- Email: [email protected]
Production of particulates from transducer erosion: Implications on food safety
Ultrasonics Sonochemistry xxx (2014) xxx–xxx
Contents lists available at ScienceDirect
Ultrasonics Sonochemistry journal homepage: www.elsevier.com/locate/ultson
Production of particulates from transducer erosion: Implications on food safety Raymond Mawson, Manoj Rout, Gabriela Ripoll, Piotr Swiergon, Tanoj Singh, Kai Knoerzer, Pablo Juliano ⇑ CSIRO Division of Animal, Food and Health Sciences, 671 Sneydes Road, Werribee, Vic 3030, Australia
a r t i c l e
i n f o
Article history: Received 1 November 2013 Received in revised form 3 April 2014 Accepted 3 April 2014 Available online xxxx Keywords: Metallic particles Cavitation Ultrasound Safety Health Food processing
a b s t r a c t The formation of metallic particulates from erosion was investigated by running a series of transducers at various frequencies in water. Two low frequency transducer sonotrodes were run for 7.5 h at 18 kHz and 20 kHz. Three high frequency plates operating at megasonic frequencies of 0.4 MHz, 1 MHz, and 2 MHz were run over a 7 days period. Electrical conductivity and pH of the solution were measured before and after each run. A portion of the non-sonicated and treated water was partially evaporated to achieve an 80-fold concentration of particles and then sieved through nano-filters of 0.1 lm, 0.05 lm, and 0.01 lm. An aliquot of the evaporated liquid was also completely dried on strips of carbon tape to determine the presence of finer particles post sieving. An aliquot was analyzed for detection of 11 trace elements by Inductively Coupled Plasma Mass Spectroscopy (ICPMS). The filters and carbon tapes were analyzed by FE-SEM imaging to track the presence of metals by EDS (Energy Dispersive Spectroscopy) and measure the particle size and approximate composition of individual particles detected. Light microscopy visualization was used to calculate the area occupied by the particles present in each filter and high resolution photography was used for visualization of sonotrode surfaces. The roughness of all transducers before and after sonication was tested through profilometry. No evidence of formation of nano-particles was found at any tested frequency. High amounts of metallic micron-sized particles at 18 kHz and 20 kHz formed within a day, while after 7 day runs only a few metallic micro particles were detected above 0.4 MHz. Erosion was corroborated by an increase in roughness in the 20 kHz tip after ultrasound. The elemental analysis showed that metal leach occurred but values remained below accepted drinking water limits, even after excessively long exposure to ultrasound. With the proviso that the particles measured here were only characterized in two dimensions and could be nanoparticulate in terms of the third dimension, this research suggests that there are no serious health implications resulting from the formation of nanoparticles under the evaluation conditions. Therefore, high frequency transducer plates can be safely operated in direct contact with foods. However, due to significant production of metallic micro-particulates, redesign of lower frequency sonotrodes and reaction chambers is advised to enable operation in various food processing direct-contact applications. Ó 2014 Elsevier B.V. All rights reserved.
1. Introduction Many food processing unit operations may release metallic particles into food from either cavitation induced processes on fluids by pumping or homogenization, or from direct erosion of equipment from pneumatic conveying of dry food ingredients and milling operations. The presence of metallic particles in food as consumed has however, been mostly associated to the use of domestic cookware [1]. Even though it has not been formally established in the published literature, there remains a concern ⇑ Corresponding author. E-mail address: [email protected] (P. Juliano).
that cavitation-related release of metallic particles into foods from sonication with ultrasonic transducers probes in direct contact with the food material. Therefore, it is important to consider firstly, the cavitation erosion of metals in direct contact with foodstuffs and secondly, the implications of the consumption of metallic particles on food safety. 1.1. Safety implications on human intake of metallic nanoparticles Most literature focuses on the medical application of metal as potential imaging contrast agents [2], the safety of inhaled metal nanoparticles [3] and skin penetration [4]. There is also some literature on the potential nanotoxicity of ingested nanoparticles based
http://dx.doi.org/10.1016/j.ultsonch.2014.04.005 1350-4177/Ó 2014 Elsevier B.V. All rights reserved.
Please cite this article in press as: R. Mawson et al., Production of particulates from transducer erosion: Implications on food safety, Ultrason. Sonochem. (2014), http://dx.doi.org/10.1016/j.ultsonch.2014.04.005
2
R. Mawson et al. / Ultrasonics Sonochemistry xxx (2014) xxx–xxx
on in vitro and in vivo studies using animal models [5–7] and post mortem examination of humans [8]. Mucus ingested creates a secondary hazard from the inhalation of nanoparticles [9]. The highest concern for human safety are bio-persistent nanoparticles [10], which include most metallic particulates. For consideration of biological consequences a metal nanoparticle ranges between 5 nm and 100 nm. The impact of nanoparticulate metals in biological systems depends on the particle characteristics such as shape, surface characteristics and surface energy differences [11]. However, the mechanisms behind the interaction of bio-persistent metal nanoparticles and the biological systems are not well understood. It is believed that biological damage from bio-persistent metal nanoparticles can extend from tissue inflammatory responses due to homeostasis disturbance by protein and peptide malformation [12] to chromosomal fragmentation and DNA cleavage with consequent apoptosis or potential cancer [3,11,13–17] and/or brain damage [18]. While generalizations about nanoparticles may not be valid [5], metal particles in particular seem to have little difficulty in penetrating the gut mucosal layer mainly through the gut lining cells. Entry into these cells appears to be limited to metal particles of less than 80 nm and only particles of less than 50 nm may find passage into the circulation system. Particles larger than 15 nm are captured by the liver and may be tagged for transport to the kidneys; in either case they can be excreted in bile or urine. Particles less than approximately 15 nm become fully systemic and may enter almost every cell in the body including the brain and bone marrow [5,6]. However, it is important to note that the majority of metal nanoparticles ingested will likely be excreted in the faeces and some predisposing condition in the gut may be required before any nanoparticles can enter the gut lining cells [5–8]. The biological impacts from exposure are many which have spawned a new science discipline ‘‘metallomics’’ the study of metallic nano- particles in biological systems and their impacts on life [19]. There is a new Journal ‘‘Metallomics’’ published under the Royal Society Proceedings series devoted to this topic. In summary, the literature reveals that metal nanoparticles or their oxides with a size less than 80 nm cannot be regarded as safe in foodstuffs. 1.2. Mechanisms of cavitation erosion in foodgrade metallic surfaces (stainless steel and titanium) Metallurgical product advancements have resulted in making stainless steel more cavitation erosion resistant; unfortunately, these improvements are largely unsuitable for food processing environments. For example, many systems for making stainless steel corrosion resistant involve addition of metals considered toxic in food. Two mechanisms are proposed for erosion by cavitation (induced by 20 kHz ultrasound at an amplitude of 6 lm): surface hardening by thermal tempering triggered by heat transferred from surface imploding cavitation bubbles [20,21] and work hardening from repeated impact by cavitation microjets from surface bubbles [22,23]. The surface hardening has been measured and there are indications of localized temperatures in excess of 300 °C, causing localized changes in the alloy composition. In either case, the effects will be confined to the surface molecular layers. Cavitation erosion is strongly influenced by the stainless steel alloy crystal size, the larger the crystals the greater the erosion, which suggests that the erosion is likely to be on a very small distance scale [24]. Growth of alloy crystals beneath the surface weakens the surface structure, making it more prone to erosion by fracture and flaking away of the hardened surface material. Little to no direct evidence of metal particle characterization from cavitation erosion by ultrasonic transducers has been reported. Evidence of silver erosion from cavitation in water
induced by pulsed laser was reported, where particles predominantly smaller than 50 nm at 50 and 100 kHz pulse frequency, at an energy density of 1.2–2.9 J/cm2 were produced [25]. Silver is more difficult to ablate than either stainless steel or titanium [9] due to its higher ductility. However, it is not acceptable to generalize erosion characteristics from cavitation. Cavitation pits self pacify and stop eroding within themselves; however, the pacified pits spawn new erosion pits in the surrounding metal. Some of these observations were made at 230 kHz using a ‘‘seed’’ pit or gas bubble [26]. The cavitation pits increase surface fluid flow turbulence and create crevices that act as nucleation sites for additional cavitation bubbles. Stainless steel ablates as the stainless steel alloy nanoparticles (not its component metals). From surface energy considerations the alloy is more prone to cavitation damage than the pure components would be [11]. Another effect caused by the generation of the bubbles while passing high intensity ultrasound through a liquid, is the decomposition of the solvent and solutes present within the bubbles, when the bubbles are formed and then collapse, generating several highly reactive radicals [27]. In the case of water, the radicals generated are H and OH [28] (Reaction (1)).
H2 OÞÞÞÞÞH þ OH
ð1Þ
The presence of these radicals can induce the sonochemical production of metal nanoparticles [29]. Ashokkumar et al. [27] found that in a water medium there is an increase of the OH yield with an increase in ultrasound frequency from 20 kHz to 358 kHz and, for higher frequencies (>450 kHz), the yield starts dropping until reaching insignificant concentrations of the radicals. The explanation for this decrease is that at higher frequencies the acoustic cycle becomes very short, limiting the size of the generated cavitation bubbles, and the bubbles contain headspace gas restricting the amount of water vapor generated within the collapsing bubble, and therefore, there are less OH radicals generated. This would indicate that no nanoparticles should be generated because of radical formation when frequencies in the MHz range are applied in a liquid medium. 1.3. Application of ultrasound in the food industry Ultrasound is an emerging technology in the food industry. It has been used for processing, preservation and extraction operations. Some of the advantages of using ultrasound are a more effective mixing and micro-mixing, faster energy and mass transfer, reduced thermal and concentration gradients, reduced processing temperature, selective extraction and reduced equipment size [30]. Currently, the food industry is using low frequency sonotrodes in applications such as: cleaning baths, emulsification [27], defoaming of carbonic beverages produced in high-speed bottling and canning lines [31], improved overall frying [32], ultrasonic food cutting, and ultrasonic osmotic dehydration [30]. There are other emerging applications utilizing high frequency transducer plates in the megasonic range (>400 kHz) for enhancing the separation of milk fat from milk and recovering palm oil in the palm oil milling process [33–36]. In light of the above summation, an attempt is made to characterize the metallic particles likely to be produced by current ultrasonic technology that could be used or is being used for food processing. 2. Materials and methods 2.1. Ultrasonic runs Fig. 1 describes the experimental protocol employed for each ultrasonic run. Double distilled water (16 L, initial pH = 5.5, initial
Please cite this article in press as: R. Mawson et al., Production of particulates from transducer erosion: Implications on food safety, Ultrason. Sonochem. (2014), http://dx.doi.org/10.1016/j.ultsonch.2014.04.005
R. Mawson et al. / Ultrasonics Sonochemistry xxx (2014) xxx–xxx
electrical conductivity 1.3 lS) was placed in food grade 316 stainless steel vessels for high frequency (plate transducers) treatment. Transducers were configured in the vessels as shown in Fig. 2. High frequency plate transducers (Submersible Transducers, Sonosys Ultraschallsysteme GmbH, Neuenbuerg, Germany, Stainless Steel AISI 303 – iron, chromium, nickel, copper) were continuously operated at 0.4 MHz, 1 MHz, and 2 MHz and matched power of 220 W in three jacketed vessels (grade 316 steel) for 7 days. All ultrasonic runs were performed in duplicate. The high frequency runs were performed under temperature controlled conditions (20 ± 3 °C, by circulating glycol in the jacketed vessel). External contamination was avoided through stainless steel covers of the same 316 grade. Attempts were made to evaluate an 80 kHz and 220 W ultrasonic system over a 7 days duration in the 16 L tanks, but on three occasions the welds in the treatment vessel eroded through and the vessel drained overnight within 24 h. Consequently no data could be collected and other than noting that these conditions promoted rapid erosion there is no further discussion of these conditions. Two vessels (grade 304 stainless steel) were set up with a low frequency transducer (or sonotrode), each operating continuously for 7.5 h at 18 kHz (Dr. Hielscher GmbH, UIP 1000, Teltow, Germany) with power input of 180 W, and 20 kHz (Branson Digital Sonifier Model 250, Danburry, CT, USA) with power input of 103 W, respectively. The low frequency runs started at room temperature reaching a maximum of 42 °C and 34 °C for the 18 kHz and 20 kHz transducers, respectively. Based on preliminary results, the experimental design was put together to achieve a worst case scenario for high frequency sonication conditions (extended sonication of 7 days and higher power). Separate controls (non-ultrasound treatment) were run in parallel to the low and high frequency vessels using the same water. A float control sensor was placed in the high frequency vessels treated for up to 7 days to ensure that the plate transducers remained below the water surface and avoid damage. A power meter (F7C005au, BELKIN, China) was utilized to measure the power drawn from the generators. At the end of the runs the electrical conductivity and pH were measured. Subsamples (4 L) were collected from the vessels and placed in glass beakers and reduced
3
to 50 mL by evaporation using an 80 °C water bath. The concentrated samples were further subsampled for trace elemental analysis (20 mL aliquots). 2.2. Characterization of transducer erosion 2.2.1. Photography After the ultrasound runs each transducer was visually observed for erosion and photographs were taken (Canon SLR 1100DSLR, Tokyo, Japan). 2.2.2. Surface profilometry The roughness was measured using an AltiSurf 500 Surface Profilometer (Bruker Biosciences, Preston, VIC) using a 300 lm optical probe. The x–y stage traversed at 500 lm/s with 0.5 lm spacing for profiles and 5 lm spacing for area tests. The profile test measured 10 lines of the active surface of the transducer. Average roughness (Ra) values were measured for every line in x and y directions. The average roughness was calculated as the integrated area by the roughness profile of each line over the evaluation length. The geometry of the Hielscher 20 kHz probe precluded any profilometry being done on this probe and there is no discussion of this parameter relating to this probe. 2.3. Characterization of dissolved and particulate matter in water 2.3.1. Sample preparation A 50 mL volume of the concentrated non-sonicated and sonicated water was filtered sequentially three times using different pore sized polycarbonate membranes in the following order: 0.1 lm (Isopore), 0.05 lm (Poretics), 0.01 lm (Poretics). Another 1 mL of the concentrated water was dried at ambient temperature in a desiccator for 24 h onto the carbon tape utilized for SEM imaging. The membranes were dried in a desiccator and both carbon tapes and membranes were stored in petri dishes. Particles trapped on each carbon tape or on membranes were later visualized by field emission scanning electron microscopy equipped with Oxford instruments energy dispersive spectroscopy for elemental analysis (FE,SEM-EDS).
