2024 Volume 58 Issue 2 Pages 71-79
Iron (Fe) in seawater is essential for marine phytoplankton. The bioavailability of Fe to phytoplankton largely depends on its chemical form in seawater. Fe(II) is an important component of surface water because photochemical reduction processes are related to the Fe acquisition mechanism of phytoplankton. However, the marine biogeochemical processes of Fe(II) have not been thoroughly investigated. This study applied the luminol chemiluminescence method to directly determine Fe(II) in acidified seawater in the Kuroshio area, subarctic Pacific, and the Bering Sea. We successfully obtained the vertical profile of Fe(II) in the Kuroshio region, where Fe oxidation has been studied; oxidation rates were consistent with previous studies. The prolonged half-life of Fe(II) in the Bering Sea suggests the possible influence of dissolved organic matter from the marginal sea on Fe(II) oxidation. The long half-life of Fe(II) in the high nutrient, low chlorophyll (HNLC) water may be crucial for supplying Fe to phytoplankton.
Iron (Fe) in seawater is essential for marine phytoplankton growth and is a limiting factor for primary production in high-nutrient, low chlorophyll (HNLC) areas (Moore et al., 2013). The bioavailability of Fe to phytoplankton largely depends on its chemical form in seawater (Breitbarth et al., 2010). Thermodynamically, Fe(III) is stable under oxic conditions (Millero and Izaguirre, 1989), and presumably 99% of the dissolved Fe (filtered with 0.2 or 0.4 μm pore-size filters) is organically complexed in the ocean (Gledhill and Buck, 2012). Among the chemical forms of dissolved iron (Fe(OH)3, Fe(III) complexed with organic ligands, colloidal Fe, and Fe(II) species), Fe(II) is a negligible minor species in oxic seawater, assuming thermodynamic equilibrium (Kester et al., 1975). If Fe(II) is produced in seawater, it is rapidly oxidized under oxic conditions (King et al., 1995). However, Fe(II) is not a minor component of the dissolved Fe under certain conditions. For example, the half-life of Fe(II) is long in the oxygen minimum zone; therefore, high concentrations of Fe are observed (Landing and Bruland, 1987; Lohan and Bruland, 2008; Kondo and Moffett, 2013; Chever et al., 2015). Moreover, photochemical processes affect the redox cycling of dissolved Fe in surface waters where detectable Fe(II) is present (Bowie et al., 2002, Croot et al., 2008; Breitbarth et al., 2009; Hansard et al., 2009; Sarthou et al., 2011). Since Fe(II) is less particle-reactive and more biologically available, the presence of even small amounts of Fe(II) in seawater can affect the residence time and biological utilization of iron in the ocean (Moffett, 2021). Reduction of Fe(III) is considered an intermediate step in the Fe uptake process by phytoplankton (e.g., Maldonado and Price, 2001; Shaked et al., 2005), thus it is important to detect Fe(II) in oxic seawater to understand the Fe acquisition mechanism.
The global distribution of dissolved Fe has been intensively studied since the beginning of the GEOTRACES project (Anderson, 2020). Despite increasing interest in Fe speciation in seawater, the marine biogeochemical cycles of Fe(II) have not yet been thoroughly investigated. Because the lifetime of Fe(II) is only a few minutes in warm, oxygenated seawater (Hansard and Landing, 2009), a sensitive and rapid analytical method for Fe(II) is required. A sensitive analytical method using luminol chemiluminescence was proposed for this purpose (King et al., 1995), which enabled picomolar detection of Fe(II) in seawater. Direct Fe(II) determination has been widely performed (e.g., Croot and Laan, 2002; Hopkinson and Barbeau, 2007). However, rapid analysis within a few minutes of sampling is not always possible because of logistical problems associated with field observations. Therefore, modified methods were examined. For example, seawater samples are acidified to pH 6 or 7 before analysis to slow the oxidation of Fe(II) (Hansard and Landing, 2009; Kondo and Moffett, 2013). Fe(II) has been successfully determined in the open ocean using these methods (Hansard et al., 2009; Kondo and Moffett, 2013, 2015).
This study applied an onboard direct determination method using luminol chemiluminescence for picomolar Fe(II) content detection in acidified seawater from three regions: the Kuroshio area, subarctic Pacific, and the Bering Sea. We also examined Fe(II) oxidation rates in oxic surface waters to investigate the marine biogeochemical cycles of Fe(II).
