Seven size‐fractionated aerosol samples were collected from Hiroshima, Japan, and were analyzed in terms of chemical composition, soluble fraction of iron (Fe), Fe species, and Fe isotope ratios. The results suggested that Fe in fine particles contained a larger fraction of anthropogenic aerosols than coarse particles did. Iron in the fine particles was more soluble in simulated seawater (up to 25%) than that in the coarse particles and was in the form of Fe (hydr)oxide species, such as ferrihydrite or hematite. The Fe isotope ratios (δ⁵⁶Fe) of the coarse particles (+0.04‰ to +0.30‰) were close to the crustal mean value (0.0‰). By contrast, the δ⁵⁶Fe values of fine particles were much lower and ranged from −2.01‰ to −0.56‰. δ⁵⁶Fe values of the soluble Fe fraction in the fine particles were remarkably low (−3.91 to −1.87‰), suggesting that anthropogenic aerosols yield soluble Fe with low δ⁵⁶Fe values. Such low values could be explained by kinetic isotope fractionation during evaporation of Fe at high temperatures, coupled with the refractory characteristics of Fe. Marine aerosols from the Northwest Pacific were also analyzed. The δ⁵⁶Fe values in the fine particles were also lower than those in the coarse particles. These results may be important to quantitatively estimate the contribution of anthropogenic Fe deposited on the surface ocean on the basis of the Fe isotopes.

1 Introduction

Iron (Fe) is an essential nutrient for phytoplankton, but its concentration in surface seawater is low, since Fe(III) is normally insoluble at the pH of modern seawater (~8). Thus, phytoplankton growth in high‐nutrient low‐chlorophyll (HNLC) regions is limited by soluble Fe concentration [Martin and Fitzwater, 1988; Martin et al., 1994; de Baar et al., 1995; Hutchins and Bruland, 1998]. The sources and budget of soluble Fe on the surface ocean should be clarified [Labatut et al., 2014; Conway and John, 2014] because phytoplankton plays an important role in carbon cycle through photosynthesis and can therefore eventually affect climate change [Watson et al., 2000; Kohfeld et al., 2005; Martínez‐Garcia et al., 2011].

Iron supplied to HNLC can be obtained from hydrothermal vents [Tagliabue et al., 2010], sediments around coasts [Lam and Bishop, 2008], and the atmosphere [Jickells et al., 2005]. Among these sources, aerosols deposited from the atmosphere may be regarded as the main source of Fe in the HNLC regions in open oceans [Fung et al., 2000; Mahowald et al., 2005; Jickells et al., 2005]. However, the source of aerosols controlling the amount of soluble Fe remains debated. In modern systems, aerosols are generally classified into two origins, namely, anthropogenic and natural. Anthropogenic aerosols are generally finer than natural aerosols because of their different formation processes. Fine anthropogenic aerosols are formed through the condensation of gaseous species emitted during combustion; the sizes of these aerosols rarely exceed 1 µm. By contrast, natural aerosols, such as windblown dust, sea spray, or volcanic ash, consist of mineral particles mostly coarser than 1 µm [Whitby, 1977; Lighty et al., 2000]. Different formation processes result in varying solubility ranges of Fe in aerosols. The seawater‐soluble fraction of anthropogenic Fe species is larger than that of Fe‐containing dust, and anthropogenic aerosols can serve as soluble Fe sources to the surface ocean [Sedwick et al., 2007; Luo et al., 2008; Takahashi et al., 2013]. However, the relative contribution of Fe in anthropogenic aerosols to the soluble Fe fraction in seawater in actual systems remains unclear, partly because the anthropogenic component of Fe is not easily determined in the actual aerosol samples supplied to surface seawater.

One of the tools that can be used to determine the contribution of Fe from different sources is the stable isotope ratio of Fe [Hoefs, 2009]. Conway and John [2014] calculated the relative contributions of Fe from aerosols (mainly dust), hydrothermal vents, and reductive and nonreductive sediments released in the North Atlantic by identifying Fe sources on the basis of their specific isotope ratios. However, anthropogenic components were not considered in their calculation, possibly because of the lack of stable isotope data for anthropogenic Fe. If the Fe stable isotope differs between anthropogenic and dust aerosols, the contribution of each type of aerosols to the surface ocean can be estimated.

Mead et al. [2013] investigated the Fe isotope ratio in anthropogenic aerosols. They compared the Fe isotope ratios (δ⁵⁶Fe) between coarse (>2.5 µm) and fine (<2.5 µm) aerosols, which can be regarded as dust and anthropogenic aerosols, respectively. Their results showed that the Fe isotope ratio of the fine aerosols collected in non‐Saharan dust season is δ⁵⁶Fe = −0.10‰, which is lower than that of coarse aerosols (δ⁵⁶Fe = +0.10‰) presumably dominated by mineral dusts. However, the difference in the δ⁵⁶Fe values of the particle sizes found by Mead et al. [2013] is not as large as the standard deviation of their measurements (2 SD = ±0.13‰). Therefore, −0.10‰ of fine particles may not clearly indicate the anthropogenic fraction, possibly because a cutoff size of 2.5 µm is excessively large to separate anthropogenic aerosol components [Whitby, 1977]. Aerosol samples with particle sizes finer than 1 µm can reflect the anthropogenic fraction in a better manner.

In the present study, the Fe isotope ratios of aerosols in seven size‐fractionated samples were measured to identify the Fe isotope ratio of the anthropogenic components. The soluble fraction and chemical species of Fe in aerosols were determined through leaching experiments and X‐ray absorption fine structure (XAFS) spectroscopy, respectively. These results can help attribute the source(s) of Fe in aerosols and identify the subsequent Fe‐related chemical processes in the atmosphere during transport [Takahashi et al., 2011]. The size dependence of the Fe species and Fe stable isotope ratios also helped to identify the anthropogenic Fe species associated with the soluble fraction of Fe into the surface seawater.

2 Methods

2.1 Samples

2.1.1 Aerosol Samples in Hiroshima, Japan

Aerosol samples were collected on filters by using a high‐volume air sampler (Kimoto, MODEL‐123SL, Osaka) with a cascade impactor (Tisch Environmental Inc., Series 230) on the roof of a building in Higashi‐Hiroshima City (latitude, 34.40°N; longitude, 132.71°E; height, approximately 10 m above the ground) from December 2012 to March 2014. The aerosols were separated into seven size fractions by using the cascade impactor (Stage 1, > 10.2 µm; Stage 2, 4.2–10.2 µm; Stage 3, 2.1–4.2 µm; Stage 4, 1.3–2.1 µm; Stage 5, 0.69–1.3 µm; Stage 6, 0.39–0.69 µm; backup filter (BF), < 0.39 µm). Cellulose filters (Stages 1–6, TE‐230WH, Tisch Environmental, Inc., US; BF, Whatman 41, 20.3 cm × 25.4 cm) were used for the sampling. “Coarse particles” and “fine particles” in the Hiroshima samples were defined as particles collected on Stages 1–4 and Stages 5–BF, respectively. The filters were weighed at constant humidity in a desiccator before and after sampling. The Fe blanks of the filters were 0.1 and 0.4 Fe µg/g filter at Stages 1–6 and BF, respectively. These values were negligible compared with the amounts of Fe in the samples (>250 Fe µg/g filter). The samples collected in March 2013 and August 2013 were selected on the basis of preliminary survey of Fe contents in 40 aerosol samples collected from December 2012 to March 2014 (Table 1). In Japan, the Fe concentration in the atmosphere is typically high in spring and low in summer because of the spring Asian dust outbreaks. The March samples were therefore expected to be strongly influenced by dust aerosols and the August samples to contain a relatively larger fraction of anthropogenic aerosols. Backward trajectory analysis was performed using the Hybrid Single‐Particle Lagrangian‐Integrated Trajectory model to identify the source of each sample [Draxler and Rolph, 2003].

Table 1. Periods for the Collection of the Hiroshima Samplesa^a Data of total precipitation was from Japan Meteorological Agency (http://www.jma.go.jp/jma/index.html).

Sample	Sampling Period	Total Suspended Particle (TSP) Concentration (µg/m³)	Total Precipitation During the Period (mm)
March 2013	4 March 2013 1:30 UTC ‐ 9 March 2013 5:30 UTC (5 days)	70.45	0.0
August 2013	22 July 2013 5:00 UTC ‐ 5 August 2013 1:30 UTC (2 weeks)	38.09	17.5

^a Data of total precipitation was from Japan Meteorological Agency (http://www.jma.go.jp/jma/index.html).

