Actinium-227 as a tracer for mixing in the Deep Northeast Pacific

PDF

Download

Open document

Flip pages

Download a page range

Download transcript

Copy asset link

Request this asset

Request accessible transcript

Transcript (if available)

Content ACTINIUM-227 AS A TRACER FOR MIXING IN THE DEEP NORTHEAST PACIFIC by Nathaniel James Kemnitz ___________________________________________________________ A Thesis Presented to the FACULTY OF THE USC GRADUATE SCHOOL UNIVERSITY OF SOUTHERN CALIFORNIA In Partial Fulfillment of the Requirements for the Degree MASTER OF SCIENCE (Geological Sciences) May 2018 Copyright 2018 Nathaniel James Kemnitz 1 ACKNOWLEDGMENTS Let me begin by expressing my sincere gratitude to my advisor, Doug Hammond. The time and energy he continuously has given me over the years has inspired me to become a better person and scientist. His dedication and selfless attitude toward others and to this department has taught me the importance of sharing my time and knowledge with others. I would like to thank the staff and students of the Veterans Resource Center at Pasadena City College, who guided me throughout my academic career and helped me to get to where I am today. I would especially like to thank Lauren Arenson and Carol Calandra for their continuous support. I would like to express my appreciation to my unbelievable family. With their guidance, unconditional love, and support I was able to overcome the most debilitating low points of my life. I would also like to extend my heartfelt gratitude to the Earth Sciences Department and the Geochemistry Lab at USC. Special thanks to Audra Bardsley and Yi Hou for their help with cruise prep and lab work. I am grateful for the friendship they have shown me throughout my time at USC. I would like to thank Will Berelson and Jess Adkins (co-Chief Scientists) for the opportunity to participate on the C-Disk-IV 2017 cruise. Special thanks to Jonny Stutsman for help with pump prep and deployment. Thank you to the crew members and scientists onboard the R/V Kilo Moana for their guidance and support during the cruise. I would also like to thank my Thesis Committee, James Moffett, Will Berelson, and Doug Hammond. Funding for this project was provided by the National Science Foundation research grants OCE1436958, awarded to Doug Hammond and OCE1220302, awarded to Will Berelson. 2 TABLE OF CONTENTS ACKNOWLEGMENTS ................................................................................................................................ 1 LIST OF TABLES ......................................................................................................................................... 3 LIST OF FIGURES ....................................................................................................................................... 4 ABSTRACT ................................................................................................................................................... 5 INTRODUCTION ......................................................................................................................................... 7 GEOCHEMISTRY OF ACTINIUM-227 .................................................................................................... 10 THEORY ..................................................................................................................................................... 11 STUDY AREA ............................................................................................................................................ 15 MATERIALS AND METHODS ................................................................................................................. 18 DISSOLVED ACTINIUM-227 ....................................................................................................... 18 SOLID PHASE ACTINIUM-227 AND LEAD-210 ......................................................................... 20 CORE INCUBATION ..................................................................................................................... 21 RESULTS .................................................................................................................................................... 24 WATER COLUMN: DISSOLVED ACTINIUM-227 AND PROTACTINIUM-231 ......................... 24 SEDIMENTS: SOLID PHASE ACTINIUM-227 AND LEAD-210 .................................................. 32 CORE INCUBATION AND ACTINIUM-227 FLUX ...................................................................... 40 DISCUSSION ............................................................................................................................................... 44 CONCLUSIONS ........................................................................................................................................... 53 REFERENCES ............................................................................................................................................. 55 APPENDICES .............................................................................................................................................. 60 APPENDIX A: OPTIMIZING METHODS AND STANDARDIZATION FOR ACTINIUM-227 USING RADECC ........................................................................................ 60 APPENDIX B: RAPID MEASUREMENT OF ACTINIUM-277 USING A HIGH PURITY GERMANIUM WELL-TYPE DETECTOR ....................................................................... 98 APPENDIX C: COMMERCIAL CARTRIDGE ABSORPTION EFFICIENCY AND COUNTING PROTOCOL ............................................................................ 106 APPENDIX D: FLUX OF A RADIONUCLIDE DURING A CORE INCUBATION .................... 110 3 LIST OF TABLES TABLE 1: Results for 227 Ac and 231 Pa Activities as a Function of Depth along C-Disk-IV Transect ....... 25 TABLE 2: 227 Ac Full Profile Results along C-Disk-IV Transect ............................................................... 27 TABLE 3: Two-Layer 227 Ac Profile Results along C-Disk-IV Transect ..................................................... 27 TABLE 4: Activities of 210 Pb, 226 Ra, 227 Ac, and Excess 210 Pb in Sediments along C-Disk-IV Transect .... 35 TABLE 5: Core incubation results along C-Disk-IV Transect ................................................................... 41 TABLE 6: Bateman Equation Fit Results for Initial Activities of 223 Ra and 227 Ac from Core Incubation ................................................................................................................................................... 42 TABLE 7: Excess 227 Ac Flux Determined from the Distribution of 227 Ac in the Water Column, Core Incubation, and Nozaki’s Transport Equation ............................................................................................ 52 TABLE A-1: Standards Prepared ................................................................................................................ 87 TABLE A-2: 219-Channel Efficiency for RaDeCC Detectors for Standards Used .................................... 