Antioxidants are frequently mentioned in everyday conversations and their health benefits are beloved by both marketers and media. Indeed, antioxidant, like natural and organic seems to now be synonymous with good health. However, few people actually know what antioxidants are and how they work. This is certainly true of the general population, but the broad appeal and widespread interest in antioxidants (and oxidative processes) have meant that there are many scientists working in this field without a knowledge of the chemistry that is fundamental to understanding antioxidant behavior or comprehension of the limitations of antioxidant activity measurements.

What has led to this situation? Antioxidants were for many years of considerable industrial and commercial interest and value because they provided protection of commercially important substances such as rubber and petroleum. Research on antioxidants was concerned with investigating the molecular mechanisms by which they exerted their protective effects.²  However, from about 1980, their commercial importance was gradually overshadowed by their putative role in human health. It is the latter role that now completely dominates the field of antioxidant research, much to the detriment of the development of the field.

An understanding of how antioxidant activity is measured provides a sound basis for studying the mechanism of antioxidant function. The determination of antioxidant activity pre-supposes some basic but detailed background knowledge. What is an antioxidant? What do we mean by antioxidant activity? In what sample matrices are we interested, and so forth. The concept of antioxidants is fairly complex, but this chapter explains it in plain, simple terms debunking the hype and removing the confusion of terminology that has proliferated around this area. Most simply, and in order to set the scene for the subsequent sections, antioxidants are substances that oppose oxidation. Oxidation is a chemical process whereby electrons are lost during the reaction by the chemical species in question i.e., the molecule, atom or ion. These electrons are gained by a different chemical species, and this process is called reduction. Oxidation and reduction reactions occur simultaneously and together, these processes are referred to as redox reactions. In undergraduate chemistry, the mnemonic OILRIG is often used as a simple summary of redox chemistry whereby Oxidation Involves Loss (of electrons) and Reduction Involves Gain (of electrons). Further detailed discussion of oxidation and antioxidants will be provided in Sections 1.2 and 1.3.

Before delving further into the chemistry of oxidation, it is informative to consider the history of antioxidants, the development of attitudes to antioxidants, and the evolution of interest from technological applications to use in foods and their impact on nutrition and health. The preservative effect of antioxidant compounds (via their inhibition of oxidative damage) was exploited by smoking of meat long before there was any understanding of the chemistry involved. The preservative action of benzoin on lard was discovered in France by Deschamps in 1843 and benzoinated lard became official in a number of the world's leading pharmacopeias. However, it was not until 1926 that the stabilizing effect of benzoin was attributed to its property as a “negative catalyst of oxidation”.³ 

The first technological application of antioxidants occurred in the late nineteenth century in the rubber industry, when it was observed that some molecules, identified empirically, could slow the degradation and allow optimization of the process of vulcanization (a process in which rubber molecules are cross-linked). From then through the early twentieth century, antioxidant research concentrated on the use of antioxidants in industrial processes such as the prevention of both metal corrosion and the polymerization of fuels in the fouling of internal combustion engines. Somewhat later, antioxidants entered the arsenal of the emerging food industry, as a key tool to curb the oxidative degradation of stored food due to oxidation of unsaturated fats, which is the cause of rancidity. However, it was the identification of the antioxidant activity of vitamins E and C in the 1920s and 1930s that revolutionized the field of antioxidants and led to the realization of their importance in the biochemistry of living organisms. Mattill's group (ref. 4 and references therein) suggested that “the successful administration of vitamin E concentrates … indicates that the function of vitamin E, … by its specific anti-oxidizing capacity, controls the progress of oxidation in the tissues.” It was the work of Mattill and his group that led to the identification of antioxidants as reducing agents that limit oxidative reactions (see ref. 4).

The next major breakthrough was the discovery of the antioxidant, superoxide dismutase (SOD) in the 1960s (see ref. 5). This enzyme is found in almost all aerobic cells and extracellular fluids and it is at the top of the list of first line defense antioxidants.⁶  It was also in the 1960s that the concept of the Mediterranean diet and its health benefits was developed. The diet first came to prominence in 1975 when it was publicized by the American biologist Ancel Keys and chemist Margaret Keys, his collaborator.⁷  Data demonstrating the health benefits of the Mediterranean diet originated from an epidemiological study⁸  that was later confirmed by the Seven Countries Study first published in 1970.^9–11  While there is no single definition of the Mediterranean diet, the most recognized version was published by Walter Willett and colleagues of Harvard University's School of Public Health (ref. 12 and references therein). It is typically high in plant foods with much higher consumption of unsaturated fat derived from olive oil but otherwise similar to many non-Mediterranean diets in other classic risk factors for coronary heart disease (CHD). The high content of plant foods led to this diet and its benefits being associated with the antioxidant content of plants.¹³  Indeed, the ATTICA study¹⁴  measured the antioxidant capacity of men and women in Greece. The antioxidant capacity of participants who followed a traditional Mediterranean diet was 11% higher than those who didn't adhere to a traditional diet. Participants who followed the traditional diet the most also had 19% lower oxidized low-density lipoprotein (LDL) cholesterol concentrations. However, conflicting results of numerous other studies (e.g., ref. 15–19) of the benefits of the diet have generated much controversy.

In parallel with the focus on the Mediterranean diet, observational studies on the typical eating habits, lifestyles, and health histories of large groups of people showed that those who ate more vegetables and fruits had lower risks of several diseases. The French paradox was formulated by French epidemiologists²⁰  in the 1980s. It is the observation of low CHD death rates amongst French people despite high intake of dietary cholesterol and saturated fat. The concepts of the Mediterranean diet and the French paradox were consolidated²¹  into a single theory on the role of nutrition in CHD. “Superfood” emerged as a marketing term to promote foods that were claimed to have exceptional health benefits. Interest focused on a common factor in these materials, they were good sources of antioxidants, and research in this area proliferated. This frenetic activity focused attention on the antioxidant hypothesis,^22,23  which posited that vitamin C, vitamin E, carotenoids and other antioxidant nutrients afforded protection against chronic diseases such as CHD and cancer by decreasing oxidative damage. Despite numerous studies and clinical trials^24–28  there is still no definitive answer on the value of increased intake of antioxidant-rich foods or of antioxidant supplements. Indeed, some studies have demonstrated deleterious effects and even toxicity of antioxidants when in excess.^29,30 

In 2012, a new phase began with the discovery of the effects of antioxidants on gene expression by mechanisms other than via an influence on the DNA sequence (i.e., epigenetics). This has been presented as a new unifying mechanism to putatively explain all the activities of antioxidants.³¹ 

There are many unanswered questions about antioxidants. Are antioxidants also pro-oxidants, i.e., can they also cause oxidation? Are antioxidants harmful or beneficial? Views on antioxidants seem to oscillate between healthy and essential to toxic, and from cure-all to waste of money. The answer to the preceding questions seems to be a definite ‘yes’ in all instances depending on a range of factors but particularly the amount. As Paracelsus (1493–1541) observed “the dose makes the poison”. Interestingly, he also believed that “nature marks each growth …. according to its curative benefit”.³²  Thus, beans were marked by nature for treatment of kidneys, walnuts to treat the brain, etc. We now recognize that shape is important to biological activity, but at the molecular rather than the macro-level.

Whatever one's attitude to antioxidants, they are ubiquitous and encountered in a broad range of situations from foods to packaging to pharmaceuticals and cosmetics to physiological situations. Reasons for the interest in antioxidants are also diverse and include their potential health impacts (both positive and negative must be considered), their effect on the shelf-life of a food, the “potency” of an antioxidant supplement, the role in cell regulation, and industrial applications, which still include fuels, plastics, pharmaceuticals, etc.

Antioxidants are, as the name implies, substances that oppose oxidation. Oxidation is therefore a good starting point for a discussion of antioxidants. An understanding of oxidation processes is essential before considering the information provided by antioxidant activity tests and examining the limitations of such tests.

Oxidation processes, not surprisingly, are also ubiquitous; for example, combustion of fuels and rusting of iron involve oxidation, batteries rely on oxidation to generate power and oxidation is fundamental to life as cells conserve energy in the form of ATP by coupling its synthesis to the release of energy via oxidation. However, oxidation cannot occur in isolation; it must be accompanied by a reduction process. In the case of rust formation, for example, it is iron that is oxidized and oxygen that is reduced; oxygen is termed the oxidizing agent or oxidant as it is the causative agent for the oxidation of the iron, and iron (the causative agent for the reduction of oxygen) is the reducing agent or reductant.

Although oxidation–reduction (redox) processes are encountered in many diverse situations, including combustion and corrosion, in pharmaceuticals,^33,34  cosmetics,³⁵  biodiesel,³⁶  petroleum and lubricants,³⁷  and automobile tyres³⁸  to name just some, our interest in these processes is limited to their occurrence in humans, animals and plants and hence in foods and beverages. Because plants and animals are often the origin of pharmaceuticals and nutraceuticals, these materials will also be mentioned at times. For convenience we discuss these systems as biological (or physiological) materials, and foods and beverages, but there is clearly much overlap. In a redox process in these systems, the substance that is oxidized is called the substrate¹ and the process is variously described, somewhat loosely (see Section 1.2.1), as either autoxidation (meaning auto-catalytic mechanism) or peroxidation. These terms are used synonymously, the former being preferred by chemists and the latter by biologists.

Fundamentally, redox reactions involve the transfer of electrons. The substrate, being oxidized, loses electrons, while the oxidant gains electrons. This can be illustrated in a simple reaction involving the oxidation of hydrogen (substrate) to water, using oxygen as the oxidant.

The transfer of electrons can be followed by considering the change in formal charges on the atoms. In H₂ and O₂, the formal charges on H and O are 0. In H₂O, the formal charge on H is +1, while on O it is −2. The increase in formal charge, 0 → +1, shows that hydrogen has been oxidized (lost electrons), while the decrease in formal charge, 0 → −2, shows that oxygen has been reduced.

Another way to view this redox process is to consider it as an example of hydrogen atom transfer, i.e., splitting the H–H bond to give two H˙ atoms (the ˙ indicates an unpaired electron or free radical), then the H˙ combines with an oxygen atom to give H₂O. Thus, reduction can involve hydrogen atom transfer (as will be seen below when considering antioxidants).

Still another way to view this process is to consider an oxygen atom, inserting itself in the H–H bond to give H–O–H. Thus, oxidation may be considered as oxygen atom insertion (as will be seen in e.g., lipid oxidation, below).

While oxygen atom insertion or hydrogen atom transfer is common in redox processes involving organic molecules, it must be stressed that electron transfer can occur without rearrangement of bonds. Commonly, these redox processes involve transition metals, e.g., iron, which can cycle between +2 and +3 oxidation states, without changing what it is bound to. This property of transition metal ions means they are frequently implicated in redox processes in systems relevant to this book.

The foregoing makes it clear that there are two essential components in a redox reaction; namely, an oxidant or oxidizing agent being the species that brings about the oxidation of the substrate (or reductant). This section examines the range of chemicals that can function as either the oxidant or substrate in foods and biological/physiological situations.