2.1.2 Marine Aerosol Samples (The Northwest Pacific)

Marine aerosol samples were collected during the R/V Hakuho Maru KH‐14‐3 cruise (24 June 2014 to 9 August 2014). Size‐fractionated aerosols (finer and coarser than 2.5 µm) were collected on Teflon filters (ADVANTEC, PF040, 90 mm ϕ) by using a high‐volume air sampler (Kimoto, Model AS‐9, Japan). The sampler was automatically controlled by a wind sector to be operated only when the relative wind direction ranged from −90° to 90° of the bow, and the relative wind speed was higher than 1 m/s. The samples were stored at −18°C until chemical analysis. Samples 1 and 2 were used (Table 2).

Table 2. Periods for the Collection of the Marine Aerosol Samples

Sample	Sampling Period	Latitude	Longitude
Sample 1	4 August 2014 22:00 UTC	47°38.69′N	167°08.23′E
	−8 August 2014 00:05 UTC (3 days)	~37°32.69′N	~145°15.61′E
Sample 2	8 August 2014 0:00	37°32.69′N	145°15.61′E
	8 August 2014 23:59 UTC (1 day)	~35°25.91′N	~139°48.20′E

2.2 Acid Digestion and Leaching Experiments

Acid digestion and leaching experiments for the size‐fractionated samples were conducted to investigate the sources and soluble fraction of Fe. Acid digestion was performed to determine the total concentrations of aluminum (Al), manganese (Mn), Fe, nickel (Ni), zinc (Zn), and lead (Pb) through inductively coupled plasma‐atomic emission spectroscopy (ICP‐AES; SII Nano Technology, Inc., SP3500, Chiba, Japan) or inductively coupled plasma‐mass spectrometry (ICPMS; Agilent 7700, Tokyo, Japan). Approximately 40 mg of each filter was treated in a closed Teflon (polytetrafluoroethylene (PTFE)) vial, digested with 6 mL of 38% HF (Tamapure AA‐100, Tama Chemicals, Tokyo, Japan) and 2 mL of 68% HNO₃ (Tamapure AA‐100) at 180°C for 1 day and then evaporated nearly to dryness at 160°C. The residue was redissolved in 2 mL of 68% HNO₃ in a closed PTFE vial at 180°C for 1 day, evaporated nearly to dryness at about 120°C, and then redissolved in 10 mL of 2% HNO₃. The evaporation procedures were conducted under the clean air condition (HEPA filter; SS‐MAC‐15, AIRTECH, Japan).

Major ions (Na⁺, NH₄⁺, K⁺, Mg²⁺, Ca²⁺, Cl⁻, NO₃⁻, and SO₄²⁻) were also analyzed. Approximately 40 mg of the filter was soaked in 5 mL of Milli‐Q water with ultrasonic treatment for 30 min. Dissolved samples were filtered through a 0.20 µm PTFE filter (DISMIC‐25HP, ADVANTEC, Japan) and analyzed through ion chromatography (Dionex, ICS‐1100). Dionex IonPac CS12A/CG12A and IonPac AS22/AG22 columns were used for cation and anion, respectively. The eluent composition was 20 mM methanesulfonic acid at a flow rate of 1.0 mL/min for cation and 1.4 mM NaHCO₃ / 4.5 mM Na₂CO₃ at a flow rate of 1.0 mL/min for anion.

Leaching experiments were conducted to determine the soluble Fe concentration ([Fe_S]). Approximately 40 mg of the filter was soaked in 15 mL of simulated seawater (0.70 M NaCl at pH 8 with 0.10 mM ethylenediaminetetraacetic acid (EDTA); pH 8 was adjusted by adding a NaOH solution without any buffers) or simulated rainwater (0.020 M oxalic acid/ammonium oxalate at pH 4.7) for 24 h at 20°C [Takahashi et al., 2011]. This condition or time was adopted to ensure that the “soluble Fe” can be dissolved completely in the solution. NaCl, EDTA, oxalic acid, and ammonium oxalate were obtained from Wako Pure Chemical Industries, Ltd. (Osaka, Japan). EDTA or oxalic acid was added as ligands, which form complexes with dissolved Fe to retain Fe in the aqueous phase after it dissolves in the solution. Otherwise, the dissolved Fe in water can be reprecipitated within the system and may consequently affect the isotope ratio of Fe remaining in the aqueous phase. EDTA is not usually found in natural seawater but was used in this study since the stability constant of Fe(III)‐EDTA is similar to that of Fe siderophores [Kraemer, 2004] in forming dissolved chelate complexes with Fe in seawater. A similar system was also employed in Takahashi et al. [2011].

Dissolved samples were filtered through a 0.20 µm PTFE filter, evaporated nearly to dryness at 120°C, and redissolved in 10 mL of 2% HNO₃. The soluble Fe concentrations were measured through ICP‐AES. The soluble fraction of Fe in the present study is defined as follows:

$urn:x-wiley:2169897X:media:jgrd53314:jgrd53314-math-0001$ (1)

2.3 XAFS Measurements

Iron K‐edge XAFS analysis was conducted to identify the Fe species. The XAFS spectra were measured at BL12C and BL4A of the Photon Factory, KEK (Ibaraki, Japan), BL37XU of SPring‐8 (Hyogo, Japan) and BL10.3.2 of Advanced Light Source (Berkeley, USA). The procedures of XAFS analysis were described in Takahashi et al. [2011]. X‐ray absorption near edge structure (XANES) and extended X‐ray absorption fine structure (EXAFS) were measured to ensure the reliability of our speciation analysis. Reference materials such as Fe(III) sulfate, Fe(II) sulfate, hematite, magnetite, and Fe(II) oxalate were obtained from Wako Pure Chemical Industries, Ltd. (Osaka, Japan) or Kanto Chemical Co., Inc. (Tokyo, Japan). Illite (IMt‐1) and chlorite (CCa‐2) were obtained from the Clay Mineral Society (Chantilly, USA); biotite and pyrite were obtained from Nichika Corp. (Japan). Suwannee River humic acid was obtained from the International Humic Substances Society. The fitting of the sample spectra was estimated by least squares analysis of each sample spectrum through the linear combination fitting of the reference spectra in REX2000 software (Rigaku Co., Ltd., Japan) and Athena [Ravel and Newville, 2005].

2.4 Iron Isotope Analysis

Iron isotope ratios were measured through multicollector ICPMS (Neptune Plus, Thermo Fisher Scientific). Data were acquired in dynamic mode with a Faraday cup setting to monitor the isotopes of ⁵²Cr, ⁵⁴(Fe + Cr), ⁵⁶Fe, ⁵⁷Fe, ⁵⁸(Fe + Ni), ⁶⁰Ni, ⁶³Cu, and ⁶⁵Cu, as described by Schoenberg and von Blanckenburg [2005]. To resolve the signals of Fe⁺ and Cr⁺ from those of ArO⁺, ArN⁺, and ArC⁺, medium resolution mode was adopted. Iron isotope data are reported as δ⁵⁶Fe relative to IRMM‐014 (Institute for Reference Material and Measurements) and described below:

$urn:x-wiley:2169897X:media:jgrd53314:jgrd53314-math-0002$ (2)

Iron isotope ratios were measured for (a) the total Fe obtained through acid digestion and (b) the soluble fraction of Fe leached into simulated rainwater recovered in the same method described in section 2.2. We used both simulated seawater and rainwater to analyze the soluble fraction of Fe. However, we used only the simulated rainwater to analyze isotope ratio of soluble Fe because (i) the separation of Fe from saline synthetic seawater was not established, which may affect the Fe isotope data and (ii) Fe in the various salts used to prepare synthetic seawater can cause artifacts for the Fe isotope data. Thus, we used the Fe isotope data for soluble fraction recovered by the simulated rainwater. The evaporated samples of the solution of (a) or (b) were redissolved in 2.5 mL of 6 M HCl (Tamapure AA‐100)/0.3 mM H₂O₂ (Tamapure AA‐100) instead of 2% HNO₃ to perform column separation for Fe isotope analysis.

