Actinium-227 ( 227 Ac) is a deep-sea mixing tracer that combines the advantages of a source in the deep-sea sediments with a half-life of 21.8 y. 227 Ac is produced by the decay of its parent nuclide protactinium-231 ( 231 Pa; T 1/2 = 32 760 y), 231 Pa being itself produced following the decay of uranium-235 ( 235 U; T 1/2 = 7.04 10 8 y). 235 U activities in the ocean are mostly constant through space and time due to its long residence time (~0.5 Ma; Ku et al., 1977), which leads to a uniform production rate of 231 Pa in the water column. Once produced, 231 Pa rapidly adsorbs onto particles. It is then transported to the seafloor and thus, accumulates slowly in the sediments (Anderson et al., 1983). In the sediment, 231 Pa decays to 227 Ac which is then released to the overlying water column due to its higher solubility (Anderson et al., 1983;Nozaki, 1984;Nozaki et al., 1990;Nozaki, 1993, p.199;Geibert et al., 2002;Kemnitz et al., 2022). 227 Ac that diffuses out of the sediments is then redistributed in the ocean by transport (advection and upwelling) and mixing (Nozaki, 1984;Geibert et al., 2002;Koch-Larrouy et al., 2015). The 227 Ac that diffuses out of the sediment thus adds to the 227 Ac produced in the water column (from 231 Pa) that is considered to be at secular equilibrium with 231 Pa, assuming steady-state conditions (Nozaki, 1984). Excess 227 Ac activities ( 227 Ac ex , the difference between the total activities and 227 Ac at secular equilibrium with 231 Pa) are thus often observed up to ca. 500 m above the seafloor due to vertical mixing (Geibert et al., 2002;Nozaki, 1984). Nozaki (1984) proposed estimating vertical eddy diffusivity coefficients (K z ) from the 227 Ac ex vertical distribution. 227 Ac ex has also been used to estimate upwelling rates, since this process also brings 227 Ac toward the surface (Geibert et al., 2002;Haskell et al., 2015;Kemnitz et al., 2022). Finally, Kipp et al. (2015) showed that hydrothermal vents constituted a source of 227 Ac to the deep Atlantic Ocean that was equivalent to ~2-6% of the deep-sea sediment flux. 227 Ac concentration in the ocean is, however, especially low. Its global oceanic inventory is only 37 mol or 8.4 kg (Geibert et al., 2008). In addition, interpretation of its distribution requires the knowledge of its 231 Pa parent distribution. As a result, the use of dissolved excess 227 Ac as an oceanic tracer has been reported in few studies to date (Dulaiova et al., 2012;Geibert et al., 2008Geibert et al., , 2002;;Geibert and Vöge, 2008;Kipp et al., 2015;Le Roy et al., 2019;Koch-Larrouy et al., 2015;Nozaki, 1984Nozaki, , 1993;;Nozaki et al., 1990;Levier et al., 2021;Kemnitz et al., 2022). In the present study, we report nine vertical profiles of 227 Ac from a large-scale section in the North Atlantic Ocean and the Labrador Sea.
Journal Pre-proof
These 227 Ac data can be compared to the 231 Pa activities reported along that same section (Deng et al., 2018). The determination of both 227 Ac and 231 Pa at the nine full-depth vertical profiles provides a unique section of 227 Ac ex activities in the North Atlantic.
The GEOVIDE section (GEOTRACES GA01; PIs: Géraldine Sarthou, LEMAR,
An Optimum Multiparameter Analysis (OMPA) was used to estimate the contributions of the different water masses found along the GEOVIDE section (García-Ibáñez et al., 2018). Lacan and Jeandel, 2005;McCartney, 1992). Then, the Irminger Subpolar Mode Water (IrSPMW) results from the transformation of the Central Waters and the IcSPMW, northwest of the Irminger Sea (Krauss, 1995). The IrSPMW is located near Greenland (Fig. 2) (García-Ibáñez et al., 2015;Lacan and Jeandel, 2004;Read, 2000). (Read, 2000;Yashayaev and Dickson, 2008). The DSOW occupies the northern end of the Irminger Sea (station 44) and the deepest part of the Greenland continental slope (stations 69 and 77) (Fig. 2). The Lower North East Atlantic Deep Water (NEADWl) is a water mass with a southern origin lying at the bottom of the West European Basin (stations 1 to 26 in Fig. 2). It recirculates into the Rockall Trough and meets ISOW in the Iceland Basin (van Aken, 2000;McCartney, 1992;Schmitz and McCartney, 1993).
