Nitrate and Nitrite Ultraviolet Actinometers

We developed nitrate and nitrite actinometers to determine radiant fluxes from 290 to 410 nm. These actinometers are based on the reaction of the photochemically generated OH radical with benzoic acid to form salicylic acid (SA) and p-hydroxybenzoic acid (pHBA). Actinometer development included determination of the temperature and wavelength dependence of the quantum yield for formation of SA and pHBA from nitrate and nitrite photolysis in air-saturated solutions. Quantum yields (at 258C) for SA production from nitrate photolysis ranged from 0.00146 to 0.00418 between 290 and 350 nm, and from 0.00185 to 0.00633 for nitrite photolysis between 290 and 405 nm. The quantum yields for SA production were approximately 50–60% greater than quantum yields for pHBA production from nitrate and nitrite photolysis. For both actinometers, SA and pHBA formation was temperature dependent, increasing by approximately a factor of 2.2 from 0 to 358C. Activation energies for SA formation varied with wavelength, ranging from 14.7 to 16.5 kJ mol21 between 290 and 330 nm for the nitrate actinometer and 12.3 to 17.8 kJ mol21 between 310 and 390 nm for the nitrite actinometer. Activation energies for pHBA formation were 2–11% higher. Wavelengthdependent changes in the quantum yield and activation energy for SA and pHBA formation from nitrate photolysis suggest multiple electronic transitions for nitrate from 290 to 350 nm. Quantum yields for OH radical formation from nitrate and nitrite photolyses were estimated from SA and pHBA quantum yields at 258C. Wavelength-dependent OH quantum yields ranged from 0.007 to 0.014 for nitrate photolysis between 290 and 330 nm and from 0.024 to 0.078 for nitrite photolysis between 298 and 390 nm. The nitrate and nitrite actinometers can maintain initial rate conditions for hours, are insensitive to laboratory lighting, easy to use and extremely sensitive; the minimum radiant energy that can be detected in our irradiation system is approximately 1029 einsteins. *To whom correspondence should be addressed at: State University of New York, College of Environmental Science and Forestry, Department of Chemistry, 1 Forestry Drive, Syracuse, NY 13210, USA. Fax: 315-470-6856; e-mail: [email protected] INTRODUCTION The determination of photochemical quantum yields is dependent on the accurate measurement of monochromatic radiant fluxes. Chemical actinometry is the most common method for determining radiant fluxes during photochemical studies, having an advantage over most instrumental methods because actinometers are able to measure the photon flux within the irradiated cell. The utility of an actinometer is based on its accuracy, sensitivity, linear dynamic range, reciprocity and ease of use. Additionally, for the study of photoprocesses occurring in water, the actinometer should be aqueous-based to minimize refractive index differences. Currently, potassium ferrioxalate actinometry (1) is the most common method used to determine monochromatic radiant fluxes. The potassium ferrioxalate actinometer (PFA)† is used to determine radiant fluxes from the UV out to visible wavelengths. Unfortunately, because PFA absorbs visible light, its analysis, which involves the complexation of ferrous iron and its subsequent absorbance detection, must be performed in a darkroom. The development of actinometers based on the photolysis of valerophenone (2) and p-nitroanisole (PNA) (3) has allowed for accurate measurement of UV radiant fluxes without the need for a darkroom. However, the sensitivity of these actinometers is limited because they are based on the loss, rather than the production, of a chromophore. To overcome the limitations of these actinometers, we developed nitrate and nitrite actinometers for radiant flux determinations from 290 to 405 nm. These actinometers are insensitive to laboratory lighting and quantify radiant fluxes based on the production of highly fluorescent and absorptive analytes. The aqueous photochemistry of nitrate has been the focus of numerous studies, with emphasis on elucidation of nitrate’s electronic transitions (4–8) and the mechanism and products of nitrate photolysis (9–15). Nitrate has two principal absorption bands, the first is an intense p → p* band (4) occurring in the far UV (emax 5 9500 M21 cm21 at 201 nm), while the second is a weak absorption band centered at 302 nm (emax 5 7.14 M21 cm21). The photolysis of nitrate in the UVB (290–320 nm) has also been examined for its †Abbreviations: BA, benzoic acid; CI, confidence interval; mHBA, m-hydroxybenzoic acid; PFA, potassium ferrioxalate actinometry; pHBA, p-hydroxybenzoic acid; PNA, p-nitroanisole; pyr, pyridine; SA, salicylic acid. 320 Joseph J. Jankowski et al. role in the marine nitrogen cycle (11) and as a source of the hydroxyl radical (12–14,16). The direct photolysis of nitrate proceeds via two pathways (10): 2 2 NO 1 hy → NO 1 O (1) 3 2 2 2 3 NO 1 hy → NO 1 O( P) (2) 3 2 Approximately 90% of nitrate photolysis proceeds through pathway 1 (14), with the O radical undergoing rapid protonation to form the hydroxyl radical (OH). Approximately 10% of nitrate photolysis occurs through the second pathway (14), where the O(3P) atom is thought to react with molecular oxygen to form ozone. Nitrite has also been studied to determine its electronic transitions (17) and mechanism of photochemical decay (13,18,19). Nitrite has three absorption bands in the UV (17). The first involves a strong p → p* absorption with a maximum at 220 nm, and the other two correspond to n → p* transitions that occur in the UVB and UVA (320–400 nm), with one defined absorption maximum at 354 nm (emax 5 22.7 M21 cm21). Nitrite undergoes direct photolysis, leading to production of the hydroxyl radical (18): 2 2 NO 1 hy → NO 1 O (3) 2 2 2 O 1 H O → OH 1 OH (4) 2 Quantum yields for hydroxyl radical production from nitrate and nitrite photolysis are low (;0.01) (12–14,19). As a consequence, nitrate and nitrite photolyses can be used for longterm irradiations (ca hours) without altering the solution composition. This is in contrast to loss-based actinometers that require sensitizers or quenchers to manipulate reaction rates (2,3). The nitrate and nitrite photolyses are extremely sensitive because the OH radical is trapped to form highly chromophoric and fluorescent products. This allows for the quantification of low radiant fluxes during short-term irradiations (ca minutes). We developed the nitrate and nitrite actinometers using benzoic acid (BA) as the scavenger of the OH radical, leading to the formation of monohydroxybenzoic acids (20,21). The mechanism and products of these reactions are well documented (22–26). A chromatographic method (27) was used to quantify ortho(salicylic acid, SA) and para-hydroxybenzoic acid (pHBA). Using this technique, wavelength-dependent quantum yields were determined for the formation of SA and pHBA from the photolysis of the nitrate or nitrite actinometer solutions. This method was also used to determine the temperature dependence of SA or pHBA formation at several wavelengths. Additionally, a batch fluorescence technique was developed at the latter stages of this study to provide an alternative method for quantifying SA production in the actinometer solutions. MATERIALS AND METHODS Chemicals. Sodium nitrate ($99.995%), pHBA and m-hydroxybenzoic acid (mHBA) ($99%), valerophenone (.99%) and potassium oxalate monohydrate ($99%) were obtained from Aldrich Chemical Company (Milwaukee, WI). Sodium nitrite (.99%) was purchased from Fluka Chemical (Buchs, Switzerland). The ACS reagent-grade phenanthroline was obtained from Fisher Scientific (Pittsburgh, PA). Salicylic acid ($99.5%) and reagent-grade sodium bicarbonate, sodium acetate, BA, monobasic potassium phosphate and phosphoric acid were purchased from J. T. Baker, Inc. (Phillipsburg, NJ). Ferric chloride was purchased from Mallinckrodt Chemical, Inc. (Paris, KY). Distilled-in-glass, high purity methanol and acetonitrile were obtained from Baxter Diagnostics, Inc. (McGaw Park, IL). All chemicals were used as received, with the exception of BA that was recrystallized three times from high purity laboratory water. All solutions were prepared with water from a Millipore water purification system (Millipore Corp., Chicago, IL) having a resistance .18 MV cm. The water purification system consisted of filtration/dechlorination cartridges followed by a Milli-RO (reverse osmosis) system, a four cartridge Milli Q system (Super C carbon cartridge, two ion-exchange cartridges, and an Organex Q cartridge) and final filtration through a 0.2 mm Whatman Polycap AS capsule (Fisher Scientific). All solutions in this study were prepared with Milli Q water in borosilicate glassware and stored in the dark at 48C. Quantum yield determinations were made in an air-saturated 2.5 mM sodium bicarbonate solution containing 1 mM BA and either 10 mM sodium nitrate or 1 mM sodium nitrite. The pH of these solutions was 7.2. These solutions were stable for at least 45 days when stored