An examination of the role of the substrate in oxidative processes would be incomplete without firstly acknowledging and discussing the importance of oxidants or pro-oxidants in foods and biological systems. Oxygen is a paradox;³⁹  on the one hand, it is essential to life but, at the same time, it causes untold damage via processes such as corrosion and, by the latest theories, also to health. The bond dissociation energy for molecular oxygen is very high so that reactions involving the rupture of this bond require high temperatures. This is fortunate as spontaneous combustion of plants and animals might otherwise be a common occurrence. The reactivity of oxygen makes interesting reading.⁴⁰  In the same way, oxidants are an enigma, leading to a dual existence in both biological and pharmacological systems. In one situation, they can exert a powerful positive impact on the system but, at other times, they are harmful to the system and exert deleterious effects. Oxygen is the terminal oxidant of the respiratory or electron transport chain⁴¹  and it has a unique combination of properties that fit it to this role.⁴² 

The oxidants of interest in the current context may be identified as reactive species of oxygen (ROS), nitrogen (RNS), and the recently identified reactive sulfur species (RSS). These reactive species can be either free radical oxidants or non-radical oxidants, depending on whether or not they contain an unpaired electron, respectively. We have retained the use of free in free radical although it is redundant. They are generated in biological materials via both endogenous (mitochondria, peroxisomes, endoplasmic reticulum, phagocytic cells, etc.) and exogenous sources. Endogenous oxidants are a result of normal aerobic respiration, metabolism, and inflammation, whereas exogenous oxidants are generally formed from environmental factors such as pollution, sunlight, strenuous exercise, X-rays, pesticides, certain drugs like halothane and paracetamol, smoking and alcohol. They are well known to induce oxidative damage to lipids, sugars, proteins, and DNA.⁴³  The damage induced by these oxidants has been linked to disease in humans and to changes in food quality and shelf life.⁴⁴ 

ROS includes all unstable metabolites of molecular oxygen that have higher reactivity than O₂. Mitochondria are the most important source of cellular ROS. This is a result of the electron transport chain located in the mitochondrial membrane, which is essential for energy production through ATP generation inside the cell. RNS are derived from reactions of nitric oxide (NO˙) with superoxide (O₂˙⁻), giving peroxynitrite (ONOO⁻). Peroxynitrite is “a potent and versatile oxidant that can attack a wide range of biological targets”, even though neither of its precursors are strong oxidants themselves.⁴⁵  RSS are omnipresent and sometimes as important as ROS and RNS, which have dominated the oxidant field for decades.⁴³  The formation and properties of free radical oxidants⁴⁶  and non-radical oxidants⁴⁷  have been discussed by a number of authors.

Examples of some typical oxidants are presented in Table 1.1. They exhibit extreme diversity in reactivity and lifetimes with half-lives varying from microseconds to many days. Free radical oxidants have attracted much attention. The superoxide anion radical and hydrogen peroxide are the most commonly formed and important oxidants. The main source of superoxide is the one-electron leakage of the mitochondrial respiratory chain and, in plant cells, of the chloroplasts redox system. The hydroxyl radical is the most reactive and can cause damage to DNA, proteins, lipids and carbohydrates and causes more severe damage to cells than any other oxidant. As a result of the reactivity and the abundance of biological targets, HO˙ primarily reacts at the site where it is generated. In contrast, the peroxyl radical (ROO˙) is relatively long lived with a considerable diffusion path length in biological systems and so can exert its effects at remote sites.

An initial oxidant can give rise to secondary species with markedly different reactivity and lifetimes. Environment is also an important factor as, for example, with the superoxide anion radical, where the protonated form is the much more powerful oxidant. The pK_a for the protonated superoxide anion radical is 4.8 so, in acidic environments such as in lysosomes, significant amounts of the more powerful protonated oxidant can be present.⁴⁸  On the other hand, in blood plasma, where the pH is around 7.4, the superoxide will exist in the deprotonated form and be less reactive.

Oxidants can be characterized by the reduction potential, E⁰, a measure of the ability of the substance to accept an electron, which provides an indication of the likely thermodynamic occurrence of a particular reaction. E⁰ values range between −3.0 V to +2.9 V, with the more positive the value, the stronger the oxidant.⁵⁶  Although reduction potentials provide thermodynamic data (i.e., will a reaction proceed or not) they tell us nothing about the rate or speed (kinetics) of a reaction. For example, in a reaction of the amino acid methionine, with either hydrogen peroxide or hypochlorous acid, the former is a more powerful oxidant (E⁰ 1.32 vs. 1.28 V) and might be expected to produce greater damage to the amino acid.⁴⁸  However, their respective rates of reaction with methionine differ by approximately nine orders of magnitude, with that of hypochlorous acid being the greater. Rate data must be interpreted carefully with due attention to the precise conditions as the rate of a reaction can vary markedly with environment. For example, the amino acid cysteine is oxidized by hydrogen peroxide. However, the rate constants for reaction of hydrogen peroxide with a cysteine residue in the protein, bovine serum albumin are about seven orders of magnitude less than those for the same amino acid residue in peroxiredoxin enzyme, another protein.⁴⁸ 

The effect of pH on oxidation kinetics and thermodynamics needs to be considered. For example, in one study on the decomposition of hydrogen peroxide, Jung et al.⁵⁷  found half-lives of hours to days depending on the pH, with lower pH values associated with longer half-lives. On the other hand, Bockris and Oldfield⁵⁸  reported that hydrogen peroxide became a weaker oxidant at higher pH. This is an instance where the pH dependency of kinetics and thermodynamics shows an inverse relationship. At lower pH values hydrogen peroxide is a stronger oxidant, but reacts more slowly and vice versa.

Oxidants play an important regulatory role in biological processes. Some oxidants are generated by enzymes specifically to perform a particular function, such as killing invading pathogens by activated neutrophils via the generation of the superoxide anion radical, for example. Other oxidants arise unintentionally as part of normal metabolic processes. The role of oxidants as regulation and redox signaling agents in various cellular pathways is now well recognized. H₂O₂ and ONOO⁻, in particular, have been implicated in a considerable number of cellular signaling cascades.⁴³  These non-radical oxidants have longer half-lives than almost all other oxidants, which allow them to migrate away from their production sites and to diffuse through membranes thus exerting their action at remote sites.

Under conditions of normal aerobic metabolism, reactive oxidative species (i.e., ROS, RNS and RSS) are formed continuously as part of normal physiological functions like ATP generation and catabolic and anabolic processes. Thus, aerobes are constantly being exposed to oxidants, and their DNA, protein and lipids are continuously damaged. Under these circumstances, the formation and elimination of oxidants are kept in balance⁴⁸  by several defense mechanisms including:

• Low molecular mass scavengers (e.g., ascorbic acid, tocopherols),

• Enzymes that remove oxidant precursors (e.g., catalases), and

• Enzymes that directly scavenge oxidants (e.g., superoxide dismutases).

The oxidized, damaged substances are either repaired by enzymes specifically tasked to repair the damage (e.g., disulfide reductases), and/or eliminated by enzymes and organelles that remove non-repairable molecules (e.g., lysosomes).

Despite the array of protective systems, overproduction of oxidants, or an imbalance between formation and elimination of oxidants in favor of higher oxidant levels, and/or depletion of the enzymatic and non-enzymatic antioxidant defense systems can occur, and this is termed oxidative stress. Reductive stress in which there is an elevation in the NAD + /NADH, NADP + /NADPH, and GSH/GSSG ratios or over-expression of antioxidant enzymatic systems can also occur. Some cellular processes that can be affected by reductive stress include cell growth, formation of disulfide bonds, mitochondrial function and metabolism.⁵⁹  Overconsumption of antioxidants in the diet can also be detrimental to redox balance, either inducing reductive stress or, at high enough levels, exhibiting pro-oxidant behavior.⁶⁰ 

In food or beverage systems, exclusion of oxygen can result in a hypoxic environment resulting in reductive defects in products. For example, the replacement of cork with plastic or metal closures can induce reductive spoilage in wine.⁶¹  In addition, just as has been found in physiological systems, antioxidants can become pro-oxidant under certain conditions. Bradshaw et al.⁶²  reported that ascorbic acid showed dual anti-/pro-oxidant behavior in a white wine model system.

Redox equilibrium is essential for cellular homeostasis as it moderates production of reactive oxidative species. The concept of redox biology, loosely defined as a homeostatic role of oxidants in cell and tissue biology, has emerged in contrast to the notion of oxidative or reductive stress.⁶³  Redox stress (oxidative or reductive) is an extremely complex process. Its impact on a human, animal or plant depends on several factors; the type of oxidant, the site and intensity of its production, the nature of the substrate and the functioning of the various defense mechanisms. In some situations, oxidants are created in very high levels throughout the body as occurs, for example, with trauma, infection, or inflammation. Prolonged high levels beyond the immediate needs of the system present a problem. There is convincing evidence that on-going oxidative stress⁶⁴  is responsible for the onset of many adverse processes including, cardiovascular diseases, neuronal degeneration, cancer and even aging in mammals, plants and microorganisms. On the other hand, if redox homeostasis is restored then substrate damage is contained and the defense systems can eliminate the damaged molecules.

There are similarities between the roles of oxidants in foods and biological systems, but also some notable differences. The raw materials for food products are the plants and animals that comprise biological/physiological systems. By way of illustration, wound response in animals mainly promotes wound healing processes, nerve cell regeneration, and immune system responses at the vicinity of the wound site. In contrast, wound response in plants is primarily directed at sealing the wound site and generating systemic signals that activate multiple defense mechanisms in remote tissues. However, oxidants play very common coordination roles in the wound responses of both plants and animals despite the differences in the responses.⁶⁵  In mammals, the oxidation process is ameliorated by the various defense systems but in general there is no automatic control system in foods.

The most significant difference in the role of oxidants between different sample types arises with processed foods where the production operations generally have been designed to stop cellular activity. Thus, generation of oxidants will not occur in processed foods via an active process but, rather, the origin of any oxidant(s) is “external” to the system. In contrast to the range of oxidants important in biological systems, the most common oxidant in foods (Table 1.2) is oxygen in the air, which penetrates foods and is dissolved in both aqueous and lipid phases. Other oxidants are of minor importance in foods and even when the reaction is catalyzed by enzymes, air oxygen is still the oxidant. In the presence of photo-sensitizers (such as chlorophylls) and in light, ordinary triplet oxygen is converted into singlet oxygen, which is 1500 times more reactive (ref. 66 and references therein).

In summary, oxidants play multiple roles, both beneficial and deleterious in biological systems, but their outcomes in food are almost always harmful, leading to loss of nutritive value, sensory quality and/or potential health risks. Every component of foods, plants and animals is a potential substrate for oxidation by oxidants. Thus, oxidation can occur anywhere in the food chain altering the economic value of plant and animal products as a raw material for food production, but also the finished food product and even the consumer.

Foods and biological systems comprise mixtures of a range of chemical compounds that include minerals, carbohydrates, proteins, peptides and amino acids, lipids including fatty acids, and DNA. These classes of biomolecules are all subject to oxidation and are potential substrates for oxidants. Indeed, oxidation of the biomolecules in physiological situations is a normal metabolic process that is under homeostatic control. Also present in these materials are vitamins, antioxidants, pigments, hormones, and numerous other small molecule minor components, but these are grouped on the basis of function and are chemically very diverse rather than a class of chemically related compounds. Although these compounds are also subject to oxidation their presence at generally lowered amounts means that they have received less attention as potential substrates for oxidants. Vitamins, because of their nutritive value, and pigments, because of their esthetic impact, are notable exceptions.

