Measurement Methods to Evaluate Engineered ... - Wiley Online Library

6 downloads 27047 Views 403KB Size Report
Life Science Inst. Center for Risk Science Innovation and Application to identify, evaluate, and develop ... become released into food indirectly from food contact materials. ..... ment, a solution of the prospective migrant in a suitable test matrix.
Measurement Methods to Evaluate Engineered Nanomaterial Release from Food Contact Materials Gregory O. Noonan, Andrew J. Whelton, David Carlander, and Timothy V. Duncan

Abstract: This article is one of a series of 4 that report on a task of the NanoRelease Food Additive project of the Intl. Life Science Inst. Center for Risk Science Innovation and Application to identify, evaluate, and develop methods that are needed to confidently detect, characterize, and quantify intentionally produced engineered nanomaterials (ENMs) released from food along the alimentary tract. This particular article focuses on the problem of detecting ENMs that become released into food indirectly from food contact materials. In this review, an in-depth analysis of the release literature is presented and relevant release mechanisms are discussed. The literature review includes discussion of articles related to the release phenomenon in general, as experimental methods to detect ENMs migrating from plastic materials into other (nonfood) complex matrices were determined to be relevant to the focus problem of food safety. From the survey of the literature, several “control points” were identified where characterization data on ENMs and materials may be most valuable. The article concludes with a summary of findings and a discussion of potential knowledge gaps and targets for method development in this area. Keywords: characterization, detection, food contact materials, food safety, measurement methods, migration, nanotechnology, release

Introduction This article is the second in a series of 4 articles related to a task of the NanoRelease Food Additive (NRFA) project of the Intl. Life Science Inst. Center for Risk Science Innovation and Application to identify, evaluate, and develop methods that are needed to confidently detect, characterize, and quantify intentionally produced engineered nanomaterials (ENMs) released from food along the alimentary tract. A full description of the project’s charge and scope, as well as an executive summary of the project’s findings, is presented in the first article in this series (Szakal and others 2014). The focus area of the present article is measurement methods, including theoretical methods, to detect the release of ENMs into foods from food contact materials. The 3rd and 4th following articles in this series, respectively, describe methods to characterize and detect ENMs in foods (including sample preparation) (Singh and others 2014) and describe methods to characterize, detect,

MS 20140332 Submitted 28/2/2014, Accepted 4/3/2014. Author Noonan is with Center for Food Safety and Applied Nutrition, United States Food and Drug Administration, 5100 Paint Branch Parkway, College Park, MD 20740, U.S.A. Author Whelton is with Dept. of Civil Engineering, Univ. of South Alabama, 150 Jaguar Drive, Shelby Hall, Suite 3142, Mobile, AL 36688, U.S.A. Author Carlander is with Nanotechnology Industries Assoc, 101 Avenue Louise, 1050 Brussels, Belgium. Author Duncan is with Center for Food Safety and Applied Nutrition, United States Food and Drug Administration, 6502 South Archer Rd, Bedford Park, IL 60516-1957, U.S.A. Direct inquiries to author Timothy V. Duncan (E-mail: [email protected]).

and study the behavior of ENMs introduced into the alimentary tract through food ingestion (Alger and others 2014). While the present article is capable of standing alone, due to the fact that some experimental methods may have utility in multiple areas relevant to the project’s overall scope, some methods discussed within this article may have additional descriptive detail offered in other articles in this series.

Background Concepts and Goals of the Article General considerations of contaminant release from food contact materials The general risk that a chemical poses to human health is dependent on 2 factors: (1) the capacity of the chemical to do physical harm if an individual is exposed to it (in other words, its toxicity) and (2) the likelihood of exposure, which in the case of oral exposure from food includes how much of the toxicant is in the food to begin with as well as its pharmacological properties (in other words, how easily it is absorbed by the alimentary tract). Regarding ENMs as the potential toxicants, the chief concern is whether their small size increases their toxicity (due to unique chemistry) or increases their bioavailability (due to a purported ability to pass more quickly through natural biological barriers). In the case of ENMs added directly to foods, exposure assessments should be relatively straightforward in the sense that it will be known with certainty that the consumed particles will be introduced to the alimentary tract. Bioavailability and the chemical fate of ENMs in the (admittedly complex) gut environment are

Published 2014. This article is a U.S. Government work and is in the public domain in the USA. doi: 10.1111/1541-4337.12079 Vol. 13, 2014 r Comprehensive Reviews in Food Science and Food Safety 679

Measuring nanomaterial release . . . the primary aspects that need to be understood. In addition, such particles are specifically engineered to exist in food matrices, so their exact morphologies and other physical characteristics in such environments should already be reasonably known. Assessing the risks of nanotechnology-enabled food contact materials introduces the additional question of whether the particles can become released into the food in the first place and, if they can, what the characteristics of such particles are once they are released into an environment for which they were not specifically designed. In the limiting situation that ENMs in food contact applications always remain attached to or dispersed within the host material under the intended conditions of use, the potential adverse toxicological properties they possess in the free state may be irrelevant. If, however, the embedded ENMs are able to diffuse through the material and then partition into the external environment during the packaging’s use or storage (this 2-part process is formally termed migration), then the risks that such migrated materials pose to consumers need to be evaluated. Determining whether a prospective migrant, ENM or otherwise, can become released into a contacted food matrix requires robust, standardized experimental methods, including sample preparation methods and chemical analytical methods, which can identify and quantify the presence of the migrant in the contacted matrix as a function of time. This enables risk assessors to estimate a typical consumer’s likely exposure to the substance per unit time based on assumptions about consumption rate, total dietary intake of all food, and so on. While such methods are fairly well established for conventional (small-molecule) migrants, an additional complication arises in the case of migrating ENMs. An exposure assessment requires information related to the following 2 questions: To how much is the consumer likely to be exposed? To what is the consumer likely to be exposed? Neither question is always straightforward to answer for an ENM. Molecular structure is the sole piece of information required to uniquely identify a small molecule. For instance, every bisphenol A or furan molecule is chemically equivalent and uniquely identified by its relative quantities and physical arrangement of specific atoms. Thus, the “To what is the consumer likely to be exposed?” question is usually self-evident and the preservation of molecular identity before and after migration of a small molecule is easily confirmed by conventional analytical techniques like gas chromatography–mass spectrometry (GC-MS). ENMs, however, are not so easily identified or classified. For example, although all AgNPs may have the same general core composition (and even here we must be careful), they can vary vastly in their physical characteristics (size and shape), surface features (charge and ligand sphere), aggregation/agglomeration or dissolution state, and level of intrinsic purity. Along this wide spectrum of properties, which are also prone to evolve over time in complex ways, the toxicological and pharmacological behavior of ENMs can change dramatically. Unlike a molecular name, the term silver nanoparticles is thus not sufficiently descriptive to identify to what a consumer is likely to be exposed in the event ENM release into a food item occurs, even if the characteristics of the pristine particles (prior to incorporation in the food contact material) are perfectly known. As a result, estimating consumer exposure to ENM-based migrants from food contact materials requires new methods that can both: (1) identify and quantify the migration of the ENMs into surrounding matrices over time and (2) simultaneously provide the type of characterization data necessary to understand how the properties of the migrated ENMs are likely to change between the time the material is made and the time the food is consumed.

The purpose of this article is to review existing methods that can help risk assessors acquire this necessary information so they can better understand release of ENMs from nanotechnologyenabled food contact materials. In particular, it is important to be able to detect particles that have already migrated into foods, as well as predict ahead of time the types of ENMs that are most likely to migrate and the conditions under which migration is most likely to occur. This will ensure that the development of nanotechnology-enabled food contact materials can be appropriately targeted toward endpoints that are least likely to pose a threat to human health. Therefore, this article focuses on methods and tools specifically dedicated to understanding the process of migration itself, including postrelease behavior and dynamics, rather than on methods that might be used simply to detect or quantify particles that have already migrated into complex matrices. The latter group would be largely identical to a larger body of experimental methods that can be used to detect nanoparticles intentionally added to food matrices, which are reviewed in the next article in this series (Singh and others 2014). As this article will show, the body of literature related to the detection of ENMs that have migrated from nanotechnologyenabled food contact materials (primarily in the form of polymer nanocomposites [PNCs]) is relatively small. In light of this fact, and because methods to trace the release of nanoscale fillers from PNCs likely have general applicability, we will also consider methods and procedures that have been used to understand release of ENMs from PNC materials intended for nonfood applications. In addition, because some of the tools we discuss rely on chemical and physical theory, both as the basis of predictive models and to put empirical results into context, we view it as a useful exercise to first review some of the basic principles of diffusion and migration as they relate to molecular species, as well as to provide a brief overview of current regulatory thinking regarding the appropriate way for packaging manufacturers to measure migration. The article will conclude with a brief discussion of current challenges with respect to detection and modeling of ENM migration as well as an overview of new methods that may become useful in the future.