in the dark and refrigerated. The potassium ferrioxalate used for chemical actinometry was prepared by adding three parts potassium oxalate (1.5 M) to one part ferric chloride (1.0 M). The resulting precipitate was recrystallized three times with Milli Q water and dried in a desiccator. An 8.7 mM aqueous potassium ferrioxalate solution in 0.05 M sulfuric acid was prepared from these crystals as needed. HPLC. The HPLC system consisted of an Eldex model B-100-S single piston pump (Eldex Laboratories, Menlo Park, CA) connected to an Anspec Bio-Merge gradient mixer (The Anspec Co., Ann Arbor, MI), a 0–5000 psi pressure gauge (C and H Sales, Vineland, NJ), a two-position, six-port injection valve (Valco Instruments Co., Houston, TX) with a 100 mL loop and a Waters RCM radial compression cartridge holder with a Waters 8 3 100 mm Radial Pak cartridge containing 4 mm diameter C18 packing (Waters Associates, Inc., Milford, MA). Fluorescence detection of SA was accomplished with an F-1050 Hitachi fluorescence spectrophotometer (Hitachi Instruments, Inc., Danbury, CT) at an excitation wavelength of 305 6 7.5 nm and an emission wavelength of 410 6 7.5 nm. The pHBA was quantified at 250 nm and mHBA at 300 nm using a Shimadzu SPD-10AV absorbance detector (Shimadzu Scientific Instruments, Inc., Columbia, MD). Peak areas (SA) and heights (pHBA and mHBA) were determined using ELAB integration software (OMS Tech, Inc., Miami, FL). For the nitrate actinometer, HPLC analyses were performed isocratically at a flow rate of 1.2 mL min21 using a 0.2 mm-filtered 45/ 55% vol/vol methanol/25 mM phosphate buffer mobile phase. The phosphate-buffered aqueous phase was prepared by combining a 25 mM phosphoric acid solution with a 25 mM potassium phosphate (monobasic) solution to a pH of 2.1. For the nitrite actinometer, a solvent program was employed due to the coelution of a contaminant peak with the pHBA. Separations were performed for 3.6 min at 22/ 78% vol/vol methanol/25 mM phosphate buffer (pH 2.1) followed by 5 min at 75/25% methanol/25 mM phosphate buffer. The column was then re-equilibrated for 10 min at the initial 22/78% mobile phase composition. All HPLC analyses were performed shortly after an irradiation with time provided only for room temperature equilibration. Analyte quantification. Salicylic acid and mHBA and pHBA were quantified by HPLC employing fluorescence and absorbance detection, respectively. A stock solution containing 20 mM SA, mHBA and pHBA was prepared in the actinometer solution previously described. Calibration curves were generated by standard additions of the stock solution to the actinometer solution, with final concentrations ranging from 10 to 1000 nM. This method of quantification yielded a coefficient of variation of 0.54% for SA and 0.64% for pHBA from multiple injections (n 5 5) of a 500 nM standard; injection of a 50 nM standard yielded a coefficient of variation (n 5 5) of 4.5% for SA and 7.1% for pHBA. The detection limit of this method, with a signal-to-noise ratio of two, was typically 0.6 nM for SA and 5 nM for pHBA. Detection limits were primarily dependent on the purity of the BA. Molar absorption coefficients of nitrate and nitrite were determined in a 2.5 mM sodium bicarbonate solution containing 1 mM BA and either 10 mM sodium nitrate or 1 mM sodium nitrite. The Photochemistry and Photobiology, 1999, 70(3) 321 nitrate concentration was increased to 100 mM for absorption measurements made above 330 nm, while nitrite was increased to 10 mM for measurements above 400 nm. All absorbance measurements were made in a 10 cm quartz cell using a Hewlett Packard UV– visible spectrophotometer (model HP 8453). Absorption measurements were referenced against a pH 7.2, 2.5 mM sodium bicarbonate solution containing 1 mM BA. The molar absorption coefficient of SA was determined in a solution containing 10 mM SA and 2.5 mM NaHCO3, referenced against a 2.5 mM sodium bicarbonate solution. Concentrations of nitrate were determined in the nitrate actinometer solution at 302 nm by Beer’s law using the experimentally determined molar absorption coefficient value of 7.14 M21 cm21. The nitrite concentration was determined in the nitrite actinometer solution at 354 nm using a molar absorption coefficient value of 22.7 M21 cm21. For the determination of nitrate and nitrite concentrations, absorbances were referenced against a pH 7.2 solution of 1 mM BA and 2.5 mM sodium bicarbonate. Batch fluorometry. Salicylic acid can be quantified by batch fluorescence without HPLC