Which class of compounds (and which specific chemical) is oxidized in a particular situation depends on a number of factors that are predictable from rate constants and concentrations of the substrates. It is a competitive process in which a higher concentration and larger rate constant favor reaction. The substrate must be accessible to the oxidant and this depends on the relative concentration of the various potential substrates and oxidants, nature of the oxidant, aqueous and lipid phase equilibria, temperature, pH, light level, physical state of the system, matrix effects (i.e., presence/absence of other components) and, in the case of in vivo situations, also on cellular compartmentalization (intracellular, extracellular, membrane-associated). The presence of an antioxidant may protect one potential substrate e.g., lipids, against oxidative damage whilst accelerating damage to other biological molecules.⁶⁷  With the assumption of a homogeneous biological system (an oversimplification), the major targets of the hydroxyl radical and singlet oxygen are cellular proteins, consuming 65–70% of the available oxidants, with much lower damage to DNA, RNA, lipids and antioxidants. The key driving force in the prediction is the abundance of proteins. Similar predictions have been made for plasma and tissues.⁴⁸  Calculations of this type provide an indication of likely targets, but must be treated with caution.

The range and general nature of potential substrates are the same in foods as in biological and physiological systems despite some obvious differences. However, in practice, the range of substrates that are important in biological systems is broader than that in foods. Indeed, in some processed foods as in the case of oils and fats, the bulk of the product fits in a single class of chemical compound, lipid in this case. A further important distinction arises in the case of processed foods where the production operations generally have been designed to stop cellular activity including metabolism.

In the case of substrates in foods, two reasons can be advanced for the interest in their oxidation; firstly, the impact on food quality and food stability during processing and storage and, secondly, the behavior of the oxidized food following ingestion. Oxidation occurs during production, storage, processing, and culinary preparation of a food and can affect its taste, aroma, appearance, and nutritional quality and consumer acceptability. In the case of biological systems which are destined for consumption, the products of oxidation may impact on safety and wholesomeness and are regarded⁶⁸  as potentially harmful to humans (and probably animals). Food intake impacts nutritional status of humans and animals and their health. “Let food be thy medicine and medicine be thy food”⁶⁹  is attributed to Hippocrates and, although evidently a misquote,⁷⁰  food was nevertheless closely linked to health and disease in Hippocrates’ thinking. There is now an even more acute awareness and understanding of the link between diet and health.

The rest of this section examines substrates from the perspective of the different chemical classes, lipids, proteins, etc. and relates the discussion back to the different sample types (food and biological materials). This is not intended as a comprehensive review, but rather it demonstrates the diversity and complexity of oxidative modifications in biomolecules plus the range of considerations that are important in understanding these changes.

All biological molecules are susceptible to damage by reactive oxidative species (Figure 1.1). However, lipids are different. They are often presented first in books relating to antioxidants and generally receive greater attention than other substrates. This is partly historical because of the history and obvious nature of rancidity. However, there is also a more fundamental reason. The activation energy for lipid oxidation is lower than that for oxidation of proteins, nucleic acids and carbohydrates. However, there is an interaction between the different oxidations as lipid oxidation may be initiated by free radicals generated by oxidation of these other biomolecules while products of lipid oxidation may modify proteins and nucleic acids.

Foods high in lipids (fats) (e.g., meats and dairy products, oils and fats) undergo oxygen-dependent deterioration leading to rancidity. Apart from the above considerations, lipid is the component of foods most susceptible to oxidation,⁷¹  and it also produces the most noticeable outcome in terms of the odor and off-flavors associated with rancidity, the most common outcome of lipid oxidation. It is not surprising that lipid is the traditional target substrate in foods that has been most extensively studied.

Lipid oxidation proceeds via three different pathways:

1. Enzymatic reaction,

2. Non-enzymatic, non-radical photo-oxidation, and

3. Non-enzymatic free-radical mediated chain reaction (called autoxidation).

Pathway 1, shown in Figure 1.2, involves the non-radical action of lipoxygenases on various lipid substrates. Lipoxygenases are a family of non-heme iron-bound enzymes with Fe in its active center that are widely distributed in plants, animals and microorganisms. Lipoxygenases use molecular oxygen to catalyze the stereospecific and regiospecific oxidation of polyunsaturated fatty acids having a 1-cis,4-cis-pentadiene moiety and oils containing linoleic, linolenic and arachidonic acids are favored substrates.

Photo-oxidation via Pathway 2 is a faster reaction than autoxidation. It involves reaction of unsaturated lipids with singlet oxygen produced by sensitizers such as myoglobin or chlorophyll. This is shown in Figure 1.3 for linoleic acid. Addition of oxygen from the activated singlet oxygen at either end of the double bonds of the linoleic acid molecule produces two conjugated and two non-conjugated hydroperoxides. In contrast, production of non-conjugated hydroperoxides does not occur in autoxidation. The reaction rate of lipid with singlet oxygen is much higher than that with triplet oxygen; the reaction rates of linoleic acid with singlet oxygen and triplet oxygen are 1.3 × 10⁵ and 8.9 × 10¹ M s⁻¹, respectively.⁷² 

Pathway 3 is the classical free-radical oxidation route termed autoxidation⁷³  that leads to initiation of rapidly progressing, destructive chain reactions. The essential features of oxidation via free radical-mediated chain reaction are initiation, propagation, branching and termination steps. The process may be initiated by the action of external agents such as heat, light or ionizing radiation or by chemical initiation involving metal ions or metalloproteins.

where LH represents the substrate molecule, for example, a lipid, with R˙ being the initiating oxidizing radical. The oxidation of the lipid generates a highly reactive allyl radical (L˙) that can rapidly react with triplet oxygen to form a lipid peroxyl radical (LOO˙)

The rate of eqn (1.4) is much lower than that of eqn (1.3) under the normal atmospheric partial pressure of oxygen.²  Therefore, the steady-state concentration of LOO˙ is very much greater than the steady-state concentration of L˙.

The peroxyl radicals are the chain carriers of the reaction that can further oxidize the lipid, producing lipid hydroperoxides (LOOH), which in turn break down to a wide range of compounds including alcohols, aldehydes, alkyl formates, ketones, and hydrocarbons and radicals including the alkoxyl radical (LO˙).

The breakdown of lipid hydroperoxides often involves transition metal ion catalysis, in reactions analogous to that with hydrogen peroxide, yielding lipid peroxyl and lipid alkoxyl radicals:

Metal ions for this reaction are found naturally in food components and arise from environmental sources or metal equipment.

Termination reactions involve a combination of radicals to form non-radical products.

When oil is heated various chemical changes occur including oxidation. Thermal oxidation involves a mechanism that is basically the same as autoxidation, although the rate of thermal oxidation is faster than that of autoxidation. Thermal oxidation also produces products such as triacylglycerol dimers and polymers that are not seen in autoxidation.

Lipid autoxidation is also the primary form of lipid damage found in biological systems. In animals and humans, the substrates are LDL, red cells and phospholipid microsomes.⁷⁴  It invariably involves lipids containing polyunsaturated fatty acids and is a particular problem for biological membranes as they contain high levels of these substances. It also causes a number of problems for the cell.⁷⁵  Although the essential features of the process of lipid autoxidation are the same in both foods and in vivo there are some notable differences. Foods may have been exposed during processing and storage to elevated temperatures that are not encountered in vivo. For example, hydroperoxides decompose readily and spontaneously at 160 °C, and the peroxyl radical (LOO˙) concentration can become relatively high under such conditions, thus leading to the formation of polymers. The range of effects of free radicals is only a few tenths of a nanometer whereas the action of the non-free-radical in vivo oxidant, hydrogen peroxide is several nanometers and moreover, hydrogen peroxide can pass biological membranes freely.

The study of protein oxidation in biological systems dates from about 1960,⁷⁶  but was not undertaken in foods until about 1990 because oxidized protein did not generate any off-flavor. Indeed, knowledge about the specific effects of protein oxidation in food⁷⁷  or the role of “oxidized” dietary protein for the human body is comparatively scarce.⁷⁸  The high abundance of proteins in most systems as well as high-rate constants for oxidation reactions⁷⁹  lead to the prediction that proteins will be the primary substrate targets for reactive oxidative species. However, in practice, this will depend on the composition of the system and the presence of antioxidants plus oxygen levels, light, etc. and a number of biological factors.

Oxidants can damage proteins by both indirect and direct mechanisms. The indirect pathway does not involve oxidative damage to the protein per se, but rather this process involves oxidative damage to the DNA molecule encoding the protein.⁸⁰  The direct pathway involves either the direct action of an oxidant on a protein or photo-oxidation of the protein or secondary (indirect) attack on the protein by the products of lipid oxidation. The peptide backbone and the functional groups located in the side chains of amino acid residues are the common targets for attack by ROS whereas attack by lipid radicals is primarily at the amino acid residues (mainly lysine, arginine, histidine, tryptophan, cysteine and cystine) of the protein to form reactive protein radicals that can propagate the reaction. Alternatively, termination of the chain reaction involves reaction of the protein radicals to produce protein–protein, protein–lipid or protein–biomolecule (in which the biomolecule is any other component) adducts.

Some amino acid residues are more susceptible to attack because of their molecular structure such as the presence of sulfur centers (cysteine and methionine), an indole group (tryptophan) or a free amino, amide or hydroxyl group (lysine, arginine and tyrosine). However, there are differences in susceptibility of amino acid residues between the various oxidation mechanisms.

Many other by-products apart from protein–protein, and protein–lipid, etc. adducts are generated by the protein oxidation processes depending on the oxidation mechanism. These by-products include oxidized amino acids, fragmentation products and carbonyl compounds (principally from oxidation of threonine, proline, arginine and lysine residues). The total protein carbonyl content is estimated to be about 1–2 nmol mg⁻¹ protein in a variety of human and animal tissues, representing modification of about 10% of the total cellular protein.⁷⁶ 

The sites responsible for the greatest production of oxygen radicals are localized on biological membranes. It is not surprising that the membrane components themselves, proteins and phospholipids, are among the principal targets for attack by the oxygen radicals. Under conditions of oxidative stress, proteins are structurally altered in vivo leading to generation of disulfide bridges and unfolding of the protein conformation. Carbonyls are also introduced into the proteins by direct oxidation of amino acid residues, or indirectly, by covalent bonding of a carbonyl-containing moiety. Repair of the damage is possible where this involves a sulfur-containing amino acid residue, but, on the whole, mechanisms for protein repair and elimination of damaged proteins within cells are limited.

The outcome of protein oxidation is the loss of the conformation and native structure of the protein molecules plus loss of protein functionality as receptors, enzymes, transport or structural proteins, etc.⁸¹  The consequences of protein oxidation, in terms of changes to biological processes, have been extensively reviewed by Hawkins and Davies⁸²  who conclude that “further development is required to fully assess the relative importance of protein oxidation and to determine whether an oxidation is a cause, or merely a consequence, of injurious processes.” In foods, the main outcome of protein oxidation is loss of nutritive value.⁸³ 

DNA is a polymer of nucleotides with each nucleotide being composed of a nitrogenous base, deoxyribose, a five-carbon sugar and a phosphate group. Four nitrogenous bases occur in DNA: two purines (adenine and guanine) and two pyrimidines (cytosine and thymine). As with all biomolecules, DNA is subject to oxidation whether in foods or biological systems. However, there is a vast difference in the knowledge and study of DNA oxidation in the two environments. Oxidation of DNA in the human body is unfortunate but a major consequence of life in an oxygen-rich environment. “Concomitantly, survival in the presence of oxygen, with the constant threat of deleterious DNA mutations and deletions, has largely been made possible through the evolution of a vast array of DNA repair enzymes.”⁸⁴  It is not surprising that the oxidation of DNA in physiological systems has been studied extensively.