To minimize the isobaric interferences from Cr and Ni, Fe in the 6 M HCl/0.3 mM H₂O₂ solution samples was loaded into a column (Poly Prep Chromatography Columns, Bio‐Rad) filled with 0.6 mL of AG MP‐1 resin (anion exchange resin; Bio‐Rad, 100‐200 mesh) by using a method similar to that of de Jong et al. [2007]. Afterward, samples were evaporated nearly to dryness to remove chloride and then finally redissolved in 2% HNO₃. The solution was diluted to obtain a solution containing 1 µg/g Fe. The typical total Fe signal intensity was approximately 10 V for the sample analysis. In addition, Cu was doped to yield 1 µg/g Cu in the final measurement solution and used as an external standard for mass bias correction. Exponential law was used for the mass bias correction [Albarède et al., 2004]. Interference correction for ⁵⁴Fe was applied following the equation below:

$urn:x-wiley:2169897X:media:jgrd53314:jgrd53314-math-0003$ (3)

where m is the mass and β(Cu) represents the correction coefficient. The values for (⁵⁴Cr/⁵²Cr)_true and m were used in reference to the National Institute of Standards and Technology. For most of the samples, the ⁵²Cr/⁵⁴Fe ratios were less than 1%.

The Fe recovery was 107% ± 11% in all the samples. JB‐2 (basalt), a geological reference material, was also analyzed to confirm the accuracy of the analysis. The average δ⁵⁶Fe value of the repeated analysis of JB‐2 throughout this study was +0.07 ± 0.18‰ (2 SD), which is consistent with the value (+0.056 ± 0.033‰) reported by Weyer et al. [2005].

3 Results and Discussion

3.1 Hiroshima Samples

3.1.1 Sources of Aerosols

The results of the backward trajectory analysis supported our prediction of the sources of the aerosols; that is, the March samples are strongly influenced by dust aerosols, and the August samples contain a relatively larger fraction of anthropogenic aerosols (Figure 1).

**Figure 1**
Open in figure viewer PowerPoint

Backward trajectories during the sampling periods of Hiroshima samples. The trajectories started at the height of 1000 m or 100 m above the sampling site in Higashi‐Hiroshima and run time is 72 h.

The size dependence of trace metals and major ions in each sample can be used to characterize the sources and solubility of aerosols (Figures 2 and S3). In particular, the enrichment factor (EF) of specific metals is often considered to infer the origin of aerosols and the contribution of anthropogenic components (Figure 3). In our study, EFs were calculated as follows:

$urn:x-wiley:2169897X:media:jgrd53314:jgrd53314-math-0004$ (4)

where M is an element evaluated in this study, and (M/Al)_dust is a reference value of Kosa dust from the Gobi Desert calculated by Nishikawa et al. [2013]. An EF higher than 10 usually indicates that the particles contain M from noncrustal sources, such as anthropogenic aerosols or sea‐salt particles.

**Figure 2**
Open in figure viewer PowerPoint

The concentrations of elements. Total concentration was described as [M_total], where M is one of the elements.

**Figure 3**
Open in figure viewer PowerPoint

Enrichment factor (EF) calculated as EF = (M/Al)_aerosol/(M/Al)_crust: (a) Na; (b) Ca; (c) Mn; (d) Fe; (e) Ni; (f) Zn; (g) Pb.

Sodium, an element abundant in seawater, was mainly present in the coarse particles. The existence of Na⁺ suggests that the coarse particles in the March and August samples were influenced by sea‐salt particles to some degree. This suggestion is consistent with the result that the size dependence of Cl⁻ concentration was similar to that of Na⁺.

The concentrations of Al, Mn, Fe, and Ca²⁺ in the March samples were much higher than those in the August samples, presumably because of the larger concentration of aluminosilicate during Kosa dust season. The EFs in all particle sizes of both samples ranged from 1 to 10. Thus, these elements mainly originated from crustal materials, such as soil or desert dust. However, the EFs of Mn and Fe were slightly higher in the fine particles than in the coarse particles. Thus, the fine particles might have partially originated from other sources, such as anthropogenic sources.

The EFs of Ni, Zn, and Pb, which are trace elements associated with fossil fuel combustion or metal production [Nriagu and Pacnya, 1988; Chen et al., 2004; Sholkovitz et al., 2009], were much higher in the fine particles than in the coarse particles. The fine particles were therefore suggested to be influenced by anthropogenic sources. The influence of anthropogenic sources was larger in August than in March, as indicated by the higher EFs values were higher in the August samples.

In a study of marine aerosols in the North Atlantic, Sholkovitz et al. [2009] showed that aerosol samples with low Fe concentration had much higher Ni/Al ratios than those of the crust, suggesting contributions of Ni from anthropogenic sources. Similarly, the Ni/Al ratios in the aerosols of the present study were higher than the ratio of the crust (Ni/Al = 3.9 × 10⁻⁴) when the Fe concentrations were low, especially in fine particles (Figure S1 in the supporting information). These results also suggest that fine particles contained an anthropogenic component.

The aerosol particles were observed by scanning electron microscopy (Hitachi S‐4500) equipped with an energy dispersive X‐ray detector. Coarse particles (Stage 2, 4.2–10.2 µm) mainly contained mineral‐like particles (Figure S2a). However, there were also fine spherical particles (Stage 6, 0.39–0.69 µm) with a high content of Fe (Figure S2b), and these particles were considered to have likely formed by combustion processes and subsequent recondensation [Choël et al., 2007; Flament et al., 2008]. The results above therefore suggest that the fine particles contained a larger amount of anthropogenic aerosols than the coarse particles.

3.1.2 Soluble Fraction of Fe

The soluble fractions of the fine particles were larger than those of the coarse particles (Figure 4). These results suggest that the fine particles, which were considered to be partially from anthropogenic sources, contained more soluble Fe than mineral dust, and this finding is consistent with previous reports [Sedwick et al., 2007; Fu et al., 2012; Takahashi et al., 2013]. In addition, the Fe in the August samples was more soluble than that in the March samples for all particle sizes. Thus, the August samples were likely influenced to a greater degree by anthropogenic components than the March samples.

**Figure 4**
Open in figure viewer PowerPoint

Concentrations of total Fe ([Fe_total]), Fe soluble in simulated seawater ([Fe_SW]) or in simulated rainwater ([Fe_RW]), and the soluble fraction of Fe to the simulated seawater calculated as Fe_sol (%) = ([Fe_SW/RW]/[Fe_total]) × 100.

3.1.3 XAFS Results

K‐edge XANES spectra of Fe in the aerosol samples and reference materials were obtained to estimate the Fe species in the samples. We found that the peak energy of the fine particles (March: 7.129 keV, August: 7.130 keV) was higher than that of the coarse particles (March: 7.128 keV, August: 7.129 keV; Figures 5a and 5b). In addition, the peak energies of the August samples were higher than those of the March samples with the same particle size. The absorption edge located at higher energy in the XANES spectra denotes the presence of more oxidized species [Srivastava and Nigam, 1972]. These results suggest that (i) the Fe in the fine particles was more oxidized than that in the coarse particles and (ii) the August samples contained a larger fraction of oxidized Fe species than the March samples. The main Fe species in the aerosol samples were ferrihydrite, hematite (α‐Fe₂O₃), and biotite (Figures 5c and 5d). The hematite or ferrihydrite fraction was larger in the fine particles, whereas biotite was relatively more important in the coarse particles. Inclusion of the spectra of Fe(II)‐oxalate and Fe(III) complex with humic acid did not improve the XANES fitting. Thus, the organic complex may not be important as a main Fe species relative to the total Fe in the aerosols studied here. The results are consistent with the fitting results revealed by the EXAFS spectra (Figure S4). The agreement of the two results showed that the Fe speciation data are reliable because the two spectra originated from different physicochemical phenomena.

**Figure 5**
Open in figure viewer PowerPoint

(a) XANES spectra of reference materials; (b) XANES spectra of samples; (c and d) iron concentration of each Fe species by XANES analysis.

The μ‐XRF‐XANES results support the bulk XANES analysis findings (Figures S5 and S6). The μ‐XRF of the aerosol particles dispersed on the substrate showed a heterogeneous distribution of Fe in specific particles. The μ‐XANES result for such spots (circle in Figure S5) in the coarse particles indicated the presence of ferrihydrite and biotite, and the existence of biotite is also supported by a relatively high signal of potassium (K) in the same spot. The presence of both ferrihydrite and biotite in the coarse particles may indicate that ferrihydrite was formed at the surface of the biotite by weathering. The formation of ferrihydrite around weathered phyllosilicates containing Fe (e.g., biotite, illite, and chlorite) is often observed in natural systems [Itai et al., 2008; Takahashi et al., 2011].

