Cations impact radical reaction dynamics in concentrated multicomponent aqueous solutions

Free radicals produced from ionizing radiation in highly alkaline and high ionic strength solutions are important drivers of speciation in complex radioactive solutions, such as in the nuclear wastes stored at the Hanford Site in Washington State, USA.^1–3 A predictive understanding of the interplay between the background ionizing radiation and complex chemistry of the waste is necessary for safe and efficient processing into a stable waste form for disposal. These wastes are multiphasic, high pH (>12), and high ionic strength (∼11 mol/l) systems^4,5 comprised of metal oxide solids (“sludge”), a liquid phase (“supernatant”), and caustic salts (“saltcake”).⁴ Each is chemically complex, with the supernatant containing molar concentrations of sodium hydroxide, nitrate, nitrite, and other oxyanions, which often precipitated to form the saltcake.⁴ The supernatant also contains multiple radionuclides, most notably ¹³⁷Cs and ⁹⁹Tc,⁵ and slow radioactive decay has exposed the waste to a flux of short-lived radiolytic radical species for decades.^2,6

An important consequence of radiolysis in nuclear waste is hydrogen production. The Hanford site has developed an empirical model to predict generation of hydrogen and other hazardous gases from radiolysis of the waste.⁷ In the model, hydrogen gas generation depends on the concentration of sodium (Na⁺), the most abundant cation species in the waste. Thus, cations are not just spectator species but may also influence radiolytic decomposition. Many cations, such as lithium (Li⁺) and Na⁺, are strongly hydrated, orienting the water to alter its structure and dynamics.^8,9 Larger alkali cations such as cesium (Cs⁺) are less strongly hydrated, causing water molecules to bind more tightly with each other.^10,11 The hydroxyl radical, one of several possible intermediates to H₂,^12,13 forms strong hydrogen bonds with water.^14–16 Hence, the nature of cations may alter radical production pathways through their influence on water structure. Although less studied, excited state lifetimes and the formation of transient species other than the hydroxyl radical are likely also influenced by the structuring of water caused by ions in solution.¹⁷ 

Nitrate (NO₃⁻) and nitrite (NO₂⁻) are also radiolytically active in these solutions, and their interaction with hydration waters is potentially important. Under highly concentrated conditions, Na⁺ can form contact ion pairs with NO₂⁻ because of a deficit of water necessary to solvate each ionic species.¹⁸ The interaction of other alkali cations with NO₂⁻ is not studied as extensively, but they are known to change the position and width of the Raman bands for NO₂⁻.^19,20 Although distinct from radiolysis, ion pairing interactions of cations with NO₃⁻ in water have been shown to directly influence its susceptibility to photolysis.²¹ NO₃⁻ has structural similarities to NO₂⁻, so it is possible that ion pairing also influences the photolysis of NO₂⁻. Ion pairing alters the reactivity of other radicals (the hydrated electron) with anions in water.²² Here, the question can be posed: What role does solution structure play in the formation, reactivity, and lifetime of reactive species generated by ionizing radiation in the complex chemical environments representative of radioactive tank waste?

The high concentrations of NO₃⁻ and NO₂⁻ present in the radioactive waste stored at Hanford results in the production of highly reactive NO_x species. These NO_x species are important for understanding the effects of radiation on the waste chemistry due to their reactivity with soluble co-disposed organics.^2,3 NO_x species also scavenge other radiolytic products, such as the hydroxyl radical. Radiolysis is dominated by energy loss via (i) ionization, (ii) direct electronic excitation, and (iii) secondary electron scattering. However, the singlet valence excited states can also be accessed via photoexcitation. If the photon energy is well above the electron affinity, photodetachment can also produce electronic excited states that dissociate into neutral radicals.²³ Through photolysis, radiation can be deposited on a particular ion in solution (i.e., NO₂⁻) through absorption of light at specific wavelength. This allows measurement of radical production as a function of solution composition without having to account for the multitude of reaction pathways and products formed during water radiolysis. The present study examines the photolysis of NO₂⁻ in highly alkaline solutions where we systematically varied the alkali metal cation. The alkali metal cations have different affinities for water and NO₂⁻,^19,20 so a difference in radical production would indicate that cations, and the cation-induced solution structures, influence the reactivity of water, NO₂⁻, or both.