DNA is a relatively reactive molecule that is readily attacked by ionizing radiation and oxidants. In vivo, oxidation of DNA occurs via either of these sources. Of the radicals found in aerobic cells, only the hydroxyl radical readily attacks DNA and this occurs at the π-bonds of the bases (Figure 1.4) or via hydrogen abstraction from deoxyribose and the methyl group of thymine.⁸⁵  The oxidative damage includes covalent modification of the bases (DNA adducts), production of alkali labile sites and strand breaks, either formed directly or as a consequence of repair processes.⁸⁶  The full complement of oxidatively damaged species in mammalian DNA may exceed 100 different types, of which 8-hydroxyguanine is one of the most abundant.⁸⁷ 

In physiological systems, there is a considerable body of evidence that these oxidatively damaged species can ultimately cause mutations resulting in a number of possible outcomes that include inheritable disease,⁸⁸  aging, carcinogenesis and neurological degeneration⁸⁷  as the end result. The maintenance of the nuclear and mitochondrial DNA genome relies on proficient mechanisms to eliminate (by repair or removal) oxidatively damaged species. When these systems fail and prior to the development of these outcomes,⁸⁹  accumulation of a single type or multiple types of oxidative lesions⁹⁰  probably occurs. These structurally modified DNA materials lead to acute effects on cell cycle progression, leading to transient cell-cycle arrest and inhibition of DNA metabolism including transcription and replication. Long-term consequences include permanent changes in the DNA sequence.^91,92  For example, 8-hydroxyguanine is considered to be a premutagenic lesion because it can mispair with adenine during DNA replication, and this mispairing results in certain base pair transversion mutations.⁸⁷  RNA should not be ignored as it is generally more susceptible to oxidation than DNA and contributes to disturbance of the translational process and impairment of protein synthesis which can cause cell deterioration⁹³  and is implicated in the progress of pathogenesis.

There has been little interest in examining the oxidative damage of DNA in foods and its effect on food stability and quality. This is presumably because of the relatively low levels of DNA in many consumed foods. Unlike protein, lipid, etc., the amount of DNA in foods is rarely measured. It is not reported on food packaging, for instance. This may change as a DNA scanning tool for measurement of the DNA content of food has recently become available⁹⁴  and this may stimulate further developments. However, it seems unlikely that there will be a flurry of data reported on the DNA content of foods unless some other major shift occurs to drive a change. The total amount of DNA in food varies according to the type of food. For example, edible offal and animal muscle tissue contain high levels of DNA (10–100 g kg⁻¹ dry matter), whereas plant foods such as grains or potatoes contain less DNA (0.5–5 g kg⁻¹ dry matter)⁹⁵  because they generally have less cell nuclei where the DNA is located than offal and muscle tissue. Food processing may lead to partial or complete degradation or removal of DNA depending on the type of processing. Thus, heating may cause fragmentation without reducing the total DNA content while refining of vegetable oils will remove most, if not all, DNA. Adult daily consumption of DNA⁹⁶  is estimated to be in the range of 0.1 to 1.0 g per day.⁹⁵ 

In comparison to the extensive research on the redox chemistry of lipids, proteins and DNA, there is little information on carbohydrates⁹⁷  except as part of the DNA molecule. Carbohydrates are biologically very important and play multiple roles in living organisms as energy stores, fuels and intermediates, structural elements in plants and as part of DNA, glycoproteins and glycolipids.⁷⁵  Indeed, the oxidation of glucose in respiration is the basis of energy production in many organisms, and as has been noted above, the by-products of this process may themselves cause oxidative damage. Nevertheless, there is limited research on oxidation of carbohydrates as being detrimental to food or biological systems, which is somewhat surprising. However, in general, carbohydrates are not as sensitive to oxidation as lipids and proteins and the end products are not volatile⁹⁸  so they do not produce odorous spoilage products as with lipids. There are however a number of processes that involve oxidation of carbohydrates that are commercially very significant. Maillard and caramelization reactions involve oxidation of carbohydrates, but generally occur at high temperatures. For example, the conversion of bread to toast involves carbohydrate oxidation and cross-linking.

The fragmentation of polysaccharides by ROS/RNS attack is important as this will alter the functionality of these molecules.⁹⁹  Indeed, non-enzymatic scission by hydroxyl radicals of polysaccharides can be detrimental (e.g., causing arthritic diseases in mammals) or beneficial (e.g., promoting the softening of ripening fruit).¹⁰⁰ 

In physiological systems, reactions of carbohydrates, especially glucose, with proteins and DNA result in advanced glycation end products (AGEs) and age-related pigments.⁷⁵  The pigments often have a brown color, which can be seen in e.g., the change in color of human cartilage – near white at birth, turning dark brown in aged individuals. Not only are AGEs the stable end products of glycation, but they can trigger oxidative stress in cells. This oxidative stress has been implicated in diseases such as diabetes, atherosclerosis and Parkinson's disease.⁷⁵ 

Free radical mechanisms of oxidation of carbohydrates are comparable to those of lipids. Simple carbohydrates such as glucose, mannitol and deoxyribose are oxidized by the hydroxyl radical but do not impact food quality.⁹⁸  Caramelization is an example of a beneficial oxidation reaction of simple carbohydrates. Also, α-dicarbonyl compounds, as products of carbohydrate oxidation, are important intermediates in Maillard reactions.

Minor components in foods and biological systems are also subject to oxidation. Enzymatic browning is an oxidation process of great commercial significance. When a fruit or vegetable is cut, the cells are ruptured releasing phenols and the enzyme phenolase, which catalyzes the oxidation of the phenols by aerial oxygen. This not only changes the color of the fruit to brown but can also potentially affect nutrient content due to the nature of the oxidation products as well as loss of potential nutritive substrates.

Plant pigments might be minor components of plant material but their color makes them highly visible and oxidative damage to pigments can impact significantly on food appearance and quality. Oxidation of vitamins is a common problem in foods and whilst it does not impact esthetic qualities of the food it is a major concern for loss of nutritional value. The presence of oxysterols in foods can be traced to oxidation reactions as a result of heat treatments, exposure to sunlight and contact with oxygen. Similarly, oxidized derivatives of cholesterol and phytosterols can be generated in humans via different oxidation processes.

This section would be incomplete without mentioning antioxidants. Antioxidants are, by definition, good substrates for oxidation. Indeed, they inhibit the oxidation of other biomolecules by sacrificial oxidation. Thus, antioxidants are highly susceptible to enzymatic and non-enzymatic oxidation, one of the requirements of being a potent antioxidant. As the antioxidants themselves show great diversity in chemical structure, the products of their oxidation can also be extremely diverse. As such, it is beyond the scope of this chapter to cover this topic in depth. Some examples that illustrate the chemical reactions of antioxidants as substrates include: tocopherols (vitamin E) in physiological¹⁰¹  and food systems;⁷²  and the role of ascorbic acid in wine.¹⁰² 

Continuous exposure to reactive oxidative species from endogenous and exogenous sources results in oxidative damage of many cell components and particularly, lipid, protein and DNA (Figure 1.5). Some of the changes can be used as markers of oxidative stress and to study putative protective effects of antioxidants.¹⁰³  Exposure to the reactive oxidant changes gene regulation patterns and protein synthesis. Oxidative damage can be assessed using a number of techniques, which can also be applied to monitor the effects of dietary antioxidants (Figure 1.6).

Lipid oxidation is commonly monitored in biological systems by markers such as thiobarbituric acid reactive substances (TBARS) and isoprostanes.¹⁰³  TBARS result from oxidation of polyunsaturated fatty acids, with malondialdehyde (MDA) being of particular interest due to its toxicity. Isoprostanes result from oxidation of arachidonic acid, with 8-prostaglandin F2α regarded as one of the most reliable markers of oxidation in vivo.

Protein oxidation is less commonly monitored than lipid oxidation, although protein oxidation is linked to several diseases.¹⁰⁴  Biomarkers associated with oxidative damage to proteins include protein-bound carbonyls and advanced oxidation protein products.¹⁰³  The latter are related to AGEs mentioned above.

Knasmüller et al.¹⁰³  describe a wide variety of methods (Figure 1.6) for determining oxidative damage to substrates including DNA: cell-based gene mutation assays; chromosome analyses in metaphase cells; micronuclei, which are formed as a consequence of chromosome breakage; and monitoring, usually in urine, of 8-oxo-7,8-dihydro-2-deoxyguanosine (8-oxodG). Using DNA oxidation markers, protective effects of lycopene and tomatoes,¹⁰⁵  cruciferous and leguminous sprouts,¹⁰⁶  and carotenoid supplementation,¹⁰⁷  have been observed.

In foods, the foci of oxidation markers are those associated with lipid oxidation. This is because oxidative rancidity is a common defect in many fatty foods, such as oils, fats, meats, and also foods such as grains. One of the earliest markers used to monitor lipid oxidation was MDA, the presence of which was correlated with defects in meat products.¹⁰⁸ 

A Google search of “antioxidant” on 5 January 2021 produced about 136 million entries. However, as observed by John Gutteridge and Barry Halliwell in 1999 “antioxidant is a term widely used but rarely defined” and this remains true today. Nevertheless, a large number, range and diversity of definitions have been used to explain what constitutes an antioxidant. In the first definition of an antioxidant, it was simply a chemical compound or substance that inhibited oxidation. A later definition as a substance that inhibited oxidative deterioration of gasoline, rubbers, plastics, soaps, etc., emphasized the role of antioxidants in protecting important commodities from deterioration. A broader definition was introduced in 1995 by Halliwell and Gutteridge¹⁰⁹  as

In other words, an antioxidant is a substance that slows the rate of an oxidation reaction.

Oxidation was later changed in this definition to oxidative damage: “any substance that delays, prevents or removes oxidative damage to a target molecule”.¹¹⁰  The inclusion of the word “prevents” is curious as it is generally acknowledged that an antioxidant delays, retards or inhibits oxidation, but cannot stop the process completely except in limited situations such as where reactants are fully consumed.

Other definitions have included:

• a substance which slows down the damage that can be caused to other substances by the effects of oxygen (given in the Collins Online Dictionary),

• a substance that when present at low concentrations compared with those of an oxidizable substrate significantly prevents or delays a pro-oxidant initiated oxidation of the substrate,¹¹¹ 

• molecules that experimentally prevent the oxidation of a biological or chemical system,¹¹² 

• an antioxidant molecule can be identified by its chemical structure that allows a free radical scavenging reaction and/or the chelation of redox-active metals,¹¹² 

• substances that delay or prevent the oxidation of cellular oxidizable substrates.¹¹³ 

In 2016, Apak et al.¹¹⁴  gave a more specific definition: “natural or synthetic substances that may prevent or delay oxidative cell damage caused by physiological oxidants having distinctly positive reduction potentials, covering reactive oxygen species (ROS)/reactive nitrogen species (RNS) and free radicals (i.e., unstable molecules or ions having unpaired electrons).” This definition has a very specific emphasis on the roles of antioxidants at cellular levels in humans as they are related to oxidative stress and free radicals.

It will be seen that the particular definition of an antioxidant typically reflects the interests of the individual with a general shift over time to a meaning that emphasizes health and in vivo situations. It is the general shift to a definition oriented to a specific sub-discipline or industry that was at the base of the statement by Frankel and German¹¹⁵  that “the term has found a home in the food and botanical industries and has assumed such a broad meaning as to be almost unusable”. Ironically, the shift now is to a very specific meaning involving cell damage and physiological oxidants and this shift also renders the term “almost unusable” except for a limited range of applications in a specific area to the exclusion of others. This is symptomatic of the fragmentation of antioxidant studies into food science and health science (with areas such as rubber technology being largely ignored). This situation is not desirable as both these and other areas are inter-related and can learn from each other.