The μ‐XANES result in the circled spot in Figure S6 shows a high intensity of Fe with (i) relatively high signals of vanadium (V) and Ni and (ii) weaker signal of K. Particles cannot be distinguished individually because of their finer size than that of the X‐ray microbeam. However, the presence of V and Ni, which are often used as indicators of aerosols produced through fossil fuel combustion [e.g., Sholkovitz et al., 2009], suggests that at least some of the particles were emitted by combustion. Ferrihydrite and hematite were the main Fe chemical species of the fine particles, which also suggests that the oxidized Fe species in these particles were partially emitted by combustion.

The results suggest that anthropogenic aerosols in fine particles contained oxidized Fe species, such as ferrihydrite or hematite. Although it cannot be said that all ferrihydrite or hematite in the fine particles are of anthropogenic origin, because they are also contained in soil, we consider it likely that anthropogenic Fe mainly forms ferrihydrite and hematite. A larger amount of ferrihydrite or hematite of anthropogenic origin is presumably contained in the fine particles than in the coarse particles. Schroth et al. [2009] also reported the presence of ferrihydrite in particles emitted from oil combustion.

Takahashi et al. [2013] suggested that Fe(III) sulfate is the main species of anthropogenic Fe. In the present study, however, Fe(III) sulfate was not detected, but ferrihydrite and hematite were identified. Ferrihydrite can be directly precipitated from Fe²⁺ or Fe³⁺ species in water, and it can be also produced by the dissolution of Fe(III) sulfate into aqueous phase in aerosols [Shi et al., 2009]. Fe(III) sulfate and ferrihydrite are formed in the presence of water, and various factors affect these Fe species. The factors include pH and [sulfate]/[Fe] ratio, the latter of which is affected by other elements that form stable compounds with sulfate. Considering that fine particles in the droplet mode are formed through the reaction with water, Fe(III) sulfate can be readily formed under low pH and/or high [sulfate]/[Fe] ratio conditions because of the relatively low stability of ferrihydrite under the conditions. This fact implies that ferrihydrite is preferentially formed from Fe(III) sulfate under high pH and/or lower [sulfate]/[Fe] ratio conditions. Indeed, the [sulfate]/[Fe] ratio of the August samples (approximately 8.7) in this study was lower than that of the samples collected in August in a previous study (approximately 11.6) [Takahashi et al., 2013].

We speculate that hematite could be formed through combustion at high temperatures because hematite is a Fe oxide formed after complete dehydration [Flament et al., 2008; Bladt et al., 2012]. Therefore, we suggest that the two oxidized Fe species identified in our fine aerosols, namely, ferrihydrite and hematite, were emitted at least in part from anthropogenic sources, such as combustion.

3.1.4 Iron Isotope Ratios

The particle size dependence of the Fe isotope ratio was determined in the March and August samples (total and soluble fractions) (Figure 6). Numerical values are presented in the supporting information (Table S1). Isotopic plots of δ⁵⁷Fe versus δ⁵⁶Fe showed that they were on a straight line with a slope of 1.457 ± 0.024, which is almost identical to the theoretical value derived in accordance with mass‐dependent fractionation laws [Young et al., 2002] (Figure S7). The ratios of ⁵⁶Fe/⁵⁴Fe versus ⁵²Cr/⁵⁴Fe exhibited no correlation, which showed that the interference of Cr did not cause any systematic Fe fractionation (Figure S8).

**Figure 6**
Open in figure viewer PowerPoint

Iron isotope ratios of (a) March samples; (b) August samples described as δ⁵⁶Fe (‰) = {(⁵⁶Fe/⁵⁴Fe)_sample/(⁵⁶Fe/⁵⁴Fe)_IRMM‐014 − 1} × 1000.

The δ⁵⁶Fe values of the coarse particles (>1.5 µm) in the total Fe ranged from +0.04 ± 0.07‰ to +0.30 ± 0.07‰, identical, within errors, to the values for terrestrial igneous rocks (0.0 ± 0.05‰) [Beard et al., 2003b] and Hiroshima weathered granite (+0.18 ± 0.22‰) collected in Higashi‐Hiroshima. This result suggests that Fe in the coarse particles is mainly of crustal origin and associated with natural soil minerals, as also indicated by the XAFS analysis in this study. By contrast, the fine particles of the total Fe showed much lower δ⁵⁶Fe values (−2.01‰ to −0.56‰). We suggest that the fine fraction of aerosols yielded lower δ⁵⁶Fe values than those of dust aerosols because most of the fine particles were probably of anthropogenic origin, which is also supported by the backward trajectory and chemical analysis.

The δ⁵⁶Fe values in the fine particles were much lower than those reported by Mead et al. [2013]. They separated aerosols into two fractions with a cutoff at 2.5 µm, and they reported that δ⁵⁶Fe values of coarse and fine particles are +0.10 ± 0.11‰ and −0.10 ± 0.13‰, respectively. By contrast, we separated aerosols into seven size fractions and found much lower δ⁵⁶Fe values of the particles, especially those finer than 0.69 µm (i.e., Stages 6 and BF). Assuming that the August samples were numerically recombined to two fractions of particles coarser and finer than 2.5 µm, the δ⁵⁶Fe values were +0.16 ± 0.12‰ and −0.87 ± 0.09‰, respectively. The δ⁵⁶Fe value of the particles finer than 2.5 µm was still lower than that reported by Mead et al. [2013], presumably because of the difference in relative contributions of Fe from dust and other sources between Bermuda and Hiroshima. The EFs of Fe, Zn, or Pb were higher in Hiroshima (August) than in Bermuda (non‐Saharan dust season), suggesting that the relative contribution of anthropogenic sources was larger in Hiroshima than in Bermuda.

The soluble fraction of Fe of the fine particles in the August samples showed remarkably low δ⁵⁶Fe values (−3.91‰ to −1.87‰). The δ⁵⁶Fe values of the soluble Fe fraction for all particle sizes were also lower than those of the total Fe. If most of the Fe in the soluble fraction is of anthropogenic origin, this accounts for the observed low δ⁵⁶Fe signature. The soluble fraction of Fe of the March samples had higher δ⁵⁶Fe values than that in the August samples, presumably because the August samples contained a larger fraction of anthropogenic Fe than the March samples in the soluble Fe.

The δ⁵⁶Fe values of the insoluble fraction of the August samples were also analyzed (Figure S9). The coarse particles yielded almost the same δ⁵⁶Fe values as those of the total Fe. These results were probably due to the fact that the soluble fractions of Fe in the coarse particles were very low (2%–10%). By contrast, the δ⁵⁶Fe values of the fine insoluble particles were higher than those in the total Fe. Thus, the particles with low δ⁵⁶Fe values dissolved preferentially. These results also support our contention that Fe in anthropogenic aerosols is soluble and exhibits lower δ⁵⁶Fe values than dust aerosols. Furthermore, the insoluble δ⁵⁶Fe values can be calculated on the basis of (i) the δ⁵⁶Fe values for the total Fe and soluble fractions and (ii) the soluble fraction of Fe (Figure S9). The consistency of the calculated and measured δ⁵⁶Fe values of the insoluble fraction suggested that our isotope analysis of Fe is reliable.

3.2 Marine Aerosols

We investigated the Fe isotope ratios in marine aerosols collected in the Northwest Pacific to estimate the possible contribution of anthropogenic Fe deposition to the surface ocean. The results of backward trajectory analysis showed that the aerosols were transported westerly from Japan, suggesting that anthropogenic aerosols were partially contained (Figure 7). On the basis of the data from the Hiroshima samples, we considered that low δ⁵⁶Fe values could help identify anthropogenic aerosols transported and supplied to the surface ocean.

**Figure 7**
Open in figure viewer PowerPoint

Backward trajectories during the sampling periods of the marine aerosols. The trajectories started at the height of 1000 m, and run time is 168 h. Arrows with double stroke show the ship track of the cruise for the sampling.

The two XANES spectra of the fine and coarse samples in Sample 1 (the Northwest Pacific; Figure 8a) showed similar spectra and fitting results. The samples can contain both ferrihydrite and biotite, as suggested by the XANES analysis. By contrast, the peak energy of the fine particles in Sample 2 (from a coastal area of northeastern Japan) was higher than that of the coarse particles in the same sample. This finding revealed that the fine particles contained larger amounts of oxidized species. Fitting results showed that the oxidized species in the fine particles are mainly ferrihydrite (Figure 8b).

**Figure 8**
Open in figure viewer PowerPoint

(a) Fe K‐edge XANES spectra of marine samples; (b) Fe concentration of each chemical species by XANES fitting and chemical analysis; (c) δ⁵⁶Fe value of each samples.

