New Insight Into the Source and Sink of 227Ac in the Ocean

Actinium‐227 (227Ac) has been used as a powerful tracer of diapycnal mixing in the ocean, assuming that it is conservative and originates mainly from deep‐sea sediments. However, here we show an unexpectedly large source (continental margin) and sink (scavenging) of 227Ac in the ocean, based on high‐resolution 227Ac distributions obtained for the first time by mooring Mn‐fibers in the East Sea (Japan Sea). Although we expected a decrease in radium‐228 (228Ra) to 227Ac ratios with depth owing to their different half‐lives, the ratios increased with depth in the upper layer, indicating efficient removal of 227Ac by particle scavenging. In addition, unusually high 227Ac activities (∼15 dpm m−3) were observed in the surface layer, likely due to the horizontal transport of 227Ac‐enriched shelf water. Thus, our results suggest refining our understanding of the geochemical cycle and application of 227Ac in the ocean.


Introduction
Naturally occurring actinium-227 ( 227 Ac; half-life = 21.77years) is produced from protactinium-231 ( 231 Pa; half-life = 32,760 years), belonging to the uranium-235 ( 235 U; half-life = 7.04 × 10 8 years) decay series.In marine environments, 231 Pa is produced at a constant rate by the decay of uniformly distributed 235 U in the entire water column.Once produced, 231 Pa is removed by particle scavenging, and it settles to the bottom sediment because of its particle affinity.Since its daughter 227 Ac is known to be soluble in seawater, it diffuses from the bottom sediment to the overlying seawater.This excess 227 Ac is vertically redistributed up to several hundred meters above the bottom sediment in the deep ocean through eddy diffusion.Based on these characteristics, 227 Ac has been used as a powerful tracer for estimating deep ocean circulation and vertical mixing of various elements on a time scale of 100 years (Geibert et al., 2002;Koch-Larrouy et al., 2015;Le Roy et al., 2023;Nozaki, 1984).
However, the activity of 227 Ac in the ocean is extremely low (0.01-5 dpm m −3 ), limiting not only its application as a tracer but also our understanding of its geochemical systematics in the ocean.Traditionally, 227 Ac has long been considered to have similar behaviors to radium (Ra) in the ocean (Geibert et al., 2002;Nozaki, 1984), although previous studies already reported that its chemical analog closely follows that of lanthanum (La) with a similar ionic radius (1.03 Å for La 3+ ; 1.12 Å for Ac 3+ ) and electron configuration (Hill, 1972;Kirby & Morss, 2006).Given that La is much more particle reactive than Ra in the ocean, the particle affinity of Ac should be considered when using it as a tracer for various mixing processes in marine environments.However, the removal of 227 Ac by particle scavenging in the ocean is still poorly understood.Concerning the sources of 227 Ac, Kipp et al. (2015) suggested that hydrothermal vents are the unrecognized source of 227 Ac in the ocean, although the magnitude is only 2%-6% of that from deep-sea sediments in the North Atlantic Ocean.Besides, the sources of 227 Ac in the ocean boundaries remain poorly understood.
In this study, for the first time, we report the high-resolution vertical distribution of 227 Ac in the ocean by mooring the MnO 2 -impregnated fiber (Mn-fiber) to overcome the major difficulty in large volume extractions.This study was mainly conducted in the East Sea (Japan Sea).We also obtained horizontal distribution data from adjacent continental shelf waters (East China Sea and Yellow Sea; ECSYS) fed to the East Sea.Ra isotope ( 226 Ra and 228 Ra) data were obtained together for the same set of samples (Cho et al., 2022).The East Sea, which has a Abstract Actinium-227 ( 227 Ac) has been used as a powerful tracer of diapycnal mixing in the ocean, assuming that it is conservative and originates mainly from deep-sea sediments.However, here we show an unexpectedly large source (continental margin) and sink (scavenging) of 227 Ac in the ocean, based on high-resolution 227 Ac distributions obtained for the first time by mooring Mn-fibers in the East Sea (Japan Sea).Although we expected a decrease in radium-228 ( 228 Ra) to 227 Ac ratios with depth owing to their different half-lives, the ratios increased with depth in the upper layer, indicating efficient removal of 227 Ac by particle scavenging.In addition, unusually high 227 Ac activities (∼15 dpm m −3 ) were observed in the surface layer, likely due to the horizontal transport of 227 Ac-enriched shelf water.Thus, our results suggest refining our understanding of the geochemical cycle and application of 227 Ac in the ocean.
Plain Language Summary Distributions of 227 Ac provide crucial information for the vertical mixing of the deep ocean on timescales of up to 100 years.However, behaviors of 227 Ac in the ocean have not been well understood to date because of its extremely low concentration.In this study, we for the first time determined high-resolution 227 Ac profiles by mooring Mn-fibers in a marginal sea of the northwestern Pacific Ocean.Our results display that the shelf source inputs as well as efficient removal by particle scavenging have been overlooked so far.In particular, we emphasize that the removal of 227 Ac by particle scavenging revealed in this study should be considered when using 227 Ac as a tracer of mixing rates in the ocean.maximum water depth of over 3,700 m, is one of the largest semi-enclosed marginal seas in the northwestern Pacific Ocean.The water exchange between the East Sea and the northwestern Pacific Ocean is limited to several sills (Tsugaru, Soya, and Korea Straits), which are shallower than 130 m (Chang et al., 2004).In the surface layer, a considerable amount of material is transported from the Asian continent to the East Sea through the atmosphere and ECSYS shelf water (Mukai & Suzuki, 1996;Seo et al., 2022;Seo & Kim, 2023), including the Tsushima Warm Current and Changjiang Diluted Water.Furthermore, the East Sea has its own deep thermohaline circulation, with a turnover time of approximately 100 years (Kumamoto et al., 1998;Tsunogai et al., 1993;Watanabe et al., 1991).Thus, the East Sea can provide an ideal natural laboratory of the global ocean for examining the terrestrial sources as well as behavior of 227 Ac under rapid scavenging conditions.

