Supplementary Information
METHODS
Unless otherwise specified, all physico-chemical characterisations were carried out at room temperature. Sample homogeneity was ensured by thoroughly blending the batches of sample received before sub-sampling. For each measurement, samples were taken from the sub-samples randomly and measurements were performed at least in duplicate. In some cases, e.g. for surface area, TGA, TEM, DLS and zeta potential, tests were performed on the same sample batch at different laboratories and by different operators, yielding comparable results.
Crystalline phase. Samples were prepared by packing approximately 0.5g of solid material in a plastic sample holder and flattening the surface with a glass slide. The crystalline phase was determined by analysis of the diffraction pattern produced upon exposure of the sample to x-rays generated by a Bruker ASX-D8 X-Ray Diffractometer (XRD) using CuKa radiation and an operating current of 40mA and voltage of 40kV. The scan ranged from 20o to 150o with a step size of 0.005o. The size of the aperture slit directing the X-ray source was 0.2mm.
Particle morphology. The morphology of each sample was assessed by analysis of Transmission Electron Microscopy (TEM) and Scanning Electron Microscopy (SEM) images. Samples for TEM were prepared by briefly sonicating a few milligrams of the solid material in approximately 20µl ethanol to form a milky dispersion. Carbon-coated grids (copper, 300 mesh) were glow discharged in nitrogen for 30s to render them hydrophilic. 5µl of dispersion was applied to the freshly-prepared grids. After 2 minutes, excess dispersion was wicked off using filter paper (Whatman 541) and the grids were dried in air for 15minutes. Grids were examined at eucentric height using an FEI Tecnai 12 TEM operating at 120kV, and micrographs were recorded using an Olympus Megaview III CCD camera running AnalySiS imaging software (Olympus). The maximum and minimum “Feret” diameters were used as a measure of particle length and width and were determined from TEM images using a semi-automated particle size analysis program (ImageJ/Fiji) (Igathinathane et al. 2012). Approximately 100 particles across the TEM grid were measured for each sample which came from more than 10 images for NM-110 and NM-111, 5 for NM-112 and >20 for NM113. This ensured that a representative set of images were acquired. To minimise uncertainty caused by overlapping particles, only particles that were separated or could be clearly defined were included in the analysis of maximum and minimum Feret diameter. In addition, two operators performed measurements using the same software. The final data reported here were the average values provided by the two operators.
Samples for SEM were prepared by sprinkling ~5mg of ZnO powder over adhesive conducting tape covering an SEM metal stub. The surface of the powder sample was flattened with a spatula, and excess powder was removed by gently tapping the stub on its side until a light coating of powder on the surface became apparent. The nanoparticles were coated with iridium using a Polaron SC570 sputter coater. Sputtering was conducted under vacuum while the passing gas was argon. The coating deposition time was 20s at a plate current of 50mA, giving a coating thickness of approximately 1nm. Images were obtained with a Philips XL30 field emission SEM. The optimal spatial resolution of the microscope was 2–5nm with varying accelerating voltage from 30kV to 1kV. Images of ZnO particles were acquired at an accelerating voltage of 5kV, a working distance of approximately10mm, and a tilt angle of 0°.
Specific surface area and porosity. Specific surface areas of the twelve ZnO samples were determined by the Brunauer-Emmett-Teller (BET) using a Micromeritics Tristar II 3020 instrument, which uses physical gas adsorption and capillary condensation principles to obtain information about the surface area. Prior to analysis, the powdered sample was transferred to a sample bulb which was de-gassed overnight at 300°C under high vacuum and subsequently weighed on an analytical balance in order to determine the sample mass after the degassing step. The sample tube containing the degassed sample was cooled by immersion in liquid nitrogen to 77K and exposed to the adsorbate gas (nitrogen, ultra high purity 99.999%) at 11 controlled pressures. Porosity was determined simultaneously with surface area by applying the Barrett-Joyner-Halenda (BJH) method for analysing gas adsorption and desorption isotherms to determine the total pore volume.
Elemental composition. To prepare samples for the determination of elemental composition, approximately 0.15g of ZnO sample was dissolved in a 1:1 HNO3:H2O2 mixture with heating for 30 minutes. The solution was diluted to 100mL, 10µgml-1 of Scandium Standard (Ultra Scientific Analytical Solutions) was added as an internal standard, and the resultant solution was analysed by Inductively Coupled Plasma-Atomic Emission Spectroscopy (Varian 730 Axial ICP-AES). Certified multi-element solutions (QCD Analysts) were used to check accuracy.