The mass concentration of the coarse particles (coarser than 2.5 µm) was higher than that of the fine particles (finer than 2.5 µm) in Sample 2, but they were almost the same in Sample 1 (Figure 8b). This finding shows that the fine particles are transported farther than the coarse particles [Duce et al., 1983].

The results of the Fe isotope analysis showed that the fine particles exhibited lower δ⁵⁶Fe values (−1.17 ± 0.22‰ and −1.72 ± 0.29‰ for Samples 1 and 2, respectively) than the coarse particles in both samples (−0.11 ± 0.27‰ and −0.32 ± 0.23‰, respectively). The low δ⁵⁶Fe values in the marine aerosols could be caused by (i) anthropogenic Fe contained in the fine particles or by (ii) different sources of Fe with low δ⁵⁶Fe values in the marine aerosols, such as ship emission plumes [Agrawal et al., 2008; Furutani et al., 2011; Ito, 2013] and/or volcanic ash [Hamme et al., 2010; Langmann et al., 2010]. Based on our seven size‐fractionated samples, the average δ⁵⁶Fe value of the particles finer than 2.5 µm in the Hiroshima samples (August) was −0.88‰, which was higher than that of the fine particles of the marine aerosols (−1.72 to −1.17‰). If anthropogenic Fe is the cause of the low δ⁵⁶Fe values in the marine aerosols, it is also possible that the ratio of anthropogenic aerosols to dust aerosols may increase during long‐range transport due to relatively longer transport distance for finer particles [Duce et al., 1983].

3.3 Origin of Remarkably Low δ⁵⁶Fe Values

The soluble Fe fraction of Stage BF aerosols in the August sample from Hiroshima yielded a δ⁵⁶Fe of −3.91‰, which is the lowest value ever reported for materials collected from the Earth's surface environment (Figure 9).

**Figure 9**
Open in figure viewer PowerPoint

δ⁵⁶Fe values of crust [*Beard et al*., 2003a], Fe‐Mn nodule [*Zhu et al*., 2000], deep sea sediments [*Rouxel et al*., 2003], hydrothermal fluids [*Beard et al*., 2003b], seawater [*Conway and John*, 2014], river water [*Fantle and De Paolo*, 2004], higher plant [*Guelke and Von Blanckenburg*, 2007], bulk black shales and pyrite in them [*Dauphas and Rouxel*, 2006], coarse aerosol (Stages 1–4), fine aerosol (Stages 5–BF; total), and fine aerosol (Stages 5–BF; soluble fraction).

Partial dissolution and subsequent precipitation during the extraction experiment are possibly responsible for the low δ⁵⁶Fe values. In our extraction experiment, however, the oxalic acid in the simulated rainwater (0.020 M oxalic acid/ammonium oxalate at pH 4.7) should prevent reprecipitation by retaining Fe in the aqueous phase through the formation of Fe‐oxalate complex. Furthermore, the concentration of oxalic acid is high enough to dissolve all soluble Fe (including ligand‐leachable Fe) within 24 h. An experiment was conducted to examine changes in the soluble Fe fraction over time (Figure S10). The concentrations of dissolved Fe were almost constant after approximately 2 h of extraction. This result suggests that soluble Fe was completely dissolved. Thus, we do not expect that any mass‐dependent isotope fractionation have occurred because of partial dissolution of certain Fe phases in the solid phase.

It is also possible that fine particles with low δ⁵⁶Fe values are formed in natural systems by dissolution and reprecipitation of Fe during weathering of Fe‐bearing clay minerals [Kiczka et al., 2010]. However, the large fractionation of δ⁵⁶Fe of up to −2‰ found by Kiczka et al. [2010] occurred when only 0.1% of Fe in the solid mineral phase was released into the aqueous phase. In our experiments, more than 10% of Fe was dissolved from the fine particles. Therefore, the low δ⁵⁶Fe in our samples cannot be explained only by the dissolution of natural aerosols because for it to be induced by weathering would require that only a very small amount of Fe was released into the aqueous phase.

Other sources of Fe in natural systems exhibit low δ⁵⁶Fe values. For example, the δ⁵⁶Fe values in higher plants are as low as −1.64‰ [Guelke and Von Blanckenburg, 2007]. Therefore, aerosols emitted during biomass burning can yield low δ⁵⁶Fe values as suggested by Mead et al. [2013]. However, the aforementioned value is still higher than those observed in fine particles of aerosols in the present study. Thus, further fractionation of Fe isotope would be required to obtain it.

The reported δ⁵⁶Fe values of black shales vary from −2‰ to +0.6‰, and for pyrite in these shales, δ⁵⁶Fe varies from −2.3‰ to +1.2‰ [Dauphas and Rouxel, 2006]. Such shales or pyrite can be contained in fossil fuels and be responsible for the low δ⁵⁶Fe values, but again, further lighter fractionation would be needed to explain the low δ⁵⁶Fe values in our aerosol samples. A very low δ⁵⁶Fe of –3.5‰ has been reported in some black shale samples collected in South Africa [Rouxel et al., 2005], but these cannot be considered as a likely major source of Fe in the atmosphere. Thus, another explanation is needed to account for the low δ⁵⁶Fe values measured in our aerosol samples.

On the basis of the results of chemical analysis and Fe speciation presented here, we consider that isotope fractionation must have occurred during evaporation of Fe species at high temperature. Some studies on kinetic isotope fractionation of Zn during evaporation have been conducted [Mattielli et al., 2009; Black et al., 2014; Ochoa Gonzalez and Weiss, 2015]. In particular, Mattielli et al. [2009] measured Zn isotope ratios (δ⁶⁶Zn) in Zn‐enriched ore and airborne particles near a refinery in France; they reported that the airborne particles exhibited significantly lower isotope ratios (−0.67 ± 0.10‰) than those of the original ores (+0.13‰); they explained the fractionation through the Rayleigh fractionation equation as follows:

$urn:x-wiley:2169897X:media:jgrd53314:jgrd53314-math-0005$ (5)

where f is the residual Zn fraction in the solid phase, α the fractionation factor, and i represents the initial state. The value of α ranged from 1.0004 to 1.0008 when f is 0.985.

Similarly, we think that Rayleigh fractionation also contributed to the very low δ⁵⁶Fe values in anthropogenic aerosols in this study, in addition to the possible low δ⁵⁶Fe values in the initial materials. The upper limit of the theoretical mass‐dependent fractionation factor α during evaporation of Fe is 1.018 (α = (⁵⁶m/⁵⁴m)^1/2, where m is the isotope mass [Wang et al., 1994]). The residual fraction (f) of Fe at any stage of evaporation is expected to be higher than that of Zn because Fe is more refractory [Allégre et al., 2001]. As a result, Fe emitted during evaporation can be fractionated by as much as −18‰ if the initial δ⁵⁶Fe value of the original materials is 0‰ (crustal value) and f is close to 1. Therefore, isotope fractionation during evaporation of Fe could result in lower δ⁵⁶Fe, which could not be explained only by the negative δ⁵⁶Fe values of the initial materials.

Anthropogenic aerosols are mixtures of particles emitted from various sources. Therefore, the low δ⁵⁶Fe value in this study is an average value, and each source can have different α, f, and initial isotope ratios. Hence, further studies on δ⁵⁶Fe from various sources are necessary [Kurisu et al., 2016].

3.4 Relationship of Fe Isotope and Fe Species

For the Hiroshima samples, the δ⁵⁶Fe values differed among the particle sizes. To explain the particle size dependence of the δ⁵⁶Fe values, we assigned plausible δ⁵⁶Fe values to each Fe species detected by XAFS. In this manner, we were able to compare the calculated δ⁵⁶Fe values at each particle size and the measured δ⁵⁶Fe values (Figure 10). The δ⁵⁶Fe value of each particle size was calculated on the basis of the ratio of Fe species and the assigned δ⁵⁶Fe values to each species. The δ⁵⁶Fe values assigned to different species were adjusted to minimize differences between the measured and calculated δ⁵⁶Fe values of the aerosol samples. The δ⁵⁶Fe value of biotite was fixed to 0‰ because it is of crustal origin regardless of particle size.

**Figure 10**
Open in figure viewer PowerPoint

Calculated δ⁵⁶Fe value of each size sample and their measured data: (a and b) δ⁵⁶Fe value for a Fe species was the same at all size particles; (c and d) δ⁵⁶Fe values were adopted to fine particles and coarse particles differently.