Sampling
Due to extremely low activity of 227 Ac in the ocean, pre-concentration of a large volume of seawater is required using an in situ pump or gravity filtration.This step is very labor intensive, and thus the sample throughput is limited by the ship time.Thus, in this study, we attempted to use a novel method for obtaining high-resolution 227 Ac profiles by attaching Mn-fibers to mooring wires and exposing them to seawater for more than 10 days.Mn-fibers housed in the polyester mesh bag (100 μm pore size) were moored for 12, 166, and 164 days at stations EC1, EC-trap1, and EC-trap2, respectively (Figure 1).In addition, 60 L of seawater samples were collected using a rosette sampler at five depths (5, 500, 1,000, 2,000, and 2,298 m) from EC1 to obtain quantitative 226 Ra and 227 Ac activities.The unfiltered seawater samples were passed through Mn-fibers at <1 L min −1 by gravity on board  (Moore, 1976).The Mn-fiber samples were rinsed with Milli-Q water to remove any sea salts or particles and then stored until analysis.Approximately 100 L of seawater samples were also collected in the ECSYS shelf using a rosette sampler (Figure 1), and then the unfiltered seawater samples were gravity-fed through Mn-fibers in the same manner as mentioned above.

Analytical Methods
To verify the accuracy of the measurement, we determined 227 Ac activities using two different analytical systems: Delayed Coincidence Counter for Ra isotopes (RaDeCC) and gamma counter.For RaDeCC, the Mn-fiber samples were stored for more than three months to allow for the secular equilibrium between 227 Ac and 223 Ra.After three months, the ratio of moisture content to Mn-fiber was readjusted to ∼1.Then, the partially dried Mn-fibers were connected to a closed circulation loop, and 223 Ra activities were measured by counting its daughter, 219 Rn.The contribution from chance coincidence counts of the system was also corrected, as described in the previous studies (Moore & Arnold, 1996;Scholten et al., 2010;Text S1 in Supporting Information S1).The counting efficiency of the RaDeCC was determined using a Mn-fiber standard containing a known amount of 227 Ac.For gamma counting, the Mn-fibers were ashed at 820°C for 16 hr in a muffle furnace.The ashed fiber samples were transferred into gamma vials and then stored for more than three months to reach the secular equilibrium of 227 Ac and 226 Ra with their daughter elements 223 Ra and 214 Pb, respectively.The activities of 227 Ac, 228 Ra, and 226 Ra were measured using a high-purity well-type germanium detector by counting their daughter nuclides 223 Ra, 228 Ac, and 214 Pb at 269, 911, and 351.9 keV, respectively.Since 228 Ac also has a gamma peak at 270 keV (branch ratio = 3.3%), the interference by 228 Ac was corrected using the counts at 911 keV of each sample and the count ratio of 270 to 911 keV for the " 227 Ac-free" 228 Ra standard (Text S1 in Supporting Information S1).