Mathematical diffusion models and prediction of migration rates The best way to determine likely consumer exposure to a substance added to a food contact material is to experimentally measure its concentration in a food that has been stored in the material under prescribed conditions. However, this is not always practical because migration experiments are not trivial to perform, and it is also difficult to generalize results. For example, if an experiment performed for a certain contact material yields migration data for that material in the presence of 1 food type, and a manufacturer decides to use the same material to store a new food with very different physical or chemical properties, the experiment may need to be performed again to ensure that the migration rate for the new intended use is low enough to meet safety standards. For a material with a number of specific intended applications, the experimental workload can quickly become unwieldy. While there are a number of experimental shortcuts that may be taken to assist in the generalization of empirical results (for example, the use of food simulants), mathematical modeling of migration is also a valuable tool. Being able to predict the migration rate for a given combination of food contact material, external environment type, and use condition (temperature, pressure, and storage time) is desirable because it alleviates the need for cumbersome experiments and can also help manufacturers predict

680 Comprehensive Reviews in Food Science and Food Safety r Vol. 13, 2014

 C 2014 Institute of Food Technologists®

Measuring nanomaterial release . . . Nevertheless, migration models have historically tended to focus on estimating diffusion constants, since diffusion is the kinetic process that is usually rate-limiting. The partition aspect of migration is usually treated as a partition coefficient, K, which scales the expected amount of migrated substance at equilibrium based on the relative solubility of the migrant in the polymer and the external matrix. Predicting partition coefficients has received less attention in the literature, and stock values for limiting “very soluble” (K = 1) and “insoluble” (K = 1000) cases are often used in conjunction with the predicted diffusion constants to estimate migration levels (Begley and others 2005). Experimentally determined partition coefficients may be used for more precise predictions of the total migration. However, because the determination of partition coefficients requires the measurement or estimation of solubility (a parameter that has ambiguous meaning for an ENM), it is not particularly clear how best to specify partition coefficients for nanoscale migrants. Even beyond the difficulties in formulating semiempirical models to estimate diffusion rates (or partition coefficients) for ENMs, the unfortunate truth is that semiempirical models are only as good as the data that support them. Even in the case of small molecule diffusion, the polymer-specific parameters are constantly being fine-tuned in light of new migration data and may need adjustment for nonpolyolefinic polymers such as PET or nylon or for temperatures below glass transition points. In particular, existing semiempirical models have been formulated for small molecules and whether they will yield accurate predictions of nanoparticle diffusion constants is unclear. Molar mass (molecular weight) certainly has little precisely defined meaning for nanoparticles and, in any case, the body of data for nanoparticle diffusion is fragmentary at best. Therefore, there are numerous challenges at this time that must be overcome before mathematical modeling becomes a viable tool set to understand nanoparticle migration. The biggest area of need may thus be the acquisition of sufficient migration rate data from well-controlled experimental model systems to build semiempirical models for ENM diffusion (or to verify that existing diffusion models are appropriate to use for ENMs). Acquiring such data is predicated, of course, on the availability of reliable experimental methods that cannot only measure ENM release rates (presumably by identifying ENMs that are released into the external environment over time), but can also provide information on what has diffused. This is particularly challenging   10454 because migration rates of ENMs are expected to be slow and ∗ 4 2/3 D p = 10 exp Ap − 0.01351M + 0.003M − (1) released quantities are anticipated to be miniscule, which places a T burden on currently available experimental methods. This model relates the “upper bound” (95% confidence limits, units of square centimeter per second) diffusion constant, D∗p , to a Experimental measurement of diffusion polynomial function of only 3 parameters: temperature, molecular Determining diffusion constants in polymers is not an easy task, weight of the diffusant, and a polymer-specific parameter, Ap . Ap even for small-molecule migrants, and there are several methods values for various polymers are set by considering experimental that are generally employed, depending primarily on the characdiffusion data (diffusion constants and diffusion activation energies) teristics of the migrant (molecular weight, volatility, and so on). for a variety of small molecules and are adjusted such that the An experiment can be performed in an actual food package if predicted diffusion constants are statistically likely to be greater available (for example, a plastic beverage bottle) or by taking a than those measured experimentally. A more thorough explanation representative section of a polymer film and assessing whether the of this process is provided in the cited literature (Begley and others migrant of interest can pass through it under a certain set of con2005). ditions. In the latter experiment, a specially designed migration or It deserves mention here that models such as that described permeation cell is usually used (Figure 1). In this particular experiabove are diffusion models and are not necessarily migration mod- ment, a solution of the prospective migrant in a suitable test matrix els. Although the terms are often mistakenly used interchange- is first loaded into the donor chamber and the receptor chamber is ably, migration involves additional processes beyond diffusion that charged with neat matrix. The cell is brought to the desired temneed to be considered in order to reliably predict the amount perature (usually chosen to represent the conditions to which food of a substance released into a food from a polymer over time. packaging materials are likely to be subjected during use/heating

the safety of a new material early in the development cycle. Risk assessors are therefore using theoretical or in silico approaches to support experimental determinations of migration (Oldring and others 2009; Hearty and others 2011). In the case of migration of an unknown substance from the interior of a polymer to the external environment under a specific set of conditions, it is usually sufficient to know the diffusion constant (related to the rate at which the migrant can move about throughout the polymer) and the relative solubilities of the substance in the polymer and the external environment (related to the degree to which a migrant can partition into the external matrix). With this information, it is, in principle, possible to predict the extent of migration at equilibrium via use of classical physical descriptions of mass transfer, such as Fick’s laws of diffusion. Complete knowledge of the molecular state of any system yields absolute predictive power over its macroscale properties; thus, in principle, one should be able to calculate a priori the diffusion constant and partition coefficient of any food contact material based on the molecular structures of the migrant and polymer as well as relevant structural information such as polymer density, crystallinity, and so forth. However, in practice, estimating the diffusion coefficient or partition coefficient from first-principles approaches is difficult given the number of factors involved. Moreover, from a risk assessment standpoint, a good predictive model should allow for enough conservatisms to ensure that any deviations of an experimentally determined diffusion value from the predicted value should err as frequently as possible on the side of caution (in other words, models should slightly overestimate the extent of migration). It is difficult to build such a degree of conservativeness into first-principles models, which again limits their usefulness to manufacturers of food contact materials. Although predictive models based on elementary physical and chemical principles have limited use in the context of risk assessment, semiempirical models (which are based largely on experimental data) can be conveniently formulated to predict an output value based on a limited number of parameters and can also be formulated with any degree of conservativeness required. Several such models to predict diffusion constants of small molecules in polymer matrices exist, the most popular of which was devised by Otto Piringer and coworkers from the Fraunhofer Inst. (Brandsch and others 2002):

 C 2014 Institute of Food Technologists®

Vol. 13, 2014 r Comprehensive Reviews in Food Science and Food Safety 681

Measuring nanomaterial release . . .

Figure 1–Schematic of a static liquid permeation cell, 1 potential experimental setup to measure migration rates of small-molecule migrants through polymer films. Reproduced from Song and others (2013), with permission from Taylor and Francis.

or storage) and aliquots are removed from the receptor chamber periodically and assayed for the migrant concentration using an analytical method of choice (LC-MS, GC-MS). The rate of permeation of the migrant through the film is determined by plotting the amount of migrant that appears in the receptor chamber as a function of time. Most migration experiments such as this make use of host substances that mimic the chemical properties of food rather than actual foods, and these food simulants can include substances such as water, dilute acetic acid (acidic foods), olive oil or coconut oil (fatty foods), and various concentrations of aqueous ethanol. Regulatory bodies typically issue documents intended to inform manufacturers what they view is the best way to perform these types of measurements. The U.S. Food and Drug Administration version of this document includes information on simulants to use, how to design a diffusion cell, conditions (temperatures and times) under which migration is most appropriately measured, the appropriate size and thickness of test films, how to ensure that the experiment assesses a “worst case scenario” for diffusion, how to properly validate tests and measure concentrations of test substances, and so forth (U.S. Food and Drug Administration 2002). This document also describes migration models that may be appropriate to use and where to find preexisting datasets to support these models, as well as additional useful information about conducting exposure assessments, proper ways to report data, how to identify an intended technical effect, and so on. In the European Union, the European Food Safety Authority (EFSA) and the European Commission have issued similar guidance documents. Despite this prodigious amount of information related to migration of small molecules from food-contact polymers, it is yet unclear to what extent these guidelines apply to the assessment of nanomaterial migration. Most of the above-described procedures were formulated from years of accumulated migration data. These data are simply not available for nanoscale migrants; thus, it is unknown whether conventional test conditions, food simulants, and migration metrics are directly applicable to the migration of nanomaterials. There are also lingering questions regarding whether conventional migration testing procedures are compatible with nanomaterials in the first place. For instance, to increase the surface area for migration and achieve better sensitivity, many laboratories manually cut test films into small pieces and submerge them in a stacked orientation into a vessel filled with the food simulant, and then measure the amount of migrant that leaches out from the surface of the cut films over time. In conventional migra-

tion experiments using thin films, the release of the migrant from the edges (as opposed to the comparatively larger surface area of the faces) is assumed to be small for thin films of sufficient diameter (Crank 1979), but whether this assumption holds for nanoparticles, which may be more likely to be manually dislodged by the cutting process and are likely to have low signal-to-background levels due to anticipated slow nanoparticle diffusion, is still an issue that needs to be resolved.

Strategy for addressing the goals of this article The previous sections presented an overview of theoretical and experimental assessment of migration and established that there are many questions to be answered with respect to proper methodology to measure release of ENMs from food contact plastics. The remainder of this article focuses on the body of ENM release literature, paying particular attention to both sampling and instrumental methods currently used to assess migration. This discussion includes a review of nanoparticle release studies related to materials that do not have direct relevance to food because they offer insight into sampling/instrumental procedures and fundamental release mechanisms that are relevant to those materials that are intended for food-related applications. We also note that in the context of this work, the term food includes drinking water, including tap water, and the term food contact materials, while predominantly packaging, also includes other materials that touch food such as potable water infrastructure, food processing equipment, cutting boards, eating utensils, appliance liners, gloves, and so forth. By broadening the scope in this way, we hope to maximize the relevance of this work as well as draw upon as many sources as possible to fully understand nanoparticle release and best survey current and emerging detection and characterization methodologies.

Literature Review of Experimental Methods Potential ENM release mechanisms In order to construct a predictive framework that can be used to predict the quantity and form of ENMs released into external media in any given situation, or at least to put acquired release data into proper context, it is necessary to understand the potential mechanisms by which ENMs could potentially be released as a function of ENM characteristics and external conditions. Once we have identified all potential mechanisms, we can begin to build a comprehensive system of knowledge related to ENM release as

682 Comprehensive Reviews in Food Science and Food Safety r Vol. 13, 2014

 C 2014 Institute of Food Technologists®

Measuring nanomaterial release . . .