Studies of NO₂⁻ photolysis have mainly pertained to the fundamental mechanisms^27,30 or processes occurring in natural water sources,^26,30–33 with emphasis on relating this behavior to the production and reactivity of ^·OH radical under various conditions. Far more studies have been devoted to the photolysis of aqueous NO₃⁻. NO₃⁻ photolysis yields similar radical species to that of NO₂⁻ photolysis but has additional possible reaction pathways that include NO₂⁻ production.^{25,27,34–36} In NO₂⁻ photolysis studies, the role of temperature,^27,37 ionizing wavelength,³⁰ ionic strength,^24,38 and pH^24,37,38 have been evaluated. Of interest to the present study is radical production in solutions with high ionic strength and high alkalinity, beyond the conditions of most photolysis studies, with emphasis on understanding the role of the alkali cation. At low ionic strengths (e.g., 4.0–7.0 mM sodium sulfate, Na₂SO₄),^{24, ·}OH production is unaffected, but variability in radical production in solutions with molar quantities of dissolved sodium nitrate, carbonate, and hydroxide has been observed.³⁸ It has also been noted that NO₂⁻ photolysis in pure water was 2.5 times faster than in solute-rich seawater,²⁶ indicating that other species can impact NO₂⁻ photolysis. Most studies have focused on neutral pH values, and it has been specifically stated that there was no observed difference in the quantum yield of ^·OH radical between the pH of 5 and 9.²⁴ However, in highly alkaline solutions, there was an observed decrease in ^·OH radical yield.³⁸ This was proposed to be due to the equilibrium between O⁻ and ^·OH shifting to O⁻ at high pH and not necessarily due to a difference in the quantum yield of photolysis products.³⁸ 

Here, we focus on understanding the roles of the cation identity and the solute concentration on the photolysis of NO₂⁻ by monitoring NO^·, ^·OH, and ^·NO₂ radicals using electron paramagnetic spectroscopy (EPR) with nitromethane as the spin trap. We examine the alkali metal series, Li⁺, Na⁺, K⁺, and Cs⁺, for which the ionic radii and charge radii increase monotonically. Solute concentration is increased through the addition of alkali hydroxides, which also serves to raise the solution pH, as well as alkali halides serving to increase the ionic strength of the solution with less impact on the pH. By increasing the solute concentration, more water is dedicated to solvation of the ions present and, eventually, competition for the limited water available in the system promotes the formation of solution structures such as ion pairs.^18,39 This can also lead to the formation of complex ion networks,⁴⁰ which can significantly alter the dynamics and reactivity of a solution.^39,41 Using the alkali cation series, this behavior can be varied systematically as each cation has different average numbers of solvating water molecules, strengths of interaction with water, and ion pairing affinities.^42,43

In situ EPR experiments were conducted to evaluate the role of cation effects on the production of radicals during the UV photolysis of aqueous NO₂⁻ under high ionic strength and high alkalinity solution conditions. The photolysis experiments can be grouped into four categories: (i) NaNO₂ + NaOH compared with KNO₂ + KOH, (ii) 1M NaNO₂ + 2M MOH (M = Li, Na, K, Cs) compared to 0.2M NaNO₂ + 2M MOH (M = Li, Na, K, Cs), (iii) 1M NaNO₂ + 0.5–3.5M MOH (M = Li, Na, K, Cs), and (iv) 1M NaNO₂ + 1M MOH + 0.5–2.3M MCl (M = Li, Na, K). In these experiments, factors that may influence radical production were explored, such as nitrite source, ion pairing, pH, and total ionic strength. Following these experiments, the role of the alkali metal cations on the solution structure was evaluated to explain observed cation effects on the production of radical species.

A series of stock solutions was prepared by dissolving the appropriate amount of NaNO₂ (Sigma-Aldrich, ≥99.0%), KNO₂ (Alfa Aesar, ≥97%), LiOH (Alfa Aesar, 99.995%), NaOH (50 wt. % solution), KOH (45 wt. % solution), CsOH (monohydrate, Alfa Aesar, 99.9%), LiCl (Sigma-Aldrich, ≥99%), NaCl (Sigma-Aldrich, ≥99%), and KCl (Sigma-Aldrich, ≥99%) into 18.2 MΩ.cm de-ionized (DI) water (Milli-Q). The NO₂⁻ and hydroxide solutions were prepared in a N₂-filled glovebox to avoid CO₂ and water adsorption to the reagent stocks. Subsequent sample preparation from the stock solutions and photolysis experiments were performed in air. A 1M stock solution of nitromethane (Sigma-Aldrich, >99.0%) was prepared to use as the spin trapping species (which was diluted to 0.2M in the final solutions). Samples were prepared by mixing appropriate amounts of stock solutions and diluting with 18.2 MΩ.cm DI water as necessary.

Electron paramagnetic resonance (EPR) measurements were performed using a Bruker ELEXSYS E580 spectrometer following the methodology presented in the work of Walter et al.³⁸ A fused silica capillary (VitroCom) with ID 0.8 mm and OD 1 mm sealed with Critoseal was used to hold the solution for EPR measurements. Samples were prepared and loaded into the capillary immediately prior to measurement to reduce the degradation of the nitromethane spin trap. The typical settings for acquiring the spectra were microwave frequency = 9.32 GHz, sweep width = 150 G, sweep time = 10.5 s, power = 20 mW, and field modulation amplitude = 0.5 G.

For the UV photolysis of NO₂⁻, samples were irradiated in situ using a Spectrum 100a UV source (Lesco) equipped with an 18 W mercury vapor lamp, a shutter-controlled switch and timer, an iris, and a 280–400 nm bandpass filter. The light was directed onto the sample using a liquid-filled light guide connected to the light access port on the EPR resonator (Bruker SHQe). For each sample, the EPR spectrometer collected 100 spectra (∼10.2 s per spectrum). Prior to irradiating the sample with UV, one spectrum was collected to serve as a blank. The shutter to the UV lamp was then opened and the sample irradiated for 120 s, at which time the shutter was closed. The EPR spectrometer continued to collect spectra following completion of the 120 s of UV irradiation.