88 TABLE A-3: 219-Channel Efficiency for Standards Used as a Function of Flow Rate ............................. 93 TABLE A-4: MF Standard Efficiencies for RaDeCC Detectors ................................................................ 94 TABLE A-5: CC Standard Efficiencies for RaDeCC Detectors ................................................................ 95 TABLE A-6: CC and MF geometry 219-channel efficiencies for USC RaDeCC detectors 1, 3, 4, 5, and 7 ...................................................................................................................................................................... 95 TABLE B-1: HPGe Total Efficiencies and Branching Ratios .................................................................. 103 4 LIST OF FIGURES FIGURE 1: Schematic of the Geochemical Behavior of 235 U Series in the Ocean .................................... 10 FIGURE 2: Activities of 231 Pa in Surface Sediments in the North Pacific along 150˚ W .......................... 14 FIGURE 3: Study Area of the Northeast Pacific Basin (NEPB). .............................................................. 15 FIGURE 4: Abyssal Circulation Pathways in Pacific Ocean ...................................................................... 17 FIGURE 5: Illustration of Core Incubation for 227 Ac ................................................................................. 23 FIGURE 6: 227 Ac and 231 Pa Activities vs. DAB at Station 4 ...................................................................... 24 FIGURE 7: Profiles of Excess 227 Ac vs. DAB for all Stations along C-Disk-IV Transect ........................ 28 FIGURE 8: Excess 227 Ac and Neutral Density vs. DAB for all Stations along C-Disk-IV Transect ......... 29 FIGURE 9: Sediment Profiles of Excess 210 Pb and 227 Ac vs. Depth for all Stations along the C-Disk-IV Transect. ..................................................................................................................................................... 37 FIGURE 10: Bateman Equation Fits to Determine 223 Ra and 227 Ac Benthic Fluxes along the C-Disk-IV Transect ........................................................................................................................................................ 43 FIGURE 11: Vertical Eddy Diffusivity (K z ) vs. Topography Roughness for all Stations along the C-Disk- IV Transect .................................................................................................................................................. 46 FIGURE 12: Upper Layer Vertical Eddy Diffusivity (K z ) vs. ∆density/∆depth ........................................ 46 FIGURE 13: Illustration of Topography Throughout the NEPB ................................................................ 48 FIGURE 14: Neutral Density Gradients along the C-Disk-IV Transect .................................................... 49 FIGURE A-1: Schematic Diagram of RaDeCC System .............................................................................. 62 FIGURE A-2: Illustration of Moore Fibers and Commercial Cartridge Geometry .................................... 65 FIGURE A-3: STD27 219-Channel Efficiency vs. Time ............................................................................ 75 FIGURE A-4: STDK-2 and STDK-6 219-Channel Efficiency vs. Time ................................................... 78 FIGURE A-5: Bateman Equation Fits for STDK-1, STDK-5, and STDK-7 Activities .............................. 81 FIGURE A-6: 219-Channel Efficiency vs. Flow Rate ............................................................................... 83 FIGURE A-7: 219-Channel Efficiency vs. % Water Content .................................................................... 86 FIGURE B-1: Schematic Illustration for the Determination of 223 Ra at 270 keV .................................... 104 FIGURE C-1: Linear Fit for Commercial Cartridges B vs. A .................................................................. 107 5 Abstract 227 Ac (t 1/2 = 22 y) is produced by decay of 231 Pa (t 1/2 = 32,800 y). 231 Pa is scavenged in the water column by falling particulates that carry it to sediments, where it decays to its more soluble daughter. A fraction of this 227 Ac diffuses out of deep-sea sediments and is transported vertically and horizontally as it decays in the water column, providing a distribution that can be used to infer transport rates for other solutes in the deep ocean. 227 Ac was measured throughout the water column and sediments at 5 stations along a cruise track from Hawaii to Alaska in the Northeast Pacific Basin. Sampling took place in August 2017 onboard the R/V Kilo Moana (C- Disk-IV cruise). The supported activity of 227 Ac in the water column was determined from previous measurements of its 231 Pa parent made by others who worked near this region. The activity of excess 227 Ac increases with depth at all 5 stations in the Northeast Pacific. Apparent vertical eddy diffusivities (K z ) derived from a one-dimensional eddy diffusion-decay model increase with depth and in areas near rough topography. Areas around the Musician Seamounts and the Mendocino Fracture Zone (MFZ) show the highest values of K z , while areas in the higher latitudes, where the seafloor is characterized by westward dipping smooth topography, show lower values of K z . Furthermore, mixing is enhanced in a benthic layer (~ 500 m thick) near the seafloor and may decrease exponentially above this layer. This is strongly correlated with column stratification, which increases markedly above the benthic mixed layer. Two independent approaches were undertaken to quantify the source function of 227 Ac from deep-sea sediments in the Northeast Pacific: direct measurement of 227 Ac fluxes via core incubation and indirect estimates based on gamma counting 210 Pb and 227 Ac in sediments. Core incubation fluxes for 227 Ac are 3-10 times greater than the fluxes of 227 Ac required to support the decay observed in the water column, but may be affected by handling artifacts. The latter approach, which uses 210 Pb and 227 Ac measurements in sediments, provides a reasonable match 6 for the benthic flux required by a 1-D model to match the water column data observed at the stations 2 and 3, which occupy a more open basin, characterized by constant depths throughout a large area. The sediment data at the higher latitude stations 4 and 5, and station 1 near the Hawaiian ridge predicts a flux that is 3-4 times greater than required by the 227 Ac inventory in the overlying water column. Horizontal transport of water far from the bottom into a dipping slope could reduce the water column inventory of 227 Ac below what is predicted by the benthic flux. This would suggest that a 1-D model may not predict a correct K z value for these regions. However, if depth and bioturbation are constant along the flow path, then 227 Ac should provide a well-constrained vertical eddy diffusivity and water column inventory, as shown by the basin stations, 2 and 3. 