Figure 2–The 4 “D” engineered nanomaterial release pathways.

a generalized phenomenon that is supportive of thorough and reliable risk assessment. The ultimate goal is a predictive framework that can be used to say “If this type of ENM is put into this type of polymer and it is subjected to these kinds of conditions, this is the likely result after a certain amount of time has elapsed.” Such an expansive picture of ENM release cannot be generated, however, without a sufficiently developed body of experimental data and this, in turn, requires a sufficiently developed tool set that can be used to distinguish between the potential release mechanisms identified and their potential endpoints (characteristics and quantity of released ENMs). With this in mind, the task group analyzed the existing body of ENM release literature with the aim of identifying possible release mechanisms, in the hope that this would enable the task group to identify the kind of tool set needed to support the predictive framework envisioned. By comparing the target tool set to the existing tool set, the task group hoped to identify and prioritize targets for method development. Our survey of the release literature has revealed 4 “D” ENM release phenomena during the life cycle of any nanocomposite material: desorption, diffusion, dissolution, and degradation of the matrix (Figure 2). (In practice, multiple processes for nanomaterial release may simultaneously occur, but these processes are treated separately herein.) The main distinctions between these phenomena pertain to where the ENM is located, the extent to which it interacts with the media (in other words, liquid and vapor), its ability to migrate through the host matrix material, and whether the particle remains an ENM or is transformed into ions (particle characteristics). It is important to note here that some of these mechanisms may be more or less important depending on the intended application of the nanocomposite; therefore, some mechanisms may be more or less important, generally, for food contact materials. Some of these distinctions will be pointed out below. Desorption. Desorption pertains to ENMs located on the material or substrate’s surface. Here, ENM adhesion is controlled by electrostatic interactions between the ENM and the substrate. Desorption would be most likely for nanocomposites in which the nanoelement is restricted to the interfacial region between the nanocomposite and the external medium (coatings, in  C 2014 Institute of Food Technologists®

other words). The textile industry, for example, has embraced this production method by dipping fabrics into a solution of suspended silver nanoparticles. Unfortunately, the actual bonding processes/interface between ENMs and substrates has been poorly characterized to date. External stimuli that would likely affect ENM–material surface bonds include liquid characteristics (pH, ionic strength, and presence of contaminants that promote bonding), temperature, fluid velocity, physical abrasion, and vibration. These external stimuli could dislodge ENMs from the food contact material surface and enable them to enter the contact medium (that is, the food). It is crucial to point out that in desorption, unlike in the diffusion mechanism described below, mobility of the ENMs within the matrix is not a limiting factor for release; therefore, particle morphology and size may not be major considerations here. Note that the fate of ENMs that are covalently bound to the material’s surface may be better described by the “degradation of the matrix” phenomena described below. Diffusion. Much of the existing food contact discussion surrounding nanomaterials pertains to the perceived risk that ENMs will migrate or diffuse out of polymer contact materials into the food substance, primarily because many of the nanocomposite materials currently under development for food contact applications are those in which the ENM elements are dispersed throughout the interior of the host matrix, rather than deposited on the surface. Diffusion has been used for decades to describe contaminant transport through materials (for example, polymers) into the contact medium (for example, water, air, or food). ENM diffusion would be expected to closely resemble molecular diffusion of other commonly studied infrastructures and food packaging additives, although if the molecular-scale interactions between ENMs and the matrix are sufficiently strong, diffusion may no longer be appropriately modeled as following Fick’s laws, upon which many migration models are based. If the principles of contaminant diffusion hold for ENMs, ENM diffusivity would be influenced by the ENM’s physicochemical properties (for example, polarity, size, and shape), the concentration gradient between the food contact material and food itself, as well as matrix properties (for example, density, polarity, and additives), and environmental conditions (for

Vol. 13, 2014 r Comprehensive Reviews in Food Science and Food Safety 683

Measuring nanomaterial release . . . example, pressure and temperature) (Comyn 1985). Note, however, the ENMs initially localized in the interior of the host matrix must migrate to the interfacial region before they are released into the external medium. Whether this partitioning process is identical to the desorption process described above remains an open question. Thus, if migration experiments are performed on a PNC, care must be taken to differentiate true diffusion and release from superficial desorption of ENMs remaining on the polymer surface after initial processing. It is also worth pointing out that ENMs are large compared to most conventional molecular-scale migrants and are therefore larger than anything that has been successfully subjected to successful diffusion models or migration experiments. Dissolution. Numerous publications in the peer-reviewed literature describe the influence of nanocomposites on metal ion levels in a contact liquid. These studies can be classified as ENM dissolution. Dissolution involves the transformation of an ENM from its native, particulate physical form into its ionic constituents. At present, there is debate as to whether nanoparticles migrate to the surface and then are dissolved into their ionic constituents, or whether the ions desorb from ENM surfaces while the ENMs are still dispersed throughout the nanocomposite matrix. Unfortunately, this is a question that cannot be answered with inductively coupled plasma mass spectrometry (ICP-MS), an elemental analysis method, but it will hopefully be elucidated in the coming years as the detection and quantification measurement methods improve. Experiments have shown that dissolution is a significant contributor to the fates of AgNPs and zinc oxide nanoparticles in the environment (Liu and Hurt 2010; Scheckel and others 2010). Available data demonstrate that nanoscale zinc oxide is much less stable in water than AgNPs. In addition, water pH, redox potential, ionic strength, particulate matter, temperature, and dissolved oxygen level are also reported to influence dissolution rates. Degradation of the matrix. In the event that ENMs are rigidly fixed to the polymer matrix (either because they are too large to be mobile or they are covalently attached to polymer molecules), ENMs could still be released if something were to happen to the integrity of the matrix itself. For example, ENMs embedded in the matrix could be exposed as the material mechanically or chemically decomposes. Decomposition could be caused by external stimuli such as physical abrasion, heating, UV exposure, and hydrolysis. Hydrolysis could change the gross properties of the polymer, thereby enabling ENM release. This aspect of matrix degradation may be particularly applicable to foods in the event that food contact materials are water- or acid-sensitive, as in the emerging class of nanocomposites fabricated from biocompatible polymers. Many foods contain water, and this water could influence polymer packaging properties (and hence diffusion rates) during long exposure times by acting as an effective plasticizing agent. Finally, matrix degradation could also be accelerated due to the physical properties of the embedded ENMs. For example, photoactive ENMs (such as TiO2 ) can generate reactive oxygen species in response to UV light exposure and thus degrade the localized area near the ENM (Wang and others 2011). This could be a desirable characteristic for degradable food contact material. ENMs released from the product could be present as individual nanoparticles or as composites, bound/surrounded by organic binders from the parent material. What release mechanisms tell us. It is important to be clear that the release mechanisms presented above are not islands unto themselves. Interrelationships do exist and some of these were alluded to above. Diffusion may not be able to proceed without desorption; dissolution and diffusion are intertwined because particles that

dissolve (partially or wholly) must still diffuse through the host medium to become released. In the case of complex (core–shell, for example) ENM architectures, some components may dissolve more readily than others; various forms of matrix degradation will impact all of the mechanisms because the basic polymer or ENM properties become attenuated and so on. Moreover, more work needs to be done to understand the conditions under which these various mechanisms are most likely to occur and, even more importantly, how the release mechanism will relate to postrelease processes like agglomeration/aggregation, particle dissolution, size changes via Ostwald ripening, and/or surface characteristic modifications. Most critical of all is a need for high-quality analytical methods that can distinguish between these mechanisms so that they can be better understood and inform safety assessments and decisionmaking. An analysis of the release mechanisms and the available release literature (see below) reveals that our ability to distinguish between dissolution and diffusion is the most significant and widely recognized gap in the current tool set. ICP-MS and ICP-OES (optical emission spectrometry) are the primary techniques utilized to monitor ENM release from nanocomposite materials, and because all information about nanoparticle characteristics is lost when test samples are fed into the plasma for analysis (indeed, information on whether the source came from a particle at all is lost), ICP-MS and ICP-OES simply cannot distinguish between these 2 mechanisms. This deficit is certainly important because although many studies have shown residual signatures of metallic ENM components, it is not known whether the eventual consumer will be exposed to ENMs or ionic salts, which potentially may have different pharmacological and toxicological profiles. Robust methods that can measure ENM–polymer surface interactions, polymer matrix integrity, and ENM characteristics, both before and after release, are crucial to understanding release mechanisms and postrelease processes. In particular, standardized sample preparation techniques for TEM imaging of ENMs while they are embedded in polymer films are needed to better study how the characteristics of ENMs change during PNC processing. Much of the release literature is based on uncharacterized or poorly characterized test materials fabricated from polymers and/or ENMs with unknown properties, which lessens their usefulness in studying release mechanisms because structure–function relationships remain largely undisclosed.

Methods used to assess release of ENMs from nanocomposites in the environment (nonfood) Purpose and scope. Because of the limited information available pertaining specifically to ENM release from food contact materials into foods, we reviewed the literature that describes ENM release from solid materials into the environment. ENMs are used in a variety of infrastructure and building-construction technologies, vehicles, consumer products, and medical applications, and a number of the release considerations for ENMs in food contact applications are shared with these other applications. Theoretically, ENM release has the potential to occur during food contact material use and disposal due to weathering or routine contact and cleaning. Airborne emission is also possible, but will not be discussed at great length here due to the dissimilarities between food contact applications and air emissions. The task group primarily focused on studies related to the release of ENMs from polymeric substrates into directly contacted liquid media, as they are expected to have the most relevance to release from food contact materials.

