Article pubs.acs.org/JPCB
Nitrate Concentration near the Surface of Frozen Aqueous Solutions Harley A. Marrocco and Rebecca R. H. Michelsen* Department of Chemistry, Randolph-Macon College, P.O. Box 5005, Ashland, Virginia 23005, United States S Supporting Information *
ABSTRACT: Photolysis of nitrate plays an important role in the emission of nitrogen oxides from snow and ice, which affects the composition of the overlying atmosphere. In order to quantify these reactions, it is necessary to know how much nitrate is available for photolysis near the surfaces of snow and ice. The concentration of nitrate excluded from frozen solutions of nitric acid, sodium nitrate, and magnesium nitrate was measured with attenuated total reflection infrared spectroscopy. Liquid water and nitrate were observed at and near the bottom surface of frozen aqueous solutions during annealing from −18 to −2 °C. At −2 °C, the nitrate concentration was determined to be ∼1.0 mol/L for frozen NaNO3 and Mg(NO3)2 solutions and ∼0.8 mol/L for frozen HNO3 solutions. At lower temperatures, nitrate concentration ranged from 1.6 to 3.7 mol/L. Ideal thermodynamics overestimates nitrate concentration at colder temperatures where the brine is highly concentrated for all solutions. The nitrate concentration at ice surfaces is well described by bulk freezing point depression data close to the melting point of ice and for nitric acid at colder temperatures. Effects of temperature and counterions and implications for modeling snow chemistry are discussed.
■
INTRODUCTION The nitrate anion plays an important role in the chemistry of snow and ice and the overlying boundary layer of Earth’s atmosphere. Photolysis of nitrate leads to emissions from snow of nitrogen oxides (NOx = NO2 + NO) and, if the snow is sufficiently acidic, HONO.1−3 Other products from nitrate photolysis in ice include OH radicals and O(3P), the latter of which can form ozone in the gas phase.4 Not only do these reactions affect the oxidative capacity and composition of the atmosphere, but they also determine in part how much nitrate is preserved in ice cores. Nitrate is often measured in ice cores, but its interpretation is difficult.5−8 A detailed understanding of nitrogen cycling in snow is required in order to deduce atmospheric NOx in the past,5 as well as to model the effect of snow chemistry on the atmosphere in the present. Ion concentrations, including nitrate, in field samples of snow and ice are measured by analyzing the melted sample. Typical values for nitrate in Arctic snow are 3−8 μM9−12 with higher concentrations in brine and frost flowers9 and somewhat lower concentration in Antarctic snow.7 Nitrate concentration at the surface is assumed to be significantly higher than the melt values, but how much resides at or near enough to the surface to undergo photolysis has not been measured. How much liquid exists at the surface of snowor even whether it covers snow grains completelyis also debatable.13 Assuming surface nitrate is solvated, the amount of liquid and the amount of nitrate together determine the nitrate concentration and thus reaction rates. In the absence of detailed parametrizations, modelers have been forced to make assumptions about the © XXXX American Chemical Society
amounts of liquid and ions. In a nitrate photochemistry model, for example, Boxe and Saiz-Lopez14 used melt nitrate concentration to calculate the average fraction of liquid water from thermodynamics.15 They then used this value to calculate the nitrate concentration at the surface by assuming the entire amount of nitrate was located in liquid at the surface. In another model, it was assumed that snow contained a fraction of 10−3 liquid by volume and this value determined the nitrate concentration.16 In the first comprehensive model of snow/air chemistry and physics, the concentration of nitrate at the surface was treated as an adjustable parameter and chosen to determine the known amount of NO in the gas phase.17 This approach resulted in only 6% of nitrate being available to react, in contrast with models in which all the nitrate is at the surface. Laboratory measurements of the local nitrate concentration near the surface of ice as a function of temperature and other relevant variables may help determine initial values in models. Direct measurement of ion concentrations at the surface of ice is challenging. The vapor pressure of ice is too high for classical surface techniques. Bulk samplesincluding liquid at both the surface and interior of frozen sampleshave been investigated with a variety of methods. Completely frozen methylene blue solutions (at −196 and −30 °C) were observed with UV−visible spectroscopy, and the increase in concentration was inferred from the amount of aggregation.18 Frozen Received: August 14, 2014 Revised: November 6, 2014
