Iron dissolution and speciation from combustion particles under environmentally relevant conditions

Material sources

BA was obtained from SC ‘Akmenės Energija’, Venta, Lithuania, where mixed biomass waste was used as fuel. LKD was obtained from JSC Naujasis Kalcitas, Naujoji Akmene, Lithuania. Samples were collected from the top of the ash pile. To avoid any large fractions of residual non-combusted biomass or partially combusted char, samples were initially processed for the laboratory experiments by sieving through a 0.063-mm mesh sieve. Bulk metal composition analysis of the combustion powder samples was carried out using X-ray fluorescence (XRF) empoying a Bruker Tracer III SD spectrometer. Accurate measurement of total iron content in BA and LKD was carried out via atomic absorption spectroscopy (AAS) using a PerkinElmer AAnalyst 800 spectrometer; approximately 0.1 g of combustion samples were acid digested in a mixture of 5 mL HNO₃ and 3 mL H₂O₂ (Borgatta et al. 2016; Kim et al. 2020) for analysis. The mineralogy of the samples was investigated using X-ray diffraction (XRD, Rigaku) and infrared spectroscopy (FTIR, Perkin Elmer). Vibrational scans were collected with a single beam Perkin-Elmer FTIR spectrometer, equipped with a ZnSe ATR element and a DLaTGS/KBr detector. Typically, samples were vacuum dried overnight at 373 K before analysis with 100 scans averaged at a resolution of 4 cm⁻¹ over the full spectroscopic range extending from 800 to 4000 cm⁻¹.

X-Ray photoelectron spectroscopy (XPS) spectra of the BA and LKD surface were collected with a SPECS Near Atmospheric Pressure X-ray Photoelectron Spectroscopy (NAP-XPS) system. The NAP-XPS system was equipped with an XR 50 MF Al Kα X-ray source with a μ-FOCUS 600 X-ray monochromator. Al K_α radiation was used with an X-ray beam energy of 1486.7 eV and power of 120 W. A PHOIBOS NAP in situ 1D-DLD hemispherical electron energy analyser (~0.85 eV energy resolution and 1 mm entrance aperture) collected the spectra. The pressure during the entire NAP-XPS experiment was always kept at ~10⁻⁸ mbar. A pass energy of 100 eV was utilised for survey spectra and 20 eV for high-resolution spectra. Data were processed using Casa XPS software (Fairley et al. 2021). Finally, the surface areas for all combustion samples were determined using an eleven-point N₂-BET adsorption isotherm, acquired with a Quantachrome Nova 1200 surface area analyser. The samples were dried under vacuum overnight before the surface area measurement.

Dissolution experiments

The acidic processing of the combustion particles was simulated in a 300 mL custom-built jacketed beaker with an airtight top at 298 K (Borgatta et al. 2016; Kim et al. 2020). Dissolution of iron and speciation analysis were performed on suspensions of 1 g L⁻¹ of the combustion particles in constantly stirred solutions acidified with hydrochloric acid. As discussed before, the acid deliquescence layer that mobilises iron from combustion particles during atmospheric transport can reach pH values between 1 and 4 (Usher et al. 2003), with a pH of around 4 also representing the extreme acidity in polluted ash ponds (Kabisch and Hemmerling 1982). Thus, in the suspension experiments the pH was controlled at 1.0 ± 0.1, 2.0 ± 0.1, 3.0 ± 0.1, and 4.0 ± 0.1. The pH was kept constant by adding microliters of concentrated HCl when needed with negligible changes in volume and concentration of dissolved iron. In a typical experiment, aliquots were periodically taken from the reaction vessel and filtered through 0.2 μm pore size polytetrafluoroethylene filters (PTEE) for iron analysis. Although most available iron partitions to the aqueous phase within the first 300 min, all suspension experiments were carried out for a minimum of 50 h to ensure that equilibrium had been reached.