684 Comprehensive Reviews in Food Science and Food Safety r Vol. 13, 2014

 C 2014 Institute of Food Technologists®

Measuring nanomaterial release . . . Table 1–Capture and detection techniques that have been used to measure characteristics of ENMs released from materials in aqueous nonfood applications. Method description Capture Centrifugation Size fractionation by filtration Direct identification ICP-MS particle number technique Laser granulometer and mastersizer TEM EDX microscopy TEM high angle angular dark field (HAADF) detector and EDX microscopy SEM EDX microscopy X-ray absorption near edge spectroscopy (XANES) Indirect identification ICP-MS acid digestate analysis ICP-OES acid digestate analysis Ion selective electrode (ISE)

ENMs

Sample integrity

Ag, ZnO Ag

Wet Wet

(Kaegi and others 2010; Scheckel and others 2010) (Benn and Westerhoff 2008; Benn and others 2010; Farkas and others 2011)

Ag TiO2 Ag, TiO2 , CNT, SiO2

Wet Wet Dry

Ag

Dry

(Farkas and others 2011) (Golanski and others 2011) (Kaegi and others 2008; Liu and Hurt 2010; Farkas and others 2011; Nguyen and others) (Kaegi and others 2008; Kaegi and others 2010)

Ag, TiO2

Dry

Ag, ZnO

Dry

Ag, TiO2

Wet

Ag, ZnO

Wet

Ag

Wet

The majority of published ENM environmental-focused studies that the task group analyzed described the fate of individual ENMs in air and water media (Chen and Elimelech 2009; O’Brien and Cummins 2010; Arvidsson and others 2011; Petersen and others 2011; Zhang and others 2011; Mudunkotuwa and others 2012; Nowack and others 2012). Using these data, modeling of predicted ENM environmental concentrations has also been carried out and will likely continue as more data become available (Gottschalk and others 2009; Arvidsson and others 2011; Gottschalk and others 2011). Through experimentally based air/water studies, investigators have discovered that some ENMs are stable in water and bind with other constituents in the water (such as organic materials), whereas other ENMs decompose into ions rendering them classic ionic constituents. This underscores the true complexity of ENM release. Methods supporting the study of ENM release into aqueous environments (nonfood). To assess ENM release from polymeric ma-

terials into aqueous but nonfood environments, a combination of particle size separation techniques, total metal quantitation methods, and microscopic identification procedures have been applied (Table 1). Application of these methods has enabled researchers to confirm release for certain ENM–material pairings and to elucidate the role of some solution, material abrasion, and environmental properties on ENM release potential. It is somewhat difficult to compare results of these studies and make generalized conclusions because there are no standardized methods to understand the form of ENMs that are released. For a limited number of studies in which nanocomposites contact liquids, only liquid metal ion levels were reported (via ICP-MS), and ENM release was implied but not confirmed. Absence of standardized methods also inhibits application of fundamental scientific principles to document ENM release (such as diffusion coefficients through polymer matrices into water). Available data for release of ENMs into aqueous environments (nonfood). Inorganic ENMs have been the most heavily scru-

tinized materials; of these, AgNP products have received the greatest attention. Release of AgNPs into water from laboratorymanufactured and commercial off-the-shelf (COTS) materials (for example, fabrics, dust masks, medical cloths, toothpaste, and building construction products) has been studied. ENM desorption has

 C 2014 Institute of Food Technologists®

References

(Benn and others 2010; Kulthong and others 2010; Golanski and others 2011) (Impellitteri and others 2009; Scheckel and others 2010) (Kaegi and others 2008; Kaegi and others 2010; Farkas and others 2011) (Benn and Westerhoff 2008; Kaegi and others 2008; Benn and others 2010; Kaegi and others 2010; Scheckel and others 2010) (Farkas and others 2011)

been documented from fabrics in distilled water, tap water, and simulated washing machine conditions (for example, tap water, biocides, and surfactants) (Geranio and others 2009; Impellitteri and others 2009; Benn and others 2010). Matrix degradation has been linked to AgNP release from COTS toothpaste, building exterior paints that contain nanofiller, and abraded PNCs that released AgNPs, metal oxide nanoparticles (silica and titania), and multiwalled carbon nanotubes (CNTs) into liquids and air (Kaegi and others 2008, 2010; Golanski and others 2011; Schlagenhauf and others 2012). Many of the same techniques that have been used to detect ENM release from PNCs into liquid environments (vide infra) have also been applied to confirm whether ENMs remained encapsulated in the matrix during and following environmental aging. For example, ICP-MS has been used to quantify metals and TEM to visually detect ENMs in digested polymers and environmental samples (Kaegi and others 2008, 2010). High-resolution SEM with energy-dispersive X-ray (EDX) has been applied for elemental analysis (Kaegi and others 2008, 2010; Nguyen and others 2011; Wohlleben and others 2011; Schlagenhauf and others 2012). Other techniques utilized in this area are perhaps more uniquely suited to study the specific release phenomenon of matrix degradation. For instance, abrasion-induced release of nanoscale particles into liquid has been investigated using a laser granulometer to indirectly identify particle sizes (Golanski and others 2011), and attenuated total reflectance Fourier transform infrared (ATR FTIR), X-ray photoelectron spectroscopy (XPS), and secondary ion mass spectrometry (SIMS) have been applied to monitor composite surface chemistry to identify changes in chemical signatures that could indirectly confirm ENM release or polymer degradation (Nguyen and others 2011; Wohlleben and others 2011). Combinations of an aerodynamic particle sizer (APS), fast mobility particle sizer (FMPS), laser aerosol particle (LAP) size spectrometer, and scanning mobility particle sizer (SMPS) devices have also been used to describe the particle size distribution of nanosized particles in air during abrasion. While no studies were found that reported ENM release due to hydrolysis of matrix materials (such as polyesters), this is also a likely release pathway and experimental methods that can study this process should be identified.

Vol. 13, 2014 r Comprehensive Reviews in Food Science and Food Safety 685

Measuring nanomaterial release . . . The dissolution of ENMs embedded within nanocomposites has not been directly studied, but some literature data imply that dissolution is significant. We identified a number of studies in which researchers exposed nanocomposites to aqueous environments and observed relevant metal ion concentrations to increase (Kumar and others 2005; Kumar and Munstedt 2005; Damm and M¨unstedt 2008). Ag and Zn ion release from AgNPs and nanoscale ZnO medical PNC devices (Yang and others 2008) into water has been documented by ICP-MS. No studies were found that reported ENM diffusion through nonfood materials into water. This is likely because metrologies needed for characterizing diffusion (or distinguishing it from other release mechanisms) are lacking and there is a limited understanding of ENM interaction and dissolution within matrices (see above).

Control points for characterization and identification of ENMs in food contact materials and experimental hurdles As stated previously, the number of studies evaluating migration of ENMs from food contact materials is relatively small. Presently, the development of nanotechnology-enabled food contact materials has primarily utilized ENMs with inorganic core compositions, including those composed of metals, metal oxides, and clays (aluminosilicate nanoplatelets). The majority of these materials utilize polymers that are either embedded or coated with ENMs and are often referred to as nanocomposites. In addition to the synthetic (or bioderived) PNCs, there are a number of reports on the development of nanomaterial-modified papers (Gottesman and others 2011) and carbohydrate-based “fabrics” such as cellulose nonwoven materials that function as pads at the bottom of meat or produce storage containers (Fernandez and others 2010a,b). Although no commercial product or application has been identified for the nanomaterial-modified papers, antimicrobial effects have been evaluated and food safety is often a targeted application. Before the current literature related to ENM release into foods from these materials is presented, it is worth reflecting first on where and when analytical methods should be employed during their life cycle. After consideration of the literature, the task group suggests that there are a number of control points during the production and use of a nanocomposite where characterization could be performed to gain a greater understanding about the physicochemical characteristics of the nanomaterial/nanocomposite and evaluate its potential risk as a food contact material. In principle, if a complete body of information on the characteristics and identities of ENMs at each of these points is obtained, we can fully specify the life cycle of the ENMs and, importantly, fully understand the 4 “D” mechanisms described earlier, which will ultimately broaden our understanding of the quantity and type of ENMs that may enter the alimentary tract via the food contact material route of exposure. Therefore, methods development experts should focus on generating a tool set that is well adapted to studying ENM properties at each of these control points. This section reviews these 3 control points (termed CP-1, CP2, and CP-3) and subsequent sections present methods that have been used to evaluate ENM characteristics and migration at these various points. Control point 1: the raw or pristine ENM. For manufacturers or researchers who are synthesizing their own nanocomposites, characterization of the raw/pristine nanomaterial (in other words, prior to addition to the polymer) is often a critical first evaluation point. Electron microscopy (for example, transmission electron microscopy [TEM] and scanning electron microscopy [SEM]), particle sizing, elemental composition (ICP, AAS, and EDS), and

