer model Nano-ZS instrument by Malvern was used to perform DLS particle size measurements. Batch DLS measurements were performed in PMMA disposable cuvettes using a backscatter reading angle (173°) while measurements under flow conditions were done using a Hellma Quartz Suprasil 3 mm flow-through cuvette adjusted to 3.90 mm measurement position and attenuator of 11. The same instrument was used to measure the zeta potential of particle solution using disposable capillary cells. In all cases, the sample cell temperature was set to 25 °C.

CLS measurements were performed with a disk centrifuge photosedimentometer DC2400UHR by CPS Instruments, Inc. The instrument was operated at 22 000 prm and samples were injected into an 8%–24% sucrose gradient.

AF4 analysis was performed in an AF2000 MT Multiflow FFF system with an on-line UV-Vis detector (Postnova Analytics). The chosen elution protocol was based on the method described by Geiss7 for separation of different sized AgNP. The AF4 channel had a 280 mm long separation channel, with a 350 μm spacer. A 10 kDa cut off membrane of regenerated cellulose and a 100 μl injection loop were used. Two types of elution buffer were used for AF4 analysis: low ionic strength solution of MQ water adjusted to pH 9.7 (50 μM NaOH) and higher ionic strength solutions containing various concentrations of ammonium carbonate (1–0.1 mM) and pH 9.1. Elution buffers were freshly prepared in MilliQ water and degassed in an ultrasonic bath before use. All samples were analyzed under the following elution conditions: 0.5 ml/min injection flow; 0.2 ml/min tip flow for 5 min; 1.3 ml/min focus flow; and a linear decrease of the cross flow from 1 to 0.1 ml/min over 35 min. The UV detector wavelength was set to 430 nm, corresponding to the maximum of the surface plasmon resonance band for 60 nm AgNP.

To assess the impact of environmentally relevant materials on the AF4 separation profile of AgNP, we have studied the separation of AgNP in the presence of alginic acid. Alginic acid (Fig. 1) is an anionic polysaccharide formed by units of 1–4 linked b-d mannuronic acid and a-l gluconic acid with a pKa around 3.5 and thus negatively charged at neutral pH. It is widely present in water environment and has a key role in the formation of biofilms that protect algae and bacteria from adverse environmental conditions.12 

Fig. 1.

Chemical structure of alginate.

Fig. 1.

Chemical structure of alginate.

Close modal

When citrate stabilized AgNP are mixed with low amounts of alginic acid, they form stable complexes with no sign of aggregation, as shown by the lack of any large aggregation peak in the DLS data measured in batch mode [Fig. 2(b)]. Figure 2(a) shows the CLS data for AgNP free and in the presence of increasing amounts of alginic acid.

Fig. 2.

Characterization of AgNP-alginate complexes. (a) CLS spectra for 60 nm AgNP, free (red), incubated with 2 μg/ml of alginic acid (green), and 4 μg/ml alginic acid (blue). (b) DLS data of the same samples.

Fig. 2.

Characterization of AgNP-alginate complexes. (a) CLS spectra for 60 nm AgNP, free (red), incubated with 2 μg/ml of alginic acid (green), and 4 μg/ml alginic acid (blue). (b) DLS data of the same samples.

Close modal

The CLS data shown in Fig. 2(a) report the detector signal intensity versus time needed for each sample to reach the detector under the influence of the centrifugal field. These raw data are usually converted to the “standard” intensity versus size plot13 (shown in supplementary material, Fig. S1),18 provided that the density of the sample is known. As it is evident from Fig. 2(a), the time needed for AgNP-alginate samples to reach the detector is longer than that of free AgNP. In addition, the CLS peaks of AgNP-alginate samples have a slightly wider distribution when compared to free AgNP. In centrifugal liquid sedimentation, the movement of particles inside the centrifugal field depends (in first approximation) on the particle size and density. Thus, the increased time needed to reach the detector for the AgNP-alginate samples can be due to a decrease in the size and/or in the density of the particles. The dynamic light scattering measurements of the different samples are shown in Fig. 2(b) and Table S1 (supplementary material). The data indicate that the size of AgNP-alginate samples are indistinguishable (in the experimental error typical of batch-mode DLS) from the free AgNP and there are no large aggregates. In fact, the DLS intensity size distribution is extremely sensitive to large particles and the presence of even 1% large agglomerates would result in a clear peak in the DLS particle size distribution. We also measured the zeta (ζ) potential of the different samples to check if the addition of the alginic acid changed the stability of silver nanoparticles colloid suspension. The results (reported in Table S2, supplementary material) show that the ζ-potential of AgNP-alginate samples are highly negatively charged (−50 and −48 mV following the addition of 2 and 4 μg/ml of alginate, respectively), thus confirming the colloidal stability of the systems.