The 226 Ra activities at different depths from the EC1 (Cho et al., 2022) showed good agreement with the 226 Ra profiles reported in the previous studies (Harada & Tsunogai, 1986;Inoue et al., 2015;Yang et al., 1992).Then, the 227 Ac activities were calculated by multiplying the measured 227 Ac to 226 Ra ratios in the Mn fibers by the 226 Ra activities (Table S1 in Supporting Information S1).In this calculation, we assumed that the depth profile of 226 Ra activity remains constant for the measurement periods ( 226 Ra for some depths are interpolated).Additionally, the adsorption efficiency of Ac in the Mn-fiber mooring method was determined by comparing 227 Ac activities measured by the onboard filtration method using discrete 60 L samples with those from the Mn-fiber mooring method at similar depths.In a previous study, Mn-fiber columns were found to extract ∼99% of both Ac and Ra when seawater is passed through at <1 L min −1 (Reid et al., 1979).In this study, 227 Ac activities measured by the onboard filtration agreed very well with those from the Mn-fiber mooring samples within the counting uncertainties (Figure 2b and Table S2 in Supporting Information S1), validating the mooring method for acquiring the Ac activities.In addition, we further confirmed that the analytical results of 227 Ac obtained from the gamma counting were consistent with the results from the RaDeCC measurement (Figure S1 in Supporting Information S1).

Distribution Characteristics of 227 Ac in the East Sea
The vertical distribution pattern of 227 Ac in the East Sea was similar to that of 228 Ra (Figures 2a and 2b).The 227 Ac activities ranged from 9 to 15 dpm m −3 (mean: 11 ± 2 dpm m −3 ) in the surface layer of the East Sea (0-100 m) and decreased to a range from 0.2 to 1.0 dpm m −3 (mean: 0.5 ± 0.3 dpm m −3 ) in the intermediate layer (1,000-1,500 m).A slight increase in 227 Ac activity near the bottom (mean: 1.8 ± 0.9 dpm m −3 ) was also observed in the East Sea (Figure 2b).However, the distributions of 227 Ac in the East Sea were significantly different from those in other open oceans, which consistently showed the lowest activities in the surface and intermediate layers, increasing toward the bottom in the deep ocean (Geibert et al., 2002;Nozaki, 1984).
Thus, although the ratios of 228 Ra to 227 Ac should decrease with depth due to the difference in the half-lives of both nuclides, they increased with depth in this sea (Figure 2c), as explained further in Section 3.3.The average 231 Pa activities (0.15 ± 0.05 dpm m −3 ; Nozaki & Yamada, 1987) previously reported in the study region were insignificant, and thus the observed 227 Ac activities were mostly present in excess over the entire water column.Additionally, it is striking that the observed 227 Ac activities in the surface layer of the East Sea were approximately four times higher than those reported in the deep layer of the major oceans (Atlantic and Pacific) (Le Roy et al., 2023;Nozaki, 1993).
To our knowledge, only Geibert et al. (2008) reported enriched 227 Ac activities in coastal regions, although most significantly high 227 Ac activities were from groundwater samples.In particular, they reported that the 227 Ac activities in the Korea Strait (up to 15 dpm m −3 ), a shallow sill that connects the East Sea and the ECSYS shelf, were the highest among the observed coastal seawater samples.Given that the direct river discharge to the East Sea and atmospheric deposition of 227 Ac are insignificant (Yanagi, 2002), the major source of excess 227 Ac in the surface layer is likely from the adjacent ECSYS shelf water, similar to 228 Ra (Cho et al., 2022).