X-ray diffraction (XRD) (in the case of nanoclays) are the most commonly used characterization methods. The limited matrix interferences allow for fairly straightforward analysis of the nanomaterial although the researcher needs to be aware of techniquedependent differences in the determination of particle size. For example, microscopic techniques such as TEM measure the physical dimensions (often different if the particle is nonspherical) of the electron-dense core, whereas light scattering techniques such as dynamic light scattering (DLS) estimate size based on the rate of movement through a fluid medium, and thus, measure a “hydrodynamic radius,” which includes the core (or its spherical abstraction if it is nonspherical), any directly attached organic ligands, and any loosely associated solvent molecules that impact the particle’s rate of movement. Some of the challenges of pristine particle analysis, as well as the importance of solving them toward measuring the properties of ENMs in more complex media, are discussed in the 1st article in this series (Szakal and others 2014). Control point 2: the polymer-distributed ENM. After production of the nanocomposite, characterization of the nanomaterial becomes more difficult and the sample matrix limits the available analytical techniques. Electron microscopies (SEM and TEM) are usually utilized for imaging the materials and acquiring such information as aggregation state and dispersion morphology; however, small particle sizes or low particle concentrations can limit the effectiveness of microscopy methods, especially SEM. In addition, sample preparation remains a challenge. For SEM, material coatings can inhibit identification of the nanomaterial within the composite. Conventional microtoming of samples for TEM analysis can be difficult for many glassy polymers or thin films such as LDPE bags, in which fixing agents do not adhere well to the polymer and the materials may be too soft (at room temperature) to get adequate shear for sectioning. It does not help that microscopic analysis of fabricated nanocomposites reported in the literature is often outsourced to contract laboratories, and so experimental procedures are often poorly documented, vague, or absent, which makes standardization of sample preparatory techniques difficult. Even if a microscopist is successful in preparing a sample and acquiring an image, the information obtained is still predominantly qualitative and may not be representative of the entire material. Therefore, other characterization techniques are required to supplement microscopy methods; however, there are few widely available options available at this point. One researcher attempted to avoid the difficulties in preparing samples for SEM and TEM by ashing the nanocomposite and then evaluating the ash for nanoparticles (Huang and others 2011). While this clearly avoids difficulties in sample preparation, it raises additional questions about the de novo formation of particles under extreme sample preparation conditions. Elemental analysis is also still used, but the elemental concentrations are determined after digestion of the nanocomposite samples and represent total elemental concentrations and not nanomaterial-specific data. Although methods of nanoparticle extraction from the PNC might allow for single-particle ICP-MS (SP-ICP-MS) analysis, successful extraction techniques have not been developed and such a method would in any case provide little information about the nature of ENM dispersion. Control point 3: the released ENM. The final control point after nanocomposite characterization is to determine the amount and form of migration of the nanomaterial from the nanocomposite. Although foods can be used to determine migration, food simulants are usually used to simplify an experimental setup and reduce matrix interferences (see above). Elemental analyses (ICPMS, AAS) are often the first methods used to assess the migration

686 Comprehensive Reviews in Food Science and Food Safety r Vol. 13, 2014

 C 2014 Institute of Food Technologists®

Measuring nanomaterial release . . . of the nanomaterial. While these methods are extremely sensitive, researchers generally must digest the sample prior to analysis; thus, while they afford information on total elemental concentrations, they provide no information on physical characteristics of migrated ENMs (size, for instance) or even whether whole ENMs migrated in the first place. EM or light scattering techniques would be the obvious choice to distinguish whole migrated nanoparticles from diffused ions, but migrated ENM concentrations are expected to be incredibly small, which is a serious problem for these techniques. SP-ICP-MS may address some of these gaps, but given the current limitations in size sensitivity as well as the fact that the technique is not widely available, there is a need for new methods to help support our understanding of ENM migration from PNC food contact materials.

Methods used to assess release of ENMs from nanocomposites into foods Theoretical modeling of ENM diffusion. As mentioned previously, diffusion or migration models assist in the prediction of migration of a substance from a food contact material to the contacted food and, in principle, theoretical modeling can be a tool adapted for each of the above-described control points. Although our understanding of the physical principles of migration for small molecules is quite advanced, there have been very few studies focused on the modeling of the diffusion of ENMs through polymer matrices, to say nothing of migration to the external environment, either in a theoretical framework or to aid in the interpretation of experimental data. Sˇ imon and others (2008) are some of the few researchers who have attempted to do so, and presented a diffusion model using the Stokes–Einstein equation for the diffusion of a spherical particle through a fluid with lamellar flow properties. Using this approach, they presented a simple relationship between particle size and predicted migration level as a function only of temperature, the polymer’s dynamic viscosity, and the available surface area for release. For example, the authors predicted that for LDPE embedded with 10 nm AgNPs at 1 kg/m3 with an exposed surface area of 0.2 m2 , the total amount of migrated silver in the surrounding medium after 1 y of storage at 25 °C would be 260 µg. Sˇ imon and others (2008) did recognize that a number of assumptions made for the model may not be entirely applicable and contended that the calculated diffusion coefficients most likely represent the highest limits and an overestimation of migration. The model also made no accounting for the chemistry of the surrounding medium and only considers the diffusion mechanism presented above. Although we can scrutinize the assumptions of any model at length, even a model based upon seemingly perfect assumptions is useless unless we have confidence that it makes accurate predictions. Such confidence is, of course, impossible to have without experimental data that can be compared with values predicted by the model. Therefore, although the lack of theoretical models of ENM diffusion certainly qualifies as a knowledge gap to be addressed in the long term, spending too much energy on this at the present moment may be premature. Until experimental methods are sufficiently developed to generate reliable ENM migration data, theoretical methods will have only academic relevance. Methods to assess migration of nanosilver from food contact materials. Although silver has been incorporated into polymers

and other textiles for some time, its application to food contact materials is relatively new. A number of early reports of the use of AgNPs did not incorporate the silver into a polymer film/nanocomposite, but created materials in which AgNP was on  C 2014 Institute of Food Technologists®

the polymer surface. For example, del Nobile and others (2004) produced a coating of silver islands of about 90 nm in polyethylene oxide on the surface of a polyethylene film by plasma-based vapor deposition. They imaged the starting materials using EM and quantified the concentration of silver in the films by XPS. Ag concentrations in a variety of solutions (water, malt extract broth, and apple juice) that had been placed in contact with the materials were determined using ICP-OES. Similarly, Fernandez and coworkers synthesized AgNPs in situ in the presence of cellulose, forming silver nanoparticles at the surface of cellulose fibers (Fernandez and others 2010a,b). These authors imaged their ENMs (in the nanocomposite stage) with TEM and quantified silver content in the meat or fruit exudates (but not in the meat or fruit itself) by graphite furnace atomic absorption spectrometry (GFAAS). While these experiments were important in determining the extent of Ag transport, they do not represent migration of silver; rather, they demonstrate simple surface desorption or dissolution or a combination thereof (see Figure 2). Unfortunately, neither of the studies attempted to determine the form (ionic compared with particle) of the silver in the solutions. Due to the in situ style of ENM generation, characterization at the CP-1 stage (pristine state) was not possible. Sample preparation for TEM analysis in both of these cases was not particularly well documented. In 2010 and 2011, there appeared a number of publications evaluating the use of silver in nanoscale form as an additive to polymeric food contact materials (Busolo and others 2010; Emamifar and others 2010; Huang and others 2011; Lin and others 2011; Song and others 2011). The materials utilized by Emamifar and Busolo did not contain AgNPs, but rather employed Ag-modified ZnO nanoparticles and Ag-modified clay platelets, respectively. The remaining researchers evaluated commercially available AgNP/PNCs. In all cases, except for that by Song and others (2011), who performed no imaging analysis of their test films, the characteristics of the particles in the host materials were evaluated to some extent by TEM or SEM imaging (with or without EDX-based confirmation of particle identity) to determine distribution patterns or aggregation extent. In the study by Busolo and others (2010), wide-angle X-ray scattering was used to determine the extent of clay exfoliation (separation of clay platelets), and differential scanning calorimetry (DSC) was used to examine the effect of ENM distribution on the thermal properties of the material. The migrated silver concentrations in aqueous media (water, food simulants, or orange juice) were determined in all cases as total silver (and zinc, in the case of the study by Emamifar and others 2010) either by elemental analysis techniques (ICP and AAS) or voltammetry, although the amount of detail provided regarding preparation of samples for migration experiments varied significantly (in 1 case, no information at all was provided about sample preparation or even the analytical technique used). In cases in which sample preparation information was provided, films or simulant were generally treated to microwave acid digestion prior to elemental analysis. It should be noted that all of the researchers detected silver content in the substance contacting the test materials, regardless of whether the silver was used as an additive to larger particles or alone. However, with the exception of Huang and others (2011), who used SEM/EDX to analyze the food simulant postmigration and reported the observation of whole AgNPs, none of these studies reported the detection of ENMs in the migration solvents, suggesting either that silver ions were the primary end point of silver migration or that the analytical tools used were inadequate to measure what was actually occurring.

Vol. 13, 2014 r Comprehensive Reviews in Food Science and Food Safety 687

Measuring nanomaterial release . . . In a departure from the approach of measuring AgNP migration through ICP or AAS analysis, von Goetz and others (2013) recently evaluated silver migration from commercially available AgNP nanocomposites, but they used SP-ICP-MS in an attempt to determine whether any of the migrated silver was in nanoparticulate form as opposed to diffused silver cations. The authors also conducted a CP-2 level analysis on the nanocomposite test materials using laser ablation directly into an ICP-MS instrument, which afforded information about the geometric distribution of the AgNPs inside the polymer that is typically lost when samples are processed by acid digestion and solution nebulization. Atomic force microscopy (AFM) was used to analyze the surface topography of the materials. From this combination of techniques, the researchers observed silver migration (into water, ethanol, acetic acid, and olive oil, to varying degrees) and concluded that most of the Ag migration could be accounted for by Ag ion migration. Nevertheless, they did report the presence of AgNPs in some of the migration solutions, which was confirmed by TEM/EDS analysis of residues left after the simulants were evaporated. The authors acknowledged that they were unable to determine whether the AgNPs migrate via the diffusion mechanism, as depicted in Figure 2, or whether they are released from the surface (desorption mechanism) or are formed postdissolution during sample-handling. The latter process may be indicated because the detected AgNPs in the simulant residuals were composed of AgCl and AgS, in agglomerated form. Methods and experiments that can distinguish between migration of particles and formation of particles from ions after the fact are sorely needed. Methods to assess migration of nanoclay residuals into foods.