Ideally, the definition of an antioxidant should be sufficiently broad in scope to allow all work in this area to be embraced. The essential features of a definition should include the retardation (by reduction in the rate of reaction) or delay (via extension of the induction period, which is effectively a change in the rate of reaction) of oxidation, the notion of a substrate and oxidant and possibly some indication of the mechanism of action. It is tempting to introduce yet another definition that encompasses these ideals. However, this would be counter-productive, as the definition given by Halliwell and Gutteridge in 1995 contains many of these features and crucially it covers all oxidation processes, both radical and non-radical ones. Various criticisms of the 1995 definition have been voiced¹¹⁶  – e.g., “a generic definition of an antioxidant is not experimentally constructive unless it is associated with the notion of the oxidant that has to be neutralized”.¹¹⁷  Tirzitis and Bartosz¹¹⁶  also raise the issue: “the validity of the term ‘antioxidant’ depends on the environment of its action, viz. whether we consider an in vitro or in vivo action.” These authors go on to assert “Outside this context, a statement that some compound is an antioxidant may not bring any biologically meaningful information.” As we have discussed above, there are pros and cons with more or less specific definitions of antioxidants. We contend that a broad definition is overall preferable, and that criticisms, as presented here, may be appropriately addressed by specifying the relevant detail (e.g., oxidant, antioxidant, temperature, in vivo, in vitro, ex vivo, etc.) in antioxidant activity measurements.

A redox reaction in the current context involves the reaction of a substrate with an oxidant to produce oxidized products as in eqn (1.12), which has been simplified to the point of having zero chemical content but provides a visual tool for picturing the role of reactants, products and other factors in the oxidation process.

The action of antioxidants in slowing down the rate of an oxidation (i.e., slowing the rate of eqn (1.12)) involves a combination of effects on the reactants (notably the oxidant) or oxidation promoters as follows:

• scavenging free radicals (oxidant in eqn (1.12)),

• chelating pro-oxidant transition metal ions,

• scavenging non-radical oxidants such as singlet oxygen and photosensitizers,and, in the case of lipid substrates

• inactivating lipoxygenase.⁷² 

These actions are the classical mechanisms by which antioxidant action was considered and are relevant to the oxidation of all biomolecules whether lipid, protein, DNA or a minor component and regardless of the system (food and beverage or biological system).

In the case of free radical scavengers (first point above), antioxidants deactivate the deleterious action of the reactive radicals by three mechanisms: hydrogen atom transfer (HAT), single electron transfer (SET) and mixed mode. In HAT, a H˙ (hydrogen atom) is donated from the antioxidant (AH) to the free radical (see eqn (1.13)–(1.15), below). The result is A˙, which is a more stable radical and hence the system is stabilized to further reactions. In a SET reaction, the electron is transferred from the antioxidant to the radical and this results in a protonated antioxidant radical (AH˙⁺). Depending on the conditions, the proton may be transferred to a base (e.g., H₂O), and A˙ is again the final product. In a mixed mode reaction, the rates of electron transfer and hydrogen atom transfer become comparable and hence it is not possible to ascribe the antioxidant action to either mechanism. Schaich et al.¹¹⁸  described the kinetics of mixed mode reactions of 2,2-diphenyl-1-picrylhydrazyl (DPPH˙).

Iron and copper are the most common redox-active metal ions in food and biological systems. Chelating agents such as ethylenediaminetetraacetic acid (EDTA) and citric acid are commonly used in foods to protect against oxidation induced by metal ions. Other chelating agents include phenolic compounds, lignans, and amino acids such as carnosine and histidine.⁷²  In physiological systems, metal ions are usually under homeostatic control, with endogenous chelators such as transferritin (for iron chelation). tpDiseases that lead to metal ion overload can also lead to oxidative damage and treatment with chelation agents such as deferoxamine. Recently, phenolic compounds have also been suggested as chelating agents for metal ions as treatments for diseases.¹¹⁹ 

Singlet oxygen is the high-energy form of molecular oxygen and can be quenched by energy transfer to another molecule. Choe and Min⁷²  reported on the ability of carotenoids to absorb energy from singlet oxygen, especially those with nine or more conjugated double bonds. Quenching of singlet oxygen can also occur via chemical reaction with antioxidants such as tocopherols, phenolic compounds, carotenoids, curcumin, urate, and ascorbate. Photosensitizers such as chlorophyll and riboflavin are important in both food and physiological systems.¹²⁰  One of the primary goals of refining vegetable oils is to remove chlorophyll to reduce photo-oxidation. Tocopherols and phenolic compounds can deactivate photo-excited riboflavin, but carotenoids are ineffective. Photosensitization of riboflavin may have implications for skin and eyes since these organs are exposed to light.

The in vivo action of antioxidants involves mechanisms additional to those listed above that may include

• inhibiting the generation of oxidants such as suppression of ROS and RNS,¹²¹ 

• up-regulation of endogenous antioxidant systems,¹²²  and

• mobilizing compounds that are able to stimulate the production of endogenous antioxidant compounds.¹²³ 

The in vivo cellular effects of antioxidants may also be mediated by their interactions with specific proteins central to intracellular signaling cascades, their modulation of the expression and activity of key proteins, modulation of gene expression and redox cell signaling, their influencing of epigenetic mechanisms or their modulation of the gut microbiota. The mechanism by which such actions occur remains the subject of active investigation.^74,124,125  The action of antioxidant enzymes which are important in many biological systems has been described as “sophisticated molecular mechanisms of redox biology and metabolic homeostasis”.¹²⁶ 

The progress of the reaction in eqn (1.12) can be monitored by measuring the decline in concentration of one or more of the reactants (i.e., substrate or oxidant) or the increase in concentration of one or more products during the course of reaction. This will produce a plot of either substrate concentration, oxidant concentration or concentration of one or more products versus time. From this plot, the slope of the tangent to the curve at time = zero gives the initial reaction rate. Changing the concentration of one reactant at a time and measuring initial rates can be used to obtain a rate law – a mathematical expression that provides information on which reactants control the rate of reaction, at least in the initial stages. Once the controlling species are known, a mechanism can be postulated to explain the rate law. It is important to remember that the experimental data are “facts”, while the mechanism is a theory used to explain the data. The mechanisms shown in this book have been obtained by interpreting experimental rate law data.

The effect of an antioxidant is evidenced by a decrease in this reaction rate and/or by extension of the induction period. These concepts are illustrated in Figure 1.7 where the time course of a typical redox reaction is illustrated for the autoxidation of LDL as monitored by the production of LOOH. The effect of increasing the concentration of antioxidant tocopherol on the production of LOOH is clearly visible. The important observations are: the increase in induction period (the relatively flat part of the curve, where no/very little LOOH is forming) as tocopherol concentration increases, and the fact that after the induction period, the [LOOH]–time curve has virtually the same slope. These data can be interpreted to explain that increasing the concentration of tocopherol delays the formation of LOOH in a concentration-dependent manner. Also, that when the tocopherol is depleted, the formation of LOOH occurs at the same rate, regardless of what the initial concentration of tocopherol was. This suggests that the product(s) of tocopherol oxidation (while it was acting as an antioxidant) do not play any further part in the formation of LOOH.

Although Figure 1.7 shows the course of reaction as monitored by the production of LOOH, and the effect of antioxidant concentration, a more useful figure for some situations is shown in Figure 1.8. Curve (1) shows the oxidation reaction occurring with a rate constant, k₁, in the absence of an antioxidant. Curve (2) shows the effect of added antioxidant, which produces an induction period (t_i). Curve (3) shows the effect of an antioxidant that behaves differently. Here there is no induction period and oxidation begins immediately, but with a rate constant k₃, which is less than k₁. After the reaction has progressed, curve (3) crosses curve (2) and at the end of the experiment there is overall less oxidative damage in reaction 3: ([LOOH]₃ < [LOOH]₂). Therefore, although the antioxidant used in experiment (3) did not produce an induction period, in the long run there was less oxidation than in (2). Does this mean that the antioxidant in (3) is better than in (2)? The answer to this question depends on when oxidation becomes problematic. If A₅₀₀² < 1.0 is a quality parameter (say in vegetable oils), then (3) has kept the absorbance below this value for longer than (2). If, on the other hand, A₅₀₀ < 0.3 is the quality parameter, then (2) is the better antioxidant.

How fast the reaction described by eqn (1.12) proceeds is an important consideration. Kinetics is the area of chemistry that studies reaction rates and mechanisms, but this is not the only consideration. It is also important to understand whether a particular oxidation reaction will occur spontaneously and this is the realm of thermodynamics, which can be approached from the standard free energy change, ΔG⁰ and standard reduction potentials, E⁰. A reaction that is predicted by thermodynamics to occur spontaneously may proceed at such a slow rate that it is considered to be under kinetic control.¹²⁸ 

Acworth⁴²  tabulated an extensive range of biologically relevant thermodynamic (E⁰) and kinetic values. These can be used to understand whether antioxidant activity is kinetically or thermodynamically driven. For example, α-tocopherol is not thermodynamically a strong antioxidant with a moderately positive E⁰ value of +0.500 V. Ascorbate (E⁰ = −0.174 V) and glutathione disulfide (E⁰ = −1.500 V) are stronger antioxidants than α-tocopherol. However, in a reaction with O₂˙⁻, α-tocopherol reacts faster (2.5 × 10⁶ M⁻¹ s⁻¹) than ascorbate (1.0 × 10⁴–2.7 × 10⁵ M⁻¹ s⁻¹).⁴²  Here, the kinetics suggest that α-tocopherol will be the better antioxidant.

For phenolic antioxidants, and flavonoids in particular, there have been various attempts to derive structure–activity relationships that relate the number and position of phenol groups to the capability of the antioxidant to inhibit oxidative damage. Typically these studies are carried out in vitro. Apak et al.¹²⁹  list various criteria for high antioxidant activity, including the presence of an ortho-diphenol group on the B-ring; the presence of electron-donating groups such as methoxy groups in ortho- or para-positions to the phenolic OH; and conjugation between the various rings. Similar criteria were found in a study by Chen et al.¹³⁰  on the antioxidant activity of phenolic acids. This study also investigated computationally derived thermodynamic parameters, bond-dissociation enthalpies (BDE), ionization potentials (IP) and electron transfer enthalpies (ETE) of the phenolic OH groups. Computed data were compared with experimentally derived antioxidant activity measurements based on DPPH and ferric ion reducing antioxidant power (FRAP) assays. In general there was a good correlation between the assay and computational data.

Studies such as those by Chen et al.¹³⁰  link the thermodynamics of a system with kinetics. For example, a weaker O–H bond (lower BDE) is likely to be easily broken, resulting in a faster antioxidant reaction, especially where hydrogen atom transfer is being measured by the assay. On the other hand, Zhong et al.¹³¹  found that tyrosol was a more effective antioxidant than caffeic acid in an accelerated oxidation study of camelia oil. Tyrosol has neither an ortho-diphenol group nor any other feature that would predict better activity than caffeic acid, which has both an ortho-diphenol group and extra conjugation extending beyond the phenolic ring. Results such as these highlight the possible differences that may be found between studies with simple assays and computation, and those where the antioxidant is in a “real” system where its function in limiting oxidation of a substrate can be evaluated.