The UV absorptivity was measured for each alkali hydroxide solution to ensure there was no absorption in the 280–400 nm region that would compete with the absorption of NO₂⁻, see the supplementary material, Fig. S1. Based upon these measurements, RbOH was found to absorb UV light in the 280–400 nm region and was excluded from the datasets reported in the main text. The other alkali hydroxides did not show absorbance in competition with the NO₂⁻ 280–400 nm absorbance.

The EPR data were analyzed by performing a least-squares fitting of the experimental spectra using a basis set of spectra for each radical adduct using the methodology listed in the work of Walter et al.³⁸ The coefficients from the least square fitting are proportional to the concentrations of the ^·NO₂, NO^·, and ^·OH radicals (detected as ^·CH₃). It should be noted that though relative concentrations for each species can be obtained, their concentration cannot be compared to one another. For example, ^·NO₂ radical production cannot be compared to ^·OH radical production. Instead, each individual radical type can be evaluated as an additional variable is changed across a series, such as cation identity or solute concentration.

Solutions contained 0.2M nitromethane, which in solutions with at least 0.5M OH⁻, converts to the aci form, where it effectively acts as a spin trap for NO^· and ^·NO₂. To quantify the OH radical, a second series of equivalent samples was prepared to contain 5 vol. % dimethyl sulfoxide (DMSO), where ^·OH radical and DMSO react to form the ^·CH₃ adduct, which can then be trapped by aci-nitromethane.

Though high pH is necessary for conversion of nitromethane to aci-nitromethane, at high pH, nitromethane also dimerizes at nitromethane also dimerizes to form methazonate.⁴⁴ Methazonate does not trap NO^·, ^·NO₂, and ^·OH radicals. Because the solutions in this study contained up to 7M alkali hydroxide, the nitromethane speciation (i.e., nitromethane, aci-nitromethane, methazonate) was evaluated via Raman spectroscopy under experimental solution conditions.

Raman spectroscopy was performed on a Horiba Labram HR spectrometer with a Nikon Ti-E inverted microscope. A 632.81 nm continuous laser light source was focused through a 40× microscope objective. Spectra were collected between 100–4000 cm⁻¹. To evaluate the nitromethane speciation, 120 individual spectra were consecutively collected, with each spectrum taking ∼50 s. Raman spectroscopy was also used to evaluate the spectral features attributed to NO₂⁻ in the 1M NaNO₂ + MOH solutions (M = Li⁺, Na⁺, K⁺, Cs⁺) for ion pairing. Spectra were obtained by averaging 10 scans with 30 s exposures per spectral region and are reported in the supplementary material, Fig. S5.

Nuclear magnetic resonance (NMR) spectroscopy was utilized to determine the local environments and interactions surrounding the dissolved solute species in solution in the absence of UV irradiation and without nitromethane. NMR measurements were performed using a 17.6185 T NMR spectrometer (Agilent/Varian) using a 5 mm broadband probe. At 17.6185 T, the ¹H, ⁷Li, ¹⁵N, ¹⁷O, ²³Na, and ¹³³Cs Larmor frequencies are 750.130, 291.529, 76.039, 101.691, 198.424, and 98.830 MHz, respectively. NMR samples were prepared in an N₂-filled glovebox using samples loaded into fluorinated ethylene polymer (FEP) coaxially inserts capped with a Teflon plug (Wilmad-labglass), placed in an N₂-gas filled 5 mm NMR tube equipped with screw caps (Wilmad-labglass). Unless stated otherwise, all spectra were collected at 20 °C. Results for ¹⁵N and ¹⁷O are presented in the subsequent text and those for ¹H, ⁷Li, ²³Na, and ¹³³Cs are presented in the supplementary material, Figs. S6-S9, Tables 3-5.

Single pulse, direct excitation ¹⁵N NMR spectra were collected with an acquisition time of 1.678 s enumerated with 131 072 complex points, a recycle delay of 5 s, a sweep width of 78 125 Hz, and 7168 transients. The pulse width of the ¹⁵N NMR spectra was 15.5 µs, equivalent to a π/4 pulse width. The chemical shift of the ¹⁵N NMR spectra and the pulse width were calibrated using an autosampler test sample composed of 0.1% ¹⁵N enriched acetonitrile, 0.1% ¹³C enriched MeOH, and 1% H₂O dissolved in D₂O (99.8% D) where the resonance was assigned to 130.2 ppm. Post-acquisition processing was performed in Mestrenova, where the ¹⁵N NMR spectra were zero-filled to 262 144 points and 2 Hz of exponential line broadening was applied.