The CLS data indicate the presence of an alginate layer around silver nanoparticles, thus forming AgNP-alginate complexes. In fact, the size of AgNP-alginate sample does not significantly change (AF4-DLS data of Fig. 4 suggest only a very small size increase) compared to free AgNP, while the layer of alginic acid (density 0.9976 g/ml) lowers the overall density of the AgNP-alginate complex. Thus, the lower density particles need longer time to reach the CLS detector [Fig. 2(a)].

Flow field flow fractionation can very efficiently separate complex mixtures of nanoparticles and the retention time from the AF4 separation channel can be used to estimate the size of the different components of polydispersed samples. Figure 3(a) shows the AF4 fractogram of a mixture of AgNP of 20, 40, 60, and 100 nm. The retention time (Tr = Tpeak – Tvoid) of each particle is a function of the size of the particles as shown in Fig. S2 (supplementary material). Fitting of the experimental data with power function of the type Size = a × TRb results in the following equation:

Size=0.595×TR1.637.
(1)
Fig. 3.

AF4 separation of AgNP and AgNP-alginate complexes. The cross flow program used is reported as a dotted red line on the right scale. (a) Mixture of AgNP 20, 40, 60, and 100 nm. (b) AgNP 60 nm free (black) and AgNP-alginate sample with 5 μg/ml AgNP and 2 μg/ml alginate (red).

Fig. 3.

AF4 separation of AgNP and AgNP-alginate complexes. The cross flow program used is reported as a dotted red line on the right scale. (a) Mixture of AgNP 20, 40, 60, and 100 nm. (b) AgNP 60 nm free (black) and AgNP-alginate sample with 5 μg/ml AgNP and 2 μg/ml alginate (red).

Close modal

With a R2 of 0.992. Tr is the retention time of each peak, and “size” is the diameter of the particle.

The AF4 separation of free AgNP 60 nm [Fig. 3(b), black curve] gives a retention time of 16.9 min that, using Eq. (1), results in a measured size of 60.8 nm, well in agreement with expectations. Repeating the same separation for the AgNP-alginate gives a quite different AF4 fractogram [Fig. 3(b), red line]. Using Eq. (1), the measured retention time of 9.55 min translates to a size of 23.9 nm for the AgNP-alginate complex, which is much smaller than the size of 60 nm for the starting free AgNP.

The smaller calculated size for AgNP-alginate complex could be due to AgNP oxidation and release of ionic silver with subsequent reduction in size (even if it seems unlikely) or to an anomalous AF4 separation process. In fact, there are reports in the literature that charge repulsive interaction of particles with one another and also with the AF4 semipermeable membrane lead to a decrease in the apparent hydrodynamic size.14 

An elegant and robust solution for the accurate measurement of the size of AgNP-alginate complex would be to couple the AF4 separation with an online size measurement technique. Two of the most used online systems for achieving this are multiangle light scattering (MALS) and DLS. Both techniques have advantages and disadvantages, and when combined together, they also give additional information on the geometry of the particles (shape factor) in addition to their size.15 In the case of silver nanoparticles only DLS can be used to sizing, as MALS does not give correct results in the case of plasmonic systems such as silver and gold nanoparticles.

Figure 4 shows the results of AF4-DLS measurement for free AgNP and AgNP-alginate complexes obtained by coupling the output from the AF4 system with a flow cell in the DLS instrument. The online measurement of the hydrodynamic diameter (Z-average, measured at the maximum intensity of the UV-Vis detector) gives a value of 71 nm for the AgNP-alginate complex [Fig. 4(b)] and 69 nm for free AgNP [Fig. 4(a)]. The increase in size from the free AgNP to the AgNP-alginate sample is in the experimental error of the AF4-DLS measurements.