Elemental analysis of the surface. The elemental composition within 1–10nm of the surface of the samples was quantitatively determined in duplicate by X-ray photoelectron spectroscopy (XPS). Powdered ZnO samples were placed in individual wells of a sample holder and irradiated with X-rays under ultra-high vacuum using a Kratos HS spectrometer, fitted with a monochromated Al Kα source, operating under standard conditions. The sampling area was approximately 0.3mm by 0.7mm. Wide scan survey spectra were recorded to identify and quantify all elements present on the surface.
Thermal stability. Thermo-gravimetric analysis (TGA) of all ZnO batches was performed on a Perkin Elmer STA6000 instrument in duplicate. Roughly 50mg of sample was heated in a ceramic crucible from 50oC to 1000oC at a rate of 10oCmin-1, under a 20mlmin-1 flow of N2 gas. Volatile species evolved during heating were passed through a heated transfer line into the gas cell of a Fourier-transform infra-red spectroscopy system (FT-IR, Perkin Elmer Frontier) for analysis.
Differential centrifugal sedimentation. To investigate the particle size distribution, differential centrifugal sedimentation (DCS) measurements were performed using a CPS Instruments 24000UHR disk centrifuge. Measurements were made at a fixed speed of 8900rpm to access the distribution from 20–1000nm, and, using the speed ramping mode, with speeds between 1000 and 12000rpm, to access the distribution from 0.02–10µm. 17ml of a gradient fluid comprising 8–24% w/w sucrose (99%, Sigma Aldrich) in ultrapure water (Milli-Q, 18.2MΩcm) with a 0.5ml dodecane evaporation cap was used for all measurements. A 0.377µm nominal mean diameter calibration standard (polyvinyl chloride, PS Instrument) was used. 0.1ml of sample was injected for each run. The samples for fixed-speed measurements were 0.015% w/w suspensions of ZnO in ultrapure water. 15mg of test powder was added to a clean, 20ml glass vial, and mixed with a spatula with ~1ml of ultrapure water. More dispersant was added to create a final volume of 15ml. This sample was then ultrasonicated using a high-power ultrasonic horn (Misonix 3000) for 20s with a 2s pulsed duty cycle with a 1.3cm tip at an amplitude of 20%. Following ultrasonication, the sample was transferred to a 150ml beaker and 85ml of ultrapure water was added to produce a 100ml dispersion. The resultant suspension was stirred with an overhead stirrer for at least 15minutes prior to measurement. The samples for speed ramping measurements were 1% w/w suspensions prepared by directly adding 1g of ZnO to 99ml of ultrapure water and stirring with an overhead stirrer for 15minutes. Probe sonication was then applied to the sample using a 1.3cm ultrasonic horn tip at 100% amplitude with a 2s pulsed duty cycle. The sample was first sonicated for 15s, measured, left to stir, sonicated for a further 15s (total sonication time: 30s), measured, left to stir, sonicated for a further 30s (total sonication time: 60s), and measured.
Dynamic light scattering (DLS). Measurements of z-average hydrodynamic particle diameter in deionised water (Milli-Q, 18.2MΩcm) by DLS were obtained for all samples using a Brookhaven particle size analyzer 90Plus equipped with a 633nm laser. The scattering angle in this instrument is 90°. Reference standards (Duke polystyrene latex with a nominal diameter of 100nm, and NIST RM8013 gold nanoparticles with a nominal diameter of 60nm) were used to verify the performance of the instrument. 10mg ZnO particles were added to a measuring cuvette containing 3ml of deionised water. The cuvette was placed in an ultrasonic bath (Branson3510, 100W, 42kHz) for 10s, and further shaken by hand to check the particles were well dispersed before starting the DLS measurements. For the uncoated products, the uniform colloidal dispersion obtained after sonication was used directly for measurements. Due to the highly hydrophobic surface on particles of Z-COTEHP1, only some were successfully dispersed while most remained floating on top of the water (as shown in Figure5). To avoid any influence from the undispersed particles, a pasteur pipette was used to transfer the dispersion to another cuvette for further measurement without additional sonication. Each measurement was based on 10 sub-runs which were averaged to produce the reported result. Duplicate experiments were performed on each sample. The temperature was maintained at 25oC during measurement. The cuvette was thoroughly washed with deionised water before and after each measurement.
Electrophoretic-mobility measurements of zeta potential (surface charge). A Malvern Zetasizer Nano Z system was used for zeta-potential measurements. 10mg of each ZnO sample was placed in a plastic tube containing 3ml dispersion medium. The dispersion medium was prepared in advance and is based on deionised water with the pH (pH=2, 4, 6, 8, 10) adjusted by adding 0.1MHCl or 0.1MNaOH. The tube was placed in an ultrasonic bath (Branson 3510, 100W, 42kHz) for 10s and then shaken manually to check dispersion. For uncoated ZnO products, the uniform colloidal dispersion after sonication was used directly for zeta potential measurements. Dispersed samples of Z-COTEHP1 were prepared as described above for DLS measurements. One ml of the dispersion was injected into a DTS1060 cuvette. Five measurements were performed at each pH to determine the average and standard deviation. The temperature of all measurements was maintained at 25oC. The cuvette was thoroughly washed with deionised water before and after each measurement.
