Measurement of Sunlight-Induced Transient Species in Surface Waters

Sunlight irradiation of natural waters results in absorption of light by dissolved organic and inorganic compounds which then generate a variety of transient species including excited state dissolved organic materials ( 3DOM), singlet oxygen ('02), peroxy radicals (ROO-), hydroxyl radicals (HO-), solvated electrons (caq), and superoxide ion (2-) 11-125. The transient nature of these species causes both the practical aspects of and philosophy behind their determination to be different from those of conventional, more stable aquatic pollutants. Firstly, transients cannot be concentrated, separated from the water matrix, or the water removed from the light source, and therefore their analysis must be performed by indirect kinetic or integrative techniques. Secondly, they do not pose a human health concern because no significant exposure route exists. Ecological effects on lower organisms are possible [13] but none has been documented to date. The primary reason for their interest is that they can affect transformation of natural and manmade compounds. Such transformation can be beneficial, such as in the detoxification of pesticides [141, harmful, such as in the production of toxic peroxidic compounds in the photo-oxidation of crude oils [15], or simply of interest for the understanding of biogeochemical cycles, such as in the cycling of sulfur, nitrogen, and humic materials on geological time scales. Their quantitation allows prediction of environmental fate dynamics, and is of interest in water treatment processes where external sources of transients are added [16,17]. In order to understand how to measure transients, it is necessary to have some understanding of the factors which control their formation and consumption. Figure I and table I give an overview of some of the main processes involved. The bulk (-99%) of sunlight absorbed by DOM is converted directly to heat. About 1% of the initially formed excited state 'DOM undergoes intersystem crossing to the longer-lived 3DOM, which transfers the energy to oxygen to form 1O2, the majority of which, in turn, decays by heating the water. A small fraction of 3DOM transfers an electron to oxygen to produce 02-, which decays by disproportionation and some unknown reactions [12]. A minute fraction of excited state DOM ejects an electron, which is consumed rapidly by dissolved oxygen or possibly by nitrate. The radical cation formed by electron ejection may react with oxygen to form peroxy radicals, or these may be formed by addition of ground state oxygen to excited carbonyls yielding a biradical

conventional, more stable aquatic pollutants.
Firstly, transients cannot be concentrated, separated from the water matrix, or the water removed from the light source, and therefore their analysis must be performed by indirect kinetic or integrative techniques. Secondly, they do not pose a human health concern because no significant exposure route exists. Ecological effects on lower organisms are possible [13] but none has been documented to date. The primary reason for their interest is that they can affect transformation of natural and manmade compounds. Such transformation can be beneficial, such as in the detoxification of pesticides [141, harmful, such as in the production of toxic peroxidic compounds in the photo-oxidation of crude oils [15], or simply of interest for the understanding of biogeochemical cycles, such as in the cycling of sulfur, nitrogen, and humic materials on geological time scales. Their quantitation allows prediction of environmental fate dynamics, and is of interest in water treatment processes where external sources of transients are added [16,17].
In order to understand how to measure transients, it is necessary to have some understanding of the factors which control their formation and consumption. Figure I and table I give an overview of some of the main processes involved. The bulk (-99%) of sunlight absorbed by DOM is converted directly to heat. About 1% of the initially formed excited state 'DOM undergoes inter-system crossing to the longer-lived 3 DOM, which transfers the energy to oxygen to form 1O2, the majority of which, in turn, decays by heating the water. A small fraction of 3 DOM transfers an electron to oxygen to produce 02-, which decays by disproportionation and some unknown reactions [12]. A minute fraction of excited state DOM ejects an electron, which is consumed rapidly by dissolved oxygen or possibly by nitrate. The radical cation formed by electron ejection may react with oxygen to form peroxy radicals, or these may be formed by addition of ground state oxygen to excited carbonyls yielding a biradical 'R 2 C=0 ±O 2 -R 2 C(O.)00 *-RC=0 +02 aheat The factors controlling peroxy radical consumption are undefined at present, but may include disproportionation, reaction with DOM, or reversal of 02 addition. Hyroxyl radicals are formed mostly by nitrate photolysis and consumed by reaction with DOM in fresh water or bromide ion in seawater.
The data in table I include only values measured or estimated thus far and therefore they do not necessarily represent all types of waters. Also, the data are of widely varying accuracy. For example, the formation and loss rates of 102 are accurately known, but the corresponding rates for ROOand 02 are crude estimates. The values for etq, HO., and 3 DOM appear to be reliable, but there is less data available than for 102 from which to judge accuracy and/or the range of values occurring under a broad variety of conditions. It may be assumed [1], however, that in well-oxygenated waters ['DOM] =0. 5[ 02] and therefore many more    The integrative approach involves addition of A in high enough concentration to trap all of the transient as it is formed, and measurement of the formation of a product or loss of a reactant such as oxygen: This method has the drawbacks that it requires precise knowledge of kd and that there is always the potential that the high concentrations of A required may affect the lifetime (concentration) of a precursor to the transient of interest. However, the method is more sensitive than the derivative technique and can give quantum yield data which can be used more generally than [T]i values in making environmental fate predictions. This approach has been used successfully in the determination of [e-q],, [22], and quantum yields of 102 production [3] and total radical production [23].
Clearly important for both methods is that the trapping agent be highly selective for the transient  The author wishes to thank J. Hoign6, T. Mill, and R. Zepp, who helped generate many of the concepts used in this paper.