All solutions were prepared using Milli-Q water (MQW) purified with a Millipore system (Millipore, USA). Ultrapure-grade aqueous ammonia, hydrochloric acid (Tamapure AA-100; Tama Chemicals, Japan), and luminol sodium salt (Sigma-Aldrich, USA) were purchased from commercial sources. Stock standard solutions of Fe(II) (100 μM) were prepared by dissolving ferrous ammonium sulfate hexahydrate (Wako, Japan) in a 0.1 M hydrochloric acid solution every three weeks, although the solution was stable in a refrigerator for one month. Working standard solutions (1 μM) were prepared by diluting the stock solutions with 0.1 M hydrochloric acid before every use. Luminol solution (0.74 mM) was made by mixing luminol, aqueous ammonia, and hydrochloric acid in MQW, heated to 60°C for 15 hours (Hansard and Landing, 2009), and left in the dark until the temperature reached room temperature. The solution was stored in a refrigerator until further use.
InstrumentsWe used a flow chemiluminescence detection system similar to that used by FeLume (King et al., 1995). A schematic of this process is shown in Fig. 1. The system consisted of a two-channel peristaltic pump (Masterflex, 7013-20), a three-way solenoid Teflon valve (Takasago, Japan), a chemiluminescence (CL) detector (EN-10; Kimoto Electronics, Japan), and a recorder. A two-channel peristaltic pump was used to send a blank seawater or seawater sample solution and a luminol solution mixed with aqueous ammonia and hydrochloric acid before the analysis. Viton tubes (Saint-Gobain, France) were used for the peristaltic pump. All flow lines were composed of 1-mm ID Teflon tubing connected to Teflon connectors (Obata et al., 1993). The mixed solutions were then immediately introduced into the flow cell of the CL detector. All polyethylene bottles (Nalge, USA), Teflon tubes, and fittings were cleaned using a method adapted for the Fe analysis of seawater (Obata et al., 1997). Viton tubes for the peristaltic pump were filled with 1% surfactant (Extran MA02; Merck, USA), 0.1 M hydrochloric acid, and MQW for one day. The Teflon tubes in the flow system were filled with 1 M hydrochloric acid and MQW for one day each.
Schematic of the flow analytical system for Fe(II) in seawater. This diagram was drawn based on Takeda et al. (2014). The system comprises a two-channel peristaltic pump (Masterflex), a three-way solenoid Teflon valve (Takasago), and a CL detector (Kimoto Electronics). “SW” means seawater.
The sample solution and blank seawater flows were switched using a solenoid Teflon valve. The seawater used as the blank solution was collected from the North Pacific Ocean and left in the dark at room temperature for more than one day. A luminol solution (0.74 mM) containing aqueous ammonia and ammonium chloride was mixed with each sample solution. The pH in the flow cell was adjusted to 10.2 by adding aqueous ammonia. The Fe(II) concentration was determined by measuring the CL intensities of seawater samples containing Fe(II) standard solutions. The solutions were analyzed at a flow rate of 7.5 mL/min. During the cruise, the detection limit of the acidification method was 13 pM (Table 2), which is comparable to the detection limit (12–15 pM) reported in a previous study (Hansard and Landing, 2009).
Sampling methodsDetails of the sampling stations are presented in Table 1. Seawater samples from the subarctic Pacific and Bering Sea were collected from the leg 1 of the R.V. Hakuho-maru KH-09-4 cruise (August 12–September 3, 2009) using Niskin-X samplers deployed on a CTD-carousel multiple sampling system connected to a titanium-armored cable (Obata et al., 1997). At stations KH-2 and KH-16 (Table 1), surface water was collected using a clean trace-metal towing fish sampling system as described by Takeda et al. (2014). Seawater samples were collected from the Kuroshio area during the R. V. Tansei-maru KT-09-13 cruise (July 24–30, 2009), using Niskin-X samplers attached to a Kevlar wire. The seawater samples were filtered with a 0.2 μm-pore size Acropak filter (Pall Corporation, USA) with gravity and collected in 125 mL low-density polyethylene (LDPE) bottles. After sampling, the seawater samples were immediately acidified with hydrochloric acid to pH 6 or used for Fe(II) determination without acidification. Hydrochloric acid was added to the LDPE bottles before sampling for the acidified samples. In the absence of acidification, the samples were analyzed within one minute of collection in LDPE bottles under dark conditions.