Due to the low 227 Ac activities in the ocean (<5 dpm m -3 , Geibert et al., 2002;0.1 -4.8 dpm m -3 , (Levier et al., 2021), the analysis of 227 Ac requires a pre-concentration step from large volumes of seawater. We used acrylic cartridges impregnated with manganese oxide (Mn-cartridges) mounted on large volume in-situ pumps, a method that is largely used to preconcentrate radionuclides from seawater such as Ra and Th isotopes, as well as 227 Ac (Henderson et al., 2013;Swarzenski and Baskaran, 2004).
Pumps were deployed at 6 to 13 depths at each of the investigated stations. Pumping was carried out for 3 to 4 hours to filter large seawater volumes (420 to 1565 L). The resulting flow rates mostly ranged from 3 to 6 L min -1 . Seawater was first filtered in situ through Supor (Pall, 0.8 µm pore size) or QMA (Sartorius, 1 µm pore size, Ø 142 mm) membranes to collect suspended particles, and then through the cartridges impregnated with MnO 2 (Mn-cartridges) to collect dissolved 227 Ac (as well as Ra isotopes).
For the deep samples, two Mn-cartridges (cartridge A and cartridge B) were placed in line to provide information on the yield of 227 Ac fixation onto the Mn-cartridges. Following collection, each Mncartridge was rinsed with Ra-free milli-Q water and slightly dried using compressed air. This protocol and the method used to determine the yield of 227 Ac fixation onto the Mn cartridges are presented in more detail in (Le Roy et al., 2019).
Following Shaw and Moore (2002), we used Radium Delayed Coincidence Counters (RaDeCC, Scientific Computer Instruments, USA; Moore and Arnold, 1996) to quantify 227 Ac activities on the Mn-cartridges. The 223 Ra in equilibrium with 227 Ac was determined by measuring the 219 Rn activities that are in secular equilibrium with 223 Ra and 227 Ac. This method is extensively described in (Le Roy et al., 2019) and briefly summarized below. The partially dried Mn-cartridges were placed in plastic cartridge holders in a closed helium circulation loop. Helium was allowed to circulate through the Mncartridges and carried the 219 Rn to the scintillation cell coated with ZnS, where alpha particles produced by the 219 Rn decaying into 215 Po were detected. A delayed coincidence system (originally developed by (Giffin et al., 1963) and then adapted by Moore and Arnold, (1996) allowed us to discriminate the signal associated to 219 Rn from those of other Rn isotopes ( 220 Rn and 222 Rn) which were not associated with 227 Ac. Corrections for chance coincidence counts were performed following Moore and Arnold
Journal Pre-proof (1996). The two RaDeCC systems used in this study were calibrated using a Mn-cartridge standard, i.e.
a Mn-cartridge containing a known amount of 227 Ac (Le Roy et al., 2019). Dissolved 227 Ac activities were then determined using the Mn-cartridge extraction efficiency determined from the two Mn cartridges placed in series (Le Roy et al., 2019). A mean extraction efficiency of 47 ± 12% was applied to all Mn-cartridges (1 SD, n=21; (Le Roy et al., 2019). Repeatability experiments allowed us to estimate the uncertainty associated with the 227 Ac analysis (19%, 1SD; Le Roy et al., 2019).
Excess 227 Ac activities ( 227 Ac ex ) were then calculated following:
This is assuming that a fraction of the 227 Ac present in the water column is at secular equilibrium with 231 Pa. In some cases, the dissolved 231 Pa activities were collected using a separate sampling system and therefore not determined at the exact same water depth as the 227 Ac activities (Deng et al., 2018).
We thus used the depth interpolated 231 Pa activity determined from samples collected above and below the 227 Ac sample.
The distribution of 227 Ac activities along the GEOVIDE section is shown in figure 3. As a comparison, the 231 Pa activities are also reported (isolines). Most samples displayed activities ranging between 0.02 and 0.4 dpm m -3 , with higher 227 Ac activities observed in deep waters near the seafloor:
deep waters at the bottom of the West European Basin and in the Irminger basin at 2500 m display 227 Ac activities of 0.6 and 1.1 dpm m -3 , respectively. In upper waters (~500 m), slightly higher activities can be found near the margins (e.g., stations 1 and 13 located on the Iberian margin or station 64 located near Greenland).