The 1993 publication by Kojima and coworkers is one of the earliest reports of improvement in mechanical and thermal characteristics of polymers after the addition of clay (Kojima and others 1993). The initial studies with clay PNCs were interested in the use of clays as flame/fire retardants (Lewin 2003; del Nobile and others 2004). In these applications, a number of researchers have evaluated the migration of clays to the surface of the polymer composite (Zammarano and others 2006; Tang and Lewin 2007). Generally, in monitoring clay migration, the researchers utilized elemental constituents of the clays (Mg, Al, and Si) and analyzed for these elements. Much like the case of AgNPs, this method does not provide information on the migration of individual clay particles, which, when fully dispersed, are highly anisotropic (a mere 1 nm thick, but often hundreds to even thousands of nanometers in each lateral dimension). In the use of clay/polymer materials as fire retardants, there is clearly an increased surface concentration measured for these elements after testing. Whatever the mechanisms responsible for this migration of clay particles to the surface, the conditions under which it occurs represent extreme temperatures, often above the melting point of the polymer, which are not applicable to evaluating the migration of clays in food contact applications. The methods used to assess the presence of clays in the clay nanocomposites or to evaluate migration into foods and food simulants are generally similar to the methods used for Ag and other metal oxide materials. Microscopy (TEM and SEM) and elemental analysis (ICP-MS, AAS, and EDX) are the most commonly used techniques. In addition, XRD has been used in a number of studies to acquire information about the dispersion of the clay within the polymer matrix (for example, interplatelet separation/degree of exfoliation). FTIR spectroscopy, usually with the benefit of ATR, and XPS have also been used to study changes in clay concentration in the surface region of the nanocomposite during annealing. As with AgNP nanocomposites, these independent techniques do

not necessarily offer any direct information about the form of the migrant, but they would be considered a critical component of the CP-2 level described above. Most commonly, elements present in the clay (Al, Si, Mg, and Fe) are detected by ICP-AES or ICPMS analysis, methods that usually cannot also provide information about the form of the migrated species. In addition, due to the common occurrence of many of the elements found in clays (Si, Al, and Mg), background interferences can be difficult to avoid or to correct for, making careful sample preparation important. Often, the use of clean-room facilities and expensive nonglass instrument components is required, which can make these analyses inaccessible to less well-equipped or less well-funded facilities. Several recent examples of clay migration studies are worth mentioning here. Mauricio-Iglesias and others (2010) fabricated their own montmorillonite/wheat gluten composite films and analyzed silicon and aluminum content in various simulants after extended storage times at 40 °C using ICP-OES. Their ICP-OES analysis was outsourced to a contract laboratory, so experimental conditions for the analysis (like sample digestion parameters) were not provided. Notably, the authors observed different results depending on whether silicon or aluminum was the target analyte, suggesting a need for standardized procedures and choice of analyte for clay migration analysis. In a follow-up study, Mauricio-Iglesias and others (2011) also used FTIR to analyze the clay structure in the film and provided information on differences in migration levels they observed during high-pressure processing of the test films. These authors did not perform any other significant analyses on clays prior to dispersion in the polymer or in the dispersed state (other than the FTIR analysis), indicating that they were primarily interested in a CP-3 level analysis. Avella and others (2005) dispersed montmorillonite clay into thermoplastic starch and analyzed release of clay residuals onto vegetables. They used magic angle spinning nuclear magnetic resonance (MAS NMR) to analyze the clay dispersion in their composite films, SEM for imaging of surface features, and a materials testing machine (Instron) to measure mechanical properties like tensile strength. They employed both flame and graphite furnace AAS for elemental characterization of the foods postmigration although they replaced aluminum with iron and magnesium as analytes, in addition to silicon. Schmidt and others (2011) investigated migration of magnesium aluminum double hydroxide clays from polylactide films and used ICP-MS to measure total aluminum migration, as well as TEM for both film characterization and characterization of the clay migrates. They also used gel permeation chromatography (GPC) to analyze the polymeric materials before and after the migration experiments, as well as SEM to characterize the clays prior to incorporation into films (a rare example of a CP-1 level analysis in the nanoclay area). Finally, Farhoodi and others (2014) prepared composites of montmorillonite in polyethyelene terephthalate (PET) by meltblending and processed these materials into bottle form by blowmolding. They analyzed the extent of clay dispersion in the films (CP-2 level analysis) by XRD, tapping mode AFM (by generating phase contrast images that are sensitive to viscoelasticity and chemical composition), and TEM; they also determined migration levels into common food simulants using an American Society for Testing and Materials (ASTM) standard migration cell and by monitoring both aluminum and silicon levels in the simulant with ICP-OES. As is often the case, the CP-3 level analysis in this study was limited to measurement of ionic or atomic residuals, which provided no information on the form of migrated clays.

688 Comprehensive Reviews in Food Science and Food Safety r Vol. 13, 2014

 C 2014 Institute of Food Technologists®

Measuring nanomaterial release . . . Using a slightly different attempt at obtaining CP-3 level information about the nature of clay migrates, Schmidt and others (2009) used a combination of asymmetric field flow fractionation (aFFF) with multiangle light scattering (MALS) and ICP-MS to characterize particle size and composition of migrants from a 5% montmorillonite/polylactide nanocomposite. Using aFFF with MALS, they detected the presence of 50- to 800-nm particles in 95% simulant after the migration experiment was completed; however, the ICP-MS component of the analysis failed to show that the migrated particles included elements characteristic of clays, suggesting that the detected nanoparticles were not associated with migration of the clay additives. Although this study presented a null result, it nevertheless demonstrated the benefit of in-line separation and characterization of ENMs. This combined technique cannot rule out the formation of nanoparticles after migration, but it does differentiate between different sizes and composition of material. One other recent approach that may assist in monitoring the clay migration profile has been to covalently label the clay particles with fluorescent tags (Diaz and others 2013). Not only does this allow monitoring of the migration experiment via fluorescence spectroscopy or microscopy (both of which are very-low-background optical techniques), but it also avoids the need to prepare samples for ICP-MS/AES or TEM/SEM analysis. Ideally, the fluorescent tag would remain covalently bound to the clay, maintain fluorescent properties, and allow in situ continuous monitoring during migration. The results of this particular study implied that migration of clays on experimental time scales was occurring because fluorescence was observed in the external media after storage and heating; however, the effect of sample preparation (for example, cutting of the film) on such release is unclear. While this approach is certainly clever and may yield valuable information about nanoparticle release from polymers as a general phenomenon, it may be less useful to evaluate migration in commercial clay/PNCs that have already been extruded without fluorescent labels. Discussion of food contact material release literature and current challenges. One of the most difficult challenges evident from

the current literature is determining the form of the nanomaterial that is migrating, that is, distinguishing between diffusion and dissolution as the release mechanism (Figure 2). Based on current migration models used for molecular species, migration of nanomaterials from nanocomposites should be quite slow: slow enough, at least, that the number of particles migrating during a 1-, 2-, or 10-d test will likely fall below the limit of detection for techniques typically used to count particles. The use of ICP-MS or other elemental analysis techniques to monitor elemental concentrations can assist with the evaluation, but information about the form of the migrant is lost when the sample passes through the 6000 K (or higher) inductively coupled plasma. SP-ICP-MS can, in some instances, help address the technical challenges; however, even with this highly sensitive technique, the size of the migrated particle may be too small to detect because of the omnipresent ionic background. Even if detection of ENMs is possible, there is often the concern that the particles formed de novo or otherwise changed their characteristics after the migration occurred. Therefore, appropriate controls, using ionic solutions and/or ENMs that do not form under simulant conditions, may be necessary to fully evaluate ENM migration. Also evident from the literature is that there are various approaches to measuring migration and, although the data for a specific product or material may be novel, the variation in exper C 2014 Institute of Food Technologists®

imental strategies complicates the comparison of results and the development of a general understanding of nanomaterial migration. The extent of migration of an ENM will be influenced by the composition and form of the ENM, the characteristics of the polymer matrix, the composition of the simulant, the time and temperature under which the experiment was carried out, and the use of single- or double-sided migration cells. The extent of migration may also be influenced by the way the sample material was prepared due to physical abrasion of particles during cutting or tearing. Unfortunately, available migration studies not only vary significantly in the sample preparation methods used, but also in the level of description provided of how these methods were carried out. In some extreme cases, this information is completely absent. A number of regulatory authorities have recommended (US FDA) or required (European Commission) migration conditions that, while developed for additive migration, could easily be applied to ENM migration. However, even if the exact conditions specified in these recommendations/requirements are not utilized in ENM migration experiments, the migration conditions should be appropriate for food contact materials and the polymer matrix should be evaluated based on a realistic intended use of the polymer. Additionally, evaluations of ENM migration should include sequential exposures of the test material to simulant in order to differentiate between surface desorption and migration. Finally, with respect to the above-identified control points, the task group found that many studies were focused primarily on CP-3 level analysis (actual migration experiments) with considerably less effort spent on CP-2 (test materials) and especially CP-1 (pristine materials) level analyses. It is true that in some cases, the CP-1 level analysis may be less important (for nanoclays, for example, in which attributes like particle size and shape are less defined) or even impossible (for ENMs incorporated into host matrices in situ, for instance, or in the case of commercial materials that were acquired from 3rd-party sources). Even so, the value of characterizing ENMs at earlier control point stages should not be discounted: A comprehensive predictive framework of ENM migration cannot be developed without a robust understanding of how the pristine ENM or host material properties, as well as their properties in the resultant nanocomposite, impact the downstream quantity, and form of migrated ENMs. Without such a framework, it will be difficult to have confidence in CP-3 level evaluations of commercial materials. Therefore, the task group recommends that significant attention be given to developing standardized tools and sample preparation methods for all identified control points, and the task group especially suggests that researchers involved in migration work characterize their starting materials with these available tools to the maximum extent possible.