Single pulse, direct excitation ¹⁷O NMR spectra were acquired with an acquisition time of 20 ms enumerated with 16 667 points, a spectral width of 833 333 Hz, a recycle delay of 0.2 s, and a π/2 pulse width of ∼30 µs, which was calibrated for each sample. The chemical shifts of the ¹⁷O NMR spectra are spectra are referenced to the chemical shift of neat H₂O (18.2 MΩ.cm), which was assigned to 0 ppm. The spectra were processed in Mestrenova, where the free induction decay was left shifted −40 points, zero-filled twice to 65 536 complex points, and 100 Hz of exponential line broadening was applied.

Radical production is impacted by competition between UV absorbing species and the aqueous NO₂⁻. Based upon the UV absorption measurements, the RbOH stock solution was found to absorb in the UV region of 280–400 nm, see the supplementary material, Fig. S1. Competition for UV between the RbOH and the NO₂⁻ resulted in the decreased radical quantities observed in the supplementary material, Fig. S2.

The effectiveness of the spin trap determines the reliability of the detection of radicals using EPR spectroscopy. Under high pH conditions, nitromethane converts to aci-nitromethane, enabling trapping and detection of radicals, see the supplementary material, Fig. S4. Under the same conditions, aci-nitromethane then dimerizes to form the methazonate ion, inhibiting radical trapping and quantification of radical concentrations. The timelines for this transformation were evaluated using Raman spectroscopy in solutions of 0.2M nitromethane + 1M NaNO₂ + 0.5, 3.5, and 7M NaOH, shown in Fig. 2. Two features emerge at ∼1170 and ∼1550 cm⁻¹, indicating dimerization of nitromethane, forming methazonate, see the supplementary material for further discussion regarding peak identification.

The maximum intensity for the peaks emerging at 1170 and 1550 cm⁻¹ in Fig. 2 were determined and plotted as a function of time in Fig. 3. Assuming the intensity of the 1170 and 1550 cm⁻¹ peaks increases linearly with conversion of nitromethane to methazonate, and that a peak intensity of ∼300 on Fig. 3 corresponds to the maximum population of methazonate, i.e., 100%, then after 15 min, ∼25% of the total methazonate has formed in solutions of 0.5M NaOH, ∼40% in solutions of 3.5M NaOH, and ∼50% in solutions of 7M NaOH. Both ^·NO₂ and ^·OH (as the ^·CH₃ adduct) are quantified during the first 120 s of the experiment while the UV light is on when their signals are at a maximum, ^·NO reaches its maximum several minutes following the UV lamp’s shutter closure.³⁸ It is likely that there is enough nitromethane in solution to trap radicals under the solution conditions investigated here, especially for radicals trapped and detected early in the experiment. Even if spin trap degradation influences how many radicals are trapped, comparison between datasets with equivalent concentrations of OH⁻ will still give a meaningful evaluation.

To evaluate the effect of alkali metal cation identity and hydroxide concentration on the evolution of radicals from the UV photolysis of NO₂⁻, solutions containing either 1M NaNO₂ + 1–7M NaOH or 1M KNO₂ + 1–7M KOH were studied. From this series, solutions containing K⁺ evolve greater radical quantities than those containing Na⁺ (Fig. 4). The production of NO^· and ^·NO₂ is maximized at hydroxide concentrations of 1.5–2.5M, or a total cation (either Na⁺ or K⁺) concentration of 2.5–3.5M. Beyond these concentrations, the radical quantities decrease. The ^·OH radical production decreases across all concentrations of added alkali hydroxide.

These initial results displayed a clear difference between sodium and potassium containing solutions prompting new experiments with other alkali cations. It should be noted that Rb⁺ was excluded from these experiments as all available sources of RbOH visibly colored the prepared solutions, and they were found to absorb in the UV, see the supplementary material, Fig. S1. A hydroxide concentration of 2M MOH was chosen as this was the approximate amount that yielded the maximum detected NO^· and ^·NO₂ radicals in prior experiments (Fig. 4). Figure 5(a) shows the radical evolution in solutions of 1M NaNO₂ + 2M MOH, with Li⁺ < Na⁺ < K⁺ a clear trend; however, it does not extend to Cs⁺. These experiments were then repeated but at a much lower nitrite concentration, Fig. 5(b), to explore the role of ion pairing. Ion pairing was also explored through Raman spectroscopy, see the supplementary material, Fig. S5. At 0.2M nitrite, any effect from ion pairing would be greatly reduced, as would secondary reactions involving nitrite such as those illustrated in Fig. 1 and in Eq. (5). Instead, the effect of the cation is even more evident. It is noted that the ratio of products between the two experiments is not merely proportional to the ratio of nitrite—since the first step involves absorption of a photon, it follows a Beer’s law (logarithmic) dependence—a complete analysis is included in the supplementary material.