7 Introduction Mixing that occurs near the seafloor plays a significant role in the global overturning circulation and is responsible for redistributing heat, salt, nutrients, and other solutes from the seafloor to the sea-surface (Mashayek et al., 2015; Talley, 2011). The variability and strength of mixing occurring near the seafloor has primarily been attributed to rough topography such as seamounts, ridges, canyons, and fracture zones (Decloedt and Luther, 2012; Munk and Wunsch, 1998; Polzin et al., 1997). Mixing will be enhanced within water masses that have relatively uniform densities in comparison to water masses that have large density gradients (Talley, 2011). Since the ocean is density-stratified, studying the movement of solutes across these diapycnal gradients are important, as this is the process that delivers cold saline water, rich in nutrients, back to the sea-surface. Previous research has determined that direct measurements of circulation and mixing near the seafloor can be a challenge without the aid of radionuclides (Ku and Luo, 2009). Radionuclides are useful due to their chemical properties and radioactive decay, which can be used to infer mixing rates, depending on their half-lives and how they behave in the water column. Another reason radionuclides are such useful tracers is that very small quantities can be accurately measured. For over a half-century now, various types of radioisotopes have been used to determine rates of chemical and physical processes occurring in the ocean (Broecker et al., 1967, 1968; Berelson et al., 1982; Hammond et al., 1990; Geibert, 2002; Shaw and Moore, 2002; Ku and Luo, 2009). These radionuclides have helped oceanographers understand fluxes and transport rates of other solutes in the ocean, facilitating ocean modeling. 8 In past studies, 222 Rn has been used to study diapycnal mixing in the oceans (Broecker et al., 1967, 1968; Chung and Craig, 1972; Chung and Kim, 1980; Sarmiento and Rooth, 1980; Berelson et al., 1982). Due to its half-life of 3.8 days, it is an excellent tracer for studying vertical mixing rates because it may not experience large lateral transport. However, due to its short half-life, it can only be used to model vertical mixing near the sea floor, often within 100 meters of the bottom. 227 Ac has shown potential as a tracer for vertical mixing on the basin scale (Nozaki, 1984; Nozaki, 1990; Geibert, 2002). Nozaki first proposed 227 Ac as a deep-sea tracer in 1984, but for years, 227 Ac went relatively unmeasured due to its low concentrations and difficulties with detection in ocean waters (Geibert, 2002). However, since the advent of Radium Delayed Coincidence Counters (RaDeCC), 227 Ac can quickly and accurately be determined (Moore and Arnold, 1996). Recently, the distribution of 227 Ac was used to infer transport rates for solutes in the deep Eastern South Pacific (Hammond et al., in prep). This method of analysis has furthered our understanding of how iron is transported throughout the Southern Ocean, which is important as this limiting nutrient helps draws drawn CO 2 from the atmosphere. The current literature suggests that circulation and mixing in the Deep Northeast Pacific is not fully understood. Tracers indicate several poorly understood water column features within this region, namely the Double Silica Maximum and intense ∆ 14 C ages around 2,500 meters depth (Talley and Joyce, 1992). The water parcel pathways which maintain these structures are attributed to a slow overturning ‘cul-de-sac’ cell which has unresolved flow paths (Hautala, 2018; Hautala and Hammond, in prep). Recent work suggests bottom topography, weak diapycnal mixing, and sloping neutral surfaces are needed to accomplish the observed vertical 9 overturning circulation in this region (Hautala, 2018; Kawabe and Fujio, 2012). To understand the aforementioned proposition more clearly, the distribution of 227 Ac has been measured on a transect through the Northeast Pacific in order to accomplish the following objectives: 1.) Describe general features of 227 Ac distribution in Northeast Pacific 2.) Apply 1-D modeling in order to determine eddy diffusivity profiles (Kz) 3.) Quantify the source function of 227 Ac from deep-sea sediments 4.) Determine if a 1-D mixing model is appropriate for the distribution of 227 Ac in the Northeast Pacific. 10 Geochemistry of Actinium-227 Figure 1 shows an illustration of the source and transport of 227 Ac in the water column. 227 Ac (t 1/2 = 22 y) is the daughter of 231 Pa (t 1/2 = 32,800 y), which is produced by 235 U decay in the water column but is then rapidly scavenged and carried to the seafloor by particles. 231 Pa decays in the sediments and releases some 227 Ac into pore waters, with some 227 Ac diffusing out of the sediments, where it is mixed vertically and horizontally in the water column (Nozaki, 1984). 227 Ac will be in excess over its parent near the SWI (sediment water interface) and approaches equilibrium with its parent near the ocean’s surface. Figure 1: Schematic of the Geochemical Behavior of 235 U Series in the Ocean. 11 Theory Figure 1 illustrates 227 Ac and 231 Pa geochemistry in deep-sea settings. The parent of 231 Pa is 235 U, which has a constant concentration throughout the world’s oceans. Because 231 Pa is strongly particle-reactive, models of reversible adsorption/desorption indicate its profiles should increase with increasing depth in the water column (Anderson et al., 1982). Removal from the water column should also result in higher concentrations in abyssal sediments (Nozaki, 1984; Geibert, 2002). 227 Ac excess is not observed in shallow marine environments because high inputs of detrital sediments strongly dilute the modest excess 231 Pa generated in the shallow overlying water column. This leads to the conclusion that the dominant source of excess 227 Ac is deep-sea sediments (Nozaki, 1984; Geibert, 2002), providing a first approximation of boundary conditions for modeling. 1-Dimensional Model The equation defining change of concentration of a dissolved radioactive solute in a 1- dimensional section in the water column is as follows: (1) where: C (dpm m -3 ) is the excess activity of the dissolved radioisotope, z (m) is the distance above the SWI (sediment water interface), K z (m 2 yr -1 ) is the vertical eddy diffusivity, € dC dt =K z d 2 C dz 2 −ω dC dz −λC 12 w (m yr -1 ) is advection, and l (yr -1 ) is the decay constant. If vertical eddy diffusivity (K z ) is assumed to be constant over the entire water column profile, steady state is invoked, and advection is assumed to be very small in the vertical direction, equation 1 can be simplified to equation 2: (2) The vertical eddy diffusivity (K z ) can be determined by solving equation 2 and fitting the following solution to the excess 227 Ac profile in the water column, assuming the boundary condition that C à 0 as z à ¥: (3) where: z is distance above bottom in meters, C is the activity of excess 227 Ac (dpm m -3 ), C 0 is the activity of excess 227 Ac (dpm m -3 ) at the SWI, and b (m -1 ) is the inverse scale length of the exponential curve. K z (m 2 yr -1 ) and b are related by the following relationship: (4) The aforementioned equation governing the vertical distributed 227 Ac excess activity can provide estimates of diapycnal mixing in the water column if horizontal transport is not significant. However, this can only be accomplished if the source of 227 Ac excess is constant € 0 =K z d 2 C dz 2 −λC € C =C 0 exp(−bz) € b = λ K z 13 throughout the study area and vertical advection is negligible. To test these assumptions, the source function and water column stratification can be studied. Nozaki et al., 1990 modified a transport equation original developed for 226 Ra (Cochran and Krishniswami, 1980) to calculate the flux 227 Ac across the SWI. The following is an expression that characterizes the benthic flux of 227 Ac: J 227Ac = fr(1-f)A 231Pa [(D s + KD b )l/(1 + K)] 0.5 (5) where: f is the fraction of 227 Ac that escapes the solid grain when it decays from 231 Pa (dpm g -1 ), f is the porosity, r (g cm -3 ) is the density of deep-marine sediments, A 231Pa is the activity of 231 Pa in the sediments, D b (cm 2 yr -1 ) is the bioturbation coefficient, D s (cm 2 yr -1 ) is the diffusion of 227 Ac through the pore water, K is the dimensionless partition coefficient for 227 Ac, and l is the decay constant for 227 Ac. The derivation assumes all parameters are independent of depth and the water column concentration is negligible. The source function of 227 Ac, in theory, should reflect the activity of 231 Pa in sediments which is largely controlled by water depth, sediment accumulation rate, and bioturbation (Nozaki, 1984; Hammond et al., in prep). Figure 2 shows the relationship of 231 Pa activity in surface sediments vs. latitude in the North Pacific (Yang et al., 1986; Lao et al, 1992; Hammond et al., in prep). Equation 5 can be a powerful tool for defining the 227 Ac flux if other terms in Eq. 5 are known. The source function of 227 Ac might also be determined by means of core incubation. Core incubations have shown to be good estimates for benthic fluxes of 222 Rn off the California margin (Wolfe, 2011). To date, determining the flux of 227 Ac via core incubation has never been 14 attempted, but establishing if this is an accurate approach is important for future measurements of 227 Ac source functions. Figure 2: Activities of 231 Pa in Surface Sediments in the North Pacific along 150˚ W. Data from Lao et al. (1992) and compilation of Yang et al. (1986). Average (±1 sdom) is 3.38±0.26 dpm/g. The point in ( ) is a site less than 3 km depth (Figure from Hammond, pers. comm.). 0 2 4 6 -20 -10 0 10 20 30 40 50 Yang et al Lao et al 231Pa (dpm/g) Latitude (°N) Ave ( ) 15 Study Area Study Area of Actinium-227 Figure 3 shows the C-Disk-IV (Carbonate Dissolution Kinetics-IV) Transect in the Northeast Pacific Basin (NEPB). 227 Ac was measured throughout the water column at 5 stations along a cruise track from Hawaii to Alaska. Sampling took place in August 2017 onboard the R/V Kilo Moana (Chief Scientists: William Berelson and Jess Adkins). Figure 3: Study Area of the Northeast Pacific Basin (NEPB). Red dots refer to stations where 227 Ac was measured on the C-Disk-IV cruise. Black triangles are stations were 231 Pa data have been measured (Hayes et al., 2013). Base map from ODV. S0202-24 S0202-32 SAFe Aloha 16 Mode Waters of the Northeast Pacific Basin The Northeast Pacific Basin (NEPB) is largely filled by Pacific Deep Water (PDW), which is predominately formed via upwelling, diffusion, and mixing of other mode waters that enter from the South Pacific (Talley, 2013; Talley, 2011; Kawabe and Fujio, 2012). The PDW is characterized by its low oxygen, high nutrients, and large 14 C age. Below the PDW is an abyssal layer ( > s q ~ 28.0 kg/m 3 ) composed of denser waters that originate from the Southern Ocean, which is largely fed into the NEPB through a northern route at a rate of 3 Sverdrup (Sv). The boundary between these two deepest layers is highly eroded (Talley, 2011). The top two ocean layers in the NEPB are the upper mixed layer, which occupies a density range between s q ~ 26.7 kg/m 3 to 27.6 kg/m 3 and the intermediate layer, which occupies a density range between s q ~ 24.5 kg/m 3 to 26.5 kg/m 3 , both which have sources at the surface in the North Pacific. Abyssal and Deep Circulation in Northeast Pacific Basin The abyssal waters that occupy the NEPB are derived from the western boundary current system in the South Pacific (Hautala, 2018). The main pathway of these abyssal waters into the NEPB is from the western side of the Emperor Seamount Chain, mostly along a northern route that travels through the Aleutian Trench at a rate of 3 Sv (Kawabe and Fujio, 2012). Another branch is funneled through the main gap of the Emperor Seamount around 40˚N and joins with a westward flowing current that then moves eastward along the Mendocino Fracture Zone at a rate of 1 Sv (Figure 4). 17 The topography in the NEPB is relatively smooth with a gentle west dipping slope between 140˚W and 180˚E, north of 40˚N. Rough topographies are present around the Hawaiian Ridge and Mendocino Fracture Zone, where the seafloor topography steps up by 1 km to the north (Hautala, 2018). Figure 4: Abyssal Circulation Pathways in the Pacific Ocean. Numbers on the figure refer to Sverdrups (Sv). Open circles are suggested areas of upwelling. Three routes of volumetric transport enter the NEPB. The black solid line refers to the C-Disk-IV Transect (Figure modified from Kawabe and Fujio, 2012). 18 Materials and Methods Sample Collection Sample material was collected during the C-Disk-IV Cruise aboard the R/V Kilo Moana in August 2017. Five stations were sampled along a transect from Hawaii to Alaska (Figure 3). Profiles of dissolved 227 Ac in the water column were collected at each station, along with sediments for measurement of solid phase 227 Ac and 210 Pb. Lastly, core incubations were carried out to determine the flux of 227 Ac from sediments. Dissolved Actinium-227 Dissolved 227 Ac was collected using a dual-flowpath in-situ pump (McLane WRT-LV). Two casts were conducted at each station (shallow and deep). Five to seven pumps were deployed on each cast. Each pump had either a single 1 µm quartz fiber (Whatman QMA) filter or two 0.8 µm polyethersulfone (Pall Supor800) filters in parallel. The QMA filter was used to capture particulate calcite/aragonite, and all flow went through the single filter. This allowed sufficient flow to pass about 1.5 m 3 during a 4-hour pump. The Supor filters restricted flow more than the QMA filters, so most pumps were fitted with 2 Supor filters mounted in parallel, whose outputs were merged to allow sufficient flow to the MnO2 cartridges. Downstream of the filter heads, flow then passed through two grooved acrylic cartridges impregnated with MnO 2 that sat in series. Having two cartridges mounted in series allows determination of cartridge absorption efficiencies of radium, actinium, and thorium from seawater. Appendix C describes the determination of cartridge efficiencies for this cruise. The 19 average absorption efficiency for all stations along the