Fig. 1. Experimental protocol followed for the untreated double distilled water samples and all transducer frequencies to evaluate erosion extent through a number of analytical methods.
Please cite this article in press as: R. Mawson et al., Production of particulates from transducer erosion: Implications on food safety, Ultrason. Sonochem. (2014), http://dx.doi.org/10.1016/j.ultsonch.2014.04.005
4
R. Mawson et al. / Ultrasonics Sonochemistry xxx (2014) xxx–xxx
Fig. 2. Schematic diagram of ultrasonic systems utilized for erosion trials (left: low frequency ultrasonics system, right: high frequency megasonics systems). The vessels were made of food grade stainless steel.
2.3.2. FE,SEM-EDS analysis The filters were mounted on an aluminium stub with doublesided conductive carbon tape. These samples and the carbon tapes carrying dried water concentrate were then carbon coated using a Polaron E6700 carbon sputter coater. The thickness of the carbon coating was approximately 20 nm. Once coated, the samples were placed into a Philips XL30 Field Emission Scanning Electron microscope (FE,SEM). The EDS (Energy Dispersive Spectroscopy) X-ray analysis system (Aztec, Oxford Instruments Pty. Ltd., Thornleigh, NSW, Australia) with an accelerating voltage of 20 kV was used for the analysis. SEM images of the filters were obtained as well as the size and chemical composition (EDS) profile of the particles detected on them. 2.3.3. Low magnification microscopy All nano sieved membranes with appropriate controls were observed under a light microscope (Leitz light microscope, soft imaging system Color view IIIu, Japan) and pictures were taken at three different locations at 40 times magnification. In brief, a small piece (1 cm 1 cm) of each membrane was cut and placed on a glass slide and a coverslip was applied. Images were taken at three different locations at a resolution of 250 lm. Image analysis was carried out via a purpose developed software (Fig. 3) programmed in MATLAB2011™ (The Mathworks Inc., Natick, MA, USA). The image was first converted into a matrix with a 2D resolution identical to the original image. Then a histogram was plotted from which a threshold value was determined to generate a binary image. From the binary image, the relative area covered by the particles collected in the membranes (area fraction) was calculated by dividing the area covered with particles by the complete area of the image. 2.3.4. Metal trace analysis The concentrated samples were filtered through a 0.45 lm filter and then measured by Inductively Coupled Plasma Mass Spectroscopy (ICPMS) for: Aluminium (Al), Chromium (Cr), Cobalt (Co), Copper (Cu), Iron (Fe), Manganese (Mn), Nickel (Ni), Tin (Sn), Sulfur (S), Molybdenum (Mb), and Silicon (Si), which are the elements present in food grade stainless steel (material of the vats and of the transducer casing). Unfiltered samples were analysed for Total carbon (TC) using the AnalytikJena Multi N/C 3100 Carbon Analyser (Wembley, UK), which oxidizes all organic and inorganic carbon at 680 °C to CO2 and carbon is then measured by a detector. Inorganic carbon (IC) is then determined by addition of phosphoric acid to the sample to convert carbonates to CO2. Concentration values are expressed on an original water sample basis by correcting with an 80-fold evaporation factor.
Fig. 3. Methodology of micrograph analysis according to the presence of particles on the nanofilter.
2.3.5. pH measurement and electrical conductivity Water samples were taken at the beginning and end of each ultrasonic run. Measurements of pH and electrical conductivity (#WP-84 Conductivity-TDS-Temp. Meter, TPS Pty Ltd., Springwood, QLD, Australia) were carried out. 3. Results Particle erosion was established after low frequency sonication in water while no erosion was detected after sonication in the high frequency systems. This was established by various methods after assessment of the transducers themselves before and after sonication and by analysing dissolved and particulate matter in the processed water.
Please cite this article in press as: R. Mawson et al., Production of particulates from transducer erosion: Implications on food safety, Ultrason. Sonochem. (2014), http://dx.doi.org/10.1016/j.ultsonch.2014.04.005
R. Mawson et al. / Ultrasonics Sonochemistry xxx (2014) xxx–xxx
3.1. Visual observations and transducer roughness Fig. 4 shows photographic evidence of pitting or indentation of the low frequency sonotrodes (18 kHz and 20 kHz) after 7.5 h continuous processing in water. However, no visual evidence of physical damage could be detected after 7 days of sonication of water with high frequency plate transducers. Eroded particles were collected in all membranes of the water concentrate obtained from the 18 kHz run. Particle deposits seen on the 0.1 lm as well as the other pore size membranes are shown in Fig. 5. Metallic particles were also visually detected on the membranes obtained from 20 kHz treated water but to a lesser extent. However, the membranes obtained from the high frequency treated water and the control water did not show any deposition residue. Visual observations of the nano-membrane filter discs were examined under an optical light microscope. The particles were visualized at three different locations of the membrane to form an impression of the particle shape and size. The presence of particles at different size ranges indicated the wide distribution of nano-materials retained during molecular sieving. The visual observations by the naked eye found that thick black powdery material was deposited on the surface of the 0.1 lm membrane
Fig. 4. Erosion occurring after ultrasound treatments carried out for 7.5 h in distilled water. (a) An 18 kHz Hielscher tip before and after ultrasound, (b) a brand new 20 kHz Branson tip after ultrasound (see demarked area on edges) and its corresponding profilometry data before and after ultrasound. Blue indicates the deeper section and red fading up to white the higher region. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
5
from ultrasonically treated water by the 18 kHz transducer. This indicated that some local erosion had taken place which was validated by FE,SEM analysis and the corresponding nano-material composition was characterized by EDS. The visual observations in the 20 kHz transducer tip (Fig. 4b) are corroborated by the changes in surface depth obtained from the profilometer analysis. The roughness value significantly increased for the 20 kHz probe from 0.36 lm to 0.80 lm (Table 1) while no major changes in roughness value were seen for any of the high frequency transducers. However, the rest of the nano-sieved ultrasonically treated water for a 7-days period did not show any deposition on the membrane surface, indicating there was no significant erosion producing nanoparticles as evident from the clear membrane filters observed after molecular sieving. 3.2. Nano-sieve imaging and water composition Low resolution magnification imaging allowed the calculation of the relative area covered by the particles collected in the membranes (area fraction). The area fraction (Fig. 5) changed significantly for both membranes obtained from the 18 kHz and 20 kHz transducer runs after 7.5 h. This indicated that significant amount of erosion material (represented by particles of various shapes and sizes) was removed from the sonotrodes. Formation of metallic particles from erosion at 18 kHz is also shown in optical images in Fig. 5. However, the area fractions for the membranes obtained from the non-sonicated controls and the high frequency did not change after the 7-day runs. Fig. 5 shows the prevalence of nonmetallic particles in the control water. FE,SEM images (Figs. 6–8) illustrate that the purely metallic particles, where identified, appeared to be in flake form rather than the expected spherical form if the surface of the ultrasonic horn had been subject to melting. For example, Particle 1 in Fig. 6 clearly demonstrates the flake like structure of particles that are clearly metallic in nature from the FE,SEM-EDS analysis. The data obtained from FE,SEM-EDS individual particle analysis are not presented but are summarized in the following discussion with respect to particular treatments. The particle analysis from the 400 kHz trial on the 0.1 lm filter reveals that Particle 1 (Fig. 7a) contains traces of iron, chromium, aluminium and magnesium. Particle 2 (Fig. 7a) contains moderate amounts of iron and traces of copper and chromium. On the 0.05 lm filter, Particle 1 (Fig. 7b) contains traces of iron, titanium, magnesium and moderately large amounts of aluminium. Particle 2 (Fig. 7b) contains traces of aluminium only. In Fig. 6, from the Hielscher 18 kHz horn, Particle 1 contains largely titanium and aluminium. Fig. 8 illustrates the particles obtained from the 20 kHz Branson ultrasonic horn collected on the 0.1 lm filter. Particle 1 (Fig. 8b) contains large amounts of titanium, moderate amounts of vanadium, and aluminium and traces of iron. Particle 3 (Fig. 8b) contains large amounts of titanium, moderate amounts of vanadium, aluminium and traces of iron and copper. However, Fig. 7c also illustrates the nature of fuzzy aggregated particles collected on the 0.01 lm filter from the 2 MHz treatment. Particle 1 (Fig. 8a) contains traces of iron in the control (non-sonicated) water sample. Particle 2 (Fig. 8a) contains traces of iron and copper magnesium and a small amount of aluminium. Table 2 shows the size ranges of particles selected from the carbon tapes and from the membranes. In general, all particles detected by FE,SEM exceeded 0.21 lm, i.e., no nanoparticles could be detected in any run. Metallic particles were detected in the controls as well as in the sonicated waters. Therefore, the fact that metallic particles were detected in the high frequency samples is not indicative of erosion.