Summary and Recommendations This article presented some of the difficulty surrounding evaluation of the release of ENMs from food contact materials into aqueous media. Such evaluations are necessary, however, to make precise assessments of likely exposure of consumers to ENMs from dietary sources, especially in light of the fact that most public perception surveys indicate that food contact materials, and particularly food packaging, are the most likely food-related applications of ENMs to be accepted by consumers in the near future. While many of the detection issues related to assessment of ENM release from food contact materials are likely to be similar to those encountered for detection of ENMs intentionally introduced into foods (as discussed in the next article in this series by Singh and others 2014), some differences in these areas justified

Vol. 13, 2014 r Comprehensive Reviews in Food Science and Food Safety 689

Measuring nanomaterial release . . . separate articles on these topics. First, the focus here has been on the prediction of exposure assessment, not the direct detection of ENMs in foods. Second, the type of ENMs likely to be encountered in food contact and in direct-food applications is likely to overlap only marginally, with the former category being primarily inorganic ENMs and the latter being more organic in nature (for example, liposomes, encapsulates, and so on). This will impact the analytical methods needed for detection. Third, the front-end sampling issues are different: the use of food simulants in the case of food contact applications as compared to the use of more complex matrices in real food analysis, as well as the need to characterize plastic materials for exposure assessments discussed above. Finally, ENMs in food contact materials are not usually intended to be released into food, so interactions with food matrices are not always known. These differences introduce unique challenges— and unique solutions—to the problem of ENM release evaluation. An evaluation of the ENM release literature revealed significant deficiencies. One of these deficiencies is the available tool set itself. While there are numerous potential candidates, most notably SP-ICP-MS, no optimal method to distinguish between the various potential methods of ENM release has been identified, particularly one that can determine whether ENMs migrate as whole particles or as dissolved ions. Even in cases in which whole particles are observed, there is still a question of whether such particles were released in a particular manner or whether they were formed from ions during postmigration handling. As a result, a majority of release studies simply ignore this question altogether and present a total migration amount by simple elemental analysis of the simulant using ICP-MS. The value of such studies toward a true understanding of the phenomenon of ENM release remains an open question. The other major deficiency uncovered by our literature analysis was a general lack of interest in thorough characterization of test materials, particularly of particles and host polymers in the pristine state. While it is true that such analyses have no obvious value with respect to a safety evaluation of a particular material, in which only the quantity and characteristics of the migrant need to be known, from a broader standpoint, the lack of such information means that existing studies shed little light on important structure–function relationships, which are necessary to understand if a predictive framework for ENM migration is to be developed. Even in studies that do undertake evaluations of their ENMs, polymers, and nanocomposite materials prior to migration experiments, the sheer variety of sample preparation techniques, analytical approaches, and style of data presentation make it difficult to compare 1 study to another, and so again development of a more expansive understanding of ENM migration is hindered. As a result of these considerations, we conclude that the following should be regarded as research priorities in this area: r Targeted development of new analytical tool sets capable of

differentiating between the various release mechanisms described here. r Efforts to research the effect of sample preparation methods on measured migration levels, particularly nanoparticulate forms. r Development of model systems, particularly those that use well-characterized reference materials, to better understand important structure–function relationships. r Formulation of standardized analytical techniques, sample preparation methods, test materials, and data reporting strategies to increase consistency in the published literature.

Allocating resources to these priorities will lend confidence to theoretical models and also experimental efforts, as well as support efficient research and development of safe nanotechnologyenabled food contact materials at the commercial level.

Acknowledgments The authors of this report are grateful to the following individuals for their expert input and support for this effort (alphabetically listed): Maurizio Avella, Joe Hotchkiss, Anil Patri, Ruud Peters, Jonathan Powell, Vicki Stone, Scott Thurmond, Jim Waldman, Stefan Weigel, and Jun Jie Yin. Experts were convened and initial framing concepts were developed for this article by the NRFA Steering Committee (http://www.ilsi.org/ResearchFoundation/ RSIA/Pages/FoodAdditiveSteeringCommittee.aspx and http:// www.ilsi.org/ResearchFoundation/RSIA/Pages/NRFA_ TaskGroup1.aspx), which operates as an independent public– private partnership. Project management and editing support was provided to the NanoRelease project experts by Richard Canady, Lyubov Tsytsikova, Christina West, Molly Bloom, and Elyse Lee of the ILSI Research Foundation. This phase of the project was funded by the Pew Charitable Trusts, the US Food and Drug Administration, Health Canada, ILSI North America, the Coca-Cola Co., the Illinois Inst. of Technology’s Inst. for Food Safety and Health, and the ILSI Research Foundation. Substantial in-kind support was provided by the Nanotechnology Industries Assoc. Furthermore, this material is partly based upon work supported by the USDA Natl. Research Initiative Agriculture and Food Research Initiative, the US Environmental Protection Agency, and the Natl. Science Foundation. This article has been reviewed in accordance with the US FDA’s peer and administrative review policies and approved for publication. The statements made in this report do not necessarily represent the official position of the US FDA or affiliated organizations. Mention of trade names or commercial products does not constitute an endorsement or recommendation for use by the US FDA.

Author Contributions Carlander serves as the task group chairperson and coordinator. Duncan serves as the task group chairperson and coordinator, provided text to the manuscript, and performed general document editing. Noonan provided text to the manuscript. Whelton provided text to the manuscript. The ordering of noncorresponding authors is alphabetical and does not reflect quantity of contribution to this article.

References Alger H, Momcilovic D, Carlander D, Duncan TV. 2014. Methods to evaluate uptake of engineered nanomaterials by the alimentary tract. Compr Rev Food Sci Food Saf 13:705–29. Arvidsson R, Molander S, Sand´en BA, Hassell¨ov M. 2011. Challenges in exposure modeling of nanoparticles in aquatic environments. Hum Ecol Risk Assess 17:245–62. Avella M, De Vlieger JJ, Errico ME, Fischer S, Vacca P, Volpe MG. 2005. Biodegradable starch/clay nanocomposite films for food packaging applications. Food Chem 93:467–74. Begley T, Castle L, Feigenbaum A, Franz R, Hinrichs K, Lickly T, Mercea P, Milana M, O’Brien A, Rebre S, Rijk R, Piringer O. 2005. Evaluation of migration models that might be used in support of regulations for food-contact plastics. Food Addit Contam 22:73–90. Benn T, Cavanagh B, Hristovski K, Posner JD, Westerhoff P. 2010. The release of nanosilver from consumer products used in the home. J Environ Qual 39:1875–82.

690 Comprehensive Reviews in Food Science and Food Safety r Vol. 13, 2014

 C 2014 Institute of Food Technologists®

Measuring nanomaterial release . . . Benn TM, Westerhoff P. 2008. Nanoparticle silver released into water from commercially available sock fabrics. Environ Sci Technol 42:4133–9. Brandsch J, Mercea P, R¨uter M, Tosa V, Piringer O. 2002. Migration modelling as a tool for quality assurance of food packaging. Food Addit Contam 19(Suppl):29–41. Busolo MA, Fernandez P, Ocio MJ, Lagaron JM. 2010. Novel silver-based nanoclay as an antimicrobial in polylactic acid food packaging coatings. Food Addit Contam A Chem Anal Control Expo Risk Assess 27:1617–26. Chen KL, Elimelech M. 2009. Relating colloidal stability of fullerene (C60) nanoparticles to nanoparticle charge and electrokinetic properties. Environ Sci Technol 43:7270–6. Comyn J, editor. 1985. Polymer permeability. New York: Chapman and Hall. Crank J. 1979. The mathematics of diffusion. Oxford, UK: Oxford University Press. Damm C, M¨unstedt H. 2008. Kinetic aspects of the silver ion release from antimicrobial polyamide/silver nanocomposites. Appl Phys A 91:479–86. Del Nobile MA, Cannarsi M, Altieri C, Sinigaglia M, Favia P, Iacoviello G, D’Agostino R. 2004. Effect of Ag-containing nanocomposite active packaging system on survival of Alicyclobacillus acidoterrestris. J Food Sci 69:E379–83. Diaz CA, Xia Y, Rubino M, Auras R, Jayaraman K, Hotchkiss J. 2013. Fluorescent labeling and tracking of nanoclay. Nanoscale 5:164–8. Emamifar A, Kadivar M, Shahedi M, Soleimanian-Zad S. 2010. Evaluation of nanocomposite packaging containing Ag and ZnO on shelf-life of fresh orange juice. Innovative Food Sci Emerg Technol 11:742–8. Farhoodi M, Mousavi SM, Sotudeh-Gharebagh R, Emam-Djomeh Z, Oromiehie A. 2014. Migration of aluminum and silicon from PET/clay nanocomposite bottles into acidic food simulant. Packag Technol Sci 27:161–8. Farkas J, Peter H, Christian P, Urrea JAG, Hassell¨ov M, Tuoriniemi J, Gustafsson S, Olsson E, Hylland K, Thomas KV. 2011. Characterization of the effluent from a nanosilver producing washing machine. Environ Intl 37:1057–62. Fern´andez A, Picouet P, Lloret E. 2010a. Cellulose-silver nanoparticle hybrid materials to control spoilage-related microflora in absorbent pads located in trays of fresh-cut melon. Intl J Food Microbiol 142:222–8. Fern´andez A, Picouet P, Lloret E. 2010b. Reduction of the spoilage-related microflora in absorbent pads by silver nanotechnology during modified atmosphere packaging of beef meat. J Food Prot 73:2263–9. Geranio L, Heuberger M, Nowack B. 2009. The behavior of silver nanotextiles during washing. Environ Sci Technol 43:8113–8. Golanski L, Gaborieau A, Guiot A, Uzu G, Chatenet J, Tardif F. 2011. Characterization of abrasion-induced nanoparticle release from paints into liquids and air. J Phys Conf Ser 304:012062. Gottesman R, Shukla S, Perkas N, Solovyov LA, Nitzan Y, Gedanken A. 2011. Sonochemical coating of paper by microbiocidal silver nanoparticles. Langmuir 27:720–6. Gottschalk F, Ort C, Scholz RW, Nowack B. 2011. Engineered nanomaterials in rivers–exposure scenarios for Switzerland at high spatial and temporal resolution. Environ Pollut 159:3439–45. Gottschalk F, Sonderer T, Scholz RW, Nowack B. 2009. Modeled environmental concentrations of engineered nanomaterials (TiO(2), ZnO, Ag, CNT, fullerenes) for different regions. Environ Sci Technol 43:9216–22. Hearty A, Gibney M, Vin K, Leclercq C, Castle L, O’Mahony C, Oldring P, Volatier J, Mckevitt A, Tenant D, Mcnamara C, Kettlitz B. 2011. The FACET project: a chemical exposure surveillance system for Europe. Food Sci Technol 25:26–9. Huang Y, Chen S, Bing X, Gao C, Wang T, Yuan B. 2011. Nanosilver migrated into food-simulating solutions from commercially available food fresh containers. Packag Technol Sci 24:291–7. Impellitteri CA, Tolaymat TM, Scheckel KG. 2009. The speciation of silver nanoparticles in antimicrobial fabric before and after exposure to a hypochlorite/detergent solution. J Environ Qual 38:1528–30. Kaegi R, Sinnet B, Zuleeg S, Hagendorfer H, Mueller E, Vonbank R, Boller M, Burkhardt M. 2010. Release of silver nanoparticles from outdoor facades. Environ Pollut 158:2900–5. Kaegi R, Ulrich A, Sinnet B, Vonbank R, Wichser A, Zuleeg S, Simmler H, Brunner S, Vonmont H, Burkhardt M, Boller M. 2008. Synthetic TiO2 nanoparticle emission from exterior facades into the aquatic environment. Environ Pollut 156:233–9.

