Nanoparticle Dissolution from the Particle Perspective: Insights from

Measured PSDs show interesting changes during the dissolution process, including a change in modality (mono to multi) with time. Although there may be...
0 downloads 0 Views 2MB Size
pubs.acs.org/Langmuir © 2010 American Chemical Society

Nanoparticle Dissolution from the Particle Perspective: Insights from Particle Sizing Measurements Sherrie Elzey† and Vicki H. Grassian*,†,‡ †

Department of Chemical and Biochemical Engineering, and ‡Department of Chemistry, University of Iowa, Iowa City, Iowa 52242 Received May 13, 2010. Revised Manuscript Received June 15, 2010

In this study, the dissolution of copper nanoparticles in aqueous low-pH suspensions is examined. The dissolution phenomenon is examined using both bulk measurements of copper ion production, as detected by inductively coupled plasma-optical emission spectroscopy (ICP/OES), and a decrease in nanoparticle size using particle-sizing instruments. For size measurements, an electrospray atomizer coupled to a scanning mobility particle sizer (ES-SMPS) was used to monitor changes in the particle size distribution (PSD) of the copper nanoparticles as they dissolved in hydrochloric acid solution in real time. Measured PSDs show interesting changes during the dissolution process, including a change in modality (mono to multi) with time. Although there may be several causes for the observed modality changes upon dissolution, it is clear that only through direct measurements of nanoparticles and nanoparticle PSDs can these dynamic details be captured as these particles change size, thus providing important insights into nanoscale processes.

Introduction The recent explosion in applications of nanomaterials requires proper methods of characterization and analysis to facilitate advancements in nanoscience and nanotechnology.1,2 Real-time detection of nanoparticles in complex environments or as they undergo temporal changes is important for many reasons, including environmental chemistry (monitoring nanoparticles in air and water), materials processing (montoring nanoparticle growth during synthesis), and in vivo (monitoring nanoparticle drugdelivery systems). In this study, we investigate the dissolution of metal-containing nanoparticles in acidic aqueous suspensions and probe changes in nanoparticle size during the dissolution process. In particular, the acid-assisted dissolution of copper nanoparticles (CuNPs) in hydrochloric acid is monitored by measuring both the production of copper ions in solution as a function of time and, more importantly, by measuring changes in particle size as the CuNPs dissolve. This latter measurement is made using an electrospray atomizer coupled with a scanning mobility particle sizer (ES-SMPS) for real-time monitoring of particles as they decrease in size. We show here that the particle size distribution changes in a complex manner and becomes multimodal as dissolution occurs. Although ES-SMPS is a technique that has been used to characterize aerosol particles and is more recently used to characterize colloidal suspensions,3-5 this is the first time that the technique has been applied to investigate and monitor nanoparticle dissolution. Other techniques used to monitor dissolution rely on light scattering. These techniques are particularly *Author to whom correspondence should be addressed. E-mail: vicki-grassian@ uiowa.edu. (1) Englert, B. C. J. Environ. Monit. 2007, 9, 1154–1161. (2) Hussain, S. M.; Braydich-Stolle, L. K.; Schrand, A. M.; Murdock, R. C.; Yu, K. O.; Mattie, D. M.; Schlager, J. J.; Terrones, M. Adv. Mater. 2009, 21, 1549– 1559. (3) Johnson, A.; Greenwood, P.; Hagstrom, M.; Abbas, Z.; Wall, S. Langmuir 2008, 24, 12798–12806. (4) Pease, L. F.; Tsai, D. H.; Zangmeister, R. A.; Zachariah, M. R.; Tarlov, M. J. J. Phys. Chem. C 2007, 111, 17155–17157. (5) Tsai, D. H.; Pease, L. F.; Zangmeister, R. A.; Tarlov, M. J.; Zachariah, M. R. Langmuir 2009, 25, 140–146.

Langmuir 2010, 26(15), 12505–12508

challenged by particles smaller than 10 nm.6 Methods that detect solution-phase ion concentrations give bulk averages and offer no detailed information on how the particle size changes as a function of the dissolution process. We note that the dynamics of these particle size changes can be delineated only by monitoring dissolution from the particle perspective. Because nanoparticles exhibit unique size-dependent behavior, particle size effects cannot be ignored and must be explicitly monitored so that a fundamental understanding of the underlying size-dependent processes on the nanoscale can be discerned.7 The current study demonstrates that particle analysis is a promising method for measuring these nanoscale processes.