When comparing radical production as a function of cation species within each dataset, the 1M NaNO₂ solution series shows that NO^· and ^·NO₂ radicals nearly double going from Li⁺ to K⁺, Cs⁺ [Fig. 5(a)]. The ^·OH radical production remains nearly constant but shows a slight variation in radical evolution with Li⁺ and Na⁺ being approximately equivalent and K⁺ showing the largest ^·OH radical production. The 0.2M NaNO₂ + 2M MOH series shows a clear cation trend in NO₂ evolution, following Li⁺ < Na⁺ < K⁺ < Cs⁺ [Fig. 5(b)]. The ^·NO₂ radical evolution doubles from Li⁺ to Cs⁺. There is far less variability in the ^·OH radical and NO^· radical production with cation in the 0.2M NaNO₂ series. Here, K⁺ has a slight increase in production compared to the other cation species. From these two datasets, it can be concluded that the most pronounced cation effect is in the ^·NO₂ radical production. This is also the first evidence that radical yields for Cs are higher than K for low concentration but lower than Na at high cation concentration.

Figure 6 shows the radical evolution in solutions with a range of alkali hydroxide concentrations (0.5–3.5M) and 1M NaNO₂. The trends in Fig. 6 are consistent with those in Fig. 4, with the maximum NO^· and ^·NO₂ radical production at ∼1.5–2.5M hydroxide concentration, and ^·OH radical production decreasing with increasing hydroxide (solute) concentration. The radical evolution follows Li⁺ < Na⁺ < K⁺ with Cs⁺ falling between Na⁺ and K⁺, depending on the radical species. There appears to be a concentration effect in Fig. 6, as in Fig. 4, but it is difficult to determine if this is an artifact of the spin trapping in high pH solutions or if there is an actual cation concentration effect. The use of a non-hydroxide alkali cation source allows for the pH effect to be eliminated while simultaneously investigating the concentration effect.

Figure 7 shows the radical evolution in solutions of 1M NaNO₂ + 1M MOH + 0–2.3M MCl, where M⁺ = Li⁺, Na⁺, or K⁺. In this series, a pH effect is avoided as all solutions have the same concentration of OH⁻ and the ionic strength is increased through the addition of alkali chloride salts. This has the added benefit that the cation effect can be interrogated without the strong interaction between water and OH⁻. The interaction between water and Cl⁻ is slightly stronger than a water–water interaction but weaker than that of water–OH⁻.⁴⁶ In the experiments with alkali chlorides, the cation effect is very evident: Li⁺ < Na⁺ < K⁺. Comparing the NO^· and ^·NO₂ radical datasets to those in Fig. 6, there does not appear to be a maximum in radical production, revealing that the increase and, then, decrease in NO^· and ^·NO₂ radical production shown are artificial and likely a pH effect. In the chloride data in Fig. 7 and the hydroxide data in Fig. 6, there is a concentration effect in ^·OH radical production; it decreases as a function of solute concentration, this is not a pH effect on the equilibrium conversion from O⁻ to ^·OH radical. There is also a concentration effect in the ^·NO₂ radical production in Fig. 7, which was not observed in the hydroxide series. In Fig. 7, the ^·NO₂ radical production amounts diverge, with K⁺ solutions increasing, Na⁺ solutions remaining relatively constant, and the Li⁺ solutions decreasing, indicating that the effect of the cation is becoming more pronounced as the overall concentration is increased relative to the NO₂⁻ in solution.

Based upon the data presented in Figs. 4–7, it can be concluded that there is a cation effect on the radical production from the UV photolysis of NO₂⁻. There are no cation species involved in the photolysis mechanism or reaction pathways previously reported.²⁵ To determine the nature of the cation-controlled solution structures, and the potential effects they may have on the dynamics of radical production during photolysis of NO₂⁻, spectroscopic analysis was performed on a subset of the solutions without nitromethane and DMSO, and in the absence of UV exposure.

¹⁵N NMR spectroscopy was used to investigate the local structure around NO₂⁻ (Fig. 8). The ¹⁵N NMR nucleus is a spin 1/2 nuclei with a low natural abundance of ∼0.4%. Nitrogen has a lone pair of electrons and is sensitive to changes in local structure, which affects shielding and the chemical shift.⁴⁷ The chemical shift of the single observed resonance depends on the identity of the alkali cation and increases as a function of the ionic radius, with Li⁺ and Na⁺ being more similar than K⁺ and Cs⁺. The line shape parameters corresponding to the fits depicted in Fig. 8 are shown in Table I.

Previous ¹⁵N NMR studies in aqueous NO₂⁻ solutions attributed the increase in chemical shift (i.e., deshielding) upon increasing NaNO₂ concentration, and upon addition of NaOH, to changes in the bulk dielectric constant, changes in hydrogen bonding (H-bonding), and ion pairing.¹⁸ An increase in chemical shift/decrease in shielding around nitrogen nuclei may be indicative of the cation effect on the strength of the interaction between the solvating water and NO₂⁻. A deshielded nitrogen nucleus in NO₂⁻ that has weaker interactions with its solvating water may (i) undergo photolysis more readily to produce NO^· and ^·OH and (ii) promote the reaction between ^·OH radical and NO₂⁻ to produce ^·NO₂. Production of ^·NO₂ shows the most noticeable cation effect, suggesting that cations play a role in increasing the electronegativity around NO₂⁻, which decrease its ability to give up an electron and be oxidized by ^·OH radical.