Excluding the enrichments observed in deep waters, 227 Ac activities determined along the section were in the lower range of the 227 Ac activities reported by Nozaki et al. (1998) in the northeast Pacific: 0.05 dpm m -3 in the upper water column to 2.68 dpm m -3 near the seafloor. However, the 227 Ac activities reported here were in the same order of magnitude as those reported by (Geibert et al., 2002) in the central Arctic ( 227 Ac activities ranging from 0.08 dpm m -3 at 1220 m to 0.28 dpm m -3 at 4220 m).
In the Atlantic Ocean, (Kipp et al., 2015) reported 227 Ac activities as low as 0.16 dpm m -3 at 1060 m and up to 0.55 dpm m -3 at 4700 m; higher 227 Ac activities (up to 1.44 dpm m -3 ) were associated with a neutrally buoyant hydrothermal vent plume.
In general, 227 Ac activities gradually increased with increasing depth at stations located in the West European Basin (stations 1, 13, 21 and 26; Fig. 4, Table 1). In contrast, the increase with depth was J o u r n a l P r e -p r o o f
Journal Pre-proof
generally less apparent at stations 38, 44, 64 and 69. These patterns followed the general patterns of the vertical distribution of 231 Pa activities that increased with increasing depth in the West European basin from 0.04 dpm m -3 in the surface waters to 0.28 dpm m -3 in the deep waters (Deng et al., 2018). In contrast, a lower and narrower range of 231 Pa activities was observed in the Irminger and Labrador basins (0.05 -0.15 dpm m -3 ; (Deng et al., 2018). At intermediate depths (below ca. 500 m and above 500 m above the seafloor), 227 Ac and 231 Pa activities were similar for most of the investigated stations.
This suggests that at these depths, 227 Ac originated from the decay of 231 Pa in the water column.
However, at several intermediate depths, 227 Ac activities were slightly lower than 231 Pa (e.g., stations 13, 21, 38).
The 227 Ac ex section is shown in Fig. 5. The 227 Ac ex distribution mirrors that of 227 Ac (Fig. 3).
Significant 227 Ac ex activities were found at several stations near the seafloor (13, 26, 32, 44, 64 and 69), but this was not the case for all stations, with excess activities reaching up to 1.0 dpm m -3 (Fig. 4 and5).
Occasional isolated peaks in 227 Ac ex activities were found at depths away from the seafloor (station 44; 2500 m; stations 1 and 21; 500 m). Surprisingly, surface waters also sometimes exhibited significant 227 Ac ex activities (stations 1, 13, 32, 44 and 64).
227 Ac is at secular equilibrium with 231 Pa for most of the stations between 500 m below the surface and ca. 500 m above the seafloor, excluding the sporadic higher 227 Ac values that were sometimes observed in the water column (Fig. 3). Indeed, 227 Ac ex is in most cases relatively low when compared to other oceanic regions (Geibert et al., 2008(Geibert et al., , 2002;;Nozaki, 1984;Levier et al., 2021). Water masses that display 227 Ac at secular equilibrium suggest that they did not enter in contact with sediments over the last 100 years (five 227 Ac half-life). East of the section, the vertical 231 Pa activities increase with increasing depth and may reflect reversible scavenging. In particular, the NEADWl, located in the eastern part of the section, has a southern component (Fig. 2). NEADWl, with an apparent age of at least 400 years, and is an old water mass (not recently ventilated) compared to the other water masses along the section that are less than 100 years old (Deng et al., 2018). NEADWl is thus characterized by relatively high 231 Pa activities, since 231 Pa activities increase with age (Deng et al., 2018). Consequently, the 227 Ac activities that are at secular equilibrium with 231 Pa also increase with increasing depth in the West European Basin (Fig. 3).
In water masses such as LSW, ISOW and DSOW (Fig. 2). Similar relatively low 227 Ac activities were found in the Arctic Ocean, even close to the seafloor (Geibert et al., 2002).
The 227 Ac ex activities increase near the seafloor at stations 13, 26, 32, 64 and 69 (Fig. 5), but this pattern is not observed at all stations as would be expected from a tracer originating from deep-sea sediments. The highest 227 Ac ex activity near the seafloor is found at station 13 (0.42 dpm m -3 ), which is the deepest station investigated along the GEOVIDE section (5500 m). Since the accumulation rate of 231 Pa in sediments increases with increasing water column depth, the flux of 227 Ac diffusing out of the sediment is also a function of the water column thickness (Geibert et al., 2002). The same explanation holds for the three deep stations (13, 21, 26) located in the deep basins (Iberian Abyssal plain; West European basin) that display significant 227 Ac ex activities near the seafloor. In addition, these stations are likely less impacted by advection and strong currents-since these basins are located away from boundary currents-than the other stations located west of the section. The 227 Ac enrichments near the seafloor are, however, restricted to the deep-sea and are rarely transported by vertical mixing further than 500 m above the seafloor (Fig. 4). This latter pattern contrasts with the profiles observed in the Pacific Ocean (Nozaki, 1984), where significant 227 Ac ex activities can be found at a greater distance from the seafloor.