Experimental Section Materials. CuNPs were purchased in powder form from Nanostructured & Amorphous Materials, Inc. (Houston, TX). These particles were previously analyzed in detail using several analysis methods including TEM, BET, XRD, and XPS as discussed in Pettibone et al.8 The results showed spherical particles with a Cu metal core and a CuO/Cu2O shell with the most oxidized phase being on the surface. Aqueous acidic solutions were prepared using Optima water and hydrochloric acid (Fisher Scientific). A copper standard was used as a 1 wt % copper ion solution (Elements). Particle Size Distribution Measurements. An electrospray aerosol generator (TSI Inc., model 3480) was coupled to a scanning mobility particle sizer (SMPS) consisting of a nanodifferential mobility analyzer (nano-DMA) (TSI Inc., model 3085) and an ultrafine condensation particle counter (UCPC) (TSI Inc., model 3776). Details of the ES-SMPS (Figure 1) have been reported previously.9 Briefly, the electrospray generates a nanoparticle aerosol stream from a nanoparticle colloidal suspension. Particles enter the SMPS, where they are charged to an equilibrium charge distribution, and their size can be classified on the basis of the electrical mobility as the charged particle stream (6) Borm, P.; Klaessig, F. C.; Landry, T. D.; Moudgil, B.; Pauluhn, J.; Thomas, K.; Trottier, R.; Wood, S. Toxicol. Sci. 2006, 90, 23–32. (7) Grassian, V. H. J. Phys. Chem. C 2008, 112, 18303–18313. (8) Pettibone, J. M.; Adamcakova-Dodd, A.; Thorne, P. S.; O’Shaughnessy, P. T.; Weydert, J. A.; Grassian, V. H. Nanotoxicology 2008, 2, 189–204. (9) Elzey, S.; Grassian, V. H. J. Nanopart. Res. 2010, 12, 1945–1958.

Published on Web 06/30/2010

DOI: 10.1021/la1019229

12505

Letter

Elzey and Grassian

Figure 1. Schematic diagram of the electrospray atomizer coupled to a scanning mobility particle sizer (ES-SMPS). flows down the high-voltage column of the nano-DMA. The particle trajectory down the column depends on the electrical mobility, Zp, of the particle in an applied electrical field and is defined by Zp ¼

neCc 3πηDp

ð1Þ

where n is the number of charges on the particle, typically single for particles below 20 nm, e is the elementary charge, Cc is the Cunningham slip correction, η is the dynamic viscosity of air, and Dp is the diameter of the particle. As a voltage range is scanned, particles with a specific electrical mobility and therefore a specific mobility diameter corresponding to each voltage value are detected. Thus, only particles within a narrow mobility diameter range are allowed to exit the nano-DMA, and the polydisperse aerosol stream at the inlet is converted into a monodisperse aerosol stream at the outlet. The monodisperse aerosol stream then enter the UCPC where the particle concentration is determined as particles are counted with a photodetector. The coupled particle sizing of the nano-DMA and particle counting of the UCPC allow the SMPS to provide a particle size distribution for a nanomaterial sample in the range of 2 to 80 nm. In these experiments, hydrochloric acid (0.5 M) was added to CuNPs for a sample concentration of 1 wt % copper. Samples were dispersed by agitation for 0. The particle number concentration varied during dissolution as smaller particles appeared to dissolve at an inherently faster rate and larger particles generated PSDs with smaller peak maxima during dissolution. A PSD was detected at significant concentrations down to a size of ∼3 to 4 nm, and there was a 12506 DOI: 10.1021/la1019229

Figure 2. PSDs with time during CuNP dissolution in 0.5 M HCl. The time (min) is shown in each PSD plot. It should be noted that the data are presented in terms of size bins and are reported as dN/d log Dp vs log Dp, where N is the number concentration of particles and Dp is the particle diameter.

continual shift in the peak maximum as the larger particles in the sample moved toward this size and rapidly dissolved and disappeared from the PSD. Below 3 to 4 nm, the particles dissolved completely between subsequent scans. After 30 min, the PSD maintained a particle number concentration, peak maximum, and polydispersity indistinguishable from those of a 1 wt % Cu ion standard solution analyzed for comparison (Figure 3). As droplets Langmuir 2010, 26(15), 12505–12508