Please cite this article in press as: R. Mawson et al., Production of particulates from transducer erosion: Implications on food safety, Ultrason. Sonochem. (2014), http://dx.doi.org/10.1016/j.ultsonch.2014.04.005
6
R. Mawson et al. / Ultrasonics Sonochemistry xxx (2014) xxx–xxx
Fig. 5. Filter photographs with digitally processed micrographs of a small section of the filter underneath. The values represent particle area fractions calculated for each micrographs. Imaging of 0.1 lm nanofilters used to follow up the formation of particles in the following water samples: control low frequency (LF, non-US, 7.5 h run), 18 kHz Hielscher transducer (7.5 h run), 20 kHz Branson transducer (7.5 h run). The ⁄ symbol indicates that the area fraction for the 18 kHz 7.5 h treatment is significantly different (P < 0.05) to the 20 kHz treatment. Control high frequency (HF, non-US, 7-days run) and the 400 kHz, 1 MHz, and 2 MHz Sonosys transducers (7-day runs). No particles are evident in any of these treatments.
Table 1 Surface profilometry indicating roughness Ra (lm, ±standard deviation) of: (a) unused 20 kHz sonotrode tip before and after a 7.5 h ultrasonic run in water and (b) transducer plates before and after 0.4 MHz, 1 MHz, and 2 MHz 7-day runs. Ra data obtained in the X and Y direction data was pooled together and taken as a total average for each transducer before and after sonication. Different superscript letters indicate significant differences obtained for the same transducer (P < 0.05). Transducer type
Ra (lm) before US run
Ra (lm) after US run
20 kHz 0.4 MHz 1 MHz 2 MHz
0.36 ± 0.04a 0.95 ± 0.05a 1.38 ± 0.06b 1.22 ± 0.12a
0.80 ± 0.08b 0.94 ± 0.06a 1.34 ± 0.04a 1.15 ± 0.15a
The trace elemental analysis shown in Table 3 demonstrates that the metals in solution, even after excessively long sonication, remain below drinking water limits [37]. Low frequency ultrasound increased the levels of Fe and Ti in solution, where 18 kHz gave higher values than 20 kHz. However, no changes in Ti were seen in the high frequency solutions compared to the 7-days control. On the other hand Al, Cr Cu, Fe, and Ni increased after high frequency ultrasound processing with respect to the 7-days control. The most marked increase occurred at 400 kHz for Fe. The pH for the control water remained unchanged over the 7days period and the electrical conductivity showed only a marginal change (Table 4). The low frequency treatment also remained unchanged in pH and electrical conductivity after sonication over a period of 7.5 h. However, at higher frequencies the pH significantly decreased after 7 days, being the most significant effect seen at 400 kHz. The electrical conductivity also changed due to a decrease in pH. The conductivity increased most markedly at 400 kHz followed by 1 MHz and 2 MHz.
Fig. 6. Scanning electron microscopy image showing an example of the flake structure of metallic particles obtained from the 18 kHz Hielscher on a 0.1 lm nanosieve.
4. Discussion All of the FE,SEM-EDS and soluble element data presented with a background contamination of silica that could be derived from the glass vessels used to evaporate the treated water samples. It is likely that the particulate matter observed on filter membranes in the critical range between 10 and 50 nm, is due to the formation of aggregated deposits that include metal ions.
Please cite this article in press as: R. Mawson et al., Production of particulates from transducer erosion: Implications on food safety, Ultrason. Sonochem. (2014), http://dx.doi.org/10.1016/j.ultsonch.2014.04.005
R. Mawson et al. / Ultrasonics Sonochemistry xxx (2014) xxx–xxx
7
Fig. 8. FE,SEM images of (a) the control (non-US) and (b) the 20 kHz water samples after filtering on a 0.1 lm nanosieve.
Fig. 7. FE,SEM of the 400 kHz (a) after filtering on a 0.1 lm and (b) the filtrate on a 0.05 lm nanosieve and of (c) the 2 MHz water samples after filtering on a 0.1 lm nanosieve.
The stainless steel used in the construction of the treatment tanks for the low frequency studies was grade 304. The high amplitude low frequency 20 kHz Branson ultrasonic horn was tipped with a screw-in plate of a harder alloy than the alloy used for the bulk of the horn, while the 18 kHz Hielscher ultrasonic horn is un-tipped and the composition of the particles eroded reflect a typical composition of an acoustic grade of titanium alloy. The treatment tanks used in the higher frequency studies were made from stainless steel alloy 316. The ultrasonic plate transducers used for the 400 kHz, 1 MHz and 2 MHz equipment operate at very low intensity per unit transducer surface area and it was not expected that there would be any significant erosion at the surfaces of these plates as is confirmed by the profilometry data (Table 1). It has been reported [21] that pitting in the surfaces of metals exposed to ultrasound is caused by the surface being locally melted by heat transferred from the contacting cavitation bubble. In our work we fail to see any evidence of this at either the surface of the ultrasonic horn or the surface of the stainless steel container. Instead, the metallic particles have a flake-like characteristic suggesting that they have been mechanically removed from a case hardened surface induced by ultrasonic peening. It is possible that the nature of this hardening promotes the variation of the alloy
crystalline structure beneath the surface hardened layer. This compositional variation may promote localized accelerated electrolytic corrosion. In our studies, conducted under a headspace of air at frequencies of 1 MHz and 2 MHz, the cavitation bubbles would have been filled with air rather than predominantly water vapor resulting in the synthesis of oxides of nitrogen and subsequently nitric acid rather than hydrogen peroxide [38], as can be seen from the declining pH and increasing conductivity of our sonicated water (Table 4). At 400 kHz, an early synthesis of nitric acid is likely to have been seen with concurrent generation of larger amounts of hydrogen peroxide so although the net corrosion was higher at 400 kHz the corrosion pathway is likely to have been rapid initially and then reducing whereas at the higher frequency the corrosion pathway would have been a slow progression towards more rapid corrosion over the duration of treatment. However, as samples were not taken during each run, this hypothesis needs to be further investigated. The presence of dissolved metal ions other than would be expected is most likely due to corrosion of the stainless steel tanks rather than erosion, especially at the frequencies where no evidence of metallic particulate material was observed. The corrosion of stainless steel is a complex topic and has been reviewed with respect to crevices and cracks [39] and at a crystal structure level [40]. Corrosion is promoted when the passivation layer on the surface of the stainless steel is breached permitting small scale electrolytic reactions to take place between the crystals of different metal species that make up the alloy. Where stainless steel is welded, zones of higher crystallization are created; consequently, welds are particularly prone to this electrolytic corrosion. The corrosion is not always obvious and can occur below the apparent surface capped by a porous layer [41]. While this is most studied with respect to chloride ions in water, conditions of low buffering capacity and low pH can also promote corrosion.
Please cite this article in press as: R. Mawson et al., Production of particulates from transducer erosion: Implications on food safety, Ultrason. Sonochem. (2014), http://dx.doi.org/10.1016/j.ultsonch.2014.04.005
8
R. Mawson et al. / Ultrasonics Sonochemistry xxx (2014) xxx–xxx
Table 2 Sizes of metallic particles (lm) found in carbon tape (evaporated, concentrated and dried water) and each of the filters (filtered water aliquots) for control, 18 kHz, 20 kHz, 0.4 MHz, 1 MHz, and 2 MHz transducers as determined from FE-SEM images.* Treatments
Carbon tapes#
0.01 lm membrane
0.05 lm membrane
0.1 lm membrane
Control low frequency (7.5 h) 18 kHz (7.5 h) 20 kHz (7.5 h) Control high frequency (7 days) 400 kHz (7 days) 1 MHz (7 days) 2 MHz (7 days)
12.9 6.76–15.10 5.08–7.98 4.24–13.80 0.99–5.62 0.21–3.98 0.25–25.7
nil nil nil nil 3.37 16.8 5.63–8.91
0.99–6.82 1.13 nil 1.59–5.76 5.73 0.84–28.8 2.54–3.87
3.71–5.06 0.66–2.10 0.43–0.69 0.36–1.97 2.40–5.55 1.20–1.40 8.02–70.1
*
Values represent the ranges of particle sizes found in two independent runs (size and composition of three particles were selected). Single values represent the value of a single particle found in carbon tapes or filters in two independent runs. # Results of three separate areas identified in the carbon tapes.
Table 3 Elemental analysis (via ICP and TC) of control and sonicated water. Trace element limits for drinking water is also included.*
*
Trace elements (lg/L original water)
Al
Cr
Co
Cu
Fe
Mn
Mo
Ni
Si
S
Ti
C
Drinking water limits LF CONTROL 18 kHz (HIELSCHER) 20 kHz (BRANSON) HF CONTROL 0.4 MHz 1 MHz 2 MHz
200 0.37 1.72 1.00 0.58 8.24 6.29 5.26
50 0.01 0.14 0.12 0.07 4.43 2.03 0.65
2.5 0.01 0.02 0.01 0.02 0.11 0.06 0.07
2000 2.15 15.76 6.31 2.97 7.15 91.89 13.14
300 2.38 9.89 3.74 2.05 182.39 63.45 19.18
100 0.28 0.79 0.23 0.18 2.83 1.27 2.77
50 0.02 0.02 0.02 0.02 0.73 0.24 0.06
20 0.29 0.76 0.37 0.28 5.42 2.55 2.00
5000 67.58 67.35 57.27 43.71 88.66 74.44 55.15
250000 12.29 12.04 7.47 7.50 11.20 12.47 8.03
700 0.04 3.87 2.57 0.07 0.18 0.92 0.05
N/A 0.61 0.90 0.71 0.44 0.36 0.84 0.39
Compiled from an international standard compilation listed in Wikipedia [37].
Table 4 Changes in pH and electrical conductivity in water during sonication at selected frequencies. Electrical conductivity (lS)
pH
Control 18 kHz (7.5 h) 20 kHz (7.5 h) 400 kHz (7 days) 1 MHz (7 days) 2 MHz (7 days) *
Start
End
Start
End
5.7 5.7 5.6 5.7 5.7 5.7
5.7* 5.5 5.2 2.5 3.0 4.4
2.4 2.4 2.4 1.2 1.2 1.2
2.1 1.5 1.3 986.0 362.0 27.7
7 days.
The role of ultrasound in accelerating corrosion is also complex and is the result of the interplay between the disruption of the passivation layer by cavitation and the impact of sonochemistry producing hydrogen peroxide which has the capacity to promote passivation [42] as demonstrated for zinc plated steel corrosion exposed to a variety of cations. The role of ultrasound on the corrosion of stainless steels specifically has been studied by [43–46] and similar effects were observed to those by [42] for zinc plating. Similarly, the combination of stainless steel corrosion at 55 kHz in the presence of nitric acid and small amounts of chloride ions [45,46] indicated an initial corrosion followed by subsequent passivation. The lack of buffering capacity in the purified water studied here is not typical of most food products which generally have significant buffering capacity. Some foods contain sodium chloride and acids which makes them aggressive in corroding stainless steel. Although in commercial practice the ultrasonic system would be set up in a continuous flow through mode at short resident times in the order of minutes, the observed corrosion in water would still be valid over the service life of the ultrasonic transducers in a continuous operation setting.