analysis. Batch quantification was performed using an ISS PC1 photon-counting spectrofluorometer (Champaign, IL). As with HPLC analysis, calibration curves were generated by standard additions of the stock solution to the actinometer solutions, with final concentrations ranging from 10 to 1000 nM. The response of the instrument was determined from the fluorescence of standards using 305 6 2 nm excitation and 410 6 2 nm emission. Photochemical experiments. The irradiation system consisted of a 1 kW xenon lamp, a GM 252 high intensity quarter meter grating monochrometer and an enclosed sample chamber (Spectral Energy Corp., Westwood, NJ). A 2.5 nm bandwidth was used for all photolyses, with the exception of nitrate irradiations at wavelengths .335 nm and nitrite irradiations at wavelengths .400 nm, where a 5.0 nm bandwidth was selected due to the low molar absorption coefficients of nitrate and nitrite, respectively. For all irradiations, a 3.1 mL aliquot of the nitrate or nitrite solution was placed in a 1 cm quartz cell with a screw cap containing a Teflon septum. The solution was equilibrated in the dark to the selected temperature for at least 20 min. The sample chamber temperature was controlled with a Fisher Scientific circulating water bath (model 9005). The sample was continually stirred throughout the irradiation. To eliminate condensation on the quartz cell at low temperatures, the sample chamber was dehumidified with a stream of dry air that was obtained by passing compressed air through a packed column containing Drierite (Fisher Scientific). Irradiation times ranged from 10 to 20 min, except for wavelengths .330 nm for nitrate irradiations and .380 nm for nitrite irradiations where samples were irradiated from 1 to 2 h. Controls were irradiated from 10 min to 2 h, depending on the wavelength. For these controls, an aluminum foil plate was placed in front of the light inlet of the sample chamber. All samples were allowed to equilibrate to room temperature and subsequently analyzed for SA and pHBA production by HPLC. Actinometry. Radiant fluxes were determined at each wavelength using the chemical actinometer potassium ferrioxalate (1). A solution of 8.7 mM potassium ferrioxalate in 0.05 M sulfuric acid was used for all wavelengths. For each wavelength, a 3.1 mL aliquot of the actinometer solution was irradiated for 3 min. After the ferrioxalate solution was irradiated, it was transferred in the dark to a redilluminated room. Subsequently, 0.5 mL of a buffer (pH 4.55) containing sodium acetate (0.60 M) and sulfuric acid (0.18 M), 2.0 mL of 0.2% wt/wt 1,10-phenanthroline aqueous solution and 6.5 mL of Milli Q water were added to a 1 mL aliquot of the irradiated sample (28). Blanks were prepared using nonirradiated samples. Samples were reacted for 45 min at room temperature. Absorbance measurements were made in a 1 cm quartz cell at 510 nm with a Spectronic Genesys absorbance detector (Spectronic Instruments, Rochester, NY). All samples were analyzed in duplicate. The radiant flux, E (einstein min21), was then calculated using the actinometry equation described in Kuhn et al. (28), with inclusion of a cell pathlength term in the denominator: 23 10 AV V 2 3 E 5 (5) eFV tl 1 where A is the absorbance of the ferrous-complexed solution at 510 nm, V1 is the volume of the irradiated solution (1.0 mL), V2 is the volume of the PFA irradiated (3.1 mL), V3 is the volume of the complexed solution (10 mL), e is the molar absorption coefficient of the ferrous phenathroline complex at 510 nm, t is the irradiation time and l is the pathlength of the cell used for the absorbance measurement. The valerophenone actinometer was prepared at a concentration of 5 3 1026 M valerophenone in water containing 0.05% acetonitrile (vol/vol). This solution was prepared from a 0.010 M stock solution of valerophenone in acetonitrile as described in Zepp et al. (2). Valerophenone concentrations were determined using HPLC with absorbance detection at 250 nm. The HPLC analyses were performed isocratically at a flow rate of 1.2 mL min21 using a 30/70% (vol/ vol) water/acetonitrile mobile phase. The HPLC column was the same as that described for SA and pHBA separations (vide supra). A standard curve was generated with concentrations ranging from 1 to 5 mM valerophenone in water with 0.05% acetonitrile (vol/vol). Quantum yield determinations. Quantum yields for SA and pHBA production were determined in air-saturated solutions at 5 nm intervals from 290 to 350 nm for nitrate photolysis and at 10 nm