Sunscreens with Titanium Dioxide (TiO₂) Nano-Particles: A Societal Experiment

The ethical aspects of the EHS issues surrounding nanoparticles in light of uncertainty will be illustrated by a case study on nanoparticles made primarily of titanium dioxide (TiO₂) used as a UV protector in cosmetic sunscreens. Nano-sized TiO₂ is different from the well-known micronized form that is used for its clear white appearance as a pigment in coatings, paints, cosmetics, paper, and food stuffs. This large-sized form of TiO₂ is known as titanium white or pigment white 6 for pigment applications, while it is often referred to as E171 in food applications. In contrast TiO₂ nanoparticles have various interesting electrical and optical properties that can be utilized in commercial applications. The photocatalytic properties of nano-sized TiO₂ are used in the photodegradation of pollutants, treatment of wastewater, and destruction of tumour cells. These applications mostly rely on the ability of TiO₂ nanoparticles to form reactive oxygen species (ROS) on its surface when excited with UV light. There are also investigations into the use of its catalytic properties for the production of hydrogen. Nano-sized TiO₂ also has photovoltaic properties which can be used in cells for producing electricity from light, while its superhydrophilicity can be applied in ‘self-cleaning’ windows and ceramics. The electrochromatic qualities of TiO₂ nanoparticles can potentially be applied in windows that colour when a small voltage is applied, while its ability to absorb substances has been investigated for hydrogen storage and sensing applications. Finally, the light scattering and absorption properties of nano-sized TiO₂ can be used for UV protection in applications such as paints, plastics, and cosmetic sunscreens.

For the case study in this paper, we focus on the cosmetic sunscreen utilisation of TiO₂ nanoparticles. The nano-sized particles are used in sunscreens as an alternative to existing chemical UV absorbers, such as p-aminobenzoic acid and benzophenones, which can cause allergic reactions in sensitive individuals. Sunscreen lotions are generally marketed as cosmetic products in most countries including the European Union. In the United States however, sunscreens are treated as over-the-counter (OTC) drugs under oversight of the US Food and Drug Administration (FDA).

This specific utilisation of nano-sized TiO₂ was selected since it has already been marketed for over a decade and has become one of the most widely-used examples of present (first generation) nanotechnological applications. Focusing upon such a specific case study may seem to narrow the case so that it loses the heterogeneity that is so prevailing in the nanotechnological field. However, the case still harbours the very specific diversity related to the nanotechnological field and the case-by-case approach is in line with the advocated way of approaching nanotechnological issues. The heterogeneity of TiO₂ nanoparticles for sunscreen applications can be found in the following factors:

The size measure of particles is generally used to define materials as nano-sized. This kind of definition based on a measured size seems straightforward; nonetheless it is subject to several difficulties. The first complication is the comparison of the size measure due to the different ways the particle size can be measured and calculated. Size measurements are based on a collection of nanoparticles and so the calculated average size of the collection can be represented on volume, weight, or area basis. Furthermore, the measurement devices generally use specific environments, such as hydrodynamic or aerodynamic, and require pre-treatment methods for measuring, which can lead to misrepresenting the actual size in applied circumstances. A second complicating factor is that the particle collections have a specific distribution of sizes. This distribution of particle sizes is the result of the production processes and is specific for the applied conditions [29]. So representing the collection by only an average particle-size measure does not indicate how varied that size distribution of the collection is. Thus the size measure alone does not give all the information about the used nanoparticle that is needed to differentiate between collections with potential different properties and risk profiles. For example a collection with an average particle size of 150 nm is considered not a nanomaterial under most accepted definitions that set the size limit at 100 nm. However, this collection can contain a certain fraction of particles that are smaller than 100 nm and thus contains nanoparticles. It is therefore difficult to compare used nanoparticles based on size measures alone. A third complication is caused by the fact that nanoparticles, such as TiO₂ nanomaterials, tend to agglomerate and aggregate into larger structures [9]. This assemblage into larger structures alters the actual properties of the material and can thus change the risk profile as well as the effectiveness of the application. The agglomeration and aggregation effects are also environmentally dependent and so sample preparation before size measurements is very critical to insure that the sample is a valid representation of the actual effective particle size in application conditions.