To prevent oxidation of Fe²⁺ by dissolved oxygen once leached from the combustion samples, all dissolution experiments were carried out in an oxygen-free atmosphere achieved with a continuous flow of 5 sccm of nitrogen above the solution surface (Borgatta et al. 2016; Kim et al. 2020). Before the suspension experiment, the acidic solution was deoxygenated by bubbling nitrogen for 15 min. This oxygen-free environment allowed for a better quantification of both Fe³⁺ and Fe²⁺ leached from the combustion samples during the suspension experiments. In addition, to control the variations in ionic strength in the suspension solution as the dissolution takes place, all acidic solutions were adjusted to 1.0 M NaCl ensuring a constant ionic strength all through the experimental time. Iron speciation and quantification started at the moment of sample loading into the acid solution, defined as t = 0 min. As reported in previous work, dissolved iron species were quantified using 1,10-phenanthroline, which forms an orange complex with Fe²⁺ with an absorbance band at 510 nm (Borgatta and Navea 2015; Borgatta et al. 2016; Kim et al. 2020). Total dissolved iron was quantified in the same samples by adding hydroxylamine to reduce all Fe³⁺ to Fe²⁺ before phenanthroline complexation. The absorbance measured via the colorimetric method was converted into concentrations using aqueous standards prepared from anhydrous ferrous chloride (Sigma-Aldrich). All colorimetric measurements were performed in triplicate using a Lambda 35 Perkin-Elmer UV/Vis spectrophotometer.

Characterisation of combustion samples

Table 1 shows the major elements detected via XRF. Overall, the bulk iron in LKD and BA was determined to be 14.9 ± 0.8 and 12.4 ± 0.5 mg g⁻¹, respectively. While iron content in both samples is relatively similar, it is comparatively low with respect of fly ash samples examined in previous work (Navea et al. 2010; Borgatta et al. 2016; Kim et al. 2020).

In Fig. 1, the XRD patterns show structural and compositional similarities between the two samples. Both LKD and BA show a fraction of mullite (3Al₂O₃·2SiO₂), typically formed during the combustion of aluminosilicate components in fuels (Navea et al. 2010; Borgatta et al. 2016; Kim et al. 2020; Drapanauskaite et al. 2021). In addition, diffraction patterns due to the presence of calcium feldspars, quartz (SiO₂), and traces of hematite (Fe₂O₃) and magnetite (Fe₃O₄) can be observed. Finally, a clear diffraction pattern for calcite (CaCO₃) is observed in both samples. Full combustion, observed in samples collected in power plant stacks or emitted directly into the atmosphere (Navea et al. 2010; Borgatta et al. 2016; Kim et al. 2020), convert all calcite into calcium oxide, removing carbonate and bicarbonate minerals from the sample. The presence of carbonate minerals in the samples has important implications for the environmental processing of these samples. These minerals will react with the acid deliquescent layer characteristic of processing during atmospheric transport, or less acidic media in water bodies, underground water, or pond reservoirs (Basu et al. 2009; Borgatta et al. 2016; Koshy and Singh 2016; Navea et al. 2017; Kim et al. 2020; Chowdhury et al. 2022). While both samples are thoroughly combusted, there are some indications of partially combusted fractions. This observation of partially combusted particles is in good agreement with the scanning electron microscopy (SEM) analysis, which shows amorphous rock-like particles in the samples. BA showed more rock-like structures than LKD, which shows some spherical particles formed during a highly efficient combustion process. We interpret this as an indication of BA having a slightly higher fraction of partially combusted particles, with most of the sample completely combusted.

Fig. 1. 

Left: Representative micrographs of combustion samples. (a) Lime kiln dust (LKD); (b) bottom ash (BA). Right: X-ray diffraction characterisation of combustion samples. (a) LKD; (b) BA. XRD legend: C, calcium carbonate; CH, calcium hydroxide; CS, calcium silicate; Q, quartz; M, mullite; H, hematite/magnetite; F, feldspar.

The presence of mullite and calcite was confirmed via total attenuated reflectance–Fourier transform infrared spectra (ATR-FTIR) of samples, as shown in Fig. 2. As in the case of the XRD analysis (Fig. 1), FTIR spectra show similarities between LKD and BA. Both samples show a noticeable band around 1410 cm⁻¹ which is characteristic of carbonates, in agreement with XRD data shown in Fig. 1. Carbonates and bicarbonates are important components in these samples as they react with acid and further breakdown particles, favouring the solubility by increasing surface area contact (Navea et al. 2010, 2017; Galhotra et al. 2009; Kim et al. 2020). Additionally, a sharp OH peak seen at 3647 cm⁻¹ suggests isolated OH terminal groups in the dry samples. The bands centred at 1118 and 1042 cm⁻¹ are assigned to vibrational absorption bands for the lattice stretching mode of Si–O (Borgatta et al. 2016; Navea et al. 2017; Galhotra et al. 2009), in agreement with the XRD bands of mullite and quartz reported in Fig. 1. While the composition of LKD and BA are similar, the band at 1042 cm⁻¹ is only observed in the LKD sample, indicating some small differences in the mineralogy of the samples. Finally, the sharp band centred at 877 cm⁻¹ is assigned to deformation modes of Fe³⁺ within the aluminium silicate structure, δ(FeAl–OH) (Navea et al. 2010, 2017; Borgatta et al. 2016). In general, elemental and chemical analysis data suggest that both samples are relatively similar in composition.