Elzey and Grassian

Letter

Figure 4. Percent copper dissolved with time based on ICP/OES analysis (black) and ES-SMPS data (red). Water and water with CuNPs removed are shown to confirm that no CuNPs were detected as ions. Figure 3. ES-SMPS PSD data for CuNPs in 0.5 M HCl after 30 min (black) and for a 1 wt % Cu ion standard solution (blue).

that exit the electrospray were dried, ions within the droplets come together to form salt particles as the liquid phase evaporates. Thus, dissolved samples are expected to generate a PSD, and the size and intensity of the PSD will depend on the concentration of ions in solution. These results indicate that the CuNPs completely dissolved, and the small signal observed after 30 min was due to particles formed from the copper ions during the generation of the aerosol. Each PSD represents a 2 min scan. Because dissolution is a dynamic process, the scan time is an important factor to consider with regard to these changes observed in the PSDs. Each PSD is generated as the voltage range is scanned, and particles with specific diameters are detected only at a specific voltage. The voltage scan allows particles with smaller diameters to exit at the start of the scan, and it moves toward larger sizes throughout the 2 min scan. Therefore, particles that were 10 nm at the start of the scan and 8 nm at the end of the scan were counted at only one diameter when the voltage allowed those particles to exit the DMA and be detected. Thus, a “smearing” effect, where particles are continuously detected as their size changes during analysis, did not result in the observed multimodal distributions. Furthermore, polydispersity not only increased but also decreased from one scan to the next (t = 12, 18) during analysis. The scans at t = 24 and 27 showed only a decrease in particle count and no continued shift in size, polydispersity, or modes. These results indicate that dissolution processes were responsible for the observed shifts in the PSDs rather than the lag time of the scan. Although each scan took 2 min, particles were detected on a shorter timescale. Specifically, particles were detected over an ∼10-12-nm-diameter range, which was ∼40 s or one-third of the full 2 min scan. Because the data show shifts in diameter modes of ∼3 to 4 nm over 3-6 min intervals (Figure 2), it is clear that the 40 s particle detection time was much faster than the timescale of dissolution. If the 40 s lag time required for particle detection was responsible for the multimodal shifts in diameter of 3 to 4 nm, then a 15 nm particle would completely dissolve before two scans were acquired. Clearly, this is not the case. Reactions of copper metal with HCl: 2Cu þ 2HCl f 2CuCl þ H2

ðR1Þ

Cu þ 2HCl f CuCl2 þ H2

ðR2Þ

are not energetically favorable, but copper oxides are known to react according to: Cu2 O þ 2HCl f 2CuCl þ H2 O

ðR3Þ

CuO þ 2HCl f CuCl2 þ H2 O

ðR4Þ

Langmuir 2010, 26(15), 12505–12508

Thus, copper oxide layers present on the surfaces of nanoparticles likely facilitated the dissolution.8 An inductively coupled plasma optical emission spectrometer was used to verify the concentration of copper ions (Figure 4). ICP/OES analysis confirmed the complete dissolution of the CuNPs via the detection of copper as solution-phase ions. ICP/OES data was compared with ES-SMPS data using the peak maxima of the PSDs at each time to calculate the particle volume and percent copper dissolved. Specifically, the particle diameter corresponding to the peak maximum at t = 0 was used to calculate the initial average volume of a particle. Subsequent diameters corresponding to the peak maxima at each time were used to calculate the average volume of a particle with time. The percent decrease in volume with time was used to estimate the percent copper dissolved. Results for percent copper dissolved using both methods agree well. However, the ES-SMPS data provides additional information about polydispersity and multimodal PSDs that cannot be extracted from bulk measurements of solution-phase ion concentration. PSD data for the CuNPs in 0.5 M HCl are shown as an example because similar multimodal results were seen with 0.25-0.75 M HCl. At 0.5 M HCl, the acid was present in slight excess for all potential reactions with copper and copper oxide layers (R1-R4). In solutions of g0.75 M HCl, dissolution occurred quickly and no PSD was observed after one or two scans. Solutions of