5. Conclusions and further recommendations Metallic nanoparticles were not detected within the size range of concern to health, that is, below 80 nm, therefore suggesting no evidence of health hazards from the use of sonotrodes and plate transducers in direct contact with food materials at a range of frequencies. However, due to the two-dimensional approach of the techniques adopted in this study, we have not totally precluded that nanoparticles may be present and warrant further investigation. Electrolytic corrosion is always a design concern with alloys such as stainless steel when water, particularly in the presence of chloride, is involved. However, the combination of sonochemistry and fine scale microstreaming could have the potential to accelerate this phenomenon and ultimately may cause ultrasonic treatment equipment failure. Nevertheless, while corrosion was evident, the dissolved metals in samples sonicated for up to 7 days were at concentrations well within international standards for water, even in a worst-case scenario at 400 kHz. In commercial practice, the ultrasonic system would be set up in a continuous flow through mode, thereby avoiding prolonged exposure of the treated material to the acoustic field, and the buffering capacity of food materials may further prevent corrosion reactions. While this study could not demonstrate the existence of any metallic nanoparticles in the size range of concern, corrosion is not a desirable outcome and some further study of the time course of metal erosion/corrosion using food materials in continuous systems over a range of ultrasonic frequencies and treatment durations is suggested. This study focuses on cavitation generated by transducers. It is important to highlight that similar cavitation phenomena can also be easily generated in hydraulic systems [47]. Hydrodynamic cavitation can occur using a constriction like an orifice plate, venturi or throttling valve in a liquid flow [48]. Moreover, equipment such as high-pressure homogenizers operate based on hydrodynamic cavitation phenomena. Hydrodynamic cavitation is not investigated here, but in light of the above medical concerns it should be
Please cite this article in press as: R. Mawson et al., Production of particulates from transducer erosion: Implications on food safety, Ultrason. Sonochem. (2014), http://dx.doi.org/10.1016/j.ultsonch.2014.04.005
R. Mawson et al. / Ultrasonics Sonochemistry xxx (2014) xxx–xxx
evaluated, although it would not be a trivial task as potentially hydrodynamic cavitation can arise in many food processing operations, e.g., pumping, mixing, wet milling and conventional homogenizing. Acknowledgements The authors would like to express their gratitude to the team at CSIRO Material Science and Engineering (Clayton, VIC, Australia) John Ward, Mark Greaves, and Karen Hands for their contributions in regards to profilometry and imaging and Malisja de Vries for her insights and contributions while reviewing this manuscript. References [1] M.S. Jellesen, A.A. Rasmussen, L.R. Hilbert, A review of metal release in the food industry, Mater. Corros. 57 (2006) 387–393. [2] T. Skotland, T.G. Iversen, K. Sandvig, New metal-based nanoparticles for intravenous use: requirements for clinical success with focus on medical imaging, Nanomedicine 6 (2010) 730–737. [3] P. Andujar, S. Lanone, P. Brochard, J. Boczkowski, Respiratory effects of manufactured nanoparticles, Rev. Mal. Respir. 26 (2009) 625–637. [4] B. Baroli, Penetration of nanoplarticles and nanomaterials in the skin: fiction or reality?, J Pharm. Sci. 99 (2010) 21–50. [5] J.W. Card, T.S. Jonaitis, S. Tafazoli, B.A. Magnuson, An appraisal of the published literature on the safety and toxicity of food-related nanomaterials, Crit. Rev. Toxicol. 41 (2011) 22–49. [6] C. Buzea, I.I. Pacheco, K. Robbie, Nanomaterials and nanoparticles: sources and toxicity, Biointerphases 2 (2007) MR17–MR71. [7] A. Elder, S. Vidyasagar, L. DeLouise, Physicochemical factors that affect metal and metal oxide nanoparticle passage across epithelial barriers, Wiley Interdiscip. Rev. Nanomed. Nanobiotechnol. 1 (2009) 434–450. [8] A.M. Gatti, Biocompatibility of micro- and nano-particles in the colon. Part II, Biomaterials 25 (2004) 385–392. [9] S. Barcikowski, A. Hahn, B.N. Chichkov, Nanoparticles as potential risk during femtosecond laser ablation, J. Laser Appl. 19 (2007) 65–73. [10] R. Landsiedel, L. Ma-Hock, A. Kroll, D. Hahn, J. Schnekenburger, K. Wiench, W. Wohlleben, Testing metal-oxide nanomaterials for human safety, Adv. Mater. 22 (2010) 2601–2627. [11] S. Barcikowski, J. Walter, A. Hahn, J. Koch, H. Haloui, T. Herrmann, A. Gatti, Picosecond and femtosecond laser machining may cause health risks related to nanoparticle emission, J. Laser Micro/Nanoeng. 4 (2009) 159–164. [12] A.A. Shemetov, I. Nabiev, A. Sukhanova, Molecular interaction of proteins and peptides with nanoparticles, ACS Nano 6 (2012) 4585–4602. [13] A.M. Alkilany, C.J. Murphy, Toxicity and cellular uptake of gold nanoparticles: what we have learned so far?, J Nanopart. Res. 12 (2010) 2313–2333. [14] J.S. Tsuji, A.D. Maynard, P.C. Howard, J.T. James, C.W. Lam, D.B. Warheit, A.B. Santamaria, Research strategies for safety evaluation of nanomaterials, part IV: risk assessment of nanoparticles, Toxicol. Sci. 89 (2006) 42–50. [15] N. Singh, B. Manshian, G.J.S. Jenkins, S.M. Griffiths, P.M. Williams, T.G.G. Maffeis, C.J. Wright, S.H. Doak, NanoGenotoxicology: the DNA damaging potential of engineered nanomaterials, Biomaterials 30 (2009) 3891–3914. [16] M.A. Maurer-Jones, Y.S. Lin, C.L. Haynes, Functional assessment of metal oxide nanoparticle toxicity in immune cells, ACS Nano 4 (2010) 3363–3373. [17] W.S. Cho, R. Duffin, C.A. Poland, S.E.M. Howie, W. MacNee, M. Bradley, I.L. Megson, K. Donaldson, Metal oxide nanoparticles induce unique inflammatory footprints in the lung: important implications for nanoparticle testing, Environ. Health Perspect. 118 (2010) 1699–1706. [18] M. Simko, M.O. Mattsson, Risks from accidental exposures to engineered nanoparticles and neurological health effects: a critical review, Part. Fibre Toxicol. 7 (2010). [19] Y.F. Li, Y.X. Gao, Z.F. Chai, C.Y. Chen, Nanometallomics: an emerging field studying the biological effects of metal-related nanomaterials, Metallomics 6 (2014) 220–232. [20] H.S. Chen, J. Li, A ring area formed around the erosion pit on 1 Cr18Ni9Ti stainless steel surface in incipient cavitation erosion, Wear 266 (2009) 884–887. [21] H.S. Chen, J. Li, S.H. Liu, Thermal effect at the incipient stage of cavitation erosion on a stainless steel in ultrasonic vibration cavitation, J. Fluids Eng. 131 (2009).
9
[22] Y.J. Chen, Strain-induced martensitic transformation of SUS304 stainless steel by cavitation micro-jet impact, J. Chin. Soc. Mechan. Eng. 30 (2009) 437–441. [23] K.Y. Chiu, F.T. Cheng, H.C. Man, Evolution of surface roughness of some metallic materials in cavitation erosion, Ultrasonics 43 (2005) 713–716. [24] G. Bregliozzi, A. Di Schino, S.I.U. Ahmed, J.M. Kenny, H. Haefke, Cavitation wear behaviour of austenitic stainless steels with different grain sizes, Wear 258 (2005) 503–510. [25] S. Barcikowski, A. Menendez-Manjon, B. Chichkov, M. Brikas, G. Raciukaitis, FlowGeneration of nanoparticle colloids by picosecond and femtosecond laser ablations in liquid, Appl. Phys. Lett. 91 (2007). 083113-1-083113-3. [26] A. Thiemann, T. Nowak, R. Mettin, F. Holsteyns, A. Lippert, Characterization of an acoustic cavitation bubble structure at 230 kHz, Ultrason. Sonochem. 18 (2011) 595–600. [27] M. Ashokkumar, D. Sunartio, S. Kentish, R. Mawson, L. Simons, K. Vilkhu, C. Versteeg, Modification of food ingredients by ultrasound to improve functionality: a preliminary study on a model system, Innovative Food Sci. Emerg. Technol. 9 (2008) 155–160. [28] A. Henglein, Contributions to various aspects of cavitation chemistry, Adv. Sonochem. 35 (1993) 17–83. [29] K. Okitsu, M. Ashokkumar, F. Grieser, Sonochemical synthesis of gold nanoparticles: effects of ultrasound frequency, J. Phys. Chem. B 109 (2005) 20673–20675. [30] F. Chemat Zill-e-Huma, M.K. Khan, Applications of ultrasound in food technology: processing, preservation and extraction, Ultrason. Sonochem. 18 (2011) 813–835. [31] E. Riera, J.A. Gallego-Juarez, T.J. Mason, Airborne ultrasound for the precipitation of smokes and powders and the destruction of foams, Ultrason. Sonochem. 13 (2006) 107–116. [32] S.H. Park, Y.R. Roh, C. Cooker, [WO 0113773] (2001). [33] P. Juliano, P. Swiergon, K.H. Lee, P.T. Gee, P.T. Clarke, Effects of pilot plant-scale ultrasound on palm oil separation and oil quality, J. Am. Oil Chem. Soc. 90 (2013) 1253–1260. [34] P. Juliano, P. Swiergon, R. Mawson, K. Knoerzer, M.A. Augustin, Application of ultrasound for oil separation and recovery of palm oil, J. Am. Oil Chem. Soc. 90 (2013) 579–588. [35] P. Juliano, A. Kutter, L.J. Cheng, P. Swiergon, R. Mawson, Enhanced creaming of milk fat globules in milk emulsions by the application of ultrasound and detection by means of optical methods, Ultrason. Sonochem. 18 (2011) 963–973. [36] P. Juliano, S. Temmel, M. Rout, P. Swiergon, R. Mawson, Creaming enhancement in a liter scale ultrasonic reactor at selected transducer configurations and frequencies, Ultrason. Sonochem. 20 (2013) 52–62. [37] Wikipedia, Drinking water quality standards (2013).
Ref Type: Internet Communication [38] Supeno, P. Kruus, Sonochemical formation of nitrate and nitrite in water, Ultrason. Sonochem. 7 (2000) 109–113. [39] A. Turnbull, Modeling of the chemistry and electrochemistry in cracks – a review, Corrosion 57 (2001) 175–189. [40] K.D. Ralston, N. Birbilis, C.H.J. Davies, Revealing the relationship between grain size and corrosion rate of metals, Scr. Mater. 63 (2010) 1201–1204. [41] N.J. Laycock, S.P. White, J.S. Noh, P.T. Wilson, R.C. Newman, Perforated covers for propagating pits, J. Electrochem. Soc. 145 (1998) 1101–1108. [42] V. Ligier, J.Y. Hihn, M. Wery, M. Tachez, J.Y. Hihn, The effects of 20 kHz and 500 kHz ultrasound on the corrosion of zinc precoated steels in [Cl] [SO2 4 ] [HCO3] [H2O2] electrolytes, J. Appl. Electrochem. 31 (2001) 213–222. [43] M. Doche, J. Hihn, A. Mandroyan, C. Maurice, O. Hervieux, X. Roizard, A novel accelerated corrosion test for exhaust systems by means of power ultrasound, Corros. Sci. 48 (2006) 4080–4093. [44] R. Wang, Influence of ultrasound on pitting corrosion and crevice corrosion of SUS304 stainless steel in chloride sodium aqueous solution, Corros. Sci. 50 (2008) 325–328. [45] G.O.H. Whillock, B.F. Harvey, Preliminary investigation of the ultrasonically enhanced corrosion of stainless steel in the nitric acid chloride system, Ultrason. Sonochem. 3 (1996) S111–S118. [46] G.O.H. Whillock, B.F. Harvey, Ultrasonically enhanced corrosion of 304L stainless steel.2. The effect of frequency, acoustic power and horn to specimen distance, Ultrason. Sonochem. 4 (1997) 33–38. [47] P.R. Gogate, Hydrodynamic cavitation for food and water processing, Food Bioprocess Technol. 4 (2011) 996–1011. [48] P.R. Gogate, V.S. Sutkar, A.B. Pandit, Sonochemical reactors: important design and scale up considerations with a special emphasis on heterogeneous systems, Chem. Eng. J. 166 (2011) 1066–1082.