Fig. 2. 

(Top) Attenuated total reflection–Fourier transform infrared (ATR-FTIR) spectra showing LKD (a, red) and BA (b, blue) samples. The inset shows a magnified view of the spectroscopic region from 2000 to 750 cm⁻¹. (Bottom) Survey and high-resolution XPS spectra of LKD and BA sample surfaces. High-resolution XPS spectra show that the sample surface of both LKD and BA is primarily comprised of C, Ca, and O, whereas high-resolution spectra also show the presence of other elements, such as K, Si, Mg, and Al with a notable S 2p peak. A peak at 289.3 eV in the C 1s region is due to calcium carbonate, a combustion product in BA (Baltrusaitis et al. 2007).

XPS compositional analysis revealed that the sample surface has significant amounts of calcium and oxygen. A marked difference was the presence of potassium in BA, in agreement with XRF analysis. Carbonates were detected in C 1s at 289.9 eV signifying a CaCO₃ structure (Usher et al. 2007). Traces of other elements, including sulfur, silicon, magnesium, and aluminium, were detected. Ultimately, XPS results suggest particle surfaces were dominated compositionally by calcium carbonates.

The surface areas for LKD and BA were found to be 6.3 ± 0.8 and 3.2 ± 0.6 m² g⁻¹, respectively. Both samples were sieved to select a maximum size of 35 µm. Key information on specific surface area and total bulk iron content in each ash sample is summarised in Table 2.

Iron dissolution

As discussed in the Experimental section, two different combustion powder samples, LKD and BA, were used to determine iron dissolution into acidic media. Fig. 3 shows the fraction of Fe species dissolved from LKD and BA from pH 1.0 to 4.0, relative to the amount of iron (both Fe²⁺ and Fe³⁺) as reported in Table 1.

Fig. 3. 

Fraction of Fe species dissolved from combustion samples at pH = 1.0 (blue circles), 2.0 (red circles), 3.0 (purple circles), and 4.0 (green circles). Left panels: LKD; Right panels: BA. Top graphs represent total iron (Fe^total = Fe²⁺ + Fe³⁺) leaching in HCl suspensions; middle graphs represent Fe³⁺; bottom graphs represent Fe²⁺.

While both combustion samples are relatively similar in mineralogy and particle size, the fraction of iron leached from BA was slightly lower than that for LKD. The fraction of total iron leached from LKD, as shown in Fig. 3, is roughly 15% at pH 1, compared to 12% for BA. At pH 2, the fraction of iron leached from both samples was statistically similar. Above pH 3, the solubility of Fe³⁺ from LKD approaches the limit of detection of the method (5 ppb), while Fe³⁺ leached from BA is also smaller but clearly above the limit of detection, with a slow solubility through the time of the experiment. Finally, Fig. 3 shows that the dissolution of iron into the aqueous phase takes place within 5–20 min of suspension at pH < 2, followed by passivation or slower dissolution. These two regimes are consistent with previous observations on iron dissolution from combustion particles in acidic media, where most available iron dissolves by a fast pathway (Borgatta et al. 2016; Kim et al. 2020).