The calculated and measured data were not consistent when the δ⁵⁶Fe value of the Fe species was the same at all particle sizes, (Figures 10a and 10b). However, the particle size dependence of δ⁵⁶Fe values could be explained when different δ⁵⁶Fe values were assigned to coarse (Stages 1–4) and fine (Stage 5–BF) aerosol particles for Fe in ferrihydrite and hematite (Figures 10c and 10d). This result shows that the same Fe species of ferrihydrite and hematite in coarse (Stages 1–4) and fine (Stages 5–BF) particles can yield different δ⁵⁶Fe values. Therefore, the same Fe species in the two components can be formed through different formation processes. The low δ⁵⁶Fe values of hematite and ferrihydrite in the finer particles were mainly attributed to anthropogenic components formed through combustion processes. By contrast, the same Fe species in the coarse particles are thought to be of crustal origin, or to have formed through the alternation of Fe‐bearing minerals such as biotite, as is also suggested by the μ‐XRF‐XANES results (Figure S5). Among the Fe species, hematite in the fine particles likely yielded the lowest δ⁵⁶Fe value. This low δ⁵⁶Fe value for hematite could explain the lowest δ⁵⁶Fe value in Stage 6 in the particle size dependence in Figure 6.

The calculated δ⁵⁶Fe values for each species in this section are considered to be the case of the Hiroshima samples. Our calculated δ⁵⁶Fe values for Fe species in the Hiroshima aerosols are not consistent with the results from the marine aerosol samples that show relatively lower δ⁵⁶Fe values, with lesser contributions of hematite. We speculate that the ferrihydrite in the fine particles of Samples 1 and 2 (marine aerosols) exhibited δ⁵⁶Fe values lower than those of the Hiroshima samples, because the sources of the particles were different.

3.5 Contribution of Anthropogenic Fe to the Surface Ocean

The very low δ⁵⁶Fe values measured in this study are useful to simulate the budget of Fe in the surface ocean of areas such as the HNLC region of the North Pacific. Conway and John [2014] calculated the contributions of Fe to the North Atlantic from various Fe sources, including the deposition of aerosols (mainly mineral dust), reductive dissolution of sediment, dissolution from nonreductive process, and hydrothermal inputs from seafloor. However, our study suggests that the relative contributions of anthropogenic and natural aerosols must be reconsidered separately because Fe isotope ratios in these two aerosol sources are presumably different. Our study also indicates that future modeling studies on the δ⁵⁶Fe values in surface seawater must include the contribution of Fe in anthropogenic aerosols, at least in the North Pacific.

4 Conclusions

This study is the first to report precise, particle size dependent δ⁵⁶Fe values in aerosols. Our results revealed that (i) the δ⁵⁶Fe values in fine particles (<1 µm) collected in Hiroshima are lower than those in coarse particles with crustal δ⁵⁶Fe values and that (ii) Fe in the fine particles is much more soluble than that in coarse particles. These findings are consistent with Fe speciation as estimated by XAFS. Coarse particles contain relatively large amounts of phyllosilicates from natural crustal sources. By contrast, fine particles also contain oxidized Fe species, such as ferrihydrite and hematite, which can partly be of anthropogenic origin such as combustion processes, with low δ⁵⁶Fe values. A very low Fe isotope ratio (δ⁵⁶Fe = −3.91‰) was observed in the soluble Fe fraction of the fine aerosols. This value was the lowest ever reported for materials collected from the Earth's surface environment and can be explained by kinetic isotope fractionation during evaporation. Low Fe isotope ratios, such as δ⁵⁶Fe = −1.17‰ or −1.72‰, were also detected in the fine fraction of marine aerosols from the northwest Pacific. The low Fe isotope ratios of fine aerosols have not been considered in modeling of the δ⁵⁶Fe values of dissolved Fe in surface seawater. Our study shows that the input of anthropogenic Fe in aerosols with low δ⁵⁶Fe values should be considered when modeling the Fe cycle in the surface ocean.

Acknowledgments

We are grateful to A. Sakaguchi, K. Sakata, A. Miyahara, Y. Yamakawa, and C. Miyamoto for collecting aerosol samples at Hiroshima. This work has been performed with the approval of KEK‐PF (proposal 2014G058), ALS (proposal ALS‐06699), and SPring‐8 (proposal 2014A1416 and 2015B0127). Raw data are provided in the supporting information or available from Y.T. (ytakaha@eps.s.u‐tokyo.ac.jp) upon request.

