WaterSubject

Pharmaceuticals are ubiquitous contaminants occurring in the aquatic environment. Concentrations of human pharmaceuticals are typically highest in wastewaters but can also be relevant in surface waters, where, despite their low levels, they might give rise to undesirable effects on living organisms. A striking example of such very low dose effects has indeed been demonstrated very recently for a mixture of therapeutic drugs affecting the physiology and morphology of human embryonic cells. Since surface waters are often used as source for drinking-water production, and treated wastewaters are utilized to recharge groundwater tables in regions affected by freshwater scarcity, there is a growing interest in understanding the fate of pharmaceuticals during drinking-water treatment.

Ultraviolet (UV) light irradiation is an established method for drinking-water disinfection and a growing technology for wastewater purification. In the case of drinking-water treatment, German and Austrian regulations prescribe a disinfection dose (fluence) of 400 J m^?2, which should be sufficient, except for some viruses, to eliminate a great variety of pathogenic microorganisms, including Cryptosporidium parvum oocysts. It has long been recognized that even at such (relatively low) UV doses organic compounds dissolved in water may undergo photochemical transformation, which raises the question about transformation products. While the UV-induced elimination of triazine and phenylurea herbicides has been systematically investigated, no such data are available for pharmaceuticals. Recently, significant advances have been made in understanding the aquatic photochemistry of some single pharmaceuticals or classes thereof  and, but tailored data on the photochemical transformation of pharmaceuticals in the aquatic environment are still largely missing.

The present study was primarily conceived to evaluate the extent of degradation (expressed as depletion of the parent compound) of four selected pharmaceuticals in UV drinking-water treatment for disinfection purposes. The investigations presented in this paper exemplify the determination of pH-dependent direct phototransformation quantum yields and how they can be used to quantify the extent of degradation of the pharmaceuticals under low-pressure (LP) and medium-pressure (MP) mercury lamp UV irradiation conditions. The results can also be used to assess the extent of degradation of pharmaceuticals at higher UV doses than those used in disinfection of drinking water, for instance in the UV treatment of wastewaters. Moreover, the possible influence of natural water components on phototransformation rates of the pharmaceuticals was also tested for water used in an actual, full-scale drinking-water treatment plant.

2.1. Reagents

The selected pharmaceuticals diclofenac and sulfamethoxazole (purity >99%) were purchased from Sigma-Aldrich, while iopromide and 17?-ethinylestradiol (EE2) were from Schering/Berlin, Germany. Atrazine, used for actinometry, was of Pestanal^® grade from Riedel-de-Haen. Phosphate sodium salts and phosphoric acid, used to make up buffers, and hydrochloric acid and sodium hydroxide, employed for pH fine adjustment, were of at least analytical grade from common commercial sources. All the above reagents were used as received. Methanol and acetonitrile applied for high-performance liquid chromatography (HPLC) were both of Multisolvent^® grade from Scharlau (Barcelona, Spain). Water (resistivity 18.2 M? cm; TOC 3 ppb), employed for solutions and as a component of HPLC eluents, was purified with a MQ-UV water device (Millipore). Water from the drinking-water treatment plant of Neuilly sur Marne (suburb of Paris, France) was sampled and characterized as described in detail elsewhere.

2.2. Analytical methods

Pharmaceuticals and atrazine were analyzed using an Agilent 1100 HPLC system equipped with a quaternary gradient pump, a degasser, an auto-sampler, a column thermostat and a diode array absorbance detector. The column was a reverse-phase Nucleosil 100, 5 ?m C18 from Machery-Nagel. Eluents consisted of methanol, acetonitrile, and water acidified with 10 mM of phosphoric acid to ensure that acid compounds were present in their neutral form. Detection limits of 50 nM for EE2 and 10 nM for diclofenac, sulfamethoxazole, iopromide, and atrazine were achieved. pH measurements were performed using a Thermo Orion Ross electrode (Hügli-Labortec AG, Switzerland) and a Metrohm 632 pH-meter (Metrohm, Herisau, Switzerland), which were calibrated with standard buffers (pH 4, 7, and 9, Merck).

2.3. UV irradiation experiments

The merry-go-round photoreactor and methodology for irradiation kinetics experiments have been described in detail elsewhere. Briefly, clear aqueous sample solutions (typically 20 mL) contained in quartz tubes are irradiated by light emitted from a lamp situated in a cooling jacket at the center of the photoreactor. The light is filtered through the material of the cooling jacket and an aqueous solution where the sample tubes are immersed and which serves both as an optical filter and as a medium to control temperature. In the present study, a low-pressure mercury (LP Hg) lamp Heraeus Noblelight model TNN 15/32 (nominal power 15 W) and a medium-pressure mercury (MP Hg) lamp Heraeus Noblelight model TQ 150 (nominal power 150 W) were employed as UV radiation sources. The cooling jacket used was made of quartz, and pure water (with negligible light absorption in the wavelength range of emitted radiation) was used as a filter solution and maintained at a temperature of 25.0±0.2 °C. To further reduce the light intensity when using the MP Hg lamp, the quartz cooling jacket was wrapped with two stainless-steel wire cloths, which had a wavelength-independent transmission of about 25%. Fluence rate values were determined by chemical actinometry at low optical using 5 ?M aqueous atrazine as an actinometer (solution buffered at pH=7.0 with 5 mM phosphate) with a quantum yield of 0.046 mol einstein^?1 and molar absorption coefficient of 3860 M^?1 cm^?1 at a wavelength of 254 nm. The fluence rate under MP Hg lamp irradiation was also determined by atrazine actinometry assuming a wavelength-independent quantum yield and using the emission spectrum of the lamp (data from the manufacturer).