A
dx.doi.org/10.1021/jp508244u | J. Phys. Chem. B XXXX, XXX, XXX−XXX
The Journal of Physical Chemistry B
Article
measured by type K thermocouples (Omega). Two thermocouples were used simultaneously to determine if there were any temperature gradients across the frozen samples. At the interface of the ice with germanium, thermocouples placed in the center and at the edge of the Ge showed no measurable temperature gradient. The temperature at the top of the ice sample compared to the interface with Ge was warmer by 0.2 ± 0.2 °C. Temperatures reported here were measured at the Ge− ice interface, where the samples were observed with ATR-IR spectroscopy. Background spectra (with a clean Ge crystal) were taken on the same day at approximately −10 °C. All spectra are an average of 32 scans and were taken with 4 cm−1 resolution for a total collection time of 19 s. Solutions of nitrate compounds were made by mixing with ACS reagent grade water (Sigma-Aldrich) and serially diluted. A 1 N solution of nitric acid (Fluka analytical) was titrated to a phenolphthalein end point with sodium hydroxide to give a concentration of 0.994 ± 0.005 M HNO3. Sodium nitrate (Sigma-Aldrich, ≥99.0%) and magnesium nitrate hexahydrate (Sigma-Aldrich, ACS reagent) were stored in a desiccator but otherwise used as received. Solutions ranged in composition from 1.00 × 10−3 to 3.000 M. A calibration curve of nitrate extinction versus concentration, shown in Figure 1, was
solutions of NaCl were analyzed using NMR, which led to the development of a thermodynamic treatment of solutes in brine that is in equilibrium with ice.15 This approach was improved upon by including activities, essential to understanding solute behavior in cold, concentrated solutions typical of cryospheric brines.19 Bower and Anastasio deduced the concentration of several salts in the brine of frozen solutions by measuring the decrease of furfuryl alcohol as it reacted with dissolved oxygen.20 They concluded that linear freezing point depression described ion concentrations under many, but not all, circumstances. These studies give a picture of the concentration of solutes in the liquid regions of frozen solutions, including both surface and interior brine. The concentration of nitrate available for reaction near the surfaces of ice and snow, however, may differ locally from the average concentration.21 Consequently, surface-sensitive techniques may offer insight into local concentrations of species available to react near the surface. Krepelova and co-workers used X-ray photoelectron spectroscopy (XPS) and near-edge Xray absorption fine structure spectroscopy (NEXAFS) to observe gas-phase nitric acid deposited on ice with a surface sensitivity of a few nanometers.22 They report nitrate to ice mole ratios at −43 °C that are equivalent to 3−15 mol/kg. This range contains the freezing point depression concentration of HNO3 in water at that temperature, which is 8.5 mol/kg.23 Wren and Donaldson probed ∼100 nm of the top of frozen nitrate salt solutions with glancing-angle Raman spectroscopy.24 While they do not report nitrate concentrations, they observe a nitrate signal which is more than an order of magnitude smaller than the concentration predicted by the phase diagram at −5 °C. These surface studies, although not quantitative, suggest that the use of the thermodynamic treatment may not always be predictive of ions found at the surface of snow and ice. In this study, we report the microscopic concentration of nitrate at the surface and near-surface region of ice samples. Attenuated total reflection infrared (ATR-IR) spectroscopy25 preferentially observes the near-surface region of a sample. We explore the influences of the counterion (H+, Na+, and Mg2+) and temperature and compare our results to thermodynamics and freezing point depression data.
Figure 1. Calibration curve of ATR-IR nitrate signal for aqueous solutions of NaNO3, HNO3, and Mg(NO3)2 at room temperature. The least-squares linear regression has a slope of 0.0146 ± 0.0002 L/ mol and an intercept of −0.0001 ± 0.0002 with an R2 value of 0.998.