Supporting Information

References

Agrawal, H., Q. G. J. Malloy, W. A. Welch, J. Wayne Miller, and D. R. Cocker (2008), In‐use gaseous and particulate matter emissions from a modern ocean going container vessel, Atmos. Environ., 42, 5504– 5510, doi:10.1016/j.atmosenv.2008.02.053.
Crossref CAS ADS Web of Science®Google Scholar
Albarède, F., P. Telouk, J. Blichert‐Toft, M. Boyet, A. Agranier, and B. Nelson (2004), Precise and accurate isotopic measurements using multiple‐collector ICPMS, Geochim. Cosmochim. Acta, 68(12), 2725– 2744, doi:10.1016/j.gca.2003.11.024.
Crossref CAS ADS Web of Science®Google Scholar
Allègre, C., G. Manhes, and E. Lewin (2001), Chemical composition of the Earth and the volatility control on planetary genetics, Earth Planet. Sci. Lett., 185, 49– 69, doi:10.1016/S0012-821X(00)00359-9.
Crossref CAS ADS Web of Science®Google Scholar
Beard, B. L., C. M. Johnson, J. L. Skulan, K. H. Nealson, L. Cox, and H. Sun (2003a), Application of Fe isotopes to tracing the geochemical and biological cycling of Fe, Chem. Geol., 195, 87– 117, doi:10.1016/S0009-2541(02)00390-X.
Crossref CAS ADS Web of Science®Google Scholar
Beard, B. L., C. M. Johnson, K. L. Von Damm, and R. L. Poulson (2003b), Iron isotope constraints on Fe cycling and mass balance in oxygenated Earth oceans, Geology, 31(7), 629– 632, doi:10.1130/0091-7613(2003)031<0629:IICOFC>2.0.CO;2.
Crossref CAS ADS Web of Science®Google Scholar
Black, J. R., S. G. John, and A. Kavner (2014), Coupled effects of temperature and mass transport on the isotope fractionation of zinc during electroplating, Geochim. Cosmochim. Acta, 124, 272– 282, doi:10.1016/j.gca.2013.09.016.
Crossref CAS ADS Web of Science®Google Scholar
Bladt, H., J. Schmid, E. D. Kireeva, O. B. Popovicheva, N. M. Perseantseva, M. A. Timofeev, K. Heister, J. Uihlein, N. P. Ivleva, and R. Niessner (2012), Impact of Fe content in laboratory‐produced soot aerosol on its composition, structure, and thermo‐chemical properties, Aerosol Sci. Technol., 46(12), 1337– 1348, doi:10.1080/02786826.2012.711917.
Crossref CAS Web of Science®Google Scholar
Chen, Y., N. Shah, F. E. Huggins, and G. P. Huffman (2004), Investigation of the microcharacteristics of PM_2.5 in residual oil fly ash by analytical transmission electron microscopy, Environ. Sci. Technol., 38(24), 6553– 6560.
Crossref CAS PubMed Web of Science®Google Scholar
Choël, M., K. Deboudt, P. Flament, L. Aimoz, and X. Mériaux (2007), Single‐particle analysis of atmospheric aerosols at Cape Gris‐Nez, English Channel: Influence of steel works on iron apportionment, Atmos. Environ., 41, 2820– 2830, doi:10.1016/j.atmosenv.2006.11.038.
Crossref CAS ADS Web of Science®Google Scholar
Conway, T. M., and S. G. John (2014), Quantification of dissolved iron sources to the North Atlantic Ocean, Nature, 511, 212– 215, doi:10.1038/nature13482.
Crossref CAS ADS PubMed Web of Science®Google Scholar
Dauphas, N., and O. Rouxel (2006), Mass spectrometry and natural variations of iron isotopes, Mass Spectrom. Rev., 25, 515– 520, doi:10.1002/mas.20078.
Wiley Online Library CAS ADS PubMed Web of Science®Google Scholar
De Baar, H. J. W., J. T. M. de Jong, D. C. E. Bakker, B. M. Löscher, C. Veth, U. Bathmann, and V. Smetacek (1995), Importance of iron for plankton blooms and carbon dioxide drawdown in the Southern Ocean, Nature, 373, 412– 415, doi:10.1038/373412a0.
Crossref CAS ADS Web of Science®Google Scholar
De Jong, J., V. Schoemann, J. L. Tison, S. Becquevort, F. Masson, D. Lannuzel, J. Petit, L. Chou, D. Weis, and N. Mattielli (2007), Precise measurement of Fe isotopes in marine samples by multi‐collector inductively coupled plasma mass spectrometry (MC‐ICP‐MS), Anal. Chim. Acta, 589, 105– 119, doi:10.1016/j.aca.2007.02.055.
Crossref CAS PubMed Web of Science®Google Scholar
Draxler, R. R., and G. D. Rolph (2003), HYSPLIT (Hybrid Single‐Particle Lagrangian Integrated Trajectory) Model access via NOAA ARL READY Website, NOAA Air Resource Laboratory, Silver Spring, Md. [Available at http://www.ready.noaa.gov/HYSPLIT.php.]
Google Scholar
Duce, R. A., R. Arimoto, B. J. Ray, C. K. Unni, and P. J. Harder (1983), Atmospheric trace elements at Enewetak Atoll: 1. Concentrations, sources, and temporal variability, J. Geophys. Res., 88, 5321– 5342, doi:10.1029/JC088iC09p05321.
Wiley Online Library CAS ADS Web of Science®Google Scholar
Fantle, M. S., and D. J. De Paolo (2004), Iron isotopic fractionation during continental weathering, Earth Planet. Sci. Lett., 228, 547– 562, doi:10.1016/j.epsl.2004.10.013.
Crossref CAS ADS Web of Science®Google Scholar
Flament, P., N. Mattielli, L. Aimoz, M. Choël, K. Deboudt, J. de Jong, J. Rimetz‐Planchon, and D. Weis (2008), Iron isotopic fractionation in industrial emissions and urban aerosols, Chemosphere, 73(11), 1793– 8, doi:10.1016/j.chemosphere.2008.08.042.
Crossref CAS PubMed Web of Science®Google Scholar
Fu, H., J. Lin, G. Shang, W. Dong, V. H. Grassian, G. R. Carmichael, Y. Li, and J. Chen (2012), Solubility of iron from combustion source particles in acidic media linked to iron speciation, Environ. Sci. Technol., 46, 11,119– 11,127, doi:10.1021/es302558m.
Crossref CAS ADS Web of Science®Google Scholar
Fung, I. Y., S. K. Meyn, I. Tegen, S. C. Doney, J. G. John, and J. K. B. Bishop (2000), Iron supply and demand in the upper ocean, Global Biogeochem. Cycles, 14(1), 281– 296, doi:10.1029/1999GB900059.
Wiley Online Library CAS ADS Web of Science®Google Scholar
Furutani, H., J. Jung, K. Miura, A. Takami, S. Kato, Y. Kajii, and M. Uematsu (2011), Single‐particle chemical characterization and source apportionment of iron‐containing atmospheric aerosols in Asian outflow, J. Geophys. Res., 116, 1– 17, doi:10.1029/2011JD015867.
Wiley Online Library CAS ADS Web of Science®Google Scholar
Guelke, M., and F. Von Blanckenburg (2007), Fractionation of stable iron isotopes in higher plants, Environ. Sci. Technol., 41(6), 1896– 1901, doi:10.1021/es062288j.
Crossref CAS ADS PubMed Web of Science®Google Scholar
Hamme, R. C., et al. (2010), Volcanic ash fuels anomalous plankton bloom in subarctic northeast Pacific, Geophys. Res. Lett., 37, L19604, doi:10.1029/2010GL044629.
Wiley Online Library CAS ADS Web of Science®Google Scholar
Hoefs, J. (2009), Stable Isotope Geochemistry, 7th ed., Springer International Publishing, Switzerland.
Google Scholar
Hutchins, D. A., and K. W. Bruland (1998), Iron‐limited diatom growth and Si: N uptake ratios in a coastal upwelling regime, Nature, 393, 561– 564, doi:10.1038/31203.
Crossref CAS ADS Web of Science®Google Scholar
Itai, T., Y. Takahashi, T. Uruga, H. Tanida, and A. Iida (2008), Selective detection of Fe and Mn species at mineral surfaces in weathered granite by conversion electron yield X‐ray absorption fine structure, Appl. Geochem., 23, 2667– 2675, doi:10.1016/j.apgeochem.2008.05.020.
Crossref CAS Web of Science®Google Scholar
Ito, A. (2013), Global modeling study of potentially bioavailable iron input from shipboard aerosol sources to the ocean, Global Biogeochem. Cycles, 27, 1– 10, doi:10.1029/2012GB004378.
Wiley Online Library CAS ADS Web of Science®Google Scholar
Jickells, T. D., et al. (2005), Global iron connections between desert dust, ocean biogeochemistry, and climate, Science, 308, 67– 71, doi:10.1126/science.1105959.
Crossref CAS ADS PubMed Web of Science®Google Scholar
Kiczka, M., J. G. Wiederhold, J. Frommer, S. M. Kraemer, B. Bourdon, and R. Kretzschmar (2010), Iron isotope fractionation during proton‐ and ligand‐promoted dissolution of primary phyllosilicates, Geochim. Cosmochim. Acta, 74, 3112– 3128, doi:10.1016/j.gca.2010.02.018.
Crossref CAS ADS Web of Science®Google Scholar
Kohfeld, K. E., S. P. Harrison, C. Le Que, and R. F. Anderson (2005), Role of marine biology in glacial‐interglacial CO₂ cycles, Science, 74, 74– 78, doi:10.1126/science.1105375.
Crossref CAS ADS Web of Science®Google Scholar
Kraemer, S. M. (2004), Iron oxide dissolution and solubility in the presence of siderophores, Aquat. Sci., 66, 3– 18, doi:10.1007/s00027-003-0690-5.
Crossref CAS Web of Science®Google Scholar
Kurisu, M., K. Sakata, C. Miyamoto, Y. Takaku, T. Iizuka, and Y. Takahashi (2016), Variation of iron isotope ratios in anthropogenic materials emitted through combustion processes, Chem. Lett., 45, 970– 972, doi:10.1246/cl.160451.
Crossref CAS Web of Science®Google Scholar
Labatut, M., A. Jain, X. Yang, H. Kheshgi, A. McGuire, W. Post, and D. Kicklighter (2014), Iron sources and dissolved‐particulate interactions in the seawater of the Western Equatorial Pacific, iron isotope perspectives, Global Biogeochem. Cycles, 23, 1044– 1065, doi:10.1002/2014GB004928.