Fig. 4.

AF4-DLS fractograms of (a) AgNP free and (b) AgNP-alginate. The intensity of the detector at 430 nm is reported on the right hand scale and the Z-average of the particles on the left scale for the AgNP-alginate complex (500 ng/ml Ag: 2 μg/ml alginate), in red, and the free AgNP (500 ng/ml), in black.

Fig. 4.

AF4-DLS fractograms of (a) AgNP free and (b) AgNP-alginate. The intensity of the detector at 430 nm is reported on the right hand scale and the Z-average of the particles on the left scale for the AgNP-alginate complex (500 ng/ml Ag: 2 μg/ml alginate), in red, and the free AgNP (500 ng/ml), in black.

Close modal

The results show that the size of the AgNP complex is much closer to the size of the free AgNP than the results obtained by using the AF4 retention time and suggest that the AF4 separation is somehow distorted by charge repulsion forces.

The results show that the AF4-DLS measurements can provide the accurate size (and even the particle size distribution) for silver nanoparticles in complex with alginate without any major optimization in the experimental parameters used for the AF4 separation. This method does not require any calibration with well-defined size standards and is independent of the different electrostatic forces that can affect the AF4 separation process when the nanoparticles surface is modified by either chemical functionalization, attachment of ligands (as in the above case with formation of AgNP-alginate complexes), or coating of the AF4 membrane (for example, by some components of environmental matrices). The main limitation of the AF4-DLS method is related to the sensitivity of the DLS online detector. Due to the dependency of the DLS scattering intensity on the sixth power of the particle diameter, small particles tend to give quite lower intensities compared to larger ones, and thus, the size measurement with online DLS can become unreliable.

To experimentally reduce the charge repulsion forces leading to the anomalous AF4 separation, we repeated some of the AF4 measurements with different elution buffers. In fact, one of the most general approaches for reducing electrostatic forces between analytes and charged membranes is to increase the ionic strength by the addition of salts or buffers.8,16,17 We have tested this by using different concentrations of ammonium carbonate, which provides both a basic pH between 8.5 and 9.0 and high ionic strength. Figure 5 shows the results of the AF4 separation of the AgNP-alginate sample in the presence of increasing concentrations of ammonium carbonate in the carrier solvent. As shown previously, with no ammonium carbonate in the solvent, AgNP-alginate complex exits much earlier (Fig. 5, blue trace) than expected (Fig. 5, free AgNP, cyan trace).

Fig. 5.

AF4 fractograms of AgNP free (cyan) and AgNP-alginate complexes with variable concentrations of ammonium carbonate: 0 (dark blue), 0.1 mM (red), 0.2 mM (violet), 0.5 mM (orange), and 1 mM (green).

Fig. 5.

AF4 fractograms of AgNP free (cyan) and AgNP-alginate complexes with variable concentrations of ammonium carbonate: 0 (dark blue), 0.1 mM (red), 0.2 mM (violet), 0.5 mM (orange), and 1 mM (green).

Close modal

When some ionic buffer is added the retention time from the AF4 tends to be much closer to that obtained for free AgNP; with 0.2 mM carbonate buffer the retention time of free AgNP and AgNP-alginate samples are almost the same. Higher concentrations of carbonate lead to an increase in the retention time and probably cause also some additional loss of sample during the AF4 separation as shown by the reduced area of the AF4 peaks (Fig. 5, orange and green traces for 0.5 and 1 mM carbonate).

Flow field flow fractionation is a very powerful and flexible method to separate nanoparticles and it can be used to estimate the size of silver nanoparticles using calibration techniques with samples of known size. Our results show that this method has to be used with the great care when the calibrating nanoparticles have different surface properties. In fact, the formation of silver-NP-alginate complexes (in conditions mimicking environmentally relevant conditions) can lead to quite misleading results due to electrostatic repulsion forces. In particular, this is very important when silver nanoparticles form complexes with dissolved organic matter, such as alginic acid. Similar considerations will be probably also very important in the case of AgNP functionalized with different ligands which may alter the surface properties of the particles which, in turn, modify the elution time by changing the interactions with the membrane.