Sampling stations
| Oceanic areas | Cruise | Station name | Location | Sampling date | Surface temperature (°C) |
|---|---|---|---|---|---|
| Kuroshio Area | KT-09-13 cruise | KT-2 | 33°32.0'N, 136°28.0'E | 25 July, 2009 | 25.10 |
| KT-5 | 31°46.1'N, 134°20.0'E | 29 July, 2009 | 28.04 | ||
| Subarctic Pacific | KH-09-4 cruise Leg. 1 | KH-1 | 51°2.1'N, 170°12.1'E | 17 August, 2009 | 10.39 |
| KH-2 | 51°22.9'N, 170°10.8'E | 18 August, 2009 | 11.40 | ||
| Subarctic Pacific near Aleutian Islands | KH-39 | 51°0.3'N, 170°40.3'W | 31 August, 2009 | 9.34 | |
| Bering Sea | KH-16 | 52°31.1'N, 175°13.3'E | 20 August, 2009 | 10.45 | |
| KH-17 | 52°22.4'N, 175°16.0'E | 20 August, 2009 | 7.84 |
Determination of Fe(II) in surface waters (10 m) with and without acidification
| Studied areas | Temperature (°C) | with acidification Fe(II) (pM) | without acidification Fe(II) pM |
|---|---|---|---|
| Subarctic Pacific (KH-2) | 11.4 | 29 | 28 |
| Bering Sea (KH-16) | 10.45 | 26 | 27 |
| Detection limits | 13 | 22 |
Samples from the sampling bottles were measured within 10 minutes with acidification and 40 seconds without acidification.
We performed experiments on Fe(II) oxidation in oxic surface waters onboard a ship. The filtered surface water was refrigerated until further use. Prior to the experiments, the seawater samples in the LDPE bottles were placed in a box filled with surface seawater obtained using the underway sampling system of R. V. Hakuho-maru under dark conditions to adjust the samples to the in situ temperature. Appropriate amounts of 1 nM Fe(II) were added to the surface water under dark conditions. Variations in the Fe(II) concentrations in seawater were investigated using a luminol chemiluminescence detection system. The oxidation rate constant (k’) was calculated using the following equation, assuming a first-order reaction for Fe(II) (Millero et al., 1987).
| (1) |
Under dark conditions, the chemical reactions of iron in natural waters include Fe(II) oxidation, iron colloidal formation, Fe(II) and Fe(III)-dissolved organic matter (DOM) complexation and dissociation, and thermal reduction (Pullin and Cabaniss, 2003). At pH 6, the oxidation of Fe(II) slows in seawater (Santana-Casiano et al., 2005), but the thermal Fe reduction rate increases (Pullin and Cabaniss, 2003). Because the effect of Fe thermal reduction overcomes that of Fe(II) oxidation at pH 6, the Fe(II) concentrations in seawater sometimes increase with time (Hansard and Landing, 2009). In this study, we examined the temporal variation in Fe(II) concentration in acidified seawater at pH 6 in some regions (Fig. 2). In the surface waters of the Kuroshio area and open ocean regions of the subarctic Pacific, the temporal variation in Fe(II) was less than 23% of the initial values during the course of the experiment. In contrast, Fe(II) concentrations increased in the surface waters of the Bering Sea (Fig. 2) at a rates of 0.4 pM/min. In seawaters collected vertically from 10 m to 2000 m in the subarctic Pacific near the Aleutian Islands (KH-39, Table 1), we also observed the increase of Fe(II) concentration (0.97–3.5 pM/min) at pH 6 with time. The variations at depths of 10 and 2000 m are shown in Fig. 2c. These rates were higher than those obtained in the open ocean in a previous study (~0.5 pM/min, Hansard and Landing, 2009). The increase in the Fe(II) concentrations in the Bering Sea and near the Aleutian Islands can be attributed to the reduction of Fe(III) in seawater. When the seawater samples were filtered, the microbial effect was minimal. As indicated previously, slow dark thermal reduction of Fe(III) was observed in the presence of Fe(III)–fluvic acid complexes (Waite and Morel, 1984; Voelker et al., 1997; Pullin and Cabaniss, 2003). In particular, acidification to pH 6 increases the reduction rate of Fe(III) complexes (Pullin and Cabaniss, 2003). In the deep layers of the Pacific Ocean, Fe(III)-solubility is strongly related to humic-type, fluorescent, dissolved organic matter (Takata et al., 2005; Kitayama et al., 2009). It implies that Fe(III) forms complexes with humic substances in open-ocean waters. Recently, it has been proposed that Fe(III) complexes with electrochemically active humic substances are present in the deep layers of the open ocean (Whitby et al., 2020). The electrochemically active humic substances have been determined with cathodic stripping voltammetry of their complexes with iron or copper in seawater (Laglera and van den Berg, 2009; Whitby