Behavior and Determination of Titanium Dioxide Nanoparticles in

Feb 27, 2014 - Biophysical characterization of functionalized titania nanoparticles and their application in dental adhesives. Jirun Sun , Elijah J. P...
6 downloads 13 Views 875KB Size
Article pubs.acs.org/ac

Behavior and Determination of Titanium Dioxide Nanoparticles in Nitric Acid and River Water by ICP Spectrometry Valerie Geertsen,*,† Michel Tabarant,‡ and Olivier Spalla† †

Commissariat à l’Energie Atomique et aux Energies Alternatives, CEA Saclay, DSM/IRAMIS/NIMBE/LIONS, CNRS UMR 3299, 91191 Gif Sur Yvette, France ‡ Commissariat à l’Energie Atomique et aux Energies Alternatives, CEA Saclay, DEN/DANS/DPC/SEARS/LISL, 91191 Gif Sur Yvette, France S Supporting Information *

ABSTRACT: ICP spectrometry (ICPMS, ICPOES) are classical techniques for the determination of solubilized or suspended elements. Unfortunately, their relevance for nanoparticles at low concentration (below 10 ppm) is rarely called into question, even if literature reports are not always coherent. This work is a systematic study based on the measurement of TiO2 nanoparticle suspensions, as a model of quasi-insoluble material, by plasma spectrometry. It studies both sample treatment and measurement in the 10 ppb to 30 ppm concentration range. Realized on a set of four engineered nanoparticles suspensions at low concentration, it shows the existence of three different regimes of stability that affect concentration measurement. Above a CS stability concentration value, suspensions are stable in time; below a low-concentration CE value, the signal loss is at a maximum, and a final partition is reached between the container walls and the suspension. Between these two regimes, the suspension aging varies with concentration. CE and CS depend on nanoparticle characteristics and the suspension medium, whereas the evolution kinetic is volume-dependent. Because TiO2 nanoparticles are present in the environment at concentrationd below CS, it is then necessary to find a way to rehomogenize the suspension between sampling and analyzing. Soft sonication, minimizing the sample temperature, and trapping of free radicals is proposed and evaluated. Homogenization is traced by the addition of an internal standard before storage. The procedure is applied to a real sample, Seine River water. The amount of total titanium found, 48.7 ppb, is in good agreement with the result of the reference method.

T

determination of exposure concentrations in the environment. This work provides a means to achieve this last point. Focused on TiO2 nanoparticles as a model of quasi-insoluble materials, it describes the required controls and the resulting implications for accurate determinations. Inductively coupled plasma spectrometers (ICPOES and ICPMS) are widely used for quantitative element determination in solution but also increasingly for that in suspension.6−8 ICP spectrometers can be used alone for direct quantification of suspended nanoparticles, as in the single particle mode,9 but also as separation technique detectors (LC−ICPMS, FFF−ICPMS,2,10 etc.). Quantitative determination of suspended TiO2 nanoparticles has already been broached by plasma spectrometry. In 2004, the determination of impurities in ground Standard Reference Materials (SRM) grade TiO2 powder (0−15 μm) and homemade nanometer particles (0−100 nm) in suspension

he last two decades have witnessed the emergence of a very large variety of engineered nanomaterials. Among those carrying the great expectations of nanotechnologies, some are new versions of materials already having dedicated applications, and others are more innovative and in a stage where massive applications are not yet in place. Oxide nanoparticles (e.g., SiO2, TiO2, CeO2) are among the first group. They were essentially chosen because of the enhanced reactivity due to their small size.1 They show applications in various fields, such as food and tire reinforcement for SiO2, soot degradation for CeO2, skin solar protection,2 and building photo catalytic coatings for TiO2. Their potential impact on the living world, 3 their dissemination,4 and the ultimate fate in the environment5 has become a social concern, although whatever transport medium and assumptions in modelization, the anticipated concentrations are relatively low6 (in the range of a few parts per billion). Because the engineered nanomaterials are designed to exhibit important chemical activities compared with larger particles, there is, despite these low concentrations, a need for risk analysis, which requires ecotoxicity studies and quantitative © 2014 American Chemical Society

Received: December 3, 2013 Accepted: February 27, 2014 Published: February 27, 2014 3453

dx.doi.org/10.1021/ac403926r | Anal. Chem. 2014, 86, 3453−3460

Analytical Chemistry

3454

>99.9

rutile powder; hydrophilic coating: Al2O3 6%, glycerin 1% Specpure Ti IV oxide powder UV-Titan M212 Sachtleben

Johnson Mattey Chemicals, Royston, UK

anatase powder PC105 Crystal Global

Composition, crystallite size, purity, and surface area were provided by the suppliers. Volume size distributions with polydispersity indexes (Pdl) and ζ potentials were measured by DLS at different suspension ages in pH 1.5 HNO3, [Ti] = 30 ppm.