This initial rate of dissolution, extracted from a linear fit from t = 0 to the passivation time (5–20 min) of the plots in Fig. 3, is measurable only at lower pH, under acidic conditions similar to the conditions found in the atmospheric aerosol deliquescent layer. The rate of iron dissolution for BA and LKD at low pH is similar to that observed for fly ashes. Considering minerology variations, iron dissolution from fly ash has been reported in a range of 5.1–0.17 at pH 1.0 and 6.0–0.203 at pH 2.0 (both rate ranges expressed in Fe ions cm⁻³ s⁻¹) (Borgatta et al. 2016; Kim et al. 2020). Both BA and LKD fall within this range. Furthermore, the total iron solubility remains statistically similar between pH 1.0 and 2.0, an effect also observed for the initial rate of iron dissolution from most fly ashes reported in the literature (Borgatta et al. 2016; Kim et al. 2020). As reported in mineral dissolution processes, surface iron species can have immediate contact with the acidic media and dissolve rapidly (Kim et al. 2020). As the pH increases, the rate of iron leach into aqueous media is not measurable, with the total iron dissolution to the aqueous phase decreasing to less than 2% of the total iron in the combustion sample. For both of the samples examined, the initial rate of dissolution at pH 1.0 and 2.0 is statistically similar. However, iron dissolution from LKD is faster than that of BA. At pH 3.0, BA shows a steady concentration of Fe³⁺, corresponding to an average of only 2% of iron dissolved from the BA, with no measurable rate. On the other hand, the rate of dissolution for LKD at pH 3.0 is measurable, as shown in Table 3, but two orders of magnitude slower than the initial rates calculated at pH 1.0 and 2.0. Overall, the rate of iron dissolution from LKD and BA is comparable to those measured for fly ash samples (Kim et al. 2020). No rate was measurable at pH 4.0, with the iron concentration reaching passivation within 5 min of suspension time and concentrations nearing the limit of detection of the analysis method.

Fig. 4 shows the iron dissolution isotherms normalised to the BET surface area, as indicated in Table 1, for both LKD and BA relative to the time of suspension, at pH 1.0 to 4.0. Contrary to the observation of iron fraction dissolved, where LKD shows a slightly higher fraction of total iron dissolved compared to BA, the amount of dissolved Fe species per surface area was statistically equivalent for both samples at all examined pHs, indicating a correlation between Fe solubility and surface area, contrary to what has been observed in comparative studies of other combustion samples (Borgatta et al. 2016; Navea et al. 2017; Kim et al. 2020). Given the relative similarity in mineralogy between LKD and BA, iron dissolution for these samples is a surface-controlled process. Specifically, when the chemical composition is relatively similar, the acid mobilisation of iron is the same when normalised by surface area, suggesting some dependency on surface properties. As in the case of iron fraction, nearly all iron partitions into the aqueous phase within the first 5–20 min of suspension. As the particle breaks down during the first dissolution time, with a concomitant increase in acid media contact with the iron-containing particle bulk, there is no significant increase in the iron concentration in the solution. This equilibrium stage suggests that most of the iron being mobilised is already on or near the surface of the particle (Marcotte et al. 2020). As the acidity of the media decreases, the particle fragmentation slows down, which is observed in the slow dissolution of iron at pH 3.0 for LKD.

Fig. 4. 

Fe species dissolved from combustion samples normalised to BET surface area of LKD and BA at pH = 1.0 (blue circles), 2.0 (red circles), 3.0 (purple circles), and 4.0 (green circles). Left panels: LKD; Right panels: BA. Top graphs represent total iron (Fe^total = Fe²⁺ + Fe³⁺) leaching in HCl suspensions; middle graphs represent Fe³⁺; bottom graphs represent Fe²⁺.

Fig. 3 shows an increase in the fraction of dissolved total iron (Fe³⁺ and Fe²⁺) with respect to suspension time for both combustion powder samples. Because suspensions take place in HCl, the main mechanism controlling the iron dissolution from the combustion particles is proton-promoted, as shown by Reactions (1) and (2) (Fu et al. 2010, 2012; Borgatta et al. 2016; Kim et al. 2020).