An important distinction can be made between short-term and long-term antioxidant protection. This is related to the reaction kinetics and the rate at which an antioxidant reacts with a specific radical versus the thermodynamics of the reaction and how completely the antioxidant reacts. For instance, disappearance of the DPPH radical followed a double-exponential equation in the presence of edible oils and oil fractions⁶⁷  which suggested the presence of a fast- and slow-acting group of antioxidants. In lipid systems, differences in antioxidant activity can be observed whether they are fast-acting chain-breaking antioxidants, which get consumed, and induce a lag phase (Figure 1.8), or whether they are slower acting, and perhaps not consumed. In the latter case, the antioxidant activity will be seen throughout the oxidation reaction (reaction 3, Figure 1.8).

Antioxidants include small molecules and complex systems that delay, inhibit or retard oxidation. They have traditionally been separated into primary antioxidants as those substances that can donate a hydrogen atom or an electron to a radical, and secondary antioxidants as those that function indirectly. However, there are subtle differences in the use of these terms in different areas of science (e.g., food science and polymer science^132,133 ) and the functionally descriptive terms of:

• chain-breaking (or, less commonly, scavenging) antioxidants for primary antioxidants, and

• preventative antioxidants instead of secondary antioxidants

are more useful (Figure 1.9). With reference to eqn (1.12), chain-breaking antioxidants react with the oxidant, where the oxidant is a free radical and thus breaks the oxidation chain, whereas preventative antioxidants react with an oxidation promoter such as metal ions to reduce the rate at which new radical chains are started.² 

Chain breaking antioxidants include small-molecule antioxidants, both hydrophilic such as butylated hydroxyanisole (BHA), polyphenols, vitamin C or glutathione, and hydrophobic such as carotene, lipoic acid, and co-enzyme Q₁₀ plus enzymatic antioxidants such as SOD and catalase. Chain-breaking antioxidants, AH, when present in trace amounts, either delay or inhibit the free-radical initiation step of the oxidation by reacting with a substrate radical, here illustrated with a lipid radical, L˙, or inhibit the free-radical propagation step by reacting with peroxyl or alkoxyl radicals:

The antioxidant free radical may further interfere with chain-propagation reactions by forming peroxyl antioxidant compounds (as noted in Section 1.2.2.5):

The activation energy of the above reactions increases with increasing A–H and L–H bond dissociation energy. Therefore, the efficiency of the antioxidant should increase with decreasing A–H bond strength, as noted above.

Secondary or preventative antioxidants retard the rate of oxidation by different mechanisms such as “removal” of the substrate, by oxygen scavenging and quenching of singlet oxygen or by chelation of pro-oxidant transition metal ions such as copper(ii) or iron(iii) preventing decomposition of lipid hydroperoxides to reactive radicals. Some secondary antioxidants are able to regenerate primary antioxidants in a synergistic manner. Compounds able to decompose peroxides/hydroperoxides or compounds able to isolate (e.g., by chelation) transition metal ions, protecting the system against initiation via a Fenton-type reaction are included in this class. (The Fenton reaction is Fe²⁺ + H₂O₂ → Fe³⁺ + OH⁻ + HO˙)

Antioxidants can also be distinguished by function as:

• Free radical scavengers e.g., ascorbic acid,

• Scavengers of non-radical oxidants e.g., catalase,

• Compounds that inhibit generation of oxidants e.g., metal chelators,

• Compounds that induce the production of antioxidants e.g., isothiocyanates.

Antioxidants exhibit a range of properties and structures that may serve as further means of classification (Figure 1.10). Antioxidants may be naturally occurring¹³⁴  or they may be produced synthetically as in the case of BHA, butylated hydroxytoluene (BHT), tert-butylhydroquinone (TBHQ) or the gallates (Figure 1.10). These synthetic antioxidants have been widely used over many years in the food industry as well as in other industries, such as cosmetics, pharmaceuticals or animal nutrition. Naturally occurring antioxidants have experienced a sustained growth over the last few years due to current trends in society with an emphasis on natural and greater expectations for food safety legislation. We prefer the term naturally occurring to natural antioxidant as the latter often infers a distinction between what occurs naturally and the identical material produced synthetically. In this connection, it is interesting that the U.S. National Library of Medicine (PubChem) states that the synthetic antioxidant, propyl gallate occurs naturally in corn seeds¹³⁵  although this is not referenced.

The increase in use of naturally occurring antioxidants has been to the detriment of synthetic antioxidants⁶⁶  as food laws of most countries further restrict their use due to concerns regarding their safety. This has promoted the use of naturally occurring antioxidants such as tocopherols, ascorbic acid, rosemary extract and plant polyphenols possessing antioxidant activity similar to or even higher than that of synthetic antioxidants. Food regulations are extremely complex and differ between jurisdictions and this can cause problems, for example, when an antioxidant is permitted by one country, but not another.

The shift to naturally occurring antioxidants should not be taken to imply that their use is without problems. Indeed, there are a number of issues associated with the use of both synthetic and naturally occurring antioxidants.⁶⁶  The carcinogenic and mutagenic effects of synthetic antioxidants have been examined extensively but many naturally occurring food compounds have not yet been tested. The fact that a substance is naturally occurring is no guarantee of safety; after all, the Greek Philosopher Socrates was dispatched to a more propitious environment with naturally occurring hemlock. Testing the “toxicity” is not a simple issue as both the potential antioxidant and its oxidized form(s) and reaction products with food constituents must be assessed. This makes no allowance for the changes that the antioxidant may undergo during food processing which further increases the scope of testing.

One of the big challenges of recent years has been finding the elusive naturally occurring super-antioxidant. Of course, once naturally occurring antioxidants have been identified then synthetic manufacture can be used for their production in larger quantities. Thus, the distinction between synthetic and natural is purely artificial, but is nevertheless a useful one. This raises an important issue and is illustrated in the case of tocopherols (Figure 1.11) where the naturally occurring materials are a complex mixture of 8 isomers (4 isomeric tocopherols and 4 isomeric tocotrienols) whereas α-tocopherol alone is used almost exclusively in vitamin E supplements and in almost all research studies on vitamin E.

Antioxidants can be distinguished on the basis of solubility as either aqueous phase (water-soluble or hydrophilic) or lipid phase (hydrophobic). This classification can be useful in guiding applications for a particular antioxidant. For example, addition of a water-soluble antioxidant to a vegetable oil would probably not be optimal for activity (notwithstanding the polar paradox theory, see below). It is also of fundamental importance pharmacologically because water-soluble antioxidants react with oxidants in the cell cytoplasm and in blood plasma, while hydrophobic antioxidants protect cell membranes from lipid oxidation. The distinction also has significance when considering consumption of lipid-soluble antioxidants such as eugenol, a major component of oil of cloves, which have toxicity limits that can be exceeded with the misuse of undiluted essential oils. Toxicity associated with high doses of water-soluble antioxidants such as ascorbic acid is less of a concern, as these compounds can be excreted rapidly in urine.¹³⁶ 

The link between antioxidant solubility and activity has been a subject of much research and conjecture. The polar paradox theory was advanced to explain findings that “polar antioxidants are more active in bulk lipids than their nonpolar counterparts, whereas nonpolar antioxidants are more effective in oil-in-water emulsion than their polar homologs.”¹³⁷  In 2015 Laguerre et al.¹³⁷  in a critique of the polar paradox theory proposed that the theory was predicated on the assumption that the critical site of lipid oxidation was the air–oil interface and reviewed evidence that contrary to this, the major site of lipid oxidation was at the oil–water interface in association with colloids within the bulk oil. Thus water-soluble antioxidants would be more likely to partition into these colloids and hence protect the oil from oxidation. In fact, the authors go further to state that when it comes to predicting antioxidant behavior in bulk oil, it is not possible to predict activity based on considering hydro- or lipophilicity, or is it possible to predict activity in emulsions. Later work by Laguerre et al.¹³⁸  has gone on to propose that a better understanding of the effectiveness of antioxidants in lipid systems will come from analyzing data “in a coherent spatiotemporal framework” and have coined the term “colloidal ecosystem” as a means of helping researchers view oxidation and antioxidant processes systemically.

Antioxidants may also be distinguished by physical location as intracellular, extracellular or membrane-located antioxidants or, in other cases, it may be appropriate to consider the activity of the antioxidant and to distinguish between enzymatic and non-enzymatic antioxidants.¹³⁹  Indeed, the main classification used pharmaceutically is to distinguish enzymatic from non-enzymatic (minerals, vitamins, carotenoids, etc.) antioxidants.¹³⁴ 

Antioxidants can also be differentiated on molecular size as small-molecule antioxidants and large-molecule antioxidants. The small-molecule antioxidants include species such as vitamin C, uric acid, glutathione, carotenoids, flavonoids, and most of the substances traditionally regarded as antioxidants. It is likely that novel small-molecule antioxidants will continue to be identified in the years ahead.¹⁴⁰  The large-molecule antioxidants are enzymes such as SOD, catalase and sacrificial proteins (albumin) that interfere with oxidant attack on other essential proteins and some hormones such as angiotensin. There has been considerable activity in recent years in identifying the peptide residues responsible for antioxidant activity in proteins.¹⁴¹ 

Based on their origin within or outside the particular system, antioxidants can be categorized as either endogenous or exogenous. Some examples of antioxidants that are endogenous or exogenous to humans are presented in Figure 1.10 but note that an antioxidant that is endogenous to one system may be exogenous to another. For example, flavonoids are endogenous in fruits but exogenous in animals. This illustrates an important point – context is always important and must be established when dealing with antioxidants. It would be easy to read information on flavonoids stating that they are endogenous antioxidants and to erroneously assume that this applies to both plant and animal samples.

Each of these systems of classifying and differentiating antioxidants is a useful way of simplifying and hence improving our understanding of these substances and their behavior. Each approach serves a purpose in that it is more useful than other systems in a particular situation. The different approaches are obviously NOT mutually exclusive (e.g., antioxidants distinguished by molecular size as large are almost exclusively enzymatic antioxidants), and finally there is no right or wrong system for classifying antioxidants.

Various terms are used to characterize antioxidant behavior. Antioxidant activity is the most widely used of these different terms and hence is included in the title of the current book, but what is the precise meaning of this phrase? Unfortunately, there is no universally accepted definition of antioxidant activity (or any of the other terms) that is used consistently throughout the literature. This situation is regrettable but understandable in the context of the diverse academic fields in which oxidation and antioxidants are important and the even more diverse interests of the people engaged in those fields.

The history of oxidation and antioxidants provides a background to understanding the terminology associated with their measurement. The deterioration of many useful materials (e.g., metal tools from the Iron Age) in the atmosphere was recognized long before any notion of oxidative reactions was developed. Much later, some of the earliest scientific investigations of oxidation were performed on natural rubber, the first technologically important polymer. The early literature uses many terms such as aging, fatigue, and poisoning to describe the loss of useful properties of rubbers caused by oxidation. Similarly, the use of antioxidants extends into prehistory, but the first recorded scientific observation on anti-oxygenic action (oxidation inhibitors) was made in 1797 by Berthollet (see ref. 142) leading to the theory of “catalyst poisoning” in oxidative reactions. The anti-oxygenic action of a number of substances in various pure compounds (e.g., chloroform, sodium sulfite, etc.) was subsequently reported during the nineteenth century (see ref. 142). The discovery of protective agents for rubber was entirely empirical, and a patent for the addition of phenol and p-cresol was granted in 1870 and for pyrogallol and hydroquinone in 1901. Systematic kinetic studies of rubbers in the 1950s¹⁴³  paved the way for an understanding of the antioxidant action of amine and phenolic substances.