The ¹⁷O NMR resonance assigned to the NO₂⁻ oxyanion can give addition information on the local NO₂⁻ structure. In Fig. 8(b), the ¹⁷O data show a similar cation dependence to that for ¹⁵N, with the chemical shift increasing as a function of the ionic radius. Radical production decreases as a function of ionic radius, suggesting that this process is inhibited by cations, e.g., Li⁺, that increase the electron density around NO₂⁻, making the interaction between NO₂⁻ and its local solvation shell stronger. The ¹⁷O NMR data show that the interaction between Li⁺ and water is stronger than that for the larger cations in the series; thus, Li⁺ effectively removes water from the bulk solution and promotes more rigid, less dynamic solution structures.

The ¹⁷O NMR nucleus is a spin 5/2 nucleus with a natural abundance of 0.038%. The chemical shift of the resonance near 0 ppm is assigned to oxygen in water and hydroxide molecules (Fig. 9), and the oxygen within those molecules is in chemical exchange, giving rise to only one resonance. The ¹⁷O resonance assigned to NO₂⁻ in Fig. 8 and the ¹⁷O NMR resonance assigned to water and hydroxide exhibits a chemical shift that is dependent on the identity of the cation. As the cation radius increases, the resonances appear at greater chemical shifts in a monotonic trend, indicating deshielding of the ¹⁷O nuclei. The line shape parameters corresponding to the fits depicted in Fig. 9 are shown in Table II.

PFG-NMR diffusometry was employed to quantify the diffusivity of the ¹H-bearing molecules and cations including ⁷Li, ²³Na, and ¹³³Cs at the millisecond timescale. The diffusion coefficients are listed in Table III. The diffusivity of ¹H increases in the order of Li⁺ ∼ Na⁺ < K⁺ < Cs⁺, and the diffusion coefficient of ²³Na also increases in the order of Li⁺ ∼ Na⁺ < K⁺ ∼ Cs⁺. In the LiOH solution, the diffusivity of the ⁷Li ions is much lower than the diffusivity of the ²³Na ions. In the CsOH solution, the diffusivity of the ¹³³Cs ions is much faster than the diffusivity of the ²³Na⁺ ions.

Many studies have shown that the diffusion rates of water, hydroxide, and other ions increase in alkali–electrolyte solutions as the radius of the alkali increases.^48–56 In the present study, the diffusion rate of both the proton and alkali increases as the size of the alkali in the electrolyte solution increased (Table III), consistent with the results reported by many earlier studies with other alkali-bearing electrolyte solutions. This is believed to be because the alkali bind water more strongly as the ionic size goes down because the charge density is higher as the radius goes down.^49,51,56 The water molecules that are bound to the alkali diffuse slower than the free water, so the average diffusion rate of the water goes down when the alkali bind water more strongly.^51,52,55 The ions that bind water strongly also diffuse slower than those that do not because they must break stronger ion–water bonds to diffuse.⁵⁶ The faster diffusion rate of species in the present study would support the formation of secondary radical species whose formation is diffusion limited. The ¹⁷O NMR data are consistent with weaker binding of water molecules as the alkali ion radius increases, consistent with the theory that diffusion rates are slowed by strong binding of water by ions. This diffusion data and ¹⁷O NMR data together are consistent with the conclusion that the smaller alkali bind water more strongly, which may influence the rate of radical production beyond just its impact on diffusion rates.

Based on the cation effect on radical production from NO₂⁻ photolysis, and the complementary NMR results, we hypothesize that cation-induced local solvation structures surrounding the NO₂⁻ anion and bulk solution structures are key to controlling radical production. This local structure influences the initial photolysis of NO₂⁻ to produce NO^· and O^·− (^·OH) radicals but has a greater impact on the later reaction of ^·OH radicals with NO₂⁻, to form ^·NO₂ radicals. Here, we evaluate the cation effect on the reactions leading to the formation (and trapping) of NO^·, ^·OH, and ^·NO₂ radicals.

In the first step of nitrite photolysis, the NO₂⁻ anion, surrounded by its solvation shell, absorbs UV light to become NO₂⁻*.^25,30 Depending upon the strength of the interaction and rigidity of the solvating water, this solvent shell can be considered a solvent cage.³⁰ The solvation geometry likely impacts the initial UV absorption by NO₂⁻, as it has been shown to do in the case of UV absorption by NO₃⁻.⁵⁷ The NO₂⁻* can then undergo photofragmentation to produce NO^· and O⁻ or can return to NO₂⁻, releasing heat in the process.³⁰ 

Figure 10 represents how the strength of interaction with the solvation shell may influence the production and subsequent trapping of these species. In the case of a rigid solvation shell surrounding the NO₂⁻ anion [Fig. 10(a)], a caged radical pair is formed upon UV absorption and photofragmentation, and the radicals subsequently undergo recombination. Due to the change in charge of the solvated species, i.e., negatively charged NO₂⁻ to neutral NO^· and negatively charged O⁻ radicals, a change in solvation geometry and orientation of the water molecules is likely necessary to accommodate these species. For the local solvent rearrangement to occur, an energy barrier must be overcome. This could be influenced by the way in which the cations control the structure of water and how strongly they bind water. In the case of kosmotropic cations, e.g., Li⁺_, this energy barrier for rearrangement cannot be readily overcome. In solutions containing chaotropic cations, e.g., K⁺, the energy barrier can be overcome, and the solvation shell undergoes a structural rearrangement such that the NO^· and O⁻ radicals react with other species in solution, or be trapped by nitromethane, as represented in Fig. 10(b).