Shelf-Water Source of 227 Ac in the East Sea
The measured activities of 227 Ac and 228 Ra in the ECSYS shelf ranged from 7 to 35 dpm m −3 (mean: 19 ± 6 dpm m −3 ) and 530 to 1,010 dpm m −3 (mean: 730 ± 140 dpm m −3 ), respectively (Figure 3), showing the highest activities in the world's oceans.To estimate the contribution of 227 Ac and 228 Ra in ECSYS shelf water to the excess activities of these elements in the upper layer of the East Sea, the total export fluxes of 227 Ac and 228 Ra from the ECSYS shelf (F export ) and the input fluxes of both radionuclides into the East Sea (F input ), assumed to be equal to their decay fluxes in the water column, were calculated as follows: (1) where A ECSYS , A excess-ES , τ ECSYS , and λ represent the water column inventory (0-70 m) of 227 Ac or 228 Ra in the ECSYS shelf (dpm m −2 ), excess inventory of 227 Ac or 228 Ra in the upper layer (0-1,000 m) of the East Sea (dpm m −2 ), residence time of ECSYS shelf water (yr), and decay constant of 227 Ac or 228 Ra, respectively (yr −1 ).The residence time of ECSYS shelf water was obtained from the previous study (4.9 ± 1.5 years; Kim et al., 2005).The F export values of 227 Ac and 228 Ra were calculated to be 260 and 10,700 dpm m −2 yr −1 , respectively.These fluxes are approximately three and two times larger than the F input of 227 Ac (90 dpm m −2 yr −1 ) and 228 Ra (4,750 dpm m −2 yr −1 ) in the upper layer (0-1,000 m) of the East Sea, respectively.Thus, this estimation suggests that approximately a third and half of ECSYS shelf-borne 227 Ac and 228 Ra are present in the East Sea, respectively.
Since the ECSYS shelf is significantly influenced by Changjiang, one of the largest rivers in the world, riverine input of freshwater could cause the high 227 Ac activities in the shelf water.Assuming that 227 Ac to 228 Ra ratios in the river are equal to the ratios of their progenitors, the maximum riverine-endmember of 227 Ac ( 227 Ac river ) can be expressed as follows: 227 Ariver = 228 Rariver × ( 238 U∕ 232 Th)river × ( 235 U∕ 238 U)river (3) where 228 Ra river , ( 238 U/ 232 Th) river , and ( 235 U/ 238 U) river are the endmember values of 228 Ra (190 dpm m −3 ), 238 U to 232 Th ratio (2.88), and 235 U to 238 U ratio (0.0073) in the Changjiang, respectively (Andersen et al., 2016;Huang et al., 2009;Kim et al., 2005).Considering that the 227 Ac activities in the ECSYS shelf water are distinctly higher than the conservative mixing line between the maximum Changjiang endmember (∼3.9 dpm m −3 ) and Kuroshio water endmember (∼0.3 dpm m −3 ; Nozaki, 1993) (Figure 3), other major sources might be responsible for the unexpectedly high 227 Ac activities in the shelf water.
Benthic diffusion of 227 Ac from 231 Pa deposited on ECSYS shelf sediments by boundary scavenging might be other potential source of 227 Ac.To investigate the contribution of benthic diffusion, we estimated the maximum diffusive flux of 227 Ac (F diff ) from the shelf sediment to the overlying water column based on previous studies (Boudreau, 1996;Krest et al., 1999;Li & Gregory, 1974) as follows: where P, k d ,  D Sed Ac ,  D SW Ac , θ, and ∅ are the production rate of 227 Ac (atoms cm −3 yr −1 ), distribution coefficient of Ac (3,000; Nozaki et al., 1990), diffusive coefficient of Ac within sediment, diffusive coefficient of Ac in seawater, tortuosity of sediment, and porosity of sediment, respectively.To calculate the production rate (P), we used previously reported 231 Pa activities in the surface sediment of the East Sea (0.5 dpm g −1 ) because its activities have not been reported in the ECSYS shelf sediment to the best of our knowledge.The  D SW Ac was assumed to be the same as that of La due to their similar chemical properties (Nozaki et al., 1990).Based on the in situ temperature of bottom water (∼10°C) and porosity within the sediment (∼0.75) in the study region (Song et al., 2016;Zhang et al., 2008;Zhou et al., 2022),  D Sed Ac was calculated to be 2.6 × 10 −6 cm 2 s −1 .Then, the F diff of 227 Ac from the ECSYS shelf sediment was estimated to be ∼2.7 dpm m −2 yr −1 , corresponding to only 1% of the export flux from the ECSYS shelf (F export ) using Equation 1.Although we even considered that bioturbation and bioirrigation could increase benthic flux up to 10 times compared with only taking the molecular diffusion into account in the study region (Cai et al., 2014), benthic flux of 227 Ac was still calculated to be ∼10% of the export flux.Thus, the benthic flux of 227 Ac from the bottom sediment seems to be insignificant for the occurrence of its high activities in the ECSYS shelf water.The ECSYS shelf is known to have significantly high submarine groundwater discharge (SGD) flux based on Ra isotopes (Kim et al., 2005), in association with high tidal ranges (average up to 6 m) (Kim et al., 2011;Park et al., 2005).Geibert et al. (2008) also reported high 227 Ac activities in groundwater from Massachusetts and Rhode Island, although the activities are variable (1.5-100 dpm m −3 ) depending on the region.Thus, if the SGD in the study region includes old waters (10-100 years), it could be the main source of the excess 227 Ac in this shelf water.