60

Anatase powder

20 nm

90

90 95

>250 91 1 L) were sampled and transferred into PFA bottles in the laboratory, where they were rapidly filtered on 0.22 μm cellulose. Filters were calcinated in the oven. The fusion was then performed as described above.

was of analytical grade (Merck), and isopropyl alcohol was from Sigma Aldrich. Seine River water used for the preparation of commercial nanoparticles suspension in river water medium was sampled in Paris near the Tolbiac bridge and filtered on a 0.22 μm filter before use. Vials, jars, and Erlenmeyers were cleaned according to the procedure proposed in the Supporting Information (SI1). Sonication was performed in a bath (Elmasonic P30H model). The samples were regularly moved in the bath for a better homogeneity of sonication. Dynamic Light Scattering. Nanoparticle size distribution and ζ-potentials were measured on TiO2 suspensions by Dynamic Light Scattering (DLS, Nano Zetasizer, Malvern Instruments). The suspensions were not sonicated prior to measurement, but they were filtered (1 mm, cellulose) to eliminate dust. Size results are reported as volume distribution. This is most consistent for an ICP spectrometer determination, as opposed to intensity or number distribution. The volume distribution implicitly assumes spherical nanoparticles. ICPOES and ICPMS Analysis. All concentrations refer to the amount of titanium (parts per billion) whether it is TiO2 nanoparticles suspensions or Ti in solution. Titanium determination was performed by ICP OES (Optima 2000DV, Perkin-Elmer, Uberlingen, Germany) on an apparatus equipped with a slurry nebulizer (Glass Expansion, Melbourne, Australia) and a glass cyclonic chamber. For direct determinations of nanoparticle suspensions, a multiposition, motorless magnetic stirrer (Variomag, ThermoScientific) was placed on the autosampler to agitate sample volumes up to 1 L. This setup ensures controlled and identical operating conditions for all the vials (identical vials, identical sample levels in vials but also identical autosampler probe aspiration positions in the suspension, identical stirring bars and speed, etc.). Two emission wavelengths were selected for the titanium measurements: 334.950 nm for concentrations lower than 10 ppm and 323.657 nm for those between 10 and 100 ppm. Titanium determination was also performed by quadrupole ICPMS (X series, Thermo Elemental). The instrument was equipped with a slurry nebulizer and an impact ball nebulization chamber. Titanium was measured on masses of 47, 49, and 50. Mass 50 was used for quantitative determination in the river water. 59Co present at a very low level in the Seine River17 was used when noted as the internal standard. For Seine River measurements, blank river water samples were obtained by filtration of river water on a 0.22 μm filter after acidification by nitric acid or aqua regia. The amount of titanium in solution in the Seine River measured elsewhere17 reached 60± 10 ppt, which was too low to be measured by the instrument. The signal observed on the blank river water sample at the mass 50 was due to the interference 13C37Cl. Blank samples were also used to prepare engineered commercial nanoparticle suspensions in river water medium. Repeatability was calculated for both ICPOES and ICPMS as the relative standard deviation (RSD) on six consecutive replicates of the same sample. Intervial reproducibility was evaluated on several suspensions of the same age. Age reproducibility was calculated by measuring the same sample at different times. Fusion Dissolution. Titanium dioxide dissolution performed here consisted of an adaptation of the dissolution procedure proposed by Radhamani.18 A polyphosphate flux was first prepared in a dry platinum crucible by mixing 3 g of sodium dihydrogen phosphate monohydrate with 3 g of



RESULTS AND DISCUSSION Nanoparticle Suspension Characterization. The four engineered commercial nanoparticles (Ana-120, Ru-104, Ana101 and Ana-102) were characterized in suspensions by DLS. These experiments were realized at very high concentrations (30 ppm in Ti or 50 ppm in TiO2) because of the low sensitivity of the technique. Suspensions were prepared in pH 1.5 nitric acid medium to impose a pH far from the isoelectric points (pH 6−8) of the TiO2 nanoparticles. The surface of the TiO2 nanoparticles was positively charged, and the suspension stabilization was favored.16 The different ζ potentials are shown in Table 1. They range between +30 and +36 mV for all the nanoparticles, highlighting the suspensions’ stabilities, except for the Ana-120, which is neutral at the pH value chosen. This difference is very probably due to the presence of a dispersant in the Ana-120 commercial suspension covering the nanoparticle surface and ensuring the dispersion. DLS volume distributions are also shown at different suspensions ages (0, 2, and 24 h) in Table 1 and in the Supporting Information (Figure S1). Ana-120 shows very low aggregation, with an average aggregate diameter of 25 nm, the suspension having been stable for 24 h. When suspended (t = 0), Ru-104 showed two aggregate populations centered on 150 nm and a few micrometers, respectively. The huge aggregates were present in the same proportion after 2 h, but clearly formed a deposit after 24 h, allowing a population of very small aggregates centered at 46 nm in diameter to appear. Both Ana102 and Ana-101 show two aggregate populations centered around 700 and 5000 nm in diameter. These distributions evolved little with time. Suspension Behavior in pH 1.5 Nitric Acid Medium. The first concentration measurements were carried out with an ICPOES on suspensions containing one or the other of the nanoparticles at various concentrations (10 ppb to 30 ppm) and ages (5 min to several hours). Whichever the nanoparticles tested, preliminary results showed poor reproducibility, especially at low concentration. They also showed uncontrolled and irregular evolution with age. More precisely, ICP signal intensities decreased with suspension age in most of the suspensions studied. Table 2 gives an illustration of the phenomenon with Ru-104. A suspension was distributed over several identical vials, and the titanium signal was measured in each vial just after preparation and 3 h later. In this example, signal repeatabilities and intervial reproducibilities were in the range of ICP performances. However, during the 3 h of aging, the average Ti signal intensity decreased by 37%. It should be pointed out that these observations seem to be in contradiction 3455

dx.doi.org/10.1021/ac403926r | Anal. Chem. 2014, 86, 3453−3460

Analytical Chemistry

Article

vials show exactly the same results; thus, the only influencing parameters here are the suspension volume and age. To go further into detail, a new suspension of Ru-104 was prepared and divided into PFA jars of different volumes (7 to 1000 mL) and analyzed through time under constant magnetic stirring to minimize possible sedimentation. (Figure 2). Signal

Table 2. Influence of Sonication and Aging on Signal Repeatability and Reproducibilitya suspension age