In addition to the strength of solvation and the structuring of water, we must also consider the amount of water available in the system. As was observed in the photolysis of NO₃⁻, an incomplete solvation shell can promote photolysis and the relative yields of radical species.³⁴ The amount of “free” water available in the system (Table IV) depends on the amount of water that is required to fully solvate the cations and anions present (Table V). There is less free water in systems with higher solute concentrations and in systems containing cations with larger solvation numbers. As the amount of free water in the system becomes scarce, the ions must either form contact ion pairs or lower their solvation number, i.e., removing one or more water molecules from the first solvation shell. Ion hydration enthalpies can be used to estimate which ions will be forced to lower their solvation number, or share solvating water molecules with counter ions, e.g., in solvent shared ion pairs (Table V). Hydration enthalpy decreases with increasing cation size; therefore, as free water becomes scarce in the system, Li⁺ is less likely to give up its solvating water compared to other cations, e.g., K⁺ and Cs⁺. Although free water is mostly consumed through the solvation of K⁺ and Cs⁺ (Table IV), it is less tightly bound and more likely to leave the solvation structures or exchange with bulk water. Thus, the solvating water of K⁺ and Cs⁺ is more available to participate in kinetically driven reactions and in the subsequent diffusion of short-lived, reactive radical species. These results demonstrate the complex interplay between kinetics and thermodynamic driving forces, often described by linear free-energy relationships.⁵⁸ Thermodynamics cannot completely describe the behavior of the system, pointing to the complex interplay of water availability, solvation structure, and solution dynamics in these systems.

Solution concentration, solvation structures, and the relative availability of water must all be considered in the formation of the ^·OH radical. To form the ^·OH radical, O⁻ must react with surrounding H₂O. Not only does the neutral ^·OH radical form, but the OH⁻ anion is also formed. This too requires a rearrangement of the local solvation structure, similar to that represented in Fig. 10(b), otherwise recombination of these species may occur.³⁰ The production of ^·OH radicals decreases with increase in both pH and ionic strength. This process is an equilibrium-controlled reaction [Eq. (2)], with ^·OH radical production favored at lower pH, and O⁻ radical production favored at higher pH (Figs. 4 and 6). In Fig. 7, where the pH is constant, ^·OH radical production decreases as a function of ionic strength because there are less available water molecules in the system for the O⁻ radical to react with and form the ^·OH radical. This reaction will also be inhibited if the cation constrains the solvent shell such that there is less bulk water available in the system.

Disruption or rearrangement of the solvation shell allows the ^·OH radical to diffuse through solution and react with remaining NO₂⁻ to form the ^·NO₂ radical.²⁵ To produce ^·NO₂, there are three primary processes that must be considered. First, the ^·OH radical must escape its local solvation shell, and diffuse through solution, which is influenced by the cation-controlled solvent structure. Second, the solvating water must enable reaction between the approaching ^·OH radical and the solvated NO₂⁻. Third, the NO₂⁻ must be oxidized by ^·OH/O⁻ radicals to produce ^·NO₂, and this reaction could reasonably be controlled by the local environment and chemistry of the NO₂⁻ in solution.⁴⁵ Thus, the nature of cation impacts the electron transfer from NO₂⁻ to the ^·OH or O⁻ radical. As ionic strength increases, the formation of shared solvation shells, contact ion pairing, or even a decrease in coordination affects the stability of NO₂⁻ complex toward oxidation. This trend is shown in Fig. 7, where the ^·NO₂ radical production decreases with increasing LiCl (no pH effect). NaCl shows only a slight decrease in ^·NO₂ radical production and KCl shows an increase in ^·NO₂ radical production, implying that the prevention of oxidation is only concentration dependent in cations that strongly bind to water. KCl, by increasing the ^·NO₂ radical production, promotes this oxidation reaction.

We have demonstrated that there is an alkali cation effect on radical production during the NO₂⁻ photolysis in concentrated alkaline nitrite solutions. K⁺ and Cs⁺ promote the formation of radicals while Li⁺ decreases the radical formation. This is especially true for the ^·NO₂ radical yield, which is dependent on the initial photolysis radical yields and their subsequent diffusion and reactivity of these species in solution. This effect is likely due to the bulk solution organization and the changes in local NO₂⁻ structure that are induced by the different cations. These effects then further control the overall reactivity of NO₂⁻ toward oxidation by ^·OH radical. We propose that cations with strong water interactions reduce the reactivity of the system, impacting the yields of radicals produced during the UV photolysis of NO₂⁻.