C-Disk-IV transect was found to be 54% ± 6%. After collection, cartridges were rinsed with DIW for several minutes to remove sea-salt and then dried to 50-120% moisture using compressed air. MnO 2 cartridges were returned to USC where they were analyzed using RaDeCC (Radium Delayed Coincidence Counter). Cartridges for stations 4 and 5 were initially analyzed 10-20 days after collection and these and all other stations were analyzed > 2 months after collection. The two-month waiting period allowed 223 Ra to grow into equilibrium with its parent, 227 Ac. Comparison of the initial and later analyses at stations 4 and 5 permitted estimations of excess 223 Ra. However, little excess was detected. For RaDeCC analysis (Moore and Arnold, 1996), dried cartridges were placed in a closed system device that re-circulates helium through the cartridges and a Lucas scintillation cell that detects alpha decays of 219 Rn and 220 Rn using a photomultiplier tube (PMT). MnO 2 cartridges are kept moist to achieve the maximum emanation efficiency of the two radon daughters (Sun and Torgersen, 1998a; Moore and Cai, 2013). The delayed coincidence system detects 219 Rn (t 1/2 = 3.96 s) and 220 Rn (t 1/2 = 55.6 s), followed by the detection of the polonium daughters’ decays ( 215 Po = 1.78 ms; 216 Po = 145 ms;), captured in one of the two time gated windows that open after detection of an initial alpha decay. The 219 Rn and 220 Rn activities reflect the activities of 223 Ra and 224 Ra that are absorbed onto the Mn-fibers. If it is assumed that 223 Ra is in secular equilibrium with 227 Ac, then the activity of 219 Rn will equal the activity of 227 Ac on the MnO 2 cartridge. Similarly, 220 Rn and 224 Ra will grow into secular equilibrium with 228 Th, and ultimately 228 Ra. Five RaDeCC detectors were operated to measure 227 Ac from the MnO 2 cartridges (Detectors 1, 3, 4, 5, and 7). Detector efficiencies ranged from 26 – 44% efficient for detecting 20 219 Rn (see Appendix A for details on RaDeCC standards and detector efficiencies). Detector efficiencies were determined by measuring standards spiked with known amounts of 227 Ac and 235 U (STDK-2 and STD27). Standard K-2 had a geometry like the cartridges used for water column samples, while STD27 was mounted on loose fibers. Furthermore, standards had activities that were similar to 227 Ac/ 223 Ra activities seen in deep-sea environments (STD27 = 2.56 dpm; STDK-2 = 7.9 dpm). Standards were routinely analyzed on all 5 detectors on a weekly basis, to monitor efficiency drifts with time (additional details in Appendix A). Solid Phase Actinium-227 and Lead-210 Sediment cores were collected at stations 1, 2, 3, 4, and 5 along the C-Disk-IV transect using a multi-coring device. The multi-cores were polycarbonate tubes (10 cm ID). Sediment cores were immediately taken to a cold room on board the R/V Kilo Moana after retrieval. Sediment cores were sectioned every 1 cm between 0-10 cm and every 2 cm between 10-30 cm. Sectioned mud was placed in a cold van after sectioning and then analyzed for porosity at the USC lab by weighing aliquots before and after drying at 50˚C. Samples were dried, gently crushed, and placed in polyethylene or polypropylene tubes. The weights and heights of the samples averaged between 0.9 - 2.0 g and 2.5 - 3.5 cm. Weights for bottom sediments were corrected for salt contribution, based on water content when collected. Solid phase 227 Ac and 210 Pb in sediments were measured using high purity intrinsic germanium well-type detectors (HPGe ORTEC, 120 cm 3 active volume). Detectors MCB1 and MCB2 were used to measure the activity. Detector efficiencies were determined from counting the activities of known standards obtained from EPA (SRM-1 diluted pitchblende and SRM-2 21 diluted monzonite) and Eckert & Zeigler (WHOI E&Z Actinium-227 CRM). Standards were 3.0 cm high geometry, and corrections were made to sample results to account for the different sample heights that were used. Samples were counted for 2-4 days and the spectra (keV) were analyzed for the following radioisotopes: 210 Pb (46.5), 234 Th (63.3), 226 Ra (186.2), 235 U (185.7), 228 Ac (338.0, 911.2), 223 Ra (269.5), 219 Rn (271.2), and 227 Th (256.3). For samples with high 226 Ra/ 235 U ratios, use of the 186 keV peak appears to provide accurate estimates for 226 Ra activity (Kemnitz, 2016). The 269.5 keV was corrected for 228 Ac interference (BR=3.34%), which appears accurate for high 223 Ra/ 228 Ac ratios (see Appendix B for 227 Ac analysis on gamma detector). Lastly, excess 210 Pb was determined by subtracting the supported 210 Pb ( 210 Pb = 226 Ra) activity from total 210 Pb activity and correcting for decay between collection and analysis. Core incubation and Actinium-227 Flux Core incubations were carried out using the methodology from Hammond et al (2004). Briefly, after retrieval of the cores from the multi-core device, the spring-loaded arms were released and rubber stoppers were placed on the bottom of the cores. Next, an incubation plug was placed on top, which was adjusted to allow 12cm of overlying water between the SWI and the incubation plug. The incubation plug had a stirring device that continuously stirred the overlying water for the entire incubation period. Incubation period varied for each site. Stations 1, 2, and 3 were incubated for 16, 17, and 13 days, while stations 4 and 5 were incubated for 10 and 5 days. Ending incubation heights in the core average around 10 cm height. Figure 5 shows an illustration of the core incubation method. 22 After the incubation period was completed for nutrient analysis, the left-over bottom water in the core was siphoned out and all water was combined from multiple cores at each site (~1.0 L). This water was passed through loose MnO 2 coated fibers at least 3 times at < 1.00 L/min to assure complete absorption of radium and actinium. MnO 2 fibers were then stored at room temp and shipped back to the USC lab. Upon arrival at USC, the MnO 2 fibers were washed several times with DI water to remove sea-salt and then dried to 80-120% moisture to be analyzed for its 223 Ra activity on RaDeCC. Core water MnO 2 fibers were measured on RaDeCC shortly after incubation period was completed. Stations 4 and 5 measurements began a week after incubation ended, while stations 1, 2, and 3 began 14-20 days after incubation ended. This immediate and the subsequent measurement of core water MnO 2 fibers allowed for both 223 Ra and 227 Ac activities to be calculated, using the Bateman equation fit to the data (See Results). The following equation was used to determine the 227 Ac flux from core incubation: 𝐽 ""#$% =𝐴 ""#$% × ) * (6) Where A 227Ac is the activity of 227 Ac in units of dpm/m 3 , which was determined from the Bateman equation fit, t is time of incubation period, and h is height of the overlaying water above the SWI in the core. Because aliquots of water were periodically removed to use for other analyses during the incubation, h/t must be summed for each incubation interval. For 223 Ra, the flux was calculated from equations developed by D. Hammond (pers. comm.) in Appendix D that considers the effect of radioactive decay on the solute accumulating in the overlying water, so that 23 +, +* = - ) − 𝜆𝐶 (7) for each interval, where C is the concentration of excess 223Ra and lambda is the decay constant. Figure 5: Illustration of Core Incubation for 227 Ac. After ending incubations to measure nutrient fluxes, remaining water above the SWI was siphoned out, filtered (0.45 µm, Acrodisc), and passed 3x through MnO 2 fibers (Figure from Hou, 2017). 