Please cite this article in press as: R. Mawson et al., Production of particulates from transducer erosion: Implications on food safety, Ultrason. Sonochem. (2014), http://dx.doi.org/10.1016/j.ultsonch.2014.04.005
Contents lists available at ScienceDirect
Ultrasonics Sonochemistry journal homepage: www.elsevier.com/locate/ultson
Production of particulates from transducer erosion: Implications on food safety Raymond Mawson, Manoj Rout, Gabriela Ripoll, Piotr Swiergon, Tanoj Singh, Kai Knoerzer, Pablo Juliano ⇑ CSIRO Division of Animal, Food and Health Sciences, 671 Sneydes Road, Werribee, Vic 3030, Australia
a r t i c l e
i n f o
Article history: Received 1 November 2013 Received in revised form 3 April 2014 Accepted 3 April 2014 Available online xxxx Keywords: Metallic particles Cavitation Ultrasound Safety Health Food processing
a b s t r a c t The formation of metallic particulates from erosion was investigated by running a series of transducers at various frequencies in water. Two low frequency transducer sonotrodes were run for 7.5 h at 18 kHz and 20 kHz. Three high frequency plates operating at megasonic frequencies of 0.4 MHz, 1 MHz, and 2 MHz were run over a 7 days period. Electrical conductivity and pH of the solution were measured before and after each run. A portion of the non-sonicated and treated water was partially evaporated to achieve an 80-fold concentration of particles and then sieved through nano-filters of 0.1 lm, 0.05 lm, and 0.01 lm. An aliquot of the evaporated liquid was also completely dried on strips of carbon tape to determine the presence of finer particles post sieving. An aliquot was analyzed for detection of 11 trace elements by Inductively Coupled Plasma Mass Spectroscopy (ICPMS). The filters and carbon tapes were analyzed by FE-SEM imaging to track the presence of metals by EDS (Energy Dispersive Spectroscopy) and measure the particle size and approximate composition of individual particles detected. Light microscopy visualization was used to calculate the area occupied by the particles present in each filter and high resolution photography was used for visualization of sonotrode surfaces. The roughness of all transducers before and after sonication was tested through profilometry. No evidence of formation of nano-particles was found at any tested frequency. High amounts of metallic micron-sized particles at 18 kHz and 20 kHz formed within a day, while after 7 day runs only a few metallic micro particles were detected above 0.4 MHz. Erosion was corroborated by an increase in roughness in the 20 kHz tip after ultrasound. The elemental analysis showed that metal leach occurred but values remained below accepted drinking water limits, even after excessively long exposure to ultrasound. With the proviso that the particles measured here were only characterized in two dimensions and could be nanoparticulate in terms of the third dimension, this research suggests that there are no serious health implications resulting from the formation of nanoparticles under the evaluation conditions. Therefore, high frequency transducer plates can be safely operated in direct contact with foods. However, due to significant production of metallic micro-particulates, redesign of lower frequency sonotrodes and reaction chambers is advised to enable operation in various food processing direct-contact applications. Ó 2014 Elsevier B.V. All rights reserved.
1. Introduction Many food processing unit operations may release metallic particles into food from either cavitation induced processes on fluids by pumping or homogenization, or from direct erosion of equipment from pneumatic conveying of dry food ingredients and milling operations. The presence of metallic particles in food as consumed has however, been mostly associated to the use of domestic cookware [1]. Even though it has not been formally established in the published literature, there remains a concern ⇑ Corresponding author. E-mail address: [email protected] (P. Juliano).
that cavitation-related release of metallic particles into foods from sonication with ultrasonic transducers probes in direct contact with the food material. Therefore, it is important to consider firstly, the cavitation erosion of metals in direct contact with foodstuffs and secondly, the implications of the consumption of metallic particles on food safety. 1.1. Safety implications on human intake of metallic nanoparticles Most literature focuses on the medical application of metal as potential imaging contrast agents [2], the safety of inhaled metal nanoparticles [3] and skin penetration [4]. There is also some literature on the potential nanotoxicity of ingested nanoparticles based
http://dx.doi.org/10.1016/j.ultsonch.2014.04.005 1350-4177/Ó 2014 Elsevier B.V. All rights reserved.
Please cite this article in press as: R. Mawson et al., Production of particulates from transducer erosion: Implications on food safety, Ultrason. Sonochem. (2014), http://dx.doi.org/10.1016/j.ultsonch.2014.04.005
2
R. Mawson et al. / Ultrasonics Sonochemistry xxx (2014) xxx–xxx
on in vitro and in vivo studies using animal models [5–7] and post mortem examination of humans [8]. Mucus ingested creates a secondary hazard from the inhalation of nanoparticles [9]. The highest concern for human safety are bio-persistent nanoparticles [10], which include most metallic particulates. For consideration of biological consequences a metal nanoparticle ranges between 5 nm and 100 nm. The impact of nanoparticulate metals in biological systems depends on the particle characteristics such as shape, surface characteristics and surface energy differences [11]. However, the mechanisms behind the interaction of bio-persistent metal nanoparticles and the biological systems are not well understood. It is believed that biological damage from bio-persistent metal nanoparticles can extend from tissue inflammatory responses due to homeostasis disturbance by protein and peptide malformation [12] to chromosomal fragmentation and DNA cleavage with consequent apoptosis or potential cancer [3,11,13–17] and/or brain damage [18]. While generalizations about nanoparticles may not be valid [5], metal particles in particular seem to have little difficulty in penetrating the gut mucosal layer mainly through the gut lining cells. Entry into these cells appears to be limited to metal particles of less than 80 nm and only particles of less than 50 nm may find passage into the circulation system. Particles larger than 15 nm are captured by the liver and may be tagged for transport to the kidneys; in either case they can be excreted in bile or urine. Particles less than approximately 15 nm become fully systemic and may enter almost every cell in the body including the brain and bone marrow [5,6]. However, it is important to note that the majority of metal nanoparticles ingested will likely be excreted in the faeces and some predisposing condition in the gut may be required before any nanoparticles can enter the gut lining cells [5–8]. The biological impacts from exposure are many which have spawned a new science discipline ‘‘metallomics’’ the study of metallic nano- particles in biological systems and their impacts on life [19]. There is a new Journal ‘‘Metallomics’’ published under the Royal Society Proceedings series devoted to this topic. In summary, the literature reveals that metal nanoparticles or their oxides with a size less than 80 nm cannot be regarded as safe in foodstuffs. 1.2. Mechanisms of cavitation erosion in foodgrade metallic surfaces (stainless steel and titanium) Metallurgical product advancements have resulted in making stainless steel more cavitation erosion resistant; unfortunately, these improvements are largely unsuitable for food processing environments. For example, many systems for making stainless steel corrosion resistant involve addition of metals considered toxic in food. Two mechanisms are proposed for erosion by cavitation (induced by 20 kHz ultrasound at an amplitude of 6 lm): surface hardening by thermal tempering triggered by heat transferred from surface imploding cavitation bubbles [20,21] and work hardening from repeated impact by cavitation microjets from surface bubbles [22,23]. The surface hardening has been measured and there are indications of localized temperatures in excess of 300 °C, causing localized changes in the alloy composition. In either case, the effects will be confined to the surface molecular layers. Cavitation erosion is strongly influenced by the stainless steel alloy crystal size, the larger the crystals the greater the erosion, which suggests that the erosion is likely to be on a very small distance scale [24]. Growth of alloy crystals beneath the surface weakens the surface structure, making it more prone to erosion by fracture and flaking away of the hardened surface material. Little to no direct evidence of metal particle characterization from cavitation erosion by ultrasonic transducers has been reported. Evidence of silver erosion from cavitation in water
induced by pulsed laser was reported, where particles predominantly smaller than 50 nm at 50 and 100 kHz pulse frequency, at an energy density of 1.2–2.9 J/cm2 were produced [25]. Silver is more difficult to ablate than either stainless steel or titanium [9] due to its higher ductility. However, it is not acceptable to generalize erosion characteristics from cavitation. Cavitation pits self pacify and stop eroding within themselves; however, the pacified pits spawn new erosion pits in the surrounding metal. Some of these observations were made at 230 kHz using a ‘‘seed’’ pit or gas bubble [26]. The cavitation pits increase surface fluid flow turbulence and create crevices that act as nucleation sites for additional cavitation bubbles. Stainless steel ablates as the stainless steel alloy nanoparticles (not its component metals). From surface energy considerations the alloy is more prone to cavitation damage than the pure components would be [11]. Another effect caused by the generation of the bubbles while passing high intensity ultrasound through a liquid, is the decomposition of the solvent and solutes present within the bubbles, when the bubbles are formed and then collapse, generating several highly reactive radicals [27]. In the case of water, the radicals generated are H and OH [28] (Reaction (1)).
H2 OÞÞÞÞÞH þ OH
ð1Þ
The presence of these radicals can induce the sonochemical production of metal nanoparticles [29]. Ashokkumar et al. [27] found that in a water medium there is an increase of the OH yield with an increase in ultrasound frequency from 20 kHz to 358 kHz and, for higher frequencies (>450 kHz), the yield starts dropping until reaching insignificant concentrations of the radicals. The explanation for this decrease is that at higher frequencies the acoustic cycle becomes very short, limiting the size of the generated cavitation bubbles, and the bubbles contain headspace gas restricting the amount of water vapor generated within the collapsing bubble, and therefore, there are less OH radicals generated. This would indicate that no nanoparticles should be generated because of radical formation when frequencies in the MHz range are applied in a liquid medium. 1.3. Application of ultrasound in the food industry Ultrasound is an emerging technology in the food industry. It has been used for processing, preservation and extraction operations. Some of the advantages of using ultrasound are a more effective mixing and micro-mixing, faster energy and mass transfer, reduced thermal and concentration gradients, reduced processing temperature, selective extraction and reduced equipment size [30]. Currently, the food industry is using low frequency sonotrodes in applications such as: cleaning baths, emulsification [27], defoaming of carbonic beverages produced in high-speed bottling and canning lines [31], improved overall frying [32], ultrasonic food cutting, and ultrasonic osmotic dehydration [30]. There are other emerging applications utilizing high frequency transducer plates in the megasonic range (>400 kHz) for enhancing the separation of milk fat from milk and recovering palm oil in the palm oil milling process [33–36]. In light of the above summation, an attempt is made to characterize the metallic particles likely to be produced by current ultrasonic technology that could be used or is being used for food processing. 2. Materials and methods 2.1. Ultrasonic runs Fig. 1 describes the experimental protocol employed for each ultrasonic run. Double distilled water (16 L, initial pH = 5.5, initial
Please cite this article in press as: R. Mawson et al., Production of particulates from transducer erosion: Implications on food safety, Ultrason. Sonochem. (2014), http://dx.doi.org/10.1016/j.ultsonch.2014.04.005
R. Mawson et al. / Ultrasonics Sonochemistry xxx (2014) xxx–xxx
electrical conductivity 1.3 lS) was placed in food grade 316 stainless steel vessels for high frequency (plate transducers) treatment. Transducers were configured in the vessels as shown in Fig. 2. High frequency plate transducers (Submersible Transducers, Sonosys Ultraschallsysteme GmbH, Neuenbuerg, Germany, Stainless Steel AISI 303 – iron, chromium, nickel, copper) were continuously operated at 0.4 MHz, 1 MHz, and 2 MHz and matched power of 220 W in three jacketed vessels (grade 316 steel) for 7 days. All ultrasonic runs were performed in duplicate. The high frequency runs were performed under temperature controlled conditions (20 ± 3 °C, by circulating glycol in the jacketed vessel). External contamination was avoided through stainless steel covers of the same 316 grade. Attempts were made to evaluate an 80 kHz and 220 W ultrasonic system over a 7 days duration in the 16 L tanks, but on three occasions the welds in the treatment vessel eroded through and the vessel drained overnight within 24 h. Consequently no data could be collected and other than noting that these conditions promoted rapid erosion there is no further discussion of these conditions. Two vessels (grade 304 stainless steel) were set up with a low frequency transducer (or sonotrode), each operating continuously for 7.5 h at 18 kHz (Dr. Hielscher GmbH, UIP 1000, Teltow, Germany) with power input of 180 W, and 20 kHz (Branson Digital Sonifier Model 250, Danburry, CT, USA) with power input of 103 W, respectively. The low frequency runs started at room temperature reaching a maximum of 42 °C and 34 °C for the 18 kHz and 20 kHz transducers, respectively. Based on preliminary results, the experimental design was put together to achieve a worst case scenario for high frequency sonication conditions (extended sonication of 7 days and higher power). Separate controls (non-ultrasound treatment) were run in parallel to the low and high frequency vessels using the same water. A float control sensor was placed in the high frequency vessels treated for up to 7 days to ensure that the plate transducers remained below the water surface and avoid damage. A power meter (F7C005au, BELKIN, China) was utilized to measure the power drawn from the generators. At the end of the runs the electrical conductivity and pH were measured. Subsamples (4 L) were collected from the vessels and placed in glass beakers and reduced
3
to 50 mL by evaporation using an 80 °C water bath. The concentrated samples were further subsampled for trace elemental analysis (20 mL aliquots). 2.2. Characterization of transducer erosion 2.2.1. Photography After the ultrasound runs each transducer was visually observed for erosion and photographs were taken (Canon SLR 1100DSLR, Tokyo, Japan). 2.2.2. Surface profilometry The roughness was measured using an AltiSurf 500 Surface Profilometer (Bruker Biosciences, Preston, VIC) using a 300 lm optical probe. The x–y stage traversed at 500 lm/s with 0.5 lm spacing for profiles and 5 lm spacing for area tests. The profile test measured 10 lines of the active surface of the transducer. Average roughness (Ra) values were measured for every line in x and y directions. The average roughness was calculated as the integrated area by the roughness profile of each line over the evaluation length. The geometry of the Hielscher 20 kHz probe precluded any profilometry being done on this probe and there is no discussion of this parameter relating to this probe. 2.3. Characterization of dissolved and particulate matter in water 2.3.1. Sample preparation A 50 mL volume of the concentrated non-sonicated and sonicated water was filtered sequentially three times using different pore sized polycarbonate membranes in the following order: 0.1 lm (Isopore), 0.05 lm (Poretics), 0.01 lm (Poretics). Another 1 mL of the concentrated water was dried at ambient temperature in a desiccator for 24 h onto the carbon tape utilized for SEM imaging. The membranes were dried in a desiccator and both carbon tapes and membranes were stored in petri dishes. Particles trapped on each carbon tape or on membranes were later visualized by field emission scanning electron microscopy equipped with Oxford instruments energy dispersive spectroscopy for elemental analysis (FE,SEM-EDS).