Photo catalytic activity. Two 62.5ml mixtures of 1:1 Mineral Oil White Light (Aldrich):Caprylic Capric C8/C10 Triglyceride (MOTG) were prepared. ZnO (31mg) was added to one mixture, and DPPH (5.2mg) to the other. Each was magnetically stirred for 1.5hours in a beaker covered on all sides with Al foil. Then the two mixtures were combined, poured into a crystallising dish (135mm diameter x 23mm height) covered on all sides with foil, and magnetically stirred for 5minutes. Before exposure to UV (t=0), 3ml of the solution was withdrawn and its UV-Vis absorption spectrum was measured using a Cary 5G UV-Vis NIR spectrophotometer. The 3ml sample was returned to the solution. The solution was then exposed to UV using a pre-warmed Spectroline UV lamp (BIB150 P/FA 365nm, 150W concentrated spot bulb, lamp diameter: 110mm). The lamp was placed 12cm above the ZnO/dye mixture, at a position where the 365nm intensity was 45mWcm2. Samples were taken at various times and absorbance at 520nm was measured. The equipment was designed with a sliding shield separating the sample and UV-lamp so that between exposures the lamp remained on, thus avoiding variations in lamp intensity.
Dissolution measurements. The dissolution of each ZnO sample was measured using the equilibrium dialysis method (Angel et al. 2013). Dialysis membranes had a molecular mass cut-off of 1kDa (Cole Parmer Spectra/Por7). The membranes were cut into 10cm lengths, washed with deionised water (Milli-Q, 18.2MΩcm) filled with 20ml of Milli-Q water and sealed with acid-washed (1%v/v HNO3) plastic dialysis clips. The total volume of the dialysis cells was kept to below 5% of the test solution in order to minimise dilution effects. At the start of the test, 50mgl-1 of each ZnO sample was prepared in triplicate by dispersing in 3L of 0.01M Ca(NO3)2 buffered with 2mM piperazine-N,N’-bisethanesulfonic acid (PIPES: Sigma-Aldrich) to pH 7.5±0.1. The dialysis cells were immediately added and the solutions were continuously stirred with PTFE magnetic stirrers for 72hours at constant temperature (21oC). At sampling times of 24, 48 and 72hours, a dialysis cell from each treatment replicate was removed from the solution, rinsed with Milli-Q water and sub-sampled for the dialysed Zn concentration. At the same time, a filtered (0.1µm) sample was also taken from the bulk solution and analysed for Zn concentration for comparison with the dialysed sample. An additional filtered (0.1µm) sample from the bulk solution was taken at 8 hours. Samples from dialysis cells and from bulk solutions (0.1µm filtered) showed no differences in levels of dissolved zinc at any time point, indicating that any agglomerated ZnO in the bulk solution was larger than 100nm. The pHs of all solutions increased marginally after 72 hours to the range 7.56 -7.68.
Dispersion of the NM-111 batch of Z-COTEHP1 in water. 20–33mg of various samples of the NM-111 batch of Z-COTEHP1 were weighed into glass bottles, and 5ml of Milli-Q water was added. The samples were bath sonicated for 3minutes. After standing for a few minutes, the samples were illuminated with a white-LED torch from the side, and images were taken with a Nikon D5000 digital camera. Scattered light in the images indicated those samples with particles in suspension. Various samples of the same batch, NM-111, had experienced slightly different storage histories:
Sample 1: Stored in original glass bottle, first opened in 2009 (only once or twice).
Sample 2: Sub-sampled (~2g) from Sample 1 in 2009, stored in plastic bottle.
Sample 3: Sub-sampled from Sample 2 on the day of the described experiment, heated at 100oC for 1.5hours and cooled.
Sample 4: Stored in original glass bottle, opened September 2012 and December 2012.
Sample 5: Stored in original glass bottle, opened September 2013, otherwise unknown history.
Sample 6: Stored in original glass bottle, opened December 2013 (once only).
Sample 7: Stored in original glass bottle, opened on the day of the described experiment only.
Figure S1: X-ray diffraction spectra from the OECD batches, NM-110, NM-111, NM-112 and NM-113, of the four respective ZnO products, Z-COTE, Z-COTEHP1, Nanosun and bulk ZnO. The spectra reveal that the only detected phase is hexagonal wurtzite zincite. Other batches of the three nano-sized ZnO products display similar characteristics. The broader peak widths for NM-112 indicate a smaller crystallite size compared with the other OECD samples.