As suggested by a proton-promoted solubility mechanism, as pH increases, total iron (Fe^total = Fe²⁺ + Fe³⁺) dissolution to the aqueous phase decreases. Most of the iron partitioned to the aqueous phase was Fe³⁺, the oxidation state expected from highly oxidised samples, such as combustion particles. Given its low concentrations in solution, conclusions pertaining to Fe²⁺ are predominantly derived from absolute values rather than comparative assessments across varying pH levels. As the pH increases from 1.0 to 4.0, the concentration of dissolved Fe²⁺, while lower than that of Fe³⁺, fluctuates within a narrower range than the uncertainty of total iron concentration (including Fe²⁺ and Fe³⁺). Therefore, Fe²⁺ concentration is bound by this uncertainty and remains statistically constant throughout the pH range examined. Since we conducted the dissolution experiments in the absence of light, dissolved Fe²⁺ must be the result of the dissolution of Fe²⁺-containing solids within the ashes iron as reported in Table 2. This has significant biogeochemical implications as Fe²⁺ is more soluble than Fe³⁺, providing trace iron even as the pH increases (Schoffman et al. 2016). Conversely, the yield of aqueous Fe³⁺ decreases as pH increases for both LKD and BA suspensions. While the yield of Fe³⁺ in solution after 100 min of suspension drops slightly from pH 1.0 to 2.0, there is a significant drop in Fe³⁺ concentrations at pH higher than 3.0, with significantly lower fractions at pH 4.0. This decrease in Fe³⁺ concentration correlates well with the solubility of iron(iii) oxide, hematite (Fe₂O₃), a component observed in the XRD analysis (Fig. 1) (Schwertmann 1991; Spokes and Jickells 1995). Thus, as pH increases the solubility of Fe₂O₃ decreases until it reaches a threshold above pH 3 (Losey et al. 2018; Freedman et al. 2019). In addition, Fe³⁺ dissolution requires a higher H⁺:Feⁿ⁺ ratio, at least three times higher for Fe³⁺ than that of Fe²⁺, as shown by Reactions (1) and (2). Thus, according to the Le Chatelier principle, higher pH will tend to affect the solubility of Fe³⁺ more with the concomitant drop in the fractional solubility of total iron from the combustion particles examined. In addition, the presence of carbonates, as shown by FTIR and XRD analysis, makes solubility more sensitive to pH, with protons dissolving structural CO32− and HCO3−, with the concomitant weakening of the physical integrity of the combustion particles (Bargar et al. 2005; Navea et al. 2010; Borgatta et al. 2016; Kim et al. 2020). While the particles show mineralogical similarities, the difference in the solubility of iron between LKD and BA is likely linked to the bulk and surface abundance of Fe oxide and Fe-containing aluminosilicates. As both FTIR and XPS analysis suggest (Fig. 2), the presence of Fe-substituted aluminosilicates may represent an important source of soluble Fe in combustion particles.

For every experiment, an equilibrium concentration of iron species was determined using the average iron concentration shown in Fig. 4 after the passivation time up to 260 min of suspension. Since a slow iron dissolution can take place, the passivation time was assumed to be 90 min for every experiment, ensuring that equilibrium concentrations had been reached. In the case of LKD suspension at pH 3.0, characterised by a slow iron dissolution, equilibrium concentrations were calculated between 160 and 260 min. As shown in Fig. 5, a comparison between equilibrium concentrations of iron normalised by surface area shows statistically similar values for both BA and LKD at low pH. Because of the oxidation process, the overall mineralogy of these particles is similar, with iron speciation likely favouring Fe³⁺. Thus, at a high concentration of H⁺, the surface area is the driving force of the iron dissolution to the aqueous phase, with the dissolution depending on the amount of surface OH terminal groups available to react, which are shown by the bands centred at 3647 and 877 cm⁻¹ in the infrared spectra of both combustion samples, likely from the lattice structure of aluminosilicate structures, as suggested by the presence of Si and Al in the XPS analysis (Fig. 2) (Galhotra et al. 2009; Fu et al. 2012; Chen and Grassian 2013). These interactions between surface OH groups and the acidic media weaken surface structures and contribute to the mobilisation of iron into the aqueous phase (Zinder et al. 1986; Sulzberger et al. 1989; Remucal and Sedlak 2011; Ostaszewski et al. 2018; Drapanauskaite et al. 2021). As pH increases, these proton-driven processes become less prevalent. As a consequence, with the decrease in the concentration of H⁺, Reaction (1) yields lower concentrations of Fe³⁺, with a more significant decrease in Fe³⁺ yield for BA than that for LKD. With a larger surface distribution and a larger amount of iron, LKD iron dissolution is still observable at pH 3.0; however, it slows down about fifty times compared to the rate of dissolution at pH 1.0 and 2.0. This effect is also dependent on the solubility of the metal oxide because the solubility of Fe₂O₃ decreases around pH 3.0 (Larsen and Vaida 2012). Conversely, for Fe²⁺ this effect is not observed, as the yield of the equilibrium concentration of Fe²⁺ is statistically similar at all pH examined, as shown in Fig. 5 in the striped section of the bars.

Fig. 5. 

Left axis: Average equilibrium concentrations of total iron (Fe^total = Fe²⁺ + Fe³⁺) normalised by the surface area of LKD (red) and BA (blue). The solid section represents the average equilibrium concentrations of Fe³⁺ normalised by surface area, the striped section represents the average equilibrium concentration of Fe²⁺ normalised by surface area. Right axis: speciation fraction of Fe²⁺ (α_FeII) in equilibrium. The red dotted line and symbol represent LKD and the blue dotted line and symbol represent BA.