This cation effect implies that the initial photolysis of NO₂⁻ and generation of radicals is controlled by short-range solute–solvent interactions. The cation effect on ^·OH and NO₂⁻ reactivity to form ^·NO₂ shows that there could be contributions from both short-range and long-range solution interactions. Short-range interactions would impact the reactivity of radicals and the solute stability. The longer-range solution interactions would control the organization of bulk water and the diffusion-controlled processes of the radical and solute species. Though the radical species investigated in this study were generated through photolysis of NO₂⁻, rather than through radiolysis, ^·OH is formed in both cases. Because the reactivity of the ^·OH radical is altered by the identity of the solute species in this photolysis study, it is likely that the same would be true in the equivalent system where ^·OH radicals were generated via radiolysis. Though the known reaction pathways do not directly involve or incorporate the dissolved ions in solution, their impacts on solution structure alter the overall yields and reactivity of radical species. In Hanford nuclear waste, solutions can contain concentrations of up to 11M dissolved sodium salts (nitrate, nitrite, hydroxide, aluminate, carbonate, etc.) and are often at the limits of solubility.⁴ It is reasonable to expect that solute-controlled short-range and long-range solution interactions impact the effects of ionizing radiation on water by altering the available reaction pathways, especially as the next nearest neighbor to the radiolytically generated reactive species may not necessarily be a water molecule.

In the present work, we have proposed the mechanisms by which cation-induced solution structures influence radical production and reactivity under high ionic strength and alkaline conditions. To obtain a comprehensive understanding of these pathways, additional studies are needed. We suggest the use of time-resolved ultrafast optical or x-ray pump-spectroscopic techniques to detect the initial formation of NO^· radical, lifetimes of ^·OH radicals, and the reaction between NO₂⁻ and ^·OH radicals to form ^·NO₂ radicals in alkali nitrites. Future work would benefit from the application of a theory component to evaluate the reactivity of NO₂⁻ with various solvating species and configurations, especially as it pertains to the reactivity toward the ^·OH radical. The fundamental investigations presented here, as well as the proposed future work, have implications for the reactivity and lifetimes of radicals in a variety of applications, including the processing of nuclear wastes that have been exposed to decades of ionizing radiation. The high solute concentrations combined with ionizing radiation have driven the waste far from equilibrium, where radiolytic processes alter the relative concentrations of NO₃⁻ to NO₂⁻, degrade co-dissolved organics promoting H₂ gas generation, and influence tank corrosion.^6,61 The ways in solution structures dictate the reactivity and dynamics of radiolytic products cannot be ignored in these complex chemical systems.

See the supplementary material for further discussion on factors impacting UV absorption and photolysis and factors impacting radical detection, and exploring solution structure through Raman and NMR spectroscopies.

This research was supported by IDREAM (Interfacial Dynamics in Radioactive Environments and Materials), an Energy Frontier Research Center funded by the U.S. Department of Energy (DOE), Office of Science, Basic Energy Science (Grant No. FWP 68932). Electron paramagnetic resonance spectroscopy and Nuclear Magnetic Resonance spectroscopy were performed using facilities at the Environmental Molecular Science Laboratory (Grant No. grid.436923.9), a DOE Office of Science User Facility sponsored by the Office of Biological and Environmental Research at Pacific Northwest National Laboratory (PNNL). PNNL is a multiprogram national laboratory operated for DOE by Battelle Memorial Institute operating under Contract No. DE AC05-76RL0-1830. The authors thank David Bazak for assistance with the solution density measurements and Gregory Schenter for discussions on ion hydration.

The authors have no conflicts to disclose.

Emily T. Nienhuis: Conceptualization (equal); Data curation (equal); Formal analysis (equal); Investigation (equal); Visualization (equal); Writing – original draft (equal); Writing – review & editing (equal). Trent R. Graham: Conceptualization (equal); Data curation (equal); Formal analysis (equal); Investigation (equal); Methodology (equal); Visualization (equal); Writing – original draft (equal); Writing – review & editing (equal). Nicolas L. D’Annunzio: Data curation (equal); Investigation (equal). Malgorzata I. Kowalska: Data curation (equal); Investigation (equal); Writing – review & editing (equal). Jay A. LaVerne: Conceptualization (equal); Formal analysis (equal); Writing – original draft (equal); Writing – review & editing (equal). Thomas M. Orlando: Conceptualization (equal); Writing – original draft (equal); Writing – review & editing (equal). Jacob G. Reynolds: Writing – original draft (equal); Writing – review & editing (equal). Donald M. Camaioni: Formal analysis (equal); Writing – original draft (equal); Writing – review & editing (equal). Kevin M. Rosso: Supervision (equal); Writing – original draft (equal); Writing – review & editing (equal). Carolyn I. Pearce: Conceptualization (equal); Formal analysis (equal); Supervision (equal); Writing – original draft (equal); Writing – review & editing (equal). Eric D. Walter: Conceptualization (equal); Data curation (equal); Formal analysis (equal); Visualization (equal); Writing – original draft (equal); Writing – review & editing (equal).

The data that support the findings of this study are available from the corresponding authors upon reasonable request.