■
EXPERIMENTAL METHODS Frozen aqueous samples were studied via ATR-IR spectroscopy to minimize the interference of water vapor in the spectra and to detect the region near the interface. Aqueous solutions of nitrate compounds were frozen on top of a cooled germanium prism embedded in a copper plate so that the bottom of the samples, i.e., the interface of ice with germanium, was probed by the spectrometer. A drawing of the experimental setup can be seen in Walker et al.26 The plate and germanium crystal were cooled and warmed via recirculating ethanol from a Huber ministat cc1 chiller. The ATR plate was positioned on a VeeMax II accessory (Pike Technologies) in the sample cell of a ThermoElectron Nicolet 6700 FT-IR spectrometer with an MCT-A detector. A constant 45° angle of incidence was used. The sample compartment, the ATR plate holder, and the area above the germanium crystal were purged with dry, CO2-free air. When the crystal was cooled to around −17 °C, solutions (∼1 mL) were deposited on the Ge surface with a micropipet. Freezing occurred in less than 1 min. The resulting polycrystalline samples were ∼10 mm in diameter and ∼3 mm in height. It was not necessary to apply pressure to the samples to obtain ATR-IR spectra. The temperature of the frozen samples was
prepared using room temperature spectra from liquid solutions of all three solutes taken on several different days. (Extinction is analogous to absorption for transmission spectroscopy.) Water was subtracted from the spectra, and the nitrate peak height at 1346.1 cm−1 was measured relative to a baseline from 1490 to 1510 cm−1. No decrease in nitrate signal for volatile HNO3 was observed over time or relative to the nonvolatile solutes Mg(NO3)2 and NaNO3 for these room temperature measurements. The nitrate signal was linear over the range of 0.0128− 3.000 M. The absorptivity of nitrate was 0.0146 ± 0.0002 L/ mol with an intercept of 0 to within the regression error. This linear calibration is possible since the nitrate ions are on average equally spaced throughout the solution, the liquid sample is much thicker (a few millimeters) than the penetration depth of the method (
Nitrate Concentration near the Surface of Frozen Aqueous Solutions Harley A. Marrocco and Rebecca R. H. Michelsen* Department of Chemistry, Randolph-Macon College, P.O. Box 5005, Ashland, Virginia 23005, United States S Supporting Information *
ABSTRACT: Photolysis of nitrate plays an important role in the emission of nitrogen oxides from snow and ice, which affects the composition of the overlying atmosphere. In order to quantify these reactions, it is necessary to know how much nitrate is available for photolysis near the surfaces of snow and ice. The concentration of nitrate excluded from frozen solutions of nitric acid, sodium nitrate, and magnesium nitrate was measured with attenuated total reflection infrared spectroscopy. Liquid water and nitrate were observed at and near the bottom surface of frozen aqueous solutions during annealing from −18 to −2 °C. At −2 °C, the nitrate concentration was determined to be ∼1.0 mol/L for frozen NaNO3 and Mg(NO3)2 solutions and ∼0.8 mol/L for frozen HNO3 solutions. At lower temperatures, nitrate concentration ranged from 1.6 to 3.7 mol/L. Ideal thermodynamics overestimates nitrate concentration at colder temperatures where the brine is highly concentrated for all solutions. The nitrate concentration at ice surfaces is well described by bulk freezing point depression data close to the melting point of ice and for nitric acid at colder temperatures. Effects of temperature and counterions and implications for modeling snow chemistry are discussed.