The course of rancidification of fats was unknown until the participation of atmospheric oxygen as the main factor in the oxidation of free fatty acids was demonstrated in 1886 by Duclaux. Meanwhile, Deschamps demonstrated in 1843¹⁴⁴  that an ointment prepared from fresh lard by addition of gum benzoin or populin (from poplar buds) did not go rancid as did ointment without the addition. For many decades, the stabilizing effect of benzoin was commonly ascribed to the antiseptic action of its constituents, until Husa and Husa in 1926 called attention to the inadequacy of this explanation and indicated that the benzoin probably functions as a negative catalyst of oxidation.^3,145  The presence of the common antioxidants, vanillin and saligenin, in benzoin and poplar buds, respectively is now well documented.

The earliest reports on antioxidants employed for food lipids originate from the work of Wright in 1852¹⁴⁶  who reported use of natural sources such as elm bark for preservation of butterfat and lard. The classic studies of Moureau and Dufraisse in the 1920s¹⁴²  introduced modern notions of antioxidants and also reported the use of synthetic compounds, principally phenolic compounds, to retard the oxidative deterioration of food lipids. They referred to the action of antioxidants as “inverse catalysis”. The term “antioxygen” was used to describe compounds which were believed to act catalytically by retarding the oxidation reaction by undergoing oxidation themselves.^142,147  They proposed that the primary role of an antioxygen was to convert substrate peroxides (at that time of indeterminate structure, but assumed to involve some association of the substrate with oxygen) to inert products. An antiperoxide theory emerged in 1926 which proposed that the antioxygen inhibitors acted by destroying active peroxides.¹⁴²  The negative catalytic mechanism of autoxidation of Moureu and Dufraisse and the chain reaction theory of negative catalysis^148,149  were superseded in 1934¹⁵⁰  by the concept of chain reactions involving free radicals.

We have not positively identified the first use of the term “antioxidant” but there are scattered uses in the 1920s (e.g., ref. 151) in relation to the use of antioxidants in controlling the aging of rubber. The terminology was firmly established prior to 1930 as Dr W.J.S. Naunton was speaking on the topic “Antioxidants” in December of that year¹⁵²  at a meeting of the Society of Chemical Industry (Yorkshire Section) to be held at the Great Northern Hotel, Leeds. Curiously, anti-reductant was also used as early as 1929¹⁵³  to describe the anomalous redox behavior of an oxazine dye with a low standard reduction potential. It is only in recent times that anti-reductants have been attracting increased attention¹⁵⁴  from the scientific community.

Interestingly, the earliest papers in this area do not use the terms “antioxidant activity” but rather discuss antioxidant power in terms of the protection afforded by the antioxidant, the effectiveness of the antioxidant or the amount of change in an observed property such as prolongation of induction period for the oxidation reaction.¹⁵⁵  The earliest uses of the term “antioxidant activity” appear to date from the 1930s¹⁵⁶  with sporadic use in the 1940s, sometimes interchangeably with antioxygenic activity,¹⁵⁷  and followed thereafter by a rapid increase in frequency of use.

Antioxidant activity as a concept presumably originated with chemists and chemical technologists but has been adopted and then adapted by food scientists, biologists, nutritionists, physiologists and many other disciplines working with antioxidants. The original meaning attached to the term (if indeed one ever existed) was lost or became clouded in this transition and a plethora of terms developed to express basically the same idea (i.e., inhibition of oxidation) including antioxidant capacity, antioxidant power and antioxidant potential, with antioxidant status, antioxidant performance and preventative antioxidant capacity¹⁵⁸  being used much more infrequently. Of the various terms, antioxidant activity and antioxidant capacity are the most commonly used. Indeed, a Google search for these two terms in July 2020 returned 7.8 million hits and 3.4 million hits, respectively. Similar searches on Web of Science returned fewer hits but typically in a similar ratio of about two or three to one.

One might expect there to exist clear definitions and an understanding of terms that are so widely used, as Socrates is reported to have said, the “precise logical definitions of concepts are a fundamental prerequisite to true knowledge”. It is doubtful that an individual can develop a full comprehension of the science of antioxidants without the appropriate language. This is one reason why we value language skills so highly. As in any area, this language (definitions, etc.) must evolve and develop with progress in the field. Nevertheless, despite the frequent use of the terms, antioxidant activity and antioxidant capacity are rarely defined. The proliferation of undefined terms to describe antioxidant behavior is counter-productive and should be discouraged. The lack of clear and consistent definitions of these terms raises questions such as what is meant when researchers measure, quote values and discuss antioxidant activity or capacity. Indeed, the failure of antioxidant clinical trials, ambiguity of conclusions about the effect of a chosen therapy in medicine and evaluation of food quality, diet, etc. have been attributed¹⁵⁹  to unreliable terminology and incompatible units of antioxidant activity/concentration expression.

Of the two main terms, it is antioxidant capacity that, although less used, is historically better documented. An early use of the term in a paper that examines the total effective concentration of chain stopping antioxidants in new and used lubricants¹⁶⁰  specifies that antioxidant capacity is the number of peroxy radicals which can be consumed by the antioxidant species in a (lubricant) sample. This concept is explored further in a review of antioxidant methods for determining chain-breaking antioxidant activity in food¹⁶¹  but the notion of this definition appears to have been largely lost in the intervening period and antioxidant capacity is now often used interchangeably with antioxidant activity although, on other occasions, with a different meaning.¹⁶² 

The term antioxidant activity, although not specifically defined in early literature, can be related to the degree to which some measurable phenomenon, associated with oxidation, is inhibited. In the earliest paper we can trace,¹⁵⁶  the phenomenon was tensile strength in rubber, which was improved through the addition of an oxidation inhibitor during vulcanization. The importance of kinetics is established in these early papers as, for example, in the autoxidation of aldehydes which exhibit an inverse proportionality between inhibitor concentration and reaction rate.¹⁴⁹  Studies from the 1950s are also notable for the prominence of kinetics and the association of antioxidant activity and rate of reaction.^163,164  A common theme in many publications from the 1950s was the correlation of antioxidant activity with chemical structure. Most of the studies from this period examined the oxidative deterioration of petrol and lubricating oils during storage with a smaller number concerned with deterioration of foodstuffs and rubbers.

In more recent times, various interpretations have been applied to antioxidant activity and antioxidant capacity. In a review paper in 2002,⁶⁷  we used antioxidant activity in a broad sense as a measure of the ability of an antioxidant to inhibit or delay oxidation although we did point out the various testing endpoints and ways of reporting activity. Antioxidant capacity was not defined except as “the sum of all antioxidant activities of a mixture containing many antioxidants, e.g., serum.” Similarly, neither antioxidant activity nor capacity are defined in Frankel and Meyer's highly cited critique of antioxidant assays.¹⁶⁵  Antioxidant capacity is not used at all by Halliwell et al.¹⁶⁶  who refer exclusively to antioxidant activity. Apak et al.¹¹⁴  avoid the capacity/activity argument by using “antioxidant activity/capacity (AOA/TAC)” although they do provide a definition of each term and state that “care must be exercised to distinguish between these two terms”, but never specifically make any further distinctions.

Others (e.g., ref. 41 and 158) hold a very different view, and MacDonald-Wicks et al.¹⁵⁸  suggested differentiating activity and capacity as:

Thus, antioxidant capacity according to this view is effectively the amount of substrate or oxidant that can be consumed by a fixed amount of antioxidant reflecting reaction stoichiometry.¹⁶¹  The definitions provided by MacDonald-Wicks et al.¹⁵⁸  are not entirely consistent with definitions from Prior's group¹²⁷  who defined antioxidant activity and antioxidant capacity thus:

We disagree with their conclusion that the other terms “are more independent of specific reactions”. Indeed, it is our position and that of others (e.g., ref. 116 and 167) that all measures of antioxidant action are highly dependent on conditions.

To introduce further confusion, TEAC “Trolox equivalent antioxidant capacity” was defined in 1993 by Rice-Evans’ group¹⁶⁸  but then referred to in later publications^169,170  as an antioxidant activity test. Some other groups (e.g., ref. 171 refer to TEAC as an antioxidant capacity test) but Cano's group, at least on occasions, describe it as an antioxidant activity assay.^172–174 

A comparison of the use of the various terms used to describe antioxidant behavior by year for the period 2000 to 2018 is available (Figure 1.12).¹⁵⁹  Of particular interest is the frequency with which the term antioxidant activity is used without associated kinetics data.¹⁶¹ 

You are probably now more confused than ever but hopefully the confusion is on a higher knowledge plane. Are the terms used to describe antioxidant behavior, and particularly antioxidant activity and antioxidant capacity, equivalent or different? How are they defined in practice?

In theory, it is our view that antioxidant activity and antioxidant capacity are distinct and different concepts that should be distinguished. Antioxidants are functional substances that inhibit oxidation; therefore the activity that should be measured is a property related to the ability to inhibit oxidation. Using an analogy to the measurement of enzymatic (or hormone or drug, etc.) activity, in which the ability to accelerate a reaction is measured, then the expression of antioxidant activity should also embrace a kinetic component and be expressed as a rate constant. This analogy must not be taken too far as there are obvious and important distinctions (e.g., antioxidants with the exception of enzymatic antioxidants are not catalysts). On the other hand, antioxidant capacity refers to an amount, specifically, the total amount of substrate or oxidant (preferably expressed in moles) scavenged per mole of antioxidant(s) contained in the sample (i.e., the stoichiometry of the reaction of the antioxidant). Apak et al.¹⁷⁵  in a separate paper to that referenced above, state that “the terms ‘antioxidant activity’ and ‘antioxidant capacity’ have different meanings: antioxidant activity deals with the kinetics of a reaction between an antioxidant and the prooxidant or radical it reduces or scavenges, whereas antioxidant capacity measures the thermodynamic conversion efficiency of an oxidant probe upon reaction with an antioxidant.” Pinchuk et al.¹⁷⁶  clearly support this notion (as do the Editors) and have studied the kinetics of lipid oxidation¹⁷⁶  in the belief that a meaningful comparison of antioxidants requires clear definitions of the experimentally observable factors derived from kinetic data.¹⁶⁷ 

Whilst it is appropriate, in theory, for antioxidant activity to be expressed as a rate constant as proposed by MacDonald-Wicks et al.¹⁵⁸  (and supported as above), many standard antioxidant activity assays in current use – e.g., TEAC, DPPH, FRAP, etc. – do not yield a rate constant (as obvious from our comments). However, the majority of these tests do involve a time element in the measurement, and the results include an inherent kinetic aspect. Apak et al.¹⁷⁷  refer to “fixed-time” tests whereby “a reproducible conversion efficiency of the measurement probe is achieved within a predetermined time.” Thus the measurements are therefore at least related to a reaction rate¹⁶⁷  insofar as conversion efficiency per unit time could be thought of as comparable to a reaction rate.

In practice, the use of both terms in an undefined colloquial sense is so firmly entrenched that change is unlikely. However, an awareness of this situation is desirable as is an understanding of its potential implications. Terms such as antioxidant activity (and capacity) are used with a very loose connotation that they say something about the ability of the specified compound(s) to act as an antioxidant. In most practical situations, this amounts to equating activity with the result that was obtained using the particular test method. Whatever term is used to describe assays and results (activity, capacity or some other descriptive term), they all need to be evaluated critically against the question – what is really being measured here and what does it mean. Careful attention must be given to the method of expressing results (Section 3.3.4) and it should be understood that the measured activity is relevant only to the specific conditions involving the particular substrate and oxidant.