Several vertical profiles of 227 Ac ex do not show an increase near the seafloor, where 227 Ac ex activities are expected (Nozaki, 1984). Benthic nepheloid layers were observed along the GEOVIDE section from the Iceland basin to the Labrador Basin (Stations 26 to 69), as deduced from decreases in the transmissiometry near the seafloor, indicative of an increase in the particle concentration (Fig. 4). These bottom nepheloid layers may impact the distribution of radionuclides in the dissolved phase. In particular, the particulate Mn concentrations (PMn) also increase near the seafloor within the bottom nepheloid layers at all these stations (Fig. 4). PMn is well known to scavenge radionuclides, including radionuclides reactive to particles such as Th or Pa. Scavenging near the seafloor has been reported by previous studies in the Labrador Basin (Bacon and Anderson, 1982;Deng et al., 2014Deng et al., , 2018)). This is potentially also the case for radionuclides such as Ra isotopes and 227 Ac, although this effect has never been investigated in past studies. It is the property of MnO 2 to adsorb radionuclides that has led to the use of fibers and cartridges impregnated with Mn to pre-concentrate radionuclides from seawater samples (Moore and Reid, 1973) and to co-p+recipitate Ra (Ghaleb et al., 2004) and Ac (Levier et al. 2021). Such impact, however, has never been clearly reported in oceanic studies. Some high manganese oxide concentrations are associated with particles resuspended from the sediment. However, we could not find, any statistical relationship between PMn concentrations and 227 Ac, 226 Ra and 231 Pa patterns near the seafloor (Fig. 4). Moreover, it is unclear why 227 Ac would be preferentially scavenged compared to 231 Pa. One hypothesis would be that recent re-suspension of surface sediments, with high
Journal Pre-proof this process may really impact the distribution of radionuclides such as Ra and Ac.
In the Irminger and Labrador Basins, some higher 227 Ac ex activities were observed in deep waters, at stations 64 and 69, or even more clearly at station 44. Peaks of 227 Ac ex may be indicative of a recent (younger than ~ 100 years) contact of the water mass with ocean boundaries such as margins and deep sediments (Nozaki, 1984;Nozaki et al., 1990) or indicative of a hydrothermal plume (Kipp et al., 2015).
The highest 227 Ac ex activities across the section (1.12 dpm m -3 ) were observed at 2500 m at station 44 in association with high dissolved iron (DFe) concentrations-up to 2.5 nmol L -1-between 2000 m and 2800 m (Fig. 4; Tonnard et al., 2018). The DFe and 227 Ac patterns observed at depth at station 44 may be attributed to the signature of a hydrothermal plume. Although there is no additional supporting evidence for hydrothermal activity (e.g. no parallel δ 3 He measurements were made), hydrothermal vents have been reported along the nearby Reykjanes Ridge (Le Roy et al., 2019). At station 64, a slight increase of DFe concentration-up to 2.1 nmol L -1 -was observed between 2000 and 3000 m (Fig. 4). The high 227 Ac ex activities observed at stations 64 and 69 in deep waters may be explained by lateral advection of waters that interacted with the margins. The Irminger and Labrador
Basins are not as deep as West European Basin and the Iberian Abyssal Plain. Second, currents are also stronger in this area, with the onset of the Deep Western Boundary Current (DWBC). These features may foster interaction between the water masses and the sediments.