Wiley Online Library CAS ADS Web of Science®Google Scholar
Lam, P. J., and J. K. B. Bishop (2008), The continental margin is a key source of iron to the HNLC North Pacific Ocean, Geophys. Res. Lett., 35, L07608, doi:10.1029/2008GL033294.
Wiley Online Library CAS ADS PubMed Web of Science®Google Scholar
Langmann, B., K. Zakšek, M. Hort, and S. Duggen (2010), Volcanic ash as fertiliser for the surface ocean, Atmos. Chem. Phys. Discuss., 10, 711– 734, doi:10.5194/acpd-10-711-2010.
Crossref Google Scholar
Lighty, J. S., J. M. Veranth, and A. F. Sarofim (2000), Combustion aerosols: Factors governing their size and composition and implications to human health, J. Air Waste Manage. Assoc., 50, 1565– 1618, doi:10.1080/10473289.2000.10464197.
Crossref CAS PubMed Web of Science®Google Scholar
Luo, C., N. Mahowald, T. Bond, P. Y. Chuang, P. Artaxo, R. Siefert, Y. Chen, and J. Schauer (2008), Combustion iron distribution and deposition, Global Biogeochem. Cycles, 22, GB1012, doi:10.1029/2007GB002964.
Wiley Online Library CAS ADS Web of Science®Google Scholar
Mahowald, N. M., A. R. Baker, G. Bergametti, N. Brooks, R. A. Duce, T. D. Jickells, N. Kubilay, J. M. Prospero, and I. Tegen (2005), Atmospheric global dust cycle and iron inputs to the ocean, Global Biogeochem. Cycles, 19, GB4025, doi:10.1029/2004GB002402.
Wiley Online Library CAS ADS Web of Science®Google Scholar
Martin, J. H., and S. E. Fitzwater (1988), Iron deficiency limits phytoplankton growth in the north‐east Pacific subarctic, Nature, 331, 341– 343, doi:10.1038/331341a0.
Crossref CAS ADS Web of Science®Google Scholar
Martin, J. H., et al. (1994), Testing the iron hypothesis in ecosystems of the equatorial Pacific Ocean, Nature, 371, 123– 129, doi:10.1038/371123a0.
Crossref CAS ADS Web of Science®Google Scholar
Martínez‐Garcia, A., A. Rosell‐Melé, S. L. Jaccard, W. Geibert, D. M. Sigman, and G. H. Haug (2011), Southern Ocean dust‐climate coupling over the past four million years, Nature, 476, 312– 315, doi:10.1038/nature10310.
Crossref CAS ADS PubMed Web of Science®Google Scholar
Mattielli, N., J. C. J. Petit, K. Deboudt, P. Flament, E. Perdrix, A. Taillez, J. Rimetz‐Planchon, and D. Weis (2009), Zn isotope study of atmospheric emissions and dry depositions within a 5 km radius of a Pb‐Zn refinery, Atmos. Environ., 43, 1265– 1272, doi:10.1016/j.atmosenv.2008.11.030.
Crossref CAS ADS Web of Science®Google Scholar
Mead, C., P. Herckes, B. J. Majestic, and A. D. Anbar (2013), Source apportionment of aerosol iron in the marine environment using iron isotope analysis, Geophys. Res. Lett., 40, 5722– 5727, doi:10.1002/2013GL057713.
Wiley Online Library CAS ADS Web of Science®Google Scholar
Nishikawa, M., D. Batdorj, M. Ukachi, K. Onishi, K. Nagano, I. Mori, I. Matsui, and T. Sano (2013), Preparation and chemical characterisation of an Asian mineral dust certified reference material, Anal. Methods, 5, 4088, doi:10.1039/c3ay40435h.
Crossref Google Scholar
Nriagu, J. O., and J. M. Pacnya (1988), Quantitative assessment of worldwide contamination of air, water and soils by trace metals, Nature, 333, 134– 139, doi:10.1038/333134a0.
Crossref CAS ADS PubMed Web of Science®Google Scholar
Ochoa Gonzalez, R., and D. Weiss (2015), Zinc isotope variability in three coal‐fired power plants: A predictive model for determining isotopic fractionation during combustion, Environ. Sci. Technol., 151006143754009, doi:10.1021/acs.est.5b02402.
Crossref ADS PubMed Web of Science®Google Scholar
Ravel, B., and M. Newville (2005), ATHENA, ARTEMIS, HEPHAESTUS: Data analysis for X‐ray absorption spectroscopy using IFEFFIT, J. Synchrotron Radiat., 12, 537– 541, doi:10.1107/S0909049505012719.
Wiley Online Library CAS PubMed Web of Science®Google Scholar
Rouxel, O., N. Dobbek, J. Ludden, and Y. Fouquet (2003), Iron isotope fractionation during oceanic crust alteration, Chem. Geol., 202, 155– 182, doi:10.1016/j.chemgeo.2003.08.011.
Crossref CAS ADS Web of Science®Google Scholar
Rouxel, O. J., A. Bekker, and K. J. Edwards (2005), Iron isotope constraints on the Archean and Paleoproterozoic ocean redox state, Science, 307, 1088– 1091, doi:10.1126/science.1105692.
Crossref CAS ADS PubMed Web of Science®Google Scholar
Schoenberg, R., and F. von Blanckenburg (2005), An assessment of the accuracy of stable Fe isotope ratio measurements on samples with organic and inorganic matrices by high‐resolution multicollector ICP‐MS, Int. J. Mass Spectrom., 242, 257– 272, doi:10.1016/j.ijms.2004.11.025.
Crossref CAS ADS Web of Science®Google Scholar
Schroth, A. W., J. Crusius, E. R. Sholkovitz, and B. C. Bostick (2009), Iron solubility driven by speciation in dust sources to the ocean, Nat. Geosci., 2(5), 337– 340, doi:10.1038/ngeo501.
Crossref CAS ADS Web of Science®Google Scholar
Sedwick, P. N., E. R. Sholkovitz, and T. M. Church (2007), Impact of anthropogenic combustion emissions on the fractional solubility of aerosol iron: Evidence from the Sargasso Sea, Geochem. Geophys. Geosyst., 8, Q10Q06, doi:10.1029/2007GC001586.
Wiley Online Library CAS ADS Web of Science®Google Scholar
Shi, Z., M. D. Krom, S. Bonneville, A. R. Baker, T. D. Jickells, and L. G. Benning (2009), Formation of iron nanoparticles and increase in iron reactivity in mineral dust during simulated cloud processing, Environ. Sci. Technol., 43(17), 6592– 6596, doi:10.1021/es901294g.
Crossref CAS ADS PubMed Web of Science®Google Scholar
Sholkovitz, E. R., P. N. Sedwick, and T. M. Church (2009), Influence of anthropogenic combustion emissions on the deposition of soluble aerosol iron to the ocean: Empirical estimates for island sites in the North Atlantic, Geochim. Cosmochim. Acta, 73, 3981– 4003, doi:10.1016/j.gca.2009.04.029.
Crossref CAS ADS Web of Science®Google Scholar
Srivastava, U. C., and H. L. Nigam (1972), X‐Ray Absorption Edge Spectrometry (XAES) as applied to Coordination Chemistry, Coord. Chem. Rev., 9, 275– 310, doi:10.1016/S0010-8545(00)82080-9.
Crossref Web of Science®Google Scholar
Tagliabue, A., et al. (2010), Hydrothermal contribution to the oceanic dissolved iron inventory, Nat. Geosci., 3(4), 252– 256, doi:10.1038/ngeo818.
Crossref CAS ADS Web of Science®Google Scholar
Takahashi, Y., M. Higashi, T. Furukawa, and S. Mitsunobu (2011), Change of iron species and iron solubility in Asian dust during the long‐range transport from western China to Japan, Atmos. Chem. Phys., 11(21), 11,237– 11,252, doi:10.5194/acp-11-11237-2011.
Crossref CAS ADS Web of Science®Google Scholar
Takahashi, Y., T. Furukawa, Y. Kanai, M. Uematsu, G. Zheng, and M. A. Marcus (2013), Seasonal changes in Fe species and soluble Fe concentration in the atmosphere in the Northwest Pacific region based on the analysis of aerosols collected in Tsukuba, Japan, Atmos. Chem. Phys., 13, 7695– 7710, doi:10.5194/acp-13-7695-2013.
Crossref ADS Web of Science®Google Scholar
Wang, J., A. M. Davis, R. N. Clayton, and T. K. Mayeda (1994), Kinetic isotopic fractionation during the evaporation of the iron oxide from liquid state, Lunar Planet. Sci., 24, 1479– 1480.
Web of Science®Google Scholar
Watson, A. J., D. C. Bakker, A. J. Ridgwell, P. W. Boyd, and C. S. Law (2000), Effect of iron supply on Southern Ocean CO₂ uptake and implications for glacial atmospheric CO₂, Nature, 407, 730– 733, doi:10.1038/35037561.
Crossref CAS ADS PubMed Web of Science®Google Scholar
Weyer, S., A. D. Anbar, G. P. Brey, C. Münker, K. Mezger, and A. B. Woodland (2005), Iron isotope fractionation during planetary differentiation, Earth Planet. Sci. Lett., 240, 251– 264, doi:10.1016/j.epsl.2005.09.023.
Crossref CAS ADS Web of Science®Google Scholar
Whitby, K. T. (1977), The physical characteristics of sulfur aerosol, Int. Symp. Sulfur Atmos., 12.
Google Scholar
Young, E. D., A. Galy, and H. Nagahara (2002), Kinetic and equilibrium mass‐dependant isotope fractionation laws in nature and their geochemical and cosmochemical significance, Geochim. Cosmochim. Acta, 66(6), 1095– 1104, doi:10.1016/S0016-7037(01)00832-8.
Crossref CAS ADS Web of Science®Google Scholar
Zhu, X., R. O'Nions, Y. Guo, and B. Reynolds (2000), Secular variation of iron isotopes in North Atlantic deep water, Science, 287, 2000– 2, doi:10.1126/science.287.5460.2000.
Crossref CAS ADS PubMed Web of Science®Google Scholar

Citing Literature

Volume121, Issue18

27 September 2016

Pages 11,119-11,136

Very low isotope ratio of iron in fine aerosols related to its contribution to the surface ocean

Abstract

1 Introduction