24 Results Water Column: Dissolved Actinium-277 and Protactinium-231 The distribution of 227 Ac along the C-Disk-IV transect in the water column decreases in activity with increasing DAB (depth above bottom). 227 Ac activity approaches equilibrium with its parent, 231 Pa, near 3000 meters DAB, assuming 231 Pa activity is equal to that measured by Hayes et al. (2013). Although 231 Pa was not measured in this study, 231 Pa data was available for 4 locations located around the C-Disk-IV Transect (Fig. 3). 231 Pa for stations SO202-24 and SO202-32 was averaged and extrapolated to C-Disk-IV stations 3, 4, and 5. 231 Pa for stations Aloha and SAFe was averaged and extrapolated to C-Disk-IV stations 2. And finally, 231 Pa for station Aloha was found to be that of C-Disk-IV station 1. This behavior of 227 Ac and 231 Pa throughout the water column is as expected. Figure 6 shows a typical profile of 227 Ac and 231 Pa activities (dpm m -3 ) vs. DAB. Table 1 summarizes all stations’ 227 Ac and estimated 231 Pa activities vs. depth along the C-Disk-IV transect. Figure 6: 227 Ac and 231 Pa Activities vs. DAB at Station 4. This is a typical profile of 227 Ac and 231 Pa activities (GEOTRACES Data Extractor) observed along C-Disk-IV transect. Activity units are in dpm (disintegration per minute) per meter cubed. 25 Table 1: Results for 227 Ac and 231 Pa Activities as a Function of Depth along C-Disk-IV Transect. Depth above the bottom (DAB) and excess 227 Ac are also shown. The excess 227 Ac was derived by subtracting total 227 Ac from 231 Pa activities. The numbers associated with filter types are the number of filters used on each cast. 231 Pa values were taken by Hayes et al., (2013). Most stations 231 Pa values were averaged between blank triangles stations in figure 1. Station 1 231 Pa values were determined at that location. Depth (m) DAB (m) Filter Type 231 Pa (dpm/m 3 ) sig Total 227 Ac (dpm/m 3 ) sig Excess 227 Ac (dpm/m 3 ) sig Station 1 150 4580 1 GFF 0.03 0.01 0.06 0.03 0.00 0.04 Bottom Depth 3200 1530 none 0.50 0.01 0.50 0.06 0.00 0.06 4730 m 3475 1255 1 Supor 0.50 0.01 0.48 0.04 0.00 0.04 4075 655 1 Supor 0.45 0.01 0.71 0.03 0.26 0.03 4375 355 1 Supor 0.40 0.01 1.32 0.11 0.92 0.11 4575 155 1 GFF 0.40 0.01 1.22 0.10 0.82 0.10 4625 105 1 Supor 0.40 0.01 1.14 0.09 0.74 0.09 Station 2 620 5020 1 GFF 0.06 0.01 0.00 0.02 0.00 0.02 Bottom Depth 1200 4440 2 Supor 0.09 0.01 0.09 0.03 0.00 0.04 5640 m 2200 3440 2 Supor 0.33 0.08 0.35 0.02 0.00 0.08 3200 2440 2 Supor 0.41 0.08 0.33 0.04 0.00 0.08 5050 650 2 Supor 0.39 0.07 1.46 0.13 1.07 0.14 5200 500 2 Supor 0.39 0.06 1.77 0.18 1.38 0.19 5275 425 2 Supor 0.39 0.06 1.48 0.11 1.10 0.12 5300 400 1 GFF 0.39 0.06 1.25 0.11 0.86 0.13 Station 3 75 5485 1 GFF 0.04 0.01 0.02 0.01 0.00 0.01 Bottom Depth 2000 3560 2 Supor 0.15 0.02 0.11 0.05 0.00 0.05 5560 m 4000 1560 2 Supor 0.35 0.03 0.43 0.06 0.07 0.07 5050 510 2 Supor 0.33 0.03 1.75 0.10 1.42 0.11 5200 360 2 Supor 0.33 0.03 1.57 0.05 1.24 0.06 5300 260 2 Supor 0.33 0.03 1.71 0.13 1.38 0.14 Station 4 150 4820 1 GFF 0.04 0.01 0.06 0.01 0.00 0.06 Bottom Depth 2000 2970 2 Supor 0.15 0.02 0.12 0.04 0.00 0.03 4970 m 3500 1470 2 Supor 0.33 0.03 0.33 0.04 0.00 0.05 4500 470 2 Supor 0.35 0.03 1.69 0.06 1.35 0.07 4600 370 2 Supor 0.35 0.03 2.14 0.24 1.79 0.25 4700 270 1 GFF 0.33 0.03 1.65 0.07 1.32 0.08 4800 170 2 Supor 0.33 0.03 2.02 0.11 1.69 0.12 Station 5 100 4625 1 GFF 0.04 0.01 0.06 0.02 0.00 0.02 Bottom Depth 2000 2725 2 Supor 0.15 0.02 0.08 0.03 0.00 0.04 4725 m 4250 475 2 Supor 0.35 0.03 0.63 0.05 0.28 0.06 4450 275 2 Supor 0.35 0.03 1.24 0.05 0.88 0.06 4500 225 1 GFF 0.35 0.03 1.18 0.15 0.83 0.16 4550 175 2 Supor 0.35 0.03 1.44 0.12 1.09 0.13 26 One-Dimensional Exponential Fits to Excess 227 Ac and Vertical Mixing Excess 227 Ac was calculated by subtracting 231 Pa from total 227 Ac activities at each depth in the water column (Table-1). The excess 227 Ac throughout the entire water column was plotted as a function of DAB and fit with a 1-D exponential model (equation 3) at each station. The values of b were then used with equation 4 to find the vertical eddy diffusivity (K z ). Figure 7 shows each station’s excess 227 Ac vs. DAB and the resulting eddy diffusivity. Furthermore, the resulting profile of excess 227 Ac vs. DAB can be integrated to obtain a 227 Ac flux: F = lC 0 /b (8) Where l is the decay constant of 227 Ac, C 0 is the concentration of excess 227 Ac at the SWI, and b is the inverse scale length of the exponential fit. The 227 Ac flux is in units of dpm/m 2 yr and describes the 227 Ac activity that is transported across the SWI, which is needed to support the 227 Ac activity in the water column (for a 1-D model). Table 2 summarizes each station’s eddy diffusivity, flux, and initial concentrations at the SWI. All results in Table 2 were obtained from the 1-D exponential model (equation 1) that was fit to the entire water column profile. Excess 227 Ac was also modeled by treating the water column profile as a two-layered structure. Figure 8 shows each station’s two layered structure along with Neutral Density vs. depth. Each profile was fit with equation 3 in the upper mixed layer and equation 4 was used to find the vertical eddy diffusivity (K z ) for this interval. Boundary conditions for this model were determined from the break in slope from the Neutral Density vs. depth graph (Figure 8). The bottom layer is assumed well-mixed and its concentrations are constant throughout the deep 27 layer interval. Mixed layer depth, thickness, activity, K z , and flux for each station are summarized in Table 3. Table 2: 227 Ac Full Profile Results along C-Disk-IV Transect. Excess 227 Ac concentrations at the SWI (C O ), vertical eddy diffusivity (K Z ), and excess 227 Ac Flux were determined by fitting equation 3 to the full profile. 