The coupling of AF4 with a sizing technique, such as DLS, provides a more general and robust method for the sizing of AgNP in environmental conditions, as we have shown for AgNP-alginate systems. The AF4-DLS method is quite general and easy to set up and should be possible to use it for different types of nanoparticles forming complexes with a wide range of molecules.

1.
C.
Levard
,
E. M.
Hotze
,
G. V.
Lowry
, and
G. E.
Brown
,
Environ. Sci. Technol.
46
,
6900
(
2012
).
2.
T. M.
Benn
and
P.
Westerhoff
,
Environ. Sci. Technol.
42
,
4133
(
2008
).
3.
R.
Kaegi
,
B.
Sinnet
,
S.
Zuleeg
,
H.
Hagendorfer
,
E.
Mueller
,
R.
Vonbank
,
M.
Boller
, and
M.
Burkhardt
,
Environ. Pollut.
158
,
2900
(
2010
).
4.
Regulation (EU) No 1169/2011 of the European Parliament and of the Council (25 October
2011
).
5.
L.
Calzolai
,
D.
Gilliland
, and
F.
Rossi
,
Food Addit. Contam.
29
,
1183
(
2012
).
6.
L.
Calzolai
,
D.
Gilliland
,
C. P.
Garcia
, and
F.
Rossi
,
J. Chromatogr., A
1218
,
4234
(
2011
).
7.
O.
Geiss
,
C.
Cascio
,
D.
Gilliland
,
F.
Franchini
, and
J.
Barrero-Moreno
,
J. Chromatogr., A
1321
,
100
(
2013
).
8.
K.
Loeschner
,
J.
Navratilova
,
S.
Legros
,
S.
Wagner
,
R.
Grombe
,
J.
Snell
,
F.
von der Kammer
, and
E. H.
Larsen
,
J. Chromatogr., A
1272
,
116
(
2013
).
9.
C.
Cascio
,
O.
Geiss
,
F.
Franchini
,
I.
Ojea-Jimenez
,
F.
Rossi
,
D.
Gilliland
, and
L.
Calzolai
,
J. Anal. At. Spectrom.
30
,
1255
(
2015
).
10.
G. R.
Aiken
,
H.
Hsu-Kim
, and
J. N.
Ryan
,
Environ. Sci. Technol.
45
,
3196
(
2011
).
11.
D. C.
António
,
C.
Cascio
,
Ž.
Jakšić
,
D.
Jurašin
,
D. M.
Lyons
,
A. J. A.
Nogueira
,
F.
Rossi
, and
L.
Calzolai
,
Mar. Environ. Res.
111
,
162
(
2015
).
12.
A.
Boyd
and
A. M.
Chakrabarty
,
J. Ind. Microbiol.
15
,
162
(
1995
).
13.
R.
Capomaccio
,
I.
Ojea Jimenez
,
P.
Colpo
,
D.
Gilliland
,
G.
Ceccone
,
F.
Rossi
, and
L.
Calzolai
,
Nanoscale
7
,
17653
(
2015
).
14.
N. M.
Thang
,
H.
Geckeis
,
J. I.
Kim
, and
H. P.
Beck
,
Colloids Surf., A
181
,
289
(
2001
).
15.
P.
Iavicoli
,
P.
Urbán
,
A.
Bella
,
M. G.
Ryadnov
,
F.
Rossi
, and
L.
Calzolai
,
J. Chromatogr., A
1422
,
260
(
2015
).
16.
K.-G.
Wahlund
,
J. Chromatogr., A
1287
,
97
(
2013
).
17.
J. R.
Runyon
,
M.
Ulmius
, and
L.
Nilsson
,
Colloids Surf., A
442
,
25
(
2014
).
18.
See supplementary material at http://dx.doi.org/10.1116/1.4972112 for CLS plot, DLS data in batch mode and zeta potential values.

Supplementary Material