Titanium dioxide nanoparticles can also be produced in various morphologies, such as spheres, tubes, rods, and wires [7]. Furthermore, titanium dioxide can exist in various crystal morphologies of which anatase and rutile are the most prominent. These different crystal structures also give various properties to nano-TiO₂. For example, the rutile has a lower density, higher refractive index, and is less photo-active than the anatase crystal structure. Another factor of the heterogeneity of TiO₂ nanoparticles for sunscreen applications can be found in the purity of the particle. Due to production methods or intentionally contaminating the crystal structure (doping) with other metals the properties of the nano-TiO₂ can be changed [1]. This doping is for example implemented by the company Oxonica-Croda, which uses manganese to reduce the unwanted photo-activity of the nano-sized TiO₂ marketed under the name Optisol.

To utilise TiO₂ nanoparticles in sunscreens, they are normally surface coated with silica, alumina and/or various polymers to (a) increase its stability in the lotion or cream and (b) reduce its photo-activity [22, 23]. Not only the material used for the surface coatings, but also its thickness, its chemical purity, and the use of multiple coatings, increase the heterogeneity of the utilised TiO₂ nanoparticles further. The use of coatings also raises questions about its effectiveness to reduce the unwanted photo-activity, durability of the coating layer in the whole life-cycle, and the recyclability of these particles. Another factor in the diversity of nano-sized TiO₂ is not the particle itself, but the matrix in which the particle resides. The nanoparticles can, for example, be fixated in a matrix, such as silica beads with nano-sized TiO₂ particle of Sunjin Chemical Corporation, or glued together by the used surface coating. Together these factors lead to a large diversity in nano-sized TiO₂ particles used in cosmetic sunscreens. As will be shown below, this heterogeneity results in more uncertainties, which forms a barrier to assess and manage the risks of nanomaterials.

To assess the risk of technological applications, the hazards should first be identified and the exposure predicted. Then the negative effect and exposure need to be correlated in such a way that the resulting risk can be characterised and classified. The relation between the impact of the undesirable event and the probability of its occurrence is generally defined as the product between the two. In this way the hazard phenomenon is specified quantitatively into risk. It is thus clear that the assessment and management of technological risk requires knowledge of the possible effects of application, the exposure, and the relationship between the effects and the exposure. This knowledge is generally based on statistical data and toxicological studies; however such information is mostly unavailable for novel technologies due to the uncertainties that surround them during conception and initial implementation.

These uncertainties are caused by a number of underlying factors, of which (a) lack of knowledge, (b) ignorance, and (c) complexity are the most prominent. In a situation where we know the effects, but cannot attribute probabilities to the likelihood of its occurrence we speak of lack of knowledge or “known unknowns”. This knowledge gap can generally be overcome by gathering more information. However, during the conception and initial implementation of a technology, not all likelihoods can be quantified. With nano-sized TiO₂, for example, there is still no consensus about the penetration of these nanoparticles through the skin. Initial studies indicated that nano-TiO₂ particles could penetrate the skin, however these results were refuted, since the methodology could not differentiate between penetrated particles and particles trapped in hair follicles [15]. Currently, studies suggest that compromised skin—such as diseased, sunburned, or physically damaged dermis—could provide an insufficient barrier [14]. Another example of such uncertainties is the conflicting results in literature, because of the different measures for exposure doses—based on mass, area, or particle number—that are utilised and in practice hard to convert into each other [16, 20, 28]. A final illustration of this lack of knowledge is that, besides the two major routes of exposure, specifically inhalation and skin contact, also alternative exposure routes, such as penetration of TiO₂ nanoparticles dust into the brain via nasal exposure [27], were and still are only marginally investigated. It is clear that in the case of TiO₂ nanoparticles for sunscreens there are still knowledge gaps in certain areas.

In some cases we even do not know that there is a specific kind of hazard. Targeted gathering of information does not remedy the situation, because we are dealing with what can be called “unknown unknowns”. Generally this cause of uncertainty is referred to as ignorance and it clouds risk assessment and management, because it is tremendously difficult—if not impossible—to anticipate the consequences of hazards that are unknown to us. Simply put, we do not know what we have to prepare ourselves for. In retrospect, we can illustrate that such a situation has already occurred with nano-sized TiO₂ in a cosmetic sunscreen application. A few years after the introduction of these sunscreens the BlueScope Steel Company investigated the appearance of defects on prepainted steel roofs. They concluded that the defects were the result of extreme weathering, caused by TiO₂ from sunscreen lotions used by workers during installation [3].