■
INTRODUCTION The nitrate anion plays an important role in the chemistry of snow and ice and the overlying boundary layer of Earth’s atmosphere. Photolysis of nitrate leads to emissions from snow of nitrogen oxides (NOx = NO2 + NO) and, if the snow is sufficiently acidic, HONO.1−3 Other products from nitrate photolysis in ice include OH radicals and O(3P), the latter of which can form ozone in the gas phase.4 Not only do these reactions affect the oxidative capacity and composition of the atmosphere, but they also determine in part how much nitrate is preserved in ice cores. Nitrate is often measured in ice cores, but its interpretation is difficult.5−8 A detailed understanding of nitrogen cycling in snow is required in order to deduce atmospheric NOx in the past,5 as well as to model the effect of snow chemistry on the atmosphere in the present. Ion concentrations, including nitrate, in field samples of snow and ice are measured by analyzing the melted sample. Typical values for nitrate in Arctic snow are 3−8 μM9−12 with higher concentrations in brine and frost flowers9 and somewhat lower concentration in Antarctic snow.7 Nitrate concentration at the surface is assumed to be significantly higher than the melt values, but how much resides at or near enough to the surface to undergo photolysis has not been measured. How much liquid exists at the surface of snowor even whether it covers snow grains completelyis also debatable.13 Assuming surface nitrate is solvated, the amount of liquid and the amount of nitrate together determine the nitrate concentration and thus reaction rates. In the absence of detailed parametrizations, modelers have been forced to make assumptions about the © XXXX American Chemical Society
amounts of liquid and ions. In a nitrate photochemistry model, for example, Boxe and Saiz-Lopez14 used melt nitrate concentration to calculate the average fraction of liquid water from thermodynamics.15 They then used this value to calculate the nitrate concentration at the surface by assuming the entire amount of nitrate was located in liquid at the surface. In another model, it was assumed that snow contained a fraction of 10−3 liquid by volume and this value determined the nitrate concentration.16 In the first comprehensive model of snow/air chemistry and physics, the concentration of nitrate at the surface was treated as an adjustable parameter and chosen to determine the known amount of NO in the gas phase.17 This approach resulted in only 6% of nitrate being available to react, in contrast with models in which all the nitrate is at the surface. Laboratory measurements of the local nitrate concentration near the surface of ice as a function of temperature and other relevant variables may help determine initial values in models. Direct measurement of ion concentrations at the surface of ice is challenging. The vapor pressure of ice is too high for classical surface techniques. Bulk samplesincluding liquid at both the surface and interior of frozen sampleshave been investigated with a variety of methods. Completely frozen methylene blue solutions (at −196 and −30 °C) were observed with UV−visible spectroscopy, and the increase in concentration was inferred from the amount of aggregation.18 Frozen Received: August 14, 2014 Revised: November 6, 2014
A
dx.doi.org/10.1021/jp508244u | J. Phys. Chem. B XXXX, XXX, XXX−XXX
The Journal of Physical Chemistry B
Article
measured by type K thermocouples (Omega). Two thermocouples were used simultaneously to determine if there were any temperature gradients across the frozen samples. At the interface of the ice with germanium, thermocouples placed in the center and at the edge of the Ge showed no measurable temperature gradient. The temperature at the top of the ice sample compared to the interface with Ge was warmer by 0.2 ± 0.2 °C. Temperatures reported here were measured at the Ge− ice interface, where the samples were observed with ATR-IR spectroscopy. Background spectra (with a clean Ge crystal) were taken on the same day at approximately −10 °C. All spectra are an average of 32 scans and were taken with 4 cm−1 resolution for a total collection time of 19 s. Solutions of nitrate compounds were made by mixing with ACS reagent grade water (Sigma-Aldrich) and serially diluted. A 1 N solution of nitric acid (Fluka analytical) was titrated to a phenolphthalein end point with sodium hydroxide to give a concentration of 0.994 ± 0.005 M HNO3. Sodium nitrate (Sigma-Aldrich, ≥99.0%) and magnesium nitrate hexahydrate (Sigma-Aldrich, ACS reagent) were stored in a desiccator but otherwise used as received. Solutions ranged in composition from 1.00 × 10−3 to 3.000 M. A calibration curve of nitrate extinction versus concentration, shown in Figure 1, was