Slightly higher 227 Ac activities were found at stations near the Iberian margin, at 500 m (stations 1 and 21) and in the upper 500 m at station 13. This pattern is similar to that reported in the Central Arctic where an input from Siberian shelves was suspected (Geibert et al., 2002). A similar increase at 500 m in the 227 Ac activities has also been recently reported in a vertical profile in the Weddell Gyre (Levier et al., 2021). Such high 227 Ac ex activities at relatively shallow depths may be explained by lateral advection of waters that interacted with shallow sediments. In coastal regions, boundary scavenging leads to preferential 231 Pa removal in areas of high particle flux and therefore accumulation of 231 Pa in margin sediments (Anderson et al., 1983;Bacon et al., 1976). However that flux maybe diluted by a higher detrital input than in deep waters. Another hypothesis is extensive irrigation of sediments by fauna that could produce high 227 Ac flux. Epping et al. (2002) showed that the geochemistry of the Nazaré Canyon sediments may be significantly affected by benthic macrofauna activity. Margin sediments are thus likely a source of 227 Ac to the water column. In particular, station 1 is located on the Iberian margin; the high 227 Ac ex activity found at relatively shallow depth could result from a lateral input of 227 Ac released by the shallow sediments deposited onto the margin, a signal that is then be advected toward offshore waters, thus reaching station 13 and station 21 via different paths (Barbot et al., 2022). Near the Iberian margin (station 1), transmissiometry data indicate a shallow nepheloid layer J o u r n a l P r e -p r o o f
Journal Pre-proof (100-450 m; Fig. 4). Based on a semi-realistic 3D tidal model and considering the sediment properties, Barbot et al. (2022) established that internal tides (ITs) generated in the Bay of Biscay and along the Iberian slope could facilitate sediment resuspension and could be the major mechanism for the resuspension happening below 300 m. As hypothesized for bottom waters, we cannot exclude that the shallow nepheloid layers impact the distribution of radionuclides such as Ra or 227 Ac (Fig. 4); in some cases, these nepheloid layers may act as a sink when particles scavenge chemical elements (e.g., radionuclides), while in other cases, they may act as a source when chemical elements are released from the particles.
When the vertical profiles of 227 Ac result from the diffusion of 227 Ac from the sediment and are not impacted by any other input or removal (e.g., bottom nepheloid layer), the vertical profiles of 227 Ac ex activities can be used to derive vertical eddy diffusivity coefficients K z (Nozaki, 1984). Vertical eddy diffusivity coefficients can be determined using a simplified vertical one-dimension model, assuming steady state conditions and that the impact of advection can be neglected (Nozaki, 1984):
Where K z is the vertical eddy diffusivity coefficient; A is the 227 Ac ex activity; z is the depth above the bottom; λ is the 227 Ac decay constant.
The solution of Equation 2 is given by:
Equation 3
Where A=A 0 at z=0 and A=0 at z→∞.
We applied this equation at station 13 located in the Iberian Abyssal Plain (Fig. 6).
Station 13 was selected because it is the deepest station (5455 m) of the section and it displays significant 227 Ac ex activities (up to 0.42 dpm m -3 ) near the seafloor. We do not observe a bottom nepheloid layer at this station (Fig. 4), as it was the case at several other stations where we cannot exclude that resuspended Mn oxides may impact 227 Ac activities near the seafloor. The K z value derived from the vertical profile of 227 Ac ex is 0.68 0.49 cm 2 s -1 at that station (Fig. 6). This value is within the lower range of K z values usually reported for the deep ocean (0.1-50 cm 2 s -1 ; Huh and Ku, 1998;Kaufman et al., 1973;Koch-Larrouy et al., 2015;Nozaki, 1984;Nozaki et al., 1990). This low K z can be explained by the topography of the Iberian Abyssal Plain since vertical mixing is linked to the underlying bathymetry and especially seafloor roughness (Polzin et al., 1997;Mauritzen et al., 2002). In
Journal Pre-proof the Atlantic Ocean, vertical diffusivity is high along the coastlines and the Mid-Atlantic Ridge compared to the interior of the basins (Hasumi and Suginohara, 1999;Mauritzen et al., 2002).
This study reports an oceanic section of 227 Ac activities built from 9 full-depth vertical profiles determined between Portugal and Canada in the North Atlantic. By combining the 227 Ac activities with the 231 Pa activities, we provide an entire section of 227 Ac ex activities. 227 Ac activities were relatively low along that section compared to other oceanic regions (e.g., Pacific Ocean). In most cases, 227 Ac was in secular equilibrium with 231 Pa. This suggests that at these depths, 227 Ac was only produced from the decay of 231 Pa in the water column and that there was no external input -younger than ~100 years -of High PMn concentrations were found in these layers and may contribute to scavenging radionuclides such as 227 Ac (and 226 Ra), a process that may impact the vertical profiles of 227 Ac ex (and Ra). At station 13 (Iberian Abyssal Plain) where no bottom nepheloid layer was observed, we estimated a vertical eddy diffusivity coefficient (K z ) of 0.670.14 cm 2 s -1 from the 227 Ac vertical distribution; this value is consistent with K z values previously documented for the deep ocean.
Journal Pre-proof