and van den Berg, 2015). As indicated in a previous study (Uchida et al., 2013), the humic-type fluorescence intensity was higher in the surface and deep waters of the Bering Sea than in the subarctic Pacific. It has been reported that the levels of humic-like components in the surface waters of the North Pacific showed clear meridional trends with a decrease toward the south (Yamashita et al., 2017). The levels were the highest in the northern high latitudes (e.g., Bering Sea) and lowest in the subtropical provinces (e.g., the Kuroshio region and North Pacific subtropical gyre), implying that the geographical distribution of humic-like components is controlled probably by input of the terrigenous humic-like components from rivers, high biological production of marine humic-like components in the northern high latitudes, and photo-degradation of humic-like components in the subtropical provinces (Tanaka et al., 2016; Yamashita et al., 2017). Photo-degradation may be an important factor shaping the geographical distribution, with low levels of humic-like components in oceanic surface waters away from terrigenous influences. In the Bering Sea and near the Aleutian Islands, Fe(III) is likely complexed with humic substances, which could promote Fe(III) reduction in the water samples from these areas.
Time variations of Fe(II) in (a) surface waters in the Kuroshio region (KT-5, red closed circle) and surface waters at the subarctic Pacific (KH-2, blue closed circle), (b) surface waters in the Bering Sea (1st experiment: green-filled circles, 2nd experiment: green open circles at KH-17), and (c) seawater at a station near the Aleutian Islands (KH-39, 10 m depth: red open square, 2000 m depth: blue open square) at pH 6 and room temperature in the laboratory.
Acidification of the samples collected in the Kuroshio area and the open ocean region of the subarctic Pacific successfully stabilized Fe(II) in seawater under dark conditions (Fig. 2a). We applied two methods for the determination of Fe(II) in the surface seawater of the open ocean subarctic Pacific and the Bering Sea, without acidification and with acidification to a pH of 6. For the acidified samples from the Bering Sea, we determined Fe(II) within 10 min because the increase in Fe(II) was relatively small (21–44% of the initial values, Fig. 2). The measured values obtained using both methods were almost identical for the surface waters of the subarctic Pacific and Bering Sea (Table 2). The acidification method has been previously applied to samples from the Pacific Ocean (Hansard et al., 2009) and mesocosm experimental samples from the Grand Canaria (Hopwood et al., 2020). On the other hand, the acidification of seawater samples was acidified to pH 7 using a 3-morpholinopropanesulfonic acid (MOPS) buffer for Fe(II) measurements in the Indian Ocean (Kondo and Moffett, 2013) and eastern tropical South Pacific (Kondo and Moffett, 2015).
Using an acidification method, we determined Fe(II) concentrations in seawater samples collected vertically from the Kuroshio area (KT-5). The oxidation of Fe(II) in surface water was delayed by the acidification of the samples to pH 6 (Fig. 2a). We did not correct the data because the oxidation was minimal to acidification. The concentration of Fe(II) ranged from 14 to 26 pM (Fig. 3), similar to that reported in the open ocean region of the Pacific Ocean (<12–130 pM) (Hansard et al., 2009). In the Kuroshio area, dissolved Fe concentrations were previously reported as 0.33 to 1.18 nM from the surface to 760 m depth (Obata et al., 2017). The fraction of Fe(II) relative to the dissolved Fe was estimated to be less than 10%. The vertical profile (Fig. 3) indicated surface maxima (22–24 pM at 0–20 m depth), as previously reported in the northeastern subarctic Pacific (Schallenberg et al., 2015), which was probably due to the photo and biological reduction of Fe(III). Once the Fe(II) concentration rapidly decreased from 20 to 50 m, we observed an increase in Fe(II) again below 50 m depth. This vertical profile pattern has been previously reported in the North Pacific (Hansard et al., 2009; Schallenberg et al., 2015). The increase in Fe(II) in the deep layers was explained by regeneration from biogenic settling particles, dissociation from settling or suspended particles, and the thermal reduction of Fe(III) (Schallenberg et al., 2015). Further studies on the process of iron production from marine particles are required to reveal the Fe(II) sources in deep waters.