Scavenging of 227 Ac in the Water Column
Although 228 Ra and 227 Ac are generally assumed to be conservative in the ocean, the 228 Ra to 227 Ac ratios increased with depth in the upper layer of the East Sea (0-1,000 m) (Figure 2c).This profile could occur by (a) slope input of 228 Ra, relative to 227 Ac, up to 1,000 m and (b) more efficient removal of 227 Ac, relative to 228 Ra, by particle scavenging.However, we rule out the possibility of slope input since 228 Ra activities in the southern part of the East Sea, close to the continental slope sediment, and those in the center of the East Sea are constant from 150 to 1,000 m depth (Kim et al., 2015;Okubo, 1980).Additionally, in the East Sea, horizontal transport occurs through shallow sills (<130 m), and the deep sea (130-3,700 m) is enclosed.
To examine the contribution of particle scavenging to the distributions of 227 Ac and 228 Ra to 227 Ac ratios in the upper layer of the East Sea (0-1,000 m), a reversible scavenging model was applied as follows (Moran et al., 2002;van der Loeff & Berger, 1993) since the vertical distribution of La, chemically most similar element to 227 Ac, has known to be dominantly governed by reversible scavenging in the ocean (Oka et al., 2009;Seo & Kim, 2020;Siddall et al., 2008).
where A, k z , z, S, and K represent the activities of 228 Ra or 227 Ac (dpm m −3 ), vertical eddy diffusion coefficient (cm 2 s −1 ), depth (m), particle sinking velocity (m yr −1 ), and ratio of particulate to dissolved phase (dimensionless), respectively.In this model, the advection, biological effect, and production from parent elements were neglected.S could be estimated from the particulate thorium-230 ( 230 Th) activity gradient with depth and production rate of 230 Th (P = 0.0262 dpm m −3 yr −1 ; ∂ 230 Th particulate /∂ z = P/S).Based on the particulate 230 Th gradient in the East Sea (2.6-3.9 × 10 −5 dpm m −4 ; Nozaki et al., 1987), S was calculated as 670-1,020 m yr −1 , which corresponds to the upper limit of previous works in other regions (300-1,000 m yr −1 ) (Bacon et al., 1985;Moran et al., 2002;Venchiarutti et al., 2011).Assuming a steady state condition, the solution to the equation above can be expressed as follows: where C 1 and C 2 are the constants that depend on the boundary conditions.In this study, C 1 and C 2 were calculated using the measured 228 Ra and 227 Ac activities at the surface (z = 0) and 1,000 m, respectively.Assuming no removal of 228 Ra by scavenging, k z was estimated to be 6.2 cm 2 s −1 by minimizing the sum of the squares of the difference between the model and the measured 228 Ra (Figure 4a).For 227 Ac, the observed activities were estimated to be 2-4 times lower than those from the model when only diffusion was considered (Figure 4b).
The observed 227 Ac activities fit closer when K is 0.10-0.15,indicating that 9%-13% of the total 227 Ac could reside in particulate form in the upper layer of the East Sea (Figures 4b and 4c).The K estimated in this study is similar to that of light rare earth elements in the Mediterranean Sea (Garcia-Solsona & Jeandel, 2020), a semi-enclosed sea with marine environmental conditions similar to the East Sea, supporting the fact that Ac has a chemical property similar to that of La rather than Ra.Kipp et al. (2015) also reported that 8%-12% of the total 227 Ac was present in the particulate phase near a hydrothermal vent in the North Atlantic Ocean, similar to the proportion of particulate La (up to 14%) in the same region (German et al., 1990).The partitioning coefficient of Ac in porewater was also observed to be 2-11 times higher than that of Ra (Kemnitz et al., 2023), indicating that Ac is much more particle-reactive than Ra.In addition, we note that if the scavenging model assumes the particle scavenging of Ra, larger amounts of 227 Ac would have to be removed (Figure S2 in Supporting Information S1).The removal of 227 Ac by scavenging could significantly vary depending on marine environments (e.g., depths of the nepheloid layer, the intensity of particle settling fluxes, and the concentrations of particulate Mn).Therefore, we suggest that scavenging effect should be considered in each study region through (a) measuring the proportions of particulate 227 Ac to total 227 Ac (Geibert et al., 2002)