Fig. 1. Experimental protocol followed for the untreated double distilled water samples and all transducer frequencies to evaluate erosion extent through a number of analytical methods.
Please cite this article in press as: R. Mawson et al., Production of particulates from transducer erosion: Implications on food safety, Ultrason. Sonochem. (2014), http://dx.doi.org/10.1016/j.ultsonch.2014.04.005
4
R. Mawson et al. / Ultrasonics Sonochemistry xxx (2014) xxx–xxx
Fig. 2. Schematic diagram of ultrasonic systems utilized for erosion trials (left: low frequency ultrasonics system, right: high frequency megasonics systems). The vessels were made of food grade stainless steel.
2.3.2. FE,SEM-EDS analysis The filters were mounted on an aluminium stub with doublesided conductive carbon tape. These samples and the carbon tapes carrying dried water concentrate were then carbon coated using a Polaron E6700 carbon sputter coater. The thickness of the carbon coating was approximately 20 nm. Once coated, the samples were placed into a Philips XL30 Field Emission Scanning Electron microscope (FE,SEM). The EDS (Energy Dispersive Spectroscopy) X-ray analysis system (Aztec, Oxford Instruments Pty. Ltd., Thornleigh, NSW, Australia) with an accelerating voltage of 20 kV was used for the analysis. SEM images of the filters were obtained as well as the size and chemical composition (EDS) profile of the particles detected on them. 2.3.3. Low magnification microscopy All nano sieved membranes with appropriate controls were observed under a light microscope (Leitz light microscope, soft imaging system Color view IIIu, Japan) and pictures were taken at three different locations at 40 times magnification. In brief, a small piece (1 cm 1 cm) of each membrane was cut and placed on a glass slide and a coverslip was applied. Images were taken at three different locations at a resolution of 250 lm. Image analysis was carried out via a purpose developed software (Fig. 3) programmed in MATLAB2011™ (The Mathworks Inc., Natick, MA, USA). The image was first converted into a matrix with a 2D resolution identical to the original image. Then a histogram was plotted from which a threshold value was determined to generate a binary image. From the binary image, the relative area covered by the particles collected in the membranes (area fraction) was calculated by dividing the area covered with particles by the complete area of the image. 2.3.4. Metal trace analysis The concentrated samples were filtered through a 0.45 lm filter and then measured by Inductively Coupled Plasma Mass Spectroscopy (ICPMS) for: Aluminium (Al), Chromium (Cr), Cobalt (Co), Copper (Cu), Iron (Fe), Manganese (Mn), Nickel (Ni), Tin (Sn), Sulfur (S), Molybdenum (Mb), and Silicon (Si), which are the elements present in food grade stainless steel (material of the vats and of the transducer casing). Unfiltered samples were analysed for Total carbon (TC) using the AnalytikJena Multi N/C 3100 Carbon Analyser (Wembley, UK), which oxidizes all organic and inorganic carbon at 680 °C to CO2 and carbon is then measured by a detector. Inorganic carbon (IC) is then determined by addition of phosphoric acid to the sample to convert carbonates to CO2. Concentration values are expressed on an original water sample basis by correcting with an 80-fold evaporation factor.
Fig. 3. Methodology of micrograph analysis according to the presence of particles on the nanofilter.
2.3.5. pH measurement and electrical conductivity Water samples were taken at the beginning and end of each ultrasonic run. Measurements of pH and electrical conductivity (#WP-84 Conductivity-TDS-Temp. Meter, TPS Pty Ltd., Springwood, QLD, Australia) were carried out. 3. Results Particle erosion was established after low frequency sonication in water while no erosion was detected after sonication in the high frequency systems. This was established by various methods after assessment of the transducers themselves before and after sonication and by analysing dissolved and particulate matter in the processed water.
Please cite this article in press as: R. Mawson et al., Production of particulates from transducer erosion: Implications on food safety, Ultrason. Sonochem. (2014), http://dx.doi.org/10.1016/j.ultsonch.2014.04.005
R. Mawson et al. / Ultrasonics Sonochemistry xxx (2014) xxx–xxx
3.1. Visual observations and transducer roughness Fig. 4 shows photographic evidence of pitting or indentation of the low frequency sonotrodes (18 kHz and 20 kHz) after 7.5 h continuous processing in water. However, no visual evidence of physical damage could be detected after 7 days of sonication of water with high frequency plate transducers. Eroded particles were collected in all membranes of the water concentrate obtained from the 18 kHz run. Particle deposits seen on the 0.1 lm as well as the other pore size membranes are shown in Fig. 5. Metallic particles were also visually detected on the membranes obtained from 20 kHz treated water but to a lesser extent. However, the membranes obtained from the high frequency treated water and the control water did not show any deposition residue. Visual observations of the nano-membrane filter discs were examined under an optical light microscope. The particles were visualized at three different locations of the membrane to form an impression of the particle shape and size. The presence of particles at different size ranges indicated the wide distribution of nano-materials retained during molecular sieving. The visual observations by the naked eye found that thick black powdery material was deposited on the surface of the 0.1 lm membrane
Fig. 4. Erosion occurring after ultrasound treatments carried out for 7.5 h in distilled water. (a) An 18 kHz Hielscher tip before and after ultrasound, (b) a brand new 20 kHz Branson tip after ultrasound (see demarked area on edges) and its corresponding profilometry data before and after ultrasound. Blue indicates the deeper section and red fading up to white the higher region. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
5
from ultrasonically treated water by the 18 kHz transducer. This indicated that some local erosion had taken place which was validated by FE,SEM analysis and the corresponding nano-material composition was characterized by EDS. The visual observations in the 20 kHz transducer tip (Fig. 4b) are corroborated by the changes in surface depth obtained from the profilometer analysis. The roughness value significantly increased for the 20 kHz probe from 0.36 lm to 0.80 lm (Table 1) while no major changes in roughness value were seen for any of the high frequency transducers. However, the rest of the nano-sieved ultrasonically treated water for a 7-days period did not show any deposition on the membrane surface, indicating there was no significant erosion producing nanoparticles as evident from the clear membrane filters observed after molecular sieving. 3.2. Nano-sieve imaging and water composition Low resolution magnification imaging allowed the calculation of the relative area covered by the particles collected in the membranes (area fraction). The area fraction (Fig. 5) changed significantly for both membranes obtained from the 18 kHz and 20 kHz transducer runs after 7.5 h. This indicated that significant amount of erosion material (represented by particles of various shapes and sizes) was removed from the sonotrodes. Formation of metallic particles from erosion at 18 kHz is also shown in optical images in Fig. 5. However, the area fractions for the membranes obtained from the non-sonicated controls and the high frequency did not change after the 7-day runs. Fig. 5 shows the prevalence of nonmetallic particles in the control water. FE,SEM images (Figs. 6–8) illustrate that the purely metallic particles, where identified, appeared to be in flake form rather than the expected spherical form if the surface of the ultrasonic horn had been subject to melting. For example, Particle 1 in Fig. 6 clearly demonstrates the flake like structure of particles that are clearly metallic in nature from the FE,SEM-EDS analysis. The data obtained from FE,SEM-EDS individual particle analysis are not presented but are summarized in the following discussion with respect to particular treatments. The particle analysis from the 400 kHz trial on the 0.1 lm filter reveals that Particle 1 (Fig. 7a) contains traces of iron, chromium, aluminium and magnesium. Particle 2 (Fig. 7a) contains moderate amounts of iron and traces of copper and chromium. On the 0.05 lm filter, Particle 1 (Fig. 7b) contains traces of iron, titanium, magnesium and moderately large amounts of aluminium. Particle 2 (Fig. 7b) contains traces of aluminium only. In Fig. 6, from the Hielscher 18 kHz horn, Particle 1 contains largely titanium and aluminium. Fig. 8 illustrates the particles obtained from the 20 kHz Branson ultrasonic horn collected on the 0.1 lm filter. Particle 1 (Fig. 8b) contains large amounts of titanium, moderate amounts of vanadium, and aluminium and traces of iron. Particle 3 (Fig. 8b) contains large amounts of titanium, moderate amounts of vanadium, aluminium and traces of iron and copper. However, Fig. 7c also illustrates the nature of fuzzy aggregated particles collected on the 0.01 lm filter from the 2 MHz treatment. Particle 1 (Fig. 8a) contains traces of iron in the control (non-sonicated) water sample. Particle 2 (Fig. 8a) contains traces of iron and copper magnesium and a small amount of aluminium. Table 2 shows the size ranges of particles selected from the carbon tapes and from the membranes. In general, all particles detected by FE,SEM exceeded 0.21 lm, i.e., no nanoparticles could be detected in any run. Metallic particles were detected in the controls as well as in the sonicated waters. Therefore, the fact that metallic particles were detected in the high frequency samples is not indicative of erosion.