Vertical profile of Fe(II) in the Kuroshio area (KT-5). During the cruise, the detection limit was approximately 12 pM.
Using this method, we determined the picomolar levels of Fe(II) in open ocean water. However, we must be careful about the risk of artifacts during water acidification in marginal seas, such as the Bering Sea.
Oxidation of Fe(II) in the Kuroshio, subarctic Pacific, and Bering Seas surface watersUnder natural seawater pH conditions, Fe(OH)2 is one of the significant species of Fe(II) that contributes to the overall oxidation rate of Fe(II) in seawater (Santana-Casiano et al., 2005). Therefore, Fe(II) oxidation was dominant in the kinetic redox process of Fe in seawater at pH 8. Temporal variations in Fe(II) concentrations at natural pH were examined in surface water (10 m) collected in the Kuroshio area, the subarctic Pacific, and Bering Sea (Fig. 4). In the surface waters of the Kuroshio area, Fe(II) rapidly decreased within 10 min; however, Fe(II) oxidation was relatively slow in the surface waters of the subarctic Pacific and Bering Sea. Based on the equation (1), we calculated the Fe(II) oxidation rate constant (k’, Table 3). Fe(II) oxidation was examined at 13 and 25°C in the subarctic Pacific, and at 8 and 18.5°C in the Bering Sea. At lower temperatures, the oxidation rate constants were lower at both stations. The effect of the water temperature on Fe(II) oxidation in the subarctic Pacific was examined as previously described (Roy et al., 2008). The k’ values were plotted as functions of 100 times reciprocal temperature (Fig. 5). The oxidation rate constants in the subarctic Pacific surface water at different water temperatures followed the relationship obtained experimentally in a previous study. The k’ value in the Kuroshio area was much higher than that in other areas, but the oxidation was too fast to obtain a detailed time course, which induced large uncertainty. The k’ value was rather close to that in UV-irradiated seawater (Roy et al., 2008), in which most of dissolved organic matter (DOM) was decomposed. The Fe(II) oxidation rate constants were lower in UV-irradiated seawater than in natural subarctic Pacific water. The oxidation rate constants in the Bering Sea were lower than those described in Fig. 5. Recently, the oxidation rate of Fe(II) in the Atlantic Ocean has been intensively investigated (e.g., Santana-Gonzalez et al., 2018, 2019; Santana-Casiano et al., 2022). The main factors controlling oxidation rates are temperature, pH, salinity, and oxygen (Santana-Casiano et al., 2005). For the Bering Sea water, other parameters probably did not lower the oxidation rates because the surface water was saturated with air, and both pH and salinity were not very low in this area (e.g., Pilcher et al., 2022). As previously discussed, the presence of DOM accelerates or inhibits Fe(II) oxidation (Santana-Casiano et al., 2022). For example, using seawater from Funka Bay, Kuma et al. (1995) demonstrated that the addition of glucaric, glucaric acid-1,4-lactone, and citric acid accelerated Fe(II) oxidation, probably due to the stabilization of the oxidized form by chelation. In contrast, the addition of glucuronic, glycolic, lactic, tartaric, glyceric, gluconic, and malic acids retarded it. Previous experiments using DOM from different sources indicated that the effect of DOM on Fe(II) oxidation rates in water was variable (Rose and Waite, 2002, 2003; Lee et al., 2017a). Slow oxidation in coastal waters is related to the presence of microbially derived autochthonous DOM (Lee et al., 2017b). Higher DOM concentrations (72–90 μM) have been reported on the continental shelf of the Bering Sea (Guo et al., 2004). The relatively slow oxidation of Fe(II) in the surface waters of the Bering Sea (Fig. 5) indicates that some DOM in the Bering Sea may be complexed with Fe(II), preventing the rapid oxidation of Fe(II).