227 Ac fluxes were determined from integration of 227 Ac excess vs. Depth Above Bottom (eq. 7). Uncertainties are derived from the exponential fit uncertainties. Station C 0 +/- K Z +/- Flux +/- # dpm/m 3 cm 2 /s dpm/m 2 -yr 1 1.1 0.2 3.7 1.4 21 9 2 1.8 0.4 10.4 5.2 60 25 3 2.1 0.4 6.0 2.3 52 20 4 2.3 0.4 5.2 2.0 50 19 5 2.2 0.4 0.6 0.2 17 3 Table 3: Two-Layer 227 Ac Profile Results along C-Disk-IV Transect. The mixed layer depth and thickness was determined from the break in slope from the Neutral Density vs. DAB graph. The mixed layer activity is assumed constant throughout the mix layer interval. The upper and lower density slopes are calculated from the neutral density vs. DAB linear fit. The upper layer K z is the exponential fit starting at the top of the mixed layer up to the sea-surface. Usually only 2 points were used so there are no uncertainties. Upper layer flux was determined from equation 8, and the deep layer flux was determined through numerical integration of the bottom mixed layer 227 Ac excess activity. Station Mixed Layer Depth Mixed Layer Thickness Mixed Layer Activity Upper Density Slope Lower Density Slope Upper Layer Kz Deep Layer Flux Upper Layer Flux Total Flux # m m dpm/m 3 kg/m 4 kg/m 4 cm 2 /s dpm/m 2 -yr dpm/m 2 -yr dpm/m 2 -yr 1 4100 630 0.83 ± 0.09 1.73E-06 7.81E-06 0.04 16 ± 2 7 ± 1 23 ± 3 2 4800 840 1.16 ± 0.13 2.43E-05 1.25E-05 1.25 31 ± 33 13 ± 14 44 ± 37 3 4900 660 1.35 ± 0.11 1.51E-05 4.87E-06 1.12 28 ± 2 14 ± 1 42 ± 3 4 4450 520 1.51 ± 0.24 2.04E-05 4.25E-06 0.37 25 ± 4 9 ± 1 34 ± 5 5 4300 425 0.91 ± 0.13 2.47E-05 8.45E-06 0.24 12 ±2 5 ± 1 17 ± 2 28 Figure 7: Profiles of Excess 227 Ac vs. DAB for all Stations along C-Disk-IV transect. Eddy diffusivities and 227 Ac flux are derived from 1-D exponential model fits to the entire profile 29 30 31 Figure 8: Excess 227 Ac and Neutral Density vs. DAB for all Stations along C-Disk-IV Transect. The exponential fits to the profiles begin at the top of the mixed layer, as determined by the break in slope from Neutral Density vs. DAB. 32 Sediments: Solid Phase Actinium-227 and Lead-210 Measurements for bottom sediments are summarized in Table 4. Figure 9 shows 227 Ac concentration (dpm/g) vs. depth (cm) for all 5 stations along the C-Disk-IV transect. For most profiles, the distribution of solid phase 227 Ac in the sediments increases in activity to a finite depth, then decreases to zero around 10-20cm. It is assumed that 227 Ac activity approaches equilibrium with its parent, 231 Pa, near 3-5 centimeters depth for most stations (Nozaki et al., 1990). The actual profile shape depends on D b and K (as defined in eq. 5). Station 3 and 4 profiles of solid phase 227 Ac are as expected. However, the other stations do not show such an obvious trend for 227 Ac activity vs. depth in the sediments. Station 1, 2, and 5 show a slight decrease in activity with depth, but the erratic nature of bioturbation in these sediments may make the 227 Ac profiles distorted (see 210 Pb profiles). Figure 9 also shows excess 210 Pb concentration (bq/kg) vs. depth (cm) for all 5 stations along the C-Disk-IV transect. Excess 210 Pb in surface sediments varied from 1200 – 50 bq kg -1 and all profiles show a decrease in activity with depth. The profiles of most stations were a bit erratic, as excess 210 Pb values drop to zero in the 2-3 cm interval but re-appear with values of 100-300 bq kg -1 at 4-5 cm depths. 33 One-Dimensional Exponential Fits to Excess 210 Pb and Mixing Coefficients Mixing coefficients were calculated in the sediments by treating bioturbation as a diffusive-like process for a solid species (Goldberg and Koide, 1962; Guinasso and Schink, 1975; Berner, 1980). Based on the assumptions of very low sedimentation rates and steady state, the vertical distribution of a strongly sorbed, particle-reactive radionuclide in deep-sea sediments is described by following equation: 0=𝐷 4 + 5 , +6 5 −𝜆𝐶 (9) where: C is the excess activity of the solid species, z is the depth in the sediment relative to the SWI (sediment water interface), D b is the mixing coefficient, and l is the decay constant. Bioturbation rate (D b ) can either be assumed constant over the entire depth profile or a specific depth range, depending on how bioturbation is distributed through the profiles (Berner, 1980). Equation 9 is easily solved if D b is constant with depth and its solution usually provides good fits for sediment concentration profiles if sediments are well-mixed (Boudreau, 1986). The mixing coefficient (D b ) was determined by solving equation 9 and fitting the following solution to the excess 210 Pb profile, assuming the boundary condition, C à 0 as z à ¥: (10) € C =C 0 exp(−bz) 34 where: z is depth in cm, C is the activity of excess 210 Pb, C 0 is the activity of excess 210 Pb at the SWI, and b is the inverse scale length of the exponential curve. D b and b are related by the following relationship: 𝑏 = 8 9 : (11) The upper sediments, instead of the entire sediment column, were modeled due to the erratic profiles of excess 210 Pb (Figure 9). The non-local transport of excess 210 Pb at depths makes the bioturbation rate unreliable, if it is determined from excess 210 Pb curve fitting below this horizon (Boudreau, 1986). Each profile of excess 210 Pb was fit with equation 10 in the upper sediments where non-local transport is assumed to be minimal, assuming constant bioturbation throughout this zone. This was usually the top 3-5 centimeters for each profile. D b ranged from 0.005 to 0.084 cm 2 /yr throughout the C-Disk-IV Transect (Fig. 9). The southernmost stations (1, 2, and 3) had low bioturbation rates (0.005 – 0.01 cm 2 /yr), while the northern stations (4 and 5) had relatively higher bioturbation rates (0.022 – 0.084 cm 2 /yr). The relative uncertainties for D b for most stations were modest, ranging from 5% - 35%. 35 Table 4: Activities of 210 Pb, 226 Ra, 227 Ac, and Excess 210 Pb in Sediments along C-Disk-IV Transect. Excess 210 Pb was determined by the difference of 210 Pb and 226 Ra. 226 Ra was determined from the 186 keV energy peak. 227 Ac was determined from the 270 keV energy peak, which is 223 Ra, the granddaughter of 227 Ac. Uncertainties for each isotope are based on counting statistics, and the uncertainty in excess 210 Pb is calculated by error propagation. Excess 210 Pb is decay corrected based on sampling and counting dates. C-Disk-IV MASS POROSITY 226 Ra +/- 210 Pb +/- 210 Pb ex +/- 227 Ac +/- Interval (cm) g Bq/kg Bq/kg Bq/kg dpm/g STATION 1 0-1 0.897 0.87 275 12 1351 25 1082 28 2.5 0.3 1-2 0.963 0.80 335 11 411 15 76 19 2.5 0.2 2-3 0.925 0.80 385 14 439 16 54 22 2.6 0.3 3-4 1.344 0.78 306 12 349 13 43 17 1.4 0.2 4-5 1.314 0.78 377 12 529 14 154 18 1.8 0.2 5-6 0.762 0.78 537 16 762 20 227 26 1.9 0.3 6-7 1.279 0.78 412 10 497 13 86 17 1.7 0.2 7-8 0.944 0.76 637 15 703 19 66 24 2.5 0.3 8-9 1.256 0.76 439 10 531 12 92 16 1.4 0.2 9-10 1.131 0.76 409 12 449 14 41 18 1.1 0.2 11-12 1.183 0.76 445 12 439 12 -6 17 1.0 0.2 13-14 1.495 0.75 508 10 598 12 90 15 0.8 0.1 15-17 1.634 0.74 543 10 546 12 3 16 0.7 0.2 19-21 0.982 0.74 722 14 719 17 -3 22 0.8 0.2 21-23 1.041 0.73 689 18 678 21 -11 28 0.7 0.2 STATION 2 0-1 0.885 0.83 285 13 463 16 181 21 1.7 0.3 1-2 0.883 0.76 360 13 395 16 35 21 2.0 0.3 2-3 0.954 0.74 455 11 452 13 -3 17 2.0 0.2 3-4 0.924 0.73 505 14 467 15 -39 21 1.9 0.2 4-5 0.967 0.73 666 14 704 16 39 21 2.6 0.2 5-6 1.024 0.73 578 13 624 17 47 21 2.1 0.3 6-7 0.971 0.72 676 15 715 17 39 23 2.0 0.2 7-8 0.918 0.72 615 11 682 15 69 19 1.0 0.2 8-9 0.914 0.72 713 15 750 19 37 25 1.7 0.3 9-10 0.991 0.73 619 14 643 18 25 23 1.7 0.2 12-14 1.101 0.72 612 15 633 18 22 24 0.1 0.2 16-18 1.009 0.71 535 10 590 14 56 18 -0.6 0.2 STATION 3 0-1 0.309 0.86 203 21 985 29 793 36 0.8 0.4 1-2 0.623 0.80 310 18 369 20 59 27 1.8 0.3 2-3 0.818 0.77 376 14 393 15 17 20 2.5 0.3 3-4 1.031 0.76 430 13 429 14 -1 19 1.1 0.2 4-5 1.017 0.76 464 16 552 17 90 24 1.5 0.3 5-6 0.486 0.75 