 C 2014 Institute of Food Technologists®

Kojima Y, Usuki A, Kawasumi M, Okada A, Fukushima Y, Kurauchi T, Kamigaito O. 1993. Mechanical properties of nylon 6-clay hybrid. J Mater Res 8:1185–9. Kulthong K, Srisung S, Boonpavanitchakul K, Kangwansupamonkon W, Maniratanachote R. 2010. Determination of silver nanoparticle release from antibacterial fabrics into artificial sweat. Part Fibre Toxicol 7:8. Kumar R, Howdle S, M¨unstedt H. 2005. Polyamide/silver antimicrobials: effect of filler types on the silver ion release. J Biomed Mater Res B Appl Biomater 75:311–9. Kumar R, M¨unstedt H. 2005. Silver ion release from antimicrobial polyamide/silver composites. Biomaterials 26:2081–8. Lewin M. 2003. Some comments on the modes of action of nanocomposites in the flame retardancy of polymers. Fire Mater 27:1–7. Lin Q-B, Li B, Song H, Wu H-J. 2011. Determination of silver in nano-plastic food packaging by microwave digestion coupled with inductively coupled plasma atomic emission spectrometry or inductively coupled plasma mass spectrometry. Food Addit Contam A Chem Anal Control Expo Risk Assess 28:1123–8. Liu J, Hurt RH. 2010. Ion release kinetics and particle persistence in aqueous nano-silver colloids. Environ Sci Technol 44:2169–75. Mauricio-Iglesias M, Gontard N, Gastaldi E. 2011. Impact of high pressure treatment on the structure of montmorillonite. Appl Clay Sci 51:174–6. Mauricio-Iglesias M, Peyron S, Guillard V, Gontard N. 2010. Wheat gluten nanocomposite films as food-contact materials: migration tests and impact of a novel food stabilization technology (high pressure). J Appl Polym Sci 116:2526–35. Mudunkotuwa IA, Pettibone JM, Grassian VH. 2012. Environmental implications of nanoparticle aging in the processing and fate of copper-based nanomaterials. Environ Sci Technol 46:7001–10. Nguyen T, Pellegrin B, Bernard C, Gu X, Gorham JM, Stutzman P, Stanley D, Shapiro A, Byrd E, Hettenhouser R, Chin J. 2011. Fate of nanoparticles during life cycle of polymer nanocomposites. J Phys Conf Ser 304:012060. Nowack B, Ranville JF, Diamond S, Gallego-Urrea JA, Metcalfe C, Rose J, Horne N, Koelmans AA, Klaine SJ. 2012. Potential scenarios for nanomaterial release and subsequent alteration in the environment. Environ Toxicol Chem 31:50–9. O’Brien N, Cummins E. 2010. Ranking initial environmental and human health risk resulting from environmentally relevant nanomaterials. J Environ Sci Health A Tox Hazard Subst Environ Eng 45:992–1007. Oldring PKJ, Castle L, Franz R. 2009. Exposure to substances from food contact materials and an introduction to the FACET project. Deutsche Lebensmittel-Rundschau 105:501–7. Petersen EJ, Zhang L, Mattison NT, O’Carroll DM, Whelton AJ, Uddin N, Nguyen T, Huang Q, Henry TB, Holbrook RD, Chen KL. 2011. Potential release pathways, environmental fate, and ecological risks of carbon nanotubes. Environ Sci Technol 45:9837–56. Scheckel KG, Luxton TP, El Badawy AM, Impellitteri CA, Tolaymat TM. 2010. Synchrotron speciation of silver and zinc oxide nanoparticles aged in a kaolin suspension. Environ Sci Technol 44:1307–12. Schlagenhauf L, Chu BTT, Buha J, N¨uesch F, Wang J. 2012. Release of carbon nanotubes from an epoxy-based nanocomposite during an abrasion process. Environ Sci Technol 46:7366–72. Schmidt B, Katiyar V, Plackett D, Larsen EH, Gerds N, Koch CB, Petersen JH. 2011. Migration of nanosized layered double hydroxide platelets from polylactide nanocomposite films. Food Addit Contam A Chem Anal Control Expo Risk Assess 28:956–66. Schmidt B, Petersen JH, Bender Koch CB, Plackett D, Johansen NR, Katiyar V, Larsen EH. 2009. Combining asymmetrical flow field-flow fractionation with light-scattering and inductively coupled plasma mass spectrometric detection for characterization of nanoclay used in biopolymer nanocomposites. Food Addit Contam A Chem Anal Control Expo Risk Assess 26:1619–27. Sˇ imon P, Chaudhry Q, Bakoˇs D. 2008. Migration of engineered nanoparticles from polymer packaging to food–a physicochemical view. J Food Nutr Res 47:105–13. Singh G, Stephan C, Westerhoff P, Carlander D, Duncan TV. 2014. Measurement methods to detect, characterize, and quantify engineered nanomaterials in foods. Compr Rev Food Sci Food Safety 13:693–704. Song H, Li B, Lin Q-B, Wu H-J, Chen Y. 2011. Migration of silver from nanosilver-polyethylene composite packaging into food simulants. Food Addit Contam A Chem Anal Control Expo Risk Assess 28:1758–62.

Vol. 13, 2014 r Comprehensive Reviews in Food Science and Food Safety 691

Measuring nanomaterial release . . . Song Y, Koontz J, Juskelis R, and Zhao Y. 2013. Static liquid permeation cell method for determining the migration parameters of low molecular weight organic compounds in polyethylene terephthalate. Food Addit Contam A Chem Anal Control Expo Risk Assess 30: 1837–48. Szakal C, Tsytsikova L, Carlander D, Duncan TV. 2014. Measurement methods for the oral uptake of engineered nanomaterials from human dietary sources: summary and outlook. Compr Rev Food Sci Food Safety 13:669–78. Tang Y, Lewin M. 2007. Maleated polypropylene OMMT nanocomposite: annealing, structural changes, exfoliated and migration. Polym Degrad Stabil 92:53–60. U.S. Food and Drug Administration. 2002. Guidance for industry: preparation of premarket submissions for food contact substances: chemistry recommendations [Online]. Available at: http://www.fda.gov/Food/ GuidanceRegulation/GuidanceDocumentsRegulatoryInformation/ IngredientsAdditivesGRASPackaging/ucm081818.htm. Accessed on 2013 Oct 29. von Goetz N, Fabricius L, Glaus R, Weitbrecht V, G¨unther D, Hungerb¨uhler K. 2013. Migration of silver from commercial plastic food containers and implications for consumer exposure assessment.

Food Addit Contam A Chem Anal Control Expo Risk Assess 30: 612–20. Wang DL, Watson SS, Sung L-P, Tseng I-H, Bouis CJ, Fernando R. 2011. Effect of TiO2 pigment type on the UV degradation of filled coatings. J Coat Technol Res 8:19–33. Wohlleben W, Brill S, Meier MW, Mertler M, Cox G, Hirth S, von Vacano B, Strauss V, Treumann S, Wiench K, Ma-Hock L, Landsiedel R. 2011. On the lifecycle of nanocomposites: comparing released fragments and their in-vivo hazards from three release mechanisms and four nanocomposites. Small 7:2384–95. Yang Z, Xie C, Xia X, Cai S. 2008. Zn(2+) release behavior and surface characteristics of Zn/LDPE nanocomposites and ZnO/LDPE nanocomposites in simulated uterine solution. J Mater Sci Mater Med 19:3319–26. Zammarano M, Gilman JW, Nyden M, Pearce EM, Lewin M. 2006. The role of oxidation in the migration mechanism of layered silicate in poly (propylene) nanocomposites. Macromol Rapid Commun 27:693–6. Zhang W, Yao Y, Sullivan N, Chen Y. 2011. Modeling the primary size effects of citrate-coated silver nanoparticles on their ion release kinetics. Environ Sci Technol 45:4422–8.

692 Comprehensive Reviews in Food Science and Food Safety r Vol. 13, 2014

 C 2014 Institute of Food Technologists®