The data from antioxidant studies are loosely applied to identifying activity as measured by one of the many antioxidant methods in terms of “method-specific” data. This is illustrated in a paper titled “Total Antioxidant Capacity of Plant Foods, Beverages and Oils Consumed in Italy Assessed by Three Different In Vitro Assays”.¹⁷⁸  The analytical data in this study were presented as TEAC units (mmol Trolox kg⁻¹), FRAP units (mmol Fe²⁺ kg⁻¹) and TRAP units (total radical trapping antioxidant power, mmol Trolox kg⁻¹) and whilst the title and the discussion generally refer to antioxidant capacity, there are occasional references to antioxidant activity. In clear contrast, the summary states that “the total antioxidant activities of 64 foods, 34 beverages and 6 oils were measured by three different methods” but these authors identify that the TRAP and FRAP assays evaluate different types of activity (chain-breaking antioxidant potential and reducing power of the sample, respectively).

The paper by Pellegrini et al.¹llustrates a further point as it introduces an additional widely used concept, that of total antioxidant capacity (TAC).^{111,179–182}  The main motivation for using the term “total antioxidant capacity” appears to be to differentiate between measuring activity/capacity of one compound at a time versus measuring a composite response. Total antioxidant capacity originated from chemistry in the 1970s in relation to use of an initiator to yield TAC information¹⁸³  on transmission fluids¹⁸⁴  and used engine oils¹⁸⁵  representing a summation of the antioxidant capacities of all the various antioxidant species present. The concept then was applied to biology and medicine, and further to nutrition and epidemiology. However, no single assay can truly measure total antioxidant capacity, and the problems associated with such measurements have been discussed.^167,186,187  The majority of TAC assays generally measure the low molecular weight, chain-breaking antioxidants, excluding the contribution of antioxidant enzymes and metal-binding proteins. Thus, in physiological systems, where antioxidant enzymes and metal-binding proteins may be significant, then TAC assays do not measure all the antioxidant activity/capacity in a sample. However, in foods and beverages (and engine oils), it is possible that the majority of antioxidants may be measurable in a TAC assay, so the term may have more validity in those situations. In an appraisal of the TAC concept,¹⁸⁶  it was concluded that neither the term “total” nor the term “capacity” is applicable to in vitro assays using an arbitrarily selected oxidant generator and assaying a sample removed from its biological context, which is characterized by enzymatic maintenance of steady state.

Antiradical activity and (free) radical scavenging activity are also used^188,189  to describe antioxidant behavior. These terms are distinct and different from both the antioxidant capacity and antioxidant activity,¹¹⁶  but similar to the latter in that they represent an activity. The antiradical activity characterizes the ability of an antioxidant to inhibit free radical-induced oxidation whereas antioxidant activity represents the ability of an antioxidant to inhibit oxidation due to reaction with any type of oxidant (free radical or non-radical). Measurement of antiradical activity is confined to those tests that involve a free radical mechanism whereas measurement of antioxidant activity should involve tests in which the reaction ideally encompasses all mechanisms (e.g., free radical quenching plus non-radical processes, metal ion chelation, etc.). It follows that all antioxidant test systems using a stable free radical (for example, DPPH˙, or the radical cation of 2,2′-azino-bis(3-ethylbenzothiazoline-6-sulfonic acid (ABTS˙⁺), etc.) give information on the (free) radical scavenging activity or antiradical activity and, in many cases, this activity does not correspond to the antioxidant activity.¹⁶¹  Indeed, the only situation in which antiradical activity and antioxidant activity correspond occurs when the sole oxidation mechanism involves free radicals. This raises the interesting question of whether the measured antioxidant activity ever represents the true or total antioxidant activity. We are not aware of any antioxidant assay that measures the inhibition of oxidation by every possible mechanism, in which case, the measured activity only ever represents a portion of the total activity although, with well-chosen methods matched to the sample type, this will be the majority of the activity.

Studying the publication patterns of an area has become an important study per se, now termed scientometrics or the generic and more recognizable bibliometrics.^190,191  The field can be traced to the work of Eugene Garfield in the 1960s, published in Current Contents. Garfield was instrumental in developing the Science Citation Index, now the Clarivate Analytics Web of Science. Significantly, scientometrics is no longer simply reflecting reality, but is actually transforming reality by changing the behavior of academics and researchers.¹⁹²  There is considerable granularity now contained in bibliometric data; for example, what country or even institution is performing the most cited work in a scientific field. However, although this is interesting it is not our concern here. There are two aspects of bibliometrics. The first and simplest answers the question, who is doing what, and where? The second aspect examines the time dependencies of the general parameters of the academic field in an endeavor to characterize the scientific and technical evolution and contemporary status of the field.¹⁹³  It is in this aspect of past and current trends in antioxidant research work that our interest lies.

Papers published in Nature (Table 1.3) and Analytical Chemistry (,Table 1.4) between 1900 and 2020 containing the keyword “antiox” in the title provide insight into this and other questions. These journals reflect “hot” topics at the forefront of science (generally), and analytical chemistry (specifically), respectively. In the 1950s, there was a distinct bias in Analytical Chemistry to technology-driven papers emphasizing gasoline and rubber with a leaning toward food science in Nature. The interest in these areas seems historically important to community needs in the immediate post-WW2 era. Interestingly, Nature uncharacteristically published two papers on oat antioxidants, in 1961 and 1964, but the second paper was more closely aligned with the photochemistry of cinnamic acid derivatives.

Since 2009, i.e., the last 10 years, all antioxidant-related papers published by Nature relate to physiological antioxidants. Analytical Chemistry, with a focus on methodological development, has papers that deal with advanced techniques for measuring antioxidants in food and physiological systems, and for general screening of antioxidant activity. However, the sample size from these two journals is small although the very selectivity of Nature does offer one advantage in this respect; published papers do focus attention very sharply on concerns and issues current at the time of publication and, in the last 10 years, this appears to be physiological antioxidants. Nevertheless, a more detailed analysis of a broader spectrum of papers would be useful.

Also noticeable in Table 1.3 is a gap in papers published between 1969 and 1999 containing “antiox” in the title in Nature. It is not clear what contributed to this gap. It may have been that there was not general interest in antioxidants in this 30 year period. An analysis of Web of Science bibliometric data is consistent with this. In the period 1969–1999, there were 13 871 papers published with “antiox” in the title, at a rate of 447 papers per year. In the next twenty years, 2000–2019, there were 90 346 papers published at a rate of 4517 papers per year; a 10-fold increase as noted below. In the same time-period there were 10 papers in Nature, a rate of one paper every two years. As the rate of papers on antioxidants increased, there were more articles in Nature.

Analysis of PubMed-generated data for 1943 to 2016 indexed using “food” showed²⁸²  that the number of publications versus year of publication followed two linear equations with the year 1998 as the time when the phase transition in food science research (or publication numbers) occurred. If the rate of publication is taken as the slope of the two linear equations for the periods pre-1998 and post-1998, then publication rate increased approximately tenfold after 1998 although the reasons for the increase at this particular juncture are unclear.

An even more detailed scientometric knowledge map of food science research for the period 2003 to 2014²⁸³  demonstrated that the growth in food science publication during this period was more rapid than the rest of world scientific production. It also predicted that the relative importance of food science will continue to accelerate. However, the main objective of the study was the establishment of the intellectual structure of food science on the basis of an analysis of the keywords present in the papers published in the field. Using Scopus as the database, download of 184 801 citable papers with 230 007 different keywords was generated. Analysis of the data revealed a thematic cognitive structure of food science comprising five clusters as follows:

1. Food composition and nutrition

2. Food processing and modification

3. Food security

4. Food preservation and shelf life

5. Antioxidants in food

Each of these clusters contained a number of second-level clusters but these need not concern us with one exception. What is important and critical to the current topic is the appearance of Cluster 5 “Antioxidants in food” with the two Level 2 clusters as:

•   5.1 Antioxidants and their effects, and

•   5.2 Plant antioxidants.

Antioxidant activity was recognized as a stand-out keyword, and the analysis clearly demonstrated the two areas of research activity as above. Interestingly, documents pertaining to Cluster 5 grew during the period at a rate faster than the other areas of food science and had the greatest scientific impact with the highest percentage of excellent documents. However, the impact of cluster 5 peaked in 2006, which coincided with a subsequent major growth in the percentage of production in 2007. The decline in impact that began in 2006 was interrupted by a slight upsurge in 2010, only for the fall to resume thereafter. This trend in publication numbers and impact may explain, in part, the decision of some journals (Section 3.5) to place restrictions on publication.

There is no study in health/nutrition/physiology comparable to that in food science although a bibliometric study of trends and performance of oxidative stress research from 1991 to 2010²⁸⁴  showed that even here antioxidant was a keyword and that the mainstream topic in the area was antioxidants in human or rat cells.

Other bibliometric studies on antioxidants examine India's contribution on antioxidants²⁸⁵  plus a paper published in 2019 with a broader scope looking specifically at the literature on “antioxidants” (and variants) with no other restrictions.²⁸⁶  There were 299 602 publications identified and analyzed from Web of Science (Figure 1.13) with a staggering 20 616 reviews with a similar number of contributions covering meeting abstracts, proceedings papers and editorial material. The first publication was published in 1957 so the period covered was the six decades from 1957 to 2018. Milestones were 100 papers per year on antioxidants achieved in 1977 and 1000 per year in 1991; the year 2007 was also pivotal as starting from that year annual publications has exceeded 10 000.

The most popular and prolific journals for antioxidant research that have each accounted for at least 1% of total publications over the period were Food Chemistry, Journal of Agricultural and Food Chemistry, Free Radical Biology and Medicine, and PLOS One. The 2011–2018 period was interesting for the emergence and rise of two multidisciplinary journals, Plos One and Scientific Reports, as avenues for antioxidant publication. Food science technology journals increased publication share over the period reflecting that the attention was shifting toward dietary antioxidants. There has also been a transition of scientific interest from research focused on antioxidant vitamins and minerals into more research attention focused on antioxidant phytochemicals (plant secondary metabolites).

Table 1.5 shows the five major Web of Science categories into which papers were classified and some clear patterns are observable. Considering total publications for the six decades, papers were mainly classified into the Web of Science categories of biochemistry/molecular biology (15.3% of total publications), food science technology (15.1%), and pharmacology/pharmacy (11.8%) with much smaller contributions in other categories. These topics have been prolific since 1990 and before, whilst polymer science was prolific before, but its publication share declined in the recent two decades. Chemistry maintained a constant share of publications except for a decline in the decade beginning 1990.

Clinical conditions with high citations included Alzheimer's disease, cancer, cardiovascular disease, and Parkinson's disease.

In summary, the bibliometric data presented above offer some insights into past and current trends in antioxidant research. The data suggest that antioxidant research continues to be considered at the forefront of topics in general science, analytical chemistry, and food science as evidenced by papers in Nature and Analytical Chemistry, and through an “antioxidants in food” cluster evident in food science journals, respectively. Moreover, the trend data in Table 1.5 show that there continues to be strong interest in antioxidant research in food science and technology and biochemistry/molecular biology. An interesting recent addition to the top 5 categories is plant sciences – possibly related to the renewed interest in natural products chemistry for the discovery of new treatments for diseases as well as naturally occurring antioxidants for use in foods, nutraceuticals, etc.

The strong current interest in antioxidant research provides justification for this book, despite the recent misgivings of some authors. As will be evident over the remaining chapters, context is key – deciding what the research question is, then choosing the most appropriate methods to answer that question – and we hope to provide the reader with the tools to answer these questions.