As the pH increases beyond pH 3.0, the equilibrium concentrations of Fe³⁺ and Fe²⁺ become statistically similar for BA. For LKD, the equilibrium concentration of Fe³⁺ is only slightly higher than that for Fe²⁺, as shown in Fig. 5 for pH 4.0 conditions. At pH 3.0 and 4.0, the small differences in the mineralogy and surface composition between the samples are responsible for this slight difference between them (Borgatta et al. 2016). Overall, there is approximately 20% less iron at pH 3 and 4.0 than that at pH 1.0 and 2.0 for Fe³⁺ and total Fe. Yet, for Fe²⁺, the concentration of aqueous Fe²⁺ remains statistically similar at pH 1.0–4.0. This effect can be observed by the change in the speciation fraction of Fe²⁺ at equilibrium (α_Feⁱⁱ), calculated with Eqn 3,

As the pH increases, the total equilibrium concentration of iron decreases, but the fraction of bioavailable iron, while slowly increasing from pH 1.0 to pH 3.0, sharply increases above pH 3.0, reaching values where α_Feⁱⁱ ≈ α_Feⁱⁱⁱ. Overall, this indicates that while pH can have a significant effect on the solubility of iron, the relative speciation fraction of Fe²⁺ at equilibrium increases. This maintains the biogeochemical impact of the iron-containing particles over the acidic pH range examined, as Fe²⁺ solubility continues to provide dissolved iron.

Conclusions

Recent studies suggest that iron-containing aerosolised combustion particles can dissolve and mobilise iron faster than mineral dust aerosols under environmentally relevant conditions (Baldo et al. 2022). As an energy process by-product with little value, iron-containing combustion particles tend to accumulate and, despite containment efforts, occasionally partition into the environment, where they can contribute to Earth’s iron mobility and budget. In particular, the smaller combustion particles used in this work (< 35 μm) are more susceptible to long range atmospheric transport. To better understand iron dissolution from these anthropogenic combustion particles, we carried out a comparative acidic iron dissolution of ash samples with relatively similar mineralogy, controlling for total iron content and specific surface area of the samples. The data presented herein show that the acidic iron solubility of samples with similar compositions has some dependency on particle composition and surface properties of the particles. The surface particle composition, such as OH surface terminal groups, and carbonate/bicarbonate content, is a driving force in iron dissolution that has been overlooked in climate and biogeochemical models (Al-Abadleh et al. 2022). The speciation of dissolved iron from the combustion particles examined is dominated by Fe³⁺, as combustion particles are highly oxidised. Thus, combustion particles dissolving in the aerosol deliquescent layer (pH 1.0 and 2.0) will primarily mobilise Fe³⁺ (Usher et al. 2003; Navea et al. 2010, 2017; Angle et al. 2021; Leonardi et al. 2020).

The relative steady concentration of Fe²⁺ across the pH examined suggests that its solubility in the aerosol deliquescent layer can be similar in acidic waters and ponds, where weak dissolved organic acids acidify the system. Compounds such as humic acid and fulvic acid are known reducing and photo-reducing agents, often found in soil and dissolved matter in rivers and lakes (Ricker et al. 2022; Mora Garcia et al. 2021). These acids have a pK_a between 4 and 6, allowing them to chelate and mobilise Fe²⁺ in environmental systems (Talbot et al. 1990; Borgatta and Navea 2015). In the atmospheric environment, where pH ranges from 1 to 3 (Ervens et al. 2004), larger amounts of iron can be mobilised from aerosolised combustion particles, with Fe³⁺ being the primary speciation of dissolved iron. However, this work shows that the contribution of dissolved Fe²⁺ tends to increase with pH for both combustion samples, suggesting that Fe²⁺ can continue to dissolve beyond atmospheric processing. While these results show the relevance of the acidic atmospheric processing on the solubility of iron from combustion particles, recent studies have shown that daylight can enhance solubility through photocatalytic processes, especially when semiconductor metal oxides are present in the particles (Fu et al. 2010, 2012; Mao et al. 2013; Hettiarachchi et al. 2018; Kim et al. 2020). Future studies should be conducted to determine if the photoinduced dissolution of iron from BA and LKD can open a pathway for further iron solubility and affect its speciation. Overall, this work shows that iron dissolution from LKD and BA is similar to that measured for fly ash, and is faster than iron dissolution from mineral dust aerosols (Borgatta et al. 2016; Baldo et al. 2022; Kim et al. 2020; Marcotte et al. 2020).