Please cite this article in press as: R. Mawson et al., Production of particulates from transducer erosion: Implications on food safety, Ultrason. Sonochem. (2014), http://dx.doi.org/10.1016/j.ultsonch.2014.04.005
6
R. Mawson et al. / Ultrasonics Sonochemistry xxx (2014) xxx–xxx
Fig. 5. Filter photographs with digitally processed micrographs of a small section of the filter underneath. The values represent particle area fractions calculated for each micrographs. Imaging of 0.1 lm nanofilters used to follow up the formation of particles in the following water samples: control low frequency (LF, non-US, 7.5 h run), 18 kHz Hielscher transducer (7.5 h run), 20 kHz Branson transducer (7.5 h run). The ⁄ symbol indicates that the area fraction for the 18 kHz 7.5 h treatment is significantly different (P < 0.05) to the 20 kHz treatment. Control high frequency (HF, non-US, 7-days run) and the 400 kHz, 1 MHz, and 2 MHz Sonosys transducers (7-day runs). No particles are evident in any of these treatments.
Table 1 Surface profilometry indicating roughness Ra (lm, ±standard deviation) of: (a) unused 20 kHz sonotrode tip before and after a 7.5 h ultrasonic run in water and (b) transducer plates before and after 0.4 MHz, 1 MHz, and 2 MHz 7-day runs. Ra data obtained in the X and Y direction data was pooled together and taken as a total average for each transducer before and after sonication. Different superscript letters indicate significant differences obtained for the same transducer (P < 0.05). Transducer type
Ra (lm) before US run
Ra (lm) after US run
20 kHz 0.4 MHz 1 MHz 2 MHz
0.36 ± 0.04a 0.95 ± 0.05a 1.38 ± 0.06b 1.22 ± 0.12a
0.80 ± 0.08b 0.94 ± 0.06a 1.34 ± 0.04a 1.15 ± 0.15a
The trace elemental analysis shown in Table 3 demonstrates that the metals in solution, even after excessively long sonication, remain below drinking water limits [37]. Low frequency ultrasound increased the levels of Fe and Ti in solution, where 18 kHz gave higher values than 20 kHz. However, no changes in Ti were seen in the high frequency solutions compared to the 7-days control. On the other hand Al, Cr Cu, Fe, and Ni increased after high frequency ultrasound processing with respect to the 7-days control. The most marked increase occurred at 400 kHz for Fe. The pH for the control water remained unchanged over the 7days period and the electrical conductivity showed only a marginal change (Table 4). The low frequency treatment also remained unchanged in pH and electrical conductivity after sonication over a period of 7.5 h. However, at higher frequencies the pH significantly decreased after 7 days, being the most significant effect seen at 400 kHz. The electrical conductivity also changed due to a decrease in pH. The conductivity increased most markedly at 400 kHz followed by 1 MHz and 2 MHz.
Fig. 6. Scanning electron microscopy image showing an example of the flake structure of metallic particles obtained from the 18 kHz Hielscher on a 0.1 lm nanosieve.
4. Discussion All of the FE,SEM-EDS and soluble element data presented with a background contamination of silica that could be derived from the glass vessels used to evaporate the treated water samples. It is likely that the particulate matter observed on filter membranes in the critical range between 10 and 50 nm, is due to the formation of aggregated deposits that include metal ions.
Please cite this article in press as: R. Mawson et al., Production of particulates from transducer erosion: Implications on food safety, Ultrason. Sonochem. (2014), http://dx.doi.org/10.1016/j.ultsonch.2014.04.005
R. Mawson et al. / Ultrasonics Sonochemistry xxx (2014) xxx–xxx
7
Fig. 8. FE,SEM images of (a) the control (non-US) and (b) the 20 kHz water samples after filtering on a 0.1 lm nanosieve.
Fig. 7. FE,SEM of the 400 kHz (a) after filtering on a 0.1 lm and (b) the filtrate on a 0.05 lm nanosieve and of (c) the 2 MHz water samples after filtering on a 0.1 lm nanosieve.
The stainless steel used in the construction of the treatment tanks for the low frequency studies was grade 304. The high amplitude low frequency 20 kHz Branson ultrasonic horn was tipped with a screw-in plate of a harder alloy than the alloy used for the bulk of the horn, while the 18 kHz Hielscher ultrasonic horn is un-tipped and the composition of the particles eroded reflect a typical composition of an acoustic grade of titanium alloy. The treatment tanks used in the higher frequency studies were made from stainless steel alloy 316. The ultrasonic plate transducers used for the 400 kHz, 1 MHz and 2 MHz equipment operate at very low intensity per unit transducer surface area and it was not expected that there would be any significant erosion at the surfaces of these plates as is confirmed by the profilometry data (Table 1). It has been reported [21] that pitting in the surfaces of metals exposed to ultrasound is caused by the surface being locally melted by heat transferred from the contacting cavitation bubble. In our work we fail to see any evidence of this at either the surface of the ultrasonic horn or the surface of the stainless steel container. Instead, the metallic particles have a flake-like characteristic suggesting that they have been mechanically removed from a case hardened surface induced by ultrasonic peening. It is possible that the nature of this hardening promotes the variation of the alloy
crystalline structure beneath the surface hardened layer. This compositional variation may promote localized accelerated electrolytic corrosion. In our studies, conducted under a headspace of air at frequencies of 1 MHz and 2 MHz, the cavitation bubbles would have been filled with air rather than predominantly water vapor resulting in the synthesis of oxides of nitrogen and subsequently nitric acid rather than hydrogen peroxide [38], as can be seen from the declining pH and increasing conductivity of our sonicated water (Table 4). At 400 kHz, an early synthesis of nitric acid is likely to have been seen with concurrent generation of larger amounts of hydrogen peroxide so although the net corrosion was higher at 400 kHz the corrosion pathway is likely to have been rapid initially and then reducing whereas at the higher frequency the corrosion pathway would have been a slow progression towards more rapid corrosion over the duration of treatment. However, as samples were not taken during each run, this hypothesis needs to be further investigated. The presence of dissolved metal ions other than would be expected is most likely due to corrosion of the stainless steel tanks rather than erosion, especially at the frequencies where no evidence of metallic particulate material was observed. The corrosion of stainless steel is a complex topic and has been reviewed with respect to crevices and cracks [39] and at a crystal structure level [40]. Corrosion is promoted when the passivation layer on the surface of the stainless steel is breached permitting small scale electrolytic reactions to take place between the crystals of different metal species that make up the alloy. Where stainless steel is welded, zones of higher crystallization are created; consequently, welds are particularly prone to this electrolytic corrosion. The corrosion is not always obvious and can occur below the apparent surface capped by a porous layer [41]. While this is most studied with respect to chloride ions in water, conditions of low buffering capacity and low pH can also promote corrosion.
Please cite this article in press as: R. Mawson et al., Production of particulates from transducer erosion: Implications on food safety, Ultrason. Sonochem. (2014), http://dx.doi.org/10.1016/j.ultsonch.2014.04.005
8
R. Mawson et al. / Ultrasonics Sonochemistry xxx (2014) xxx–xxx
Table 2 Sizes of metallic particles (lm) found in carbon tape (evaporated, concentrated and dried water) and each of the filters (filtered water aliquots) for control, 18 kHz, 20 kHz, 0.4 MHz, 1 MHz, and 2 MHz transducers as determined from FE-SEM images.* Treatments
Carbon tapes#
0.01 lm membrane
0.05 lm membrane
0.1 lm membrane
Control low frequency (7.5 h) 18 kHz (7.5 h) 20 kHz (7.5 h) Control high frequency (7 days) 400 kHz (7 days) 1 MHz (7 days) 2 MHz (7 days)
12.9 6.76–15.10 5.08–7.98 4.24–13.80 0.99–5.62 0.21–3.98 0.25–25.7
nil nil nil nil 3.37 16.8 5.63–8.91
0.99–6.82 1.13 nil 1.59–5.76 5.73 0.84–28.8 2.54–3.87
3.71–5.06 0.66–2.10 0.43–0.69 0.36–1.97 2.40–5.55 1.20–1.40 8.02–70.1
*
Values represent the ranges of particle sizes found in two independent runs (size and composition of three particles were selected). Single values represent the value of a single particle found in carbon tapes or filters in two independent runs. # Results of three separate areas identified in the carbon tapes.
Table 3 Elemental analysis (via ICP and TC) of control and sonicated water. Trace element limits for drinking water is also included.*
*
Trace elements (lg/L original water)
Al
Cr
Co
Cu
Fe
Mn
Mo
Ni
Si
S
Ti
C
Drinking water limits LF CONTROL 18 kHz (HIELSCHER) 20 kHz (BRANSON) HF CONTROL 0.4 MHz 1 MHz 2 MHz
200 0.37 1.72 1.00 0.58 8.24 6.29 5.26
50 0.01 0.14 0.12 0.07 4.43 2.03 0.65
2.5 0.01 0.02 0.01 0.02 0.11 0.06 0.07
2000 2.15 15.76 6.31 2.97 7.15 91.89 13.14
300 2.38 9.89 3.74 2.05 182.39 63.45 19.18
100 0.28 0.79 0.23 0.18 2.83 1.27 2.77
50 0.02 0.02 0.02 0.02 0.73 0.24 0.06
20 0.29 0.76 0.37 0.28 5.42 2.55 2.00
5000 67.58 67.35 57.27 43.71 88.66 74.44 55.15
250000 12.29 12.04 7.47 7.50 11.20 12.47 8.03
700 0.04 3.87 2.57 0.07 0.18 0.92 0.05
N/A 0.61 0.90 0.71 0.44 0.36 0.84 0.39
Compiled from an international standard compilation listed in Wikipedia [37].
Table 4 Changes in pH and electrical conductivity in water during sonication at selected frequencies. Electrical conductivity (lS)
pH
Control 18 kHz (7.5 h) 20 kHz (7.5 h) 400 kHz (7 days) 1 MHz (7 days) 2 MHz (7 days) *
Start
End
Start
End
5.7 5.7 5.6 5.7 5.7 5.7
5.7* 5.5 5.2 2.5 3.0 4.4
2.4 2.4 2.4 1.2 1.2 1.2
2.1 1.5 1.3 986.0 362.0 27.7
7 days.
The role of ultrasound in accelerating corrosion is also complex and is the result of the interplay between the disruption of the passivation layer by cavitation and the impact of sonochemistry producing hydrogen peroxide which has the capacity to promote passivation [42] as demonstrated for zinc plated steel corrosion exposed to a variety of cations. The role of ultrasound on the corrosion of stainless steels specifically has been studied by [43–46] and similar effects were observed to those by [42] for zinc plating. Similarly, the combination of stainless steel corrosion at 55 kHz in the presence of nitric acid and small amounts of chloride ions [45,46] indicated an initial corrosion followed by subsequent passivation. The lack of buffering capacity in the purified water studied here is not typical of most food products which generally have significant buffering capacity. Some foods contain sodium chloride and acids which makes them aggressive in corroding stainless steel. Although in commercial practice the ultrasonic system would be set up in a continuous flow through mode at short resident times in the order of minutes, the observed corrosion in water would still be valid over the service life of the ultrasonic transducers in a continuous operation setting.