Thirdly, uncertainty can also be caused by the complexity of causal relations. The causal relation between the agent and the specific effect can be very hard to express due to diffuse system interactions, long term cumulative effects, system adaptability, incoherent dynamics, and subject heterogeneity. The large diversity in utilised TiO₂ nanoparticles, as described above, creates such complexity and so forms a barrier for effective risk assessment and management. This problem can be illustrated by the fact that there are some issues surrounding the interpretation and relevance of toxicological studies. Many toxicological studies of nano-sized TiO₂ use P25 of Evonik Degussa as their object of research. However, the usability of these studies for TiO₂ nanoparticles for sunscreens is at least questionable, because P25 is sold as a catalyst and not for cosmetic application. Moreover P25 is not coated to reduce photo-activity and it is of the anatase crystal type, whereas almost all of the TiO₂ nanoparticles utilised in sunscreens are coated and of the less photo-active rutile type. Another issue is that most toxicological studies on TiO₂ are short term and only look for acute toxicity. For instance Griffitt et al. [11] studied the effects of nano-sized TiO₂ on Zebrafish in 48 h experiments. They concluded that it is almost non-toxic in the timeframe studied, however, they also found some alterations in the expression of genes that might lead to negative effect upon long-term exposure. A further hurdle in setting up needed causal relation is that the studies are done under laboratory conditions and on test animals, such as rats, mice and hamsters, which makes it questionable if these represent the actual circumstances in the product’s life cycle and if the results can be directly translated to other species. For example, it is already established that the various test rodents have species differentiated responses to TiO₂ nanoparticles [4].

Finally, ambiguity can contribute to the uncertainties in the EHS debate. Ambiguity refers here to the type of uncertainty in the interpretation or assessment that can be given to EHS data [18]. For example, one can interpret the conclusion of a study reporting that a certain nanoparticle can cause observable harm above a certain level as a negative effect and claims that avoidance of the material would be wise. In contrast, one could also regard this nanomaterial as safe, because the level is above that of a reference material which is already used in practice. Although, these ambiguity issues raise ethical questions, the other three factors of uncertainty (lack of knowledge, ignorance, and complexity) also raise ethical issues which we discuss in further detail below.

From the current debate about the impact of nanoparticles on (EHS) issues debate it is clear that stakeholders want to have a complete assessment of the possible risks, before the novel technology, such a nanotechnology are being introduced into the market. However, the uncertainties surrounding the novel technology make it essentially impossible to make an early complete risk assessment, because not all information can be available prior to introduction. In principle, only after its introduction the hazards of the technology can be fully assessed. In such case, the introduction would amount to a large scale societal experiment, to test the hazards of the novel technology [13].

The moral acceptability of such a societal experiment has been investigated by van de Poel [26]. He describes four prima facie moral conditions that are helpful to judge the acceptability of a societal experiment in the light of uncertain risks. These seemingly reasonable conditions are: (1) the absence of alternatives, (2) the controllability of the experiment, (3) informed consent, and (4) the proportionality of hazard and benefits. To these acceptability conditions van de Poel adds four responsible set-up requirements: monitoring, feedback, scale, and containment. We will use those requirements as part of the controllability condition, since that is their main aim.¹ In the case of nano-sized TiO₂ in cosmetic sunscreen we will show that the introduction and current practice do not adhere to these four conditions for the moral acceptability of societal experimentation under uncertainty.

We will show in this section that the four modified acceptability conditions of societal experimenting could be used to support and guide the introduction of innovative technologies. In general this requires that novel technologies can be monitored and are flexible; so that the technology can be adapted after introduction when new information becomes available via continued monitoring and evaluation. Furthermore, the conditions can also be used to point the way for already applied technologies that could be considered morally unacceptable, such as the TiO₂ nanoparticles in cosmetic sunscreens. In the TiO₂ case one could conceive of five complementing routes to do so, knowingly: (1) closing the existing knowledge gap, (2) monitoring the societal experiment, (3) systematic evaluation of quantifiable risk and benefits, (4) designing for safety , and (5) incorporation the experimental nature of technological introduction into regulation. These five routes based upon the four modified conditions will be discussed below.