intervals from 290 to 405 nm for nitrite photolysis. The equation used to determine the quantum yield for SA or pHBA production at a given wavelength (l) under optically thin conditions is [X]V F 5 (6) l 2 2.303E e lt[NO ] l l x where [X] is the pHBA or SA concentration produced during an irradiation (M), V is the volume of the irradiated solution (L), E is the radiant flux (determined by ferrioxalate actinometry) (einstein min21), e is the molar absorption coefficient of nitrate or nitrite at the selected wavelength (M21 cm21), l is the pathlength of the irradiation cell (cm), t is the irradiation time (min) and [NOx] is the concentration of nitrate or nitrite in the irradiated solution (M). The radiant flux in the 1 cm quartz cell varied from 1.0 to 4.0 3 1027 einstein min21 over the wavelength range considered, as determined by PFA. For all quantum yield determinations, initial rate conditions were maintained with no loss of nitrate or nitrite observed during irradiations as determined by spectrophotometry. Temperature dependence. The temperature dependence of the quantum yields for SA or pHBA production was determined at 5 nm intervals for the nitrate actinometer and at 20 nm intervals for the nitrite actinometer. For nitrate, the temperature dependence was determined from 290 to 330 nm, while for nitrite, the temperature dependence was determined from 310 to 390 nm. For each wavelength band examined, quantum yields were determined at 5, 15, 25 and 358C. Nonlinear Arrhenius plots of the quantum yield versus reciprocal temperature were generated to determine the activation energy (Ea) for each wavelength using nonlinear least-squares regression analysis. Determination of actinometer response as a function of time and intensity. Experiments were conducted to demonstrate that the production of SA and pHBA in the actinometers was dependent on the photon flux only, and not on the radiant intensity or time of irradiation (i.e. the condition of reciprocity was met). To demonstrate linearity of response with time, the actinometers were irradiated at 310 nm and 258C for 1–120 min. The response of the actinometers was then plotted as a function of time. To demonstrate that the quantum yields for SA and pHBA production were independent of radiant intensity, a series of screens and neutral density filters were employed. By varying the combinations of screens and filters, irradiations at 310 nm were conducted at radiant intensities ranging from 36.5 to 380 3 10212 einsteins s21 as determined by PFA. Salicyclic acid and pHBA production were then plotted as a function of the radiant intensity. Determination of percentage of OH reacting with BA. The relative proportion of photochemically produced OH reacting with BA was determined in the 10 mM nitrate actinometer solution that was irradiated for 15 min at 305 nm and 258C. Solutions were prepared with BA concentrations ranging from 1 to 1000 mM. The pH of these solutions ranged from 7.2 to 8.2. A plot of production rate of the products (SA and pHBA) versus BA concentration was generated to demonstrate saturation for the nitrate actinometer. Complete scavenging of the OH radical was expected based on published sec322 Joseph J. Jankowski et al. Figure 1. The molar absorption coefficient (e) plotted as a function of wavelength for (A) 10 mM nitrate and (B) 1 mM nitrite in a 2.5 mM bicarbonate solution containing 1 mM BA. The inset in panel A is a plot of eNO32 versus wavelength from 330 to 360 nm for 100 mM nitrate. The inset in panel B is a plot of eNO22 versus wavelength from 400 to 425 nm for 10 mM nitrite. All spectra were referenced against a 2.5 mM bicarbonate/1 mM BA solution (pH 7.2). Figure 2. The wavelength-dependent quantum yield for SA and pHBA production from the photolysis of 10 mM nitrate in the actinometer solution. Error bars denote the 95% confidence interval (n 5 3–12). ond-order rate constants for the reaction of the OH radical with BA, nitrate and the bicarbonate ion. In the nitrite actinometer solution, a large fraction of the OH radical was expected to react with the nitrite ion based on the large value of its second-order rate constant (1.1 3 1010 M21 s21) (29). Because of this high rate constant and the limited solubility of BA in water, complete scavenging of the OH radical was not achieved. Therefore, literature values for the second-order rate constants for the reaction of OH with BA, nitrite and bicarbonate ion were used to estimate the percentage of OH that reacts with BA in the 1 mM nitrite actinometer.