solutions of NaCl were analyzed using NMR, which led to the development of a thermodynamic treatment of solutes in brine that is in equilibrium with ice.15 This approach was improved upon by including activities, essential to understanding solute behavior in cold, concentrated solutions typical of cryospheric brines.19 Bower and Anastasio deduced the concentration of several salts in the brine of frozen solutions by measuring the decrease of furfuryl alcohol as it reacted with dissolved oxygen.20 They concluded that linear freezing point depression described ion concentrations under many, but not all, circumstances. These studies give a picture of the concentration of solutes in the liquid regions of frozen solutions, including both surface and interior brine. The concentration of nitrate available for reaction near the surfaces of ice and snow, however, may differ locally from the average concentration.21 Consequently, surface-sensitive techniques may offer insight into local concentrations of species available to react near the surface. Krepelova and co-workers used X-ray photoelectron spectroscopy (XPS) and near-edge Xray absorption fine structure spectroscopy (NEXAFS) to observe gas-phase nitric acid deposited on ice with a surface sensitivity of a few nanometers.22 They report nitrate to ice mole ratios at −43 °C that are equivalent to 3−15 mol/kg. This range contains the freezing point depression concentration of HNO3 in water at that temperature, which is 8.5 mol/kg.23 Wren and Donaldson probed ∼100 nm of the top of frozen nitrate salt solutions with glancing-angle Raman spectroscopy.24 While they do not report nitrate concentrations, they observe a nitrate signal which is more than an order of magnitude smaller than the concentration predicted by the phase diagram at −5 °C. These surface studies, although not quantitative, suggest that the use of the thermodynamic treatment may not always be predictive of ions found at the surface of snow and ice. In this study, we report the microscopic concentration of nitrate at the surface and near-surface region of ice samples. Attenuated total reflection infrared (ATR-IR) spectroscopy25 preferentially observes the near-surface region of a sample. We explore the influences of the counterion (H+, Na+, and Mg2+) and temperature and compare our results to thermodynamics and freezing point depression data.
Figure 1. Calibration curve of ATR-IR nitrate signal for aqueous solutions of NaNO3, HNO3, and Mg(NO3)2 at room temperature. The least-squares linear regression has a slope of 0.0146 ± 0.0002 L/ mol and an intercept of −0.0001 ± 0.0002 with an R2 value of 0.998.
■
EXPERIMENTAL METHODS Frozen aqueous samples were studied via ATR-IR spectroscopy to minimize the interference of water vapor in the spectra and to detect the region near the interface. Aqueous solutions of nitrate compounds were frozen on top of a cooled germanium prism embedded in a copper plate so that the bottom of the samples, i.e., the interface of ice with germanium, was probed by the spectrometer. A drawing of the experimental setup can be seen in Walker et al.26 The plate and germanium crystal were cooled and warmed via recirculating ethanol from a Huber ministat cc1 chiller. The ATR plate was positioned on a VeeMax II accessory (Pike Technologies) in the sample cell of a ThermoElectron Nicolet 6700 FT-IR spectrometer with an MCT-A detector. A constant 45° angle of incidence was used. The sample compartment, the ATR plate holder, and the area above the germanium crystal were purged with dry, CO2-free air. When the crystal was cooled to around −17 °C, solutions (∼1 mL) were deposited on the Ge surface with a micropipet. Freezing occurred in less than 1 min. The resulting polycrystalline samples were ∼10 mm in diameter and ∼3 mm in height. It was not necessary to apply pressure to the samples to obtain ATR-IR spectra. The temperature of the frozen samples was
prepared using room temperature spectra from liquid solutions of all three solutes taken on several different days. (Extinction is analogous to absorption for transmission spectroscopy.) Water was subtracted from the spectra, and the nitrate peak height at 1346.1 cm−1 was measured relative to a baseline from 1490 to 1510 cm−1. No decrease in nitrate signal for volatile HNO3 was observed over time or relative to the nonvolatile solutes Mg(NO3)2 and NaNO3 for these room temperature measurements. The nitrate signal was linear over the range of 0.0128− 3.000 M. The absorptivity of nitrate was 0.0146 ± 0.0002 L/ mol with an intercept of 0 to within the regression error. This linear calibration is possible since the nitrate ions are on average equally spaced throughout the solution, the liquid sample is much thicker (a few millimeters) than the penetration depth of the method (