Conclusions
This study measured high-resolution vertical profiles of 227 Ac in the marginal sea of the northwestern Pacific Ocean (East Sea), using a novel Mn-fiber mooring method for the first time.The distribution of 227 Ac in the East Sea showed surprisingly high activities in the surface layer, even approximately four times higher than those in the deep major oceans.This excess 227 Ac might be associated with the input of the adjacent East China Sea and Yellow Sea shelf water, mainly resulting from SGD.In addition, the distinct removal of 227 Ac by scavenging is revealed in the upper layer of the East Sea, based on the vertical distributions of 228 Ra to 227 Ac ratios and the reversible scavenging model.This result suggests that Ac is not as soluble as Ra in the ocean, which has been conventionally assumed, and it has a similar particle affinity to that of La.Thus, our results shed light on the unexamined large source and sink of 227 Ac and suggest that more careful consideration of 227 Ac scavenging is necessary for its application as an oceanic tracer.

Figure 1 .
Figure 1.A map showing the sampling stations of 227 Ac and 228 Ra, bottom topography, and patterns of surface currents in the East Sea, East China Sea, and Yellow Sea.The red and blue arrows indicate the warm and cold current, respectively.The dotted arrows represent the currents that can vary in intensity and direction according to the seasons.

Figure 2 .
Figure 2. Vertical profiles of (a) 228 Ra (dpm m −3 ), (b)227 Ac (dpm m −3 ), and (c) the activity ratios of 228 Ra to 227 Ac in the East Sea.The 227 Ac activities obtained from the onboard Mn-fiber filtration were compared with those obtained from the mooring Mn-fibers at similar depths (5, 500, 1,000, 2,000, and 2,298 m), respectively.The 228 Ra data are from the previous study byCho et al. (2022).

Figure 3 .
Figure 3. Plots of salinity versus (a) 228 Ra (dpm m −3 ) and (b) 227 Ac (dpm m −3 ) activities in the East China Sea and Yellow Sea shelf (ECSYS) water.The 228 Ra data in the ECSYS shelf are from the previous study by Kim et al. (2005).Red dashed lines indicate the theoretical mixing lines between the riverine endmember from the Changjiang and the Kuroshio endmember.

Figure 4 .
Figure 4. Vertical profiles of (a) 228 Ra (dpm m −3 ), (b) 227 Ac (dpm m −3 ), and (c) the activity ratios of 228 Ra to 227 Ac in the upper layer of the East Sea by fitting to different ratios of particulate to dissolved 227 Ac (K) using a reversible scavenging model.
, (b) measuring the ratios of 228 Ra to 227 Ac (this study), or at least (c) excluding the possible nepheloid layers based on transmission data (Le Roy et al., 2023) for using 227 Ac as a tracer.