5. Conclusions and further recommendations Metallic nanoparticles were not detected within the size range of concern to health, that is, below 80 nm, therefore suggesting no evidence of health hazards from the use of sonotrodes and plate transducers in direct contact with food materials at a range of frequencies. However, due to the two-dimensional approach of the techniques adopted in this study, we have not totally precluded that nanoparticles may be present and warrant further investigation. Electrolytic corrosion is always a design concern with alloys such as stainless steel when water, particularly in the presence of chloride, is involved. However, the combination of sonochemistry and fine scale microstreaming could have the potential to accelerate this phenomenon and ultimately may cause ultrasonic treatment equipment failure. Nevertheless, while corrosion was evident, the dissolved metals in samples sonicated for up to 7 days were at concentrations well within international standards for water, even in a worst-case scenario at 400 kHz. In commercial practice, the ultrasonic system would be set up in a continuous flow through mode, thereby avoiding prolonged exposure of the treated material to the acoustic field, and the buffering capacity of food materials may further prevent corrosion reactions. While this study could not demonstrate the existence of any metallic nanoparticles in the size range of concern, corrosion is not a desirable outcome and some further study of the time course of metal erosion/corrosion using food materials in continuous systems over a range of ultrasonic frequencies and treatment durations is suggested. This study focuses on cavitation generated by transducers. It is important to highlight that similar cavitation phenomena can also be easily generated in hydraulic systems [47]. Hydrodynamic cavitation can occur using a constriction like an orifice plate, venturi or throttling valve in a liquid flow [48]. Moreover, equipment such as high-pressure homogenizers operate based on hydrodynamic cavitation phenomena. Hydrodynamic cavitation is not investigated here, but in light of the above medical concerns it should be
Please cite this article in press as: R. Mawson et al., Production of particulates from transducer erosion: Implications on food safety, Ultrason. Sonochem. (2014), http://dx.doi.org/10.1016/j.ultsonch.2014.04.005
R. Mawson et al. / Ultrasonics Sonochemistry xxx (2014) xxx–xxx
evaluated, although it would not be a trivial task as potentially hydrodynamic cavitation can arise in many food processing operations, e.g., pumping, mixing, wet milling and conventional homogenizing. Acknowledgements The authors would like to express their gratitude to the team at CSIRO Material Science and Engineering (Clayton, VIC, Australia) John Ward, Mark Greaves, and Karen Hands for their contributions in regards to profilometry and imaging and Malisja de Vries for her insights and contributions while reviewing this manuscript. References [1] M.S. Jellesen, A.A. Rasmussen, L.R. Hilbert, A review of metal release in the food industry, Mater. Corros. 57 (2006) 387–393. [2] T. Skotland, T.G. Iversen, K. Sandvig, New metal-based nanoparticles for intravenous use: requirements for clinical success with focus on medical imaging, Nanomedicine 6 (2010) 730–737. [3] P. Andujar, S. Lanone, P. Brochard, J. Boczkowski, Respiratory effects of manufactured nanoparticles, Rev. Mal. Respir. 26 (2009) 625–637. [4] B. Baroli, Penetration of nanoplarticles and nanomaterials in the skin: fiction or reality?, J Pharm. Sci. 99 (2010) 21–50. [5] J.W. Card, T.S. Jonaitis, S. Tafazoli, B.A. Magnuson, An appraisal of the published literature on the safety and toxicity of food-related nanomaterials, Crit. Rev. Toxicol. 41 (2011) 22–49. [6] C. Buzea, I.I. Pacheco, K. Robbie, Nanomaterials and nanoparticles: sources and toxicity, Biointerphases 2 (2007) MR17–MR71. [7] A. Elder, S. Vidyasagar, L. DeLouise, Physicochemical factors that affect metal and metal oxide nanoparticle passage across epithelial barriers, Wiley Interdiscip. Rev. Nanomed. Nanobiotechnol. 1 (2009) 434–450. [8] A.M. Gatti, Biocompatibility of micro- and nano-particles in the colon. Part II, Biomaterials 25 (2004) 385–392. [9] S. Barcikowski, A. Hahn, B.N. Chichkov, Nanoparticles as potential risk during femtosecond laser ablation, J. Laser Appl. 19 (2007) 65–73. [10] R. Landsiedel, L. Ma-Hock, A. Kroll, D. Hahn, J. Schnekenburger, K. Wiench, W. Wohlleben, Testing metal-oxide nanomaterials for human safety, Adv. Mater. 22 (2010) 2601–2627. [11] S. Barcikowski, J. Walter, A. Hahn, J. Koch, H. Haloui, T. Herrmann, A. Gatti, Picosecond and femtosecond laser machining may cause health risks related to nanoparticle emission, J. Laser Micro/Nanoeng. 4 (2009) 159–164. [12] A.A. Shemetov, I. Nabiev, A. Sukhanova, Molecular interaction of proteins and peptides with nanoparticles, ACS Nano 6 (2012) 4585–4602. [13] A.M. Alkilany, C.J. Murphy, Toxicity and cellular uptake of gold nanoparticles: what we have learned so far?, J Nanopart. Res. 12 (2010) 2313–2333. [14] J.S. Tsuji, A.D. Maynard, P.C. Howard, J.T. James, C.W. Lam, D.B. Warheit, A.B. Santamaria, Research strategies for safety evaluation of nanomaterials, part IV: risk assessment of nanoparticles, Toxicol. Sci. 89 (2006) 42–50. [15] N. Singh, B. Manshian, G.J.S. Jenkins, S.M. Griffiths, P.M. Williams, T.G.G. Maffeis, C.J. Wright, S.H. Doak, NanoGenotoxicology: the DNA damaging potential of engineered nanomaterials, Biomaterials 30 (2009) 3891–3914. [16] M.A. Maurer-Jones, Y.S. Lin, C.L. Haynes, Functional assessment of metal oxide nanoparticle toxicity in immune cells, ACS Nano 4 (2010) 3363–3373. [17] W.S. Cho, R. Duffin, C.A. Poland, S.E.M. Howie, W. MacNee, M. Bradley, I.L. Megson, K. Donaldson, Metal oxide nanoparticles induce unique inflammatory footprints in the lung: important implications for nanoparticle testing, Environ. Health Perspect. 118 (2010) 1699–1706. [18] M. Simko, M.O. Mattsson, Risks from accidental exposures to engineered nanoparticles and neurological health effects: a critical review, Part. Fibre Toxicol. 7 (2010). [19] Y.F. Li, Y.X. Gao, Z.F. Chai, C.Y. Chen, Nanometallomics: an emerging field studying the biological effects of metal-related nanomaterials, Metallomics 6 (2014) 220–232. [20] H.S. Chen, J. Li, A ring area formed around the erosion pit on 1 Cr18Ni9Ti stainless steel surface in incipient cavitation erosion, Wear 266 (2009) 884–887. [21] H.S. Chen, J. Li, S.H. Liu, Thermal effect at the incipient stage of cavitation erosion on a stainless steel in ultrasonic vibration cavitation, J. Fluids Eng. 131 (2009).
9
[22] Y.J. Chen, Strain-induced martensitic transformation of SUS304 stainless steel by cavitation micro-jet impact, J. Chin. Soc. Mechan. Eng. 30 (2009) 437–441. [23] K.Y. Chiu, F.T. Cheng, H.C. Man, Evolution of surface roughness of some metallic materials in cavitation erosion, Ultrasonics 43 (2005) 713–716. [24] G. Bregliozzi, A. Di Schino, S.I.U. Ahmed, J.M. Kenny, H. Haefke, Cavitation wear behaviour of austenitic stainless steels with different grain sizes, Wear 258 (2005) 503–510. [25] S. Barcikowski, A. Menendez-Manjon, B. Chichkov, M. Brikas, G. Raciukaitis, FlowGeneration of nanoparticle colloids by picosecond and femtosecond laser ablations in liquid, Appl. Phys. Lett. 91 (2007). 083113-1-083113-3. [26] A. Thiemann, T. Nowak, R. Mettin, F. Holsteyns, A. Lippert, Characterization of an acoustic cavitation bubble structure at 230 kHz, Ultrason. Sonochem. 18 (2011) 595–600. [27] M. Ashokkumar, D. Sunartio, S. Kentish, R. Mawson, L. Simons, K. Vilkhu, C. Versteeg, Modification of food ingredients by ultrasound to improve functionality: a preliminary study on a model system, Innovative Food Sci. Emerg. Technol. 9 (2008) 155–160. [28] A. Henglein, Contributions to various aspects of cavitation chemistry, Adv. Sonochem. 35 (1993) 17–83. [29] K. Okitsu, M. Ashokkumar, F. Grieser, Sonochemical synthesis of gold nanoparticles: effects of ultrasound frequency, J. Phys. Chem. B 109 (2005) 20673–20675. [30] F. Chemat Zill-e-Huma, M.K. Khan, Applications of ultrasound in food technology: processing, preservation and extraction, Ultrason. Sonochem. 18 (2011) 813–835. [31] E. Riera, J.A. Gallego-Juarez, T.J. Mason, Airborne ultrasound for the precipitation of smokes and powders and the destruction of foams, Ultrason. Sonochem. 13 (2006) 107–116. [32] S.H. Park, Y.R. Roh, C. Cooker, [WO 0113773] (2001). [33] P. Juliano, P. Swiergon, K.H. Lee, P.T. Gee, P.T. Clarke, Effects of pilot plant-scale ultrasound on palm oil separation and oil quality, J. Am. Oil Chem. Soc. 90 (2013) 1253–1260. [34] P. Juliano, P. Swiergon, R. Mawson, K. Knoerzer, M.A. Augustin, Application of ultrasound for oil separation and recovery of palm oil, J. Am. Oil Chem. Soc. 90 (2013) 579–588. [35] P. Juliano, A. Kutter, L.J. Cheng, P. Swiergon, R. Mawson, Enhanced creaming of milk fat globules in milk emulsions by the application of ultrasound and detection by means of optical methods, Ultrason. Sonochem. 18 (2011) 963–973. [36] P. Juliano, S. Temmel, M. Rout, P. Swiergon, R. Mawson, Creaming enhancement in a liter scale ultrasonic reactor at selected transducer configurations and frequencies, Ultrason. Sonochem. 20 (2013) 52–62. [37] Wikipedia, Drinking water quality standards (2013).
Ref Type: Internet Communication [38] Supeno, P. Kruus, Sonochemical formation of nitrate and nitrite in water, Ultrason. Sonochem. 7 (2000) 109–113. [39] A. Turnbull, Modeling of the chemistry and electrochemistry in cracks – a review, Corrosion 57 (2001) 175–189. [40] K.D. Ralston, N. Birbilis, C.H.J. Davies, Revealing the relationship between grain size and corrosion rate of metals, Scr. Mater. 63 (2010) 1201–1204. [41] N.J. Laycock, S.P. White, J.S. Noh, P.T. Wilson, R.C. Newman, Perforated covers for propagating pits, J. Electrochem. Soc. 145 (1998) 1101–1108. [42] V. Ligier, J.Y. Hihn, M. Wery, M. Tachez, J.Y. Hihn, The effects of 20 kHz and 500 kHz ultrasound on the corrosion of zinc precoated steels in [Cl] [SO2 4 ] [HCO3] [H2O2] electrolytes, J. Appl. Electrochem. 31 (2001) 213–222. [43] M. Doche, J. Hihn, A. Mandroyan, C. Maurice, O. Hervieux, X. Roizard, A novel accelerated corrosion test for exhaust systems by means of power ultrasound, Corros. Sci. 48 (2006) 4080–4093. [44] R. Wang, Influence of ultrasound on pitting corrosion and crevice corrosion of SUS304 stainless steel in chloride sodium aqueous solution, Corros. Sci. 50 (2008) 325–328. [45] G.O.H. Whillock, B.F. Harvey, Preliminary investigation of the ultrasonically enhanced corrosion of stainless steel in the nitric acid chloride system, Ultrason. Sonochem. 3 (1996) S111–S118. [46] G.O.H. Whillock, B.F. Harvey, Ultrasonically enhanced corrosion of 304L stainless steel.2. The effect of frequency, acoustic power and horn to specimen distance, Ultrason. Sonochem. 4 (1997) 33–38. [47] P.R. Gogate, Hydrodynamic cavitation for food and water processing, Food Bioprocess Technol. 4 (2011) 996–1011. [48] P.R. Gogate, V.S. Sutkar, A.B. Pandit, Sonochemical reactors: important design and scale up considerations with a special emphasis on heterogeneous systems, Chem. Eng. J. 166 (2011) 1066–1082.
Please cite this article in press as: R. Mawson et al., Production of particulates from transducer erosion: Implications on food safety, Ultrason. Sonochem. (2014), http://dx.doi.org/10.1016/j.ultsonch.2014.04.005