Time variation of Fe(II) in surface water at various temperatures. Green-filled circles and open squares: at 8°C and 19°C and in the Bering Sea (KH-17), blue-filled circles and open squares at 13°C and 25°C in the subarctic Pacific (KH-1), and red-filled circles at 28°C in the Kuroshio area (KT-5).
Oxidation rate constants of Fe(II) in seawaters
| Oceanic areas | Temperature (°C) | Oxidation rate constant (k’) (min–1) |
|---|---|---|
| Kuroshio Area (KT-5) | 28.2 | 1.09 ± 0.42 |
| Subarctic Pacific (KH-1) | 13.0 | 0.20 ± 0.03 |
| 25.0 | 0.33 ± 0.06 | |
| Bering Sea (KH-17) | 8.0 | 0.05 ± 0.001 |
| 18.5 | 0.10 ± 0.004 |
The oxidation rate constant (k’) was calculated using the following equation, assuming a first-order reaction for Fe(II), d[Fe(II)]/dt = –k’[Fe(II)].
Relationship between pseudo-first-order rate constants k’ (min–1) for Fe(II) oxidation in seawater and 100 times reciprocal temperature (K–1). Rate constants were experimentally obtained for the Kuroshio area (red squares), the subarctic Pacific (blue triangles), and Bering Sea (green triangles). The lines indicate the relationships obtained experimentally using natural subarctic Pacific surface water (NSPSW, blue line) and UV-treated seawater (UVSW, red line) from a previous study (Roy et al., 2008).
Based on the oxidation rate constant of Fe(II), we calculated the half-life of Fe(II) in surface waters. At the in-situ temperatures, the half-life of Fe(II) in surface waters was estimated as 0.6 min at the Kuroshio area, 3.8 min at the subarctic Pacific, and 14.8 min at the Bering Sea, respectively. In the coastal areas of the Atlantic islands, half-lives of Fe(II) were around 1.8–3.5 min (Santana-Casiano et al., 2022), which was close to those obtained at the Kuroshio area and subarctic Pacific. At high latitude areas, the half-lives were 7.6–38.5 min in surface waters of the Labrador Sea (Santana-Gonzalez et al., 2019) and 3.3–43.6 min in surface waters of the subarctic Atlantic (Santana-Gonzalez et al., 2018). The half-life of Fe(II) in the Bering Sea was within this range.
During summer, the HNLC surface water is characteristic of the oceanic region of the Bering Sea, where relatively low dissolved Fe concentrations have been observed (Aguilar-Islas et al., 2007; Uchida et al., 2013). Iron deficiency likely limits primary production in this region. Under these conditions, iron recycling is critical for phytoplankton growth. The long half-life of Fe(II) on the Bering Sea surface is an important factor in the bioavailability of Fe to phytoplankton in the Bering Sea. We need to investigate the mechanism that maintains the Fe(II) abundance in HNLC waters in more detail.
We successfully applied the luminol chemiluminescence method to directly determine Fe(II) in seawater from the Kuroshio area after the samples were acidified to pH 6. The vertical profile of Fe(II) was obtained from surface to a depth of 800 m. The applied method is useful for open-ocean water, but we must be careful about its application to marginal seawater samples because acidification to pH 6 easily causes an overestimation due to Fe(II) increase with time.
The Fe(II) oxidation rates in the Bering Sea, the subarctic Pacific, and Kuroshio areas were consistent with those reported in previous studies. The prolonged half-life of Fe(II) in the Bering Sea suggests the possible influence of DOM from the marginal sea on Fe(II) oxidation. The long half-life of Fe(II) in the HNLC water may be important considering the supply process of Fe to phytoplankton. Therefore, it is necessary to study the mechanism of the slow oxidation of Fe(II) in relation to the interaction between Fe and DOM in seawater.
The authors declare no conflict of interest.
We thank all researchers, captains, and crew members of the R. V. Hakuho-maru and R. V. Tansei-maru cruises. We also appreciate the anonymous reviewers whose comments have greatly improved our manuscript. This study was supported by Grant-in-Aid for Scientific Research (B) (No. 22H03728).