572 19 644 19 73 27 1.6 0.3 6-7 1.013 0.74 466 13 539 17 74 22 1.7 0.2 7-8 1.206 0.74 488 13 560 16 73 21 1.7 0.3 8-9 0.387 0.73 453 23 494 28 42 37 0.6 0.4 9-10 1.076 0.73 499 13 518 16 20 21 1.2 0.2 12-14 1.056 0.73 452 13 552 16 102 21 1.0 0.2 16-18 1.122 0.73 468 14 503 15 35 20 0.1 0.2 36 STATION 4 0-1 0.509 0.89 391 20 988 28 601 35 1.8 0.4 1-2 0.248 0.86 510 29 765 37 257 47 2.6 0.6 2-3 0.474 0.85 440 21 668 27 229 35 3.2 0.5 3-4 1.137 0.84 524 15 843 21 322 26 3.1 0.3 4-5 0.564 0.81 652 24 945 30 295 38 3.4 0.5 5-6 0.416 0.80 545 21 995 27 454 34 2.9 0.4 6-7 0.756 0.77 635 15 605 18 -29 24 2.8 0.3 7-8 1.081 0.77 640 13 677 15 37 20 2.8 0.3 8-9 0.931 0.77 580 14 770 19 191 24 2.8 0.3 9-10 0.510 0.77 573 21 767 24 196 33 1.9 0.4 10-12 0.494 0.77 447 23 774 31 330 39 0.9 0.4 12-14 0.538 0.78 466 21 611 25 146 33 1.2 0.3 14-16 0.521 0.79 379 18 571 21 194 28 1.3 0.3 18-20 0.473 0.79 403 19 490 20 88 28 0.8 0.4 20-22 0.830 0.78 445 15 588 17 144 23 0.6 0.2 STATION 5 0-1 0.708 0.79 200 14 1035 24 843 28 1.0 0.3 1-2 0.706 0.80 175 8 383 12 210 15 0.9 0.2 2-3 0.800 0.78 206 13 343 16 138 21 0.6 0.2 3-4 0.714 0.75 189 13 243 15 54 20 0.7 0.2 4-5 1.110 0.76 254 11 381 14 128 18 1.1 0.2 5-6 0.763 0.74 242 13 379 15 138 21 1.5 0.3 6-7 1.072 0.70 270 10 260 13 -11 16 1.3 0.2 7-8 0.512 0.73 356 19 318 22 -38 30 1.4 0.3 8-9 0.916 0.75 427 19 472 23 46 31 3.0 0.4 9-10 0.810 0.72 320 14 349 16 29 21 1.9 0.3 11-12 0.997 0.73 339 14 398 16 60 21 1.4 0.3 13-14 0.805 0.72 233 13 323 16 91 21 0.9 0.3 19-20 1.096 0.69 104 11 171 16 68 20 0.3 0.3 37 38 39 Figure 9: Sediment Profiles of Excess 210 Pb and 227 Ac vs. Depth for all Stations along the C- Disk-IV Transect. 227 Ac was measured via its granddaughter, 223 Ra, where secular equilibrium is assumed between its progenitors. Each profile of excess 210 Pb were fit with equation 10 in the upper sediments where non-local transport is absent, assuming constant bioturbation throughout this zone. Stations 1, 2, 3, and 5 excess 210 Pb profiles were fit with equation 10 in the top 4 cm, while station 4 was fitted in the top 3 cm. D b is the bioturbation rate in units of cm 2 /yr. 40 Core incubation and Actinium-277 Flux The source function of 227 Ac was determined indirectly by following the decay of 223 Ra activity over 120 days after core incubation ended. 223 Ra is the granddaughter of 227 Ac and because it is more soluble in seawater compared to 227 Ac, it will be in excess over its parents ( 227 Th and 227 Ac) during core incubation. This behavior of 223 Ra, 227 Th, and 227 Ac in seawater makes these radioisotopes useful for mathematical modeling by means of the Bateman equation (Bateman, 1910). Below is an expression for 227 Ac, 227 Th, and 223 Ra decay chain series. +; 55<=> +* =−𝜆 ""#$% 𝑁 ""#$% + 𝐴 "@ABC +; 55<DE +* =−𝜆 ""#F) 𝑁 ""#F) +𝜆 ""#$% 𝑁 ""#$% (12) +; 55HIJ +* =−𝜆 ""@KC 𝑁 ""@KC +𝜆 ""#F) 𝑁 ""#F) Equation 13 is a numerical solution for the above 3 radioactive parent-daughter pairs, which was fit to the data (D. Hammond, pers. comm). A 219Rn = m1*(1.0038*exp(-8.717e-5*m0)-2.58662*exp(-3.711e-2*m0) + 1.58239*exp(-0.06059*m0)) + m2*(2.58011*(exp(-3.711e-2*m0)-exp(0.06059*m0))) + m3*exp(-0.06059*m0) (13) Where m0 is the time since the incubation ended and at that time m1 = 227 Ac, m2 = 227 Th, and m3 = 223 Ra (dpm). Table 5 and 6 summarizes the results of the 120-day counting period for all 5 stations along the C-Disk-IV transect. Figure 10 shows the Bateman equation fits to the 223 Ra activity for each station. 41 Table 5: Core Incubation Results along C-Disk-IV Transect. 223 Ra activity was measured continuously on RaDeCC for 125 days after core incubation ended. Errors are counting statistics. 223 Ra (dpm) +/- days since incubated 223 Ra (dpm) +/- days since incubated Station 1 0.31 0.08 17 Station 2 0.49 0.07 14 0.32 0.06 18 0.42 0.06 19 0.24 0.05 25 0.39 0.08 21 0.16 0.04 40 0.30 0.04 33 0.09 0.04 42 0.06 0.03 47 0.07 0.03 60 0.08 0.03 61 0.08 0.02 71 0.10 0.02 65 0.08 0.02 79 0.08 0.02 77 0.04 0.02 83 0.11 0.03 79 0.03 0.02 84 0.16 0.04 82 0.04 0.02 93 0.32 0.08 83 0.03 0.02 127 0.29 0.06 84 0.22 0.05 125 Station 3 0.43 0.06 11 Station 4 0.28 0.08 11 0.49 0.06 16 0.18 0.05 11 0.26 0.04 18 0.18 0.04 14 0.22 0.05 25 0.16 0.05 14 0.14 0.04 56 0.06 0.03 39 0.09 0.02 63 0.05 0.02 42 0.05 0.02 76 0.03 0.02 44 0.13 0.04 77 0.03 0.02 54 0.10 0.03 78 0.03 0.02 75 0.10 0.05 79 0.01 0.01 84 0.18 0.07 125 0.01 0.01 84 0.05 0.02 127 Station 5 0.35 0.07 6 0.31 0.06 7 0.15 0.04 16 0.22 0.04 16 0.12 0.04 30 0.14 0.05 31 0.09 0.03 42 0.09 0.03 49 0.05 0.03 53 0.16 0.03 60 0.05 0.03 74 0.07 0.03 84 0.09 0.03 125 42 Table 6: Bateman Equation Fit Results for Initial Activities of 223 Ra and 227 Ac from Core Incubation. Equation 13 was fit to the 219 Rn activity over the 120-counting interval to find 223 Ra and 227 Ac at the end of the incubation. Time/height is the summation of time incubated and height of the remaining water in the core after each analysis and volume indicates the seawater that was passed through the Mn-fibers. a i is the effective inverse height needed for a short-lived radioisotope (see Appendix D). 223 Ra and 227 Ac fluxes were calculated by equation 7 and 6. Station time/height Volume 227 Ac (m1) sig 223 Ra (m3) sig days/m L dpm dpm 1 142.1 1.35 0.032 0.013 0.651 0.105 2 171.3 1.29 0.213 0.055 1.436 0.398 3 128.1 1.16 0.126 0.042 0.932 0.234 4 94.9 1.00 0.033 0.020 0.455 0.074 5 50.9 1.00 0.086 0.022 0.471 0.049 Station 227 Ac Flux sig a i 223 Ra sig dpm/m 2 -yr m -1 dpm/m 2 -yr 1 60 24 5.60 1905 305 2 352 91 6.49 3807 1066 3 309 103 5.49 3235 809 4 125 77 4.35 2312 370 5 618 160 2.70 3806 381 43 Figure 10: Bateman Equation Fits to Determine 223 Ra and 227 Ac Benthic Fluxes for all Stations along the C-Disk-IV Transect. 223 Ra activity found from 219 Rn and is in dpm. 223 Ra decays to its grandparent activity, 227 Ac by 125 days. 227 Ac Fluxes were determined by taking 227 Ac activity at the end of the incubation and dividing it by the seawater that was incubated (~ 1.0 L) and incubation period. 223 Ra flux calculation is discussed in Appendix D. 44 Discussion The distribution of 227 Ac in the water column can give rates of transport for other solutes that are essential for understanding chemical, biological, and physical processes occurring in the ocean (Hammond et al., in prep). Applying a one-dimensional eddy diffusion-decay model to the distribution of 227 Ac throughout the water column can constrain these transport rates. The goal of this thesis is to characterize the distribution of 227 Ac throughout the NEPB and quantify the source function of 227 Ac in deep-sea sediments. These two objectives, coupled together, can indicate if a 1-D model is appropriate for the distribution of 227 Ac in the water column, or if horizontal transport must be considered. In addition, bottom topography, density gradients, and water parcel pathways were correlated with 227 Ac to provide insight into what controls the distribution of 227 Ac throughout the NEPB. The activity of 227 Ac was shown to increase with depth at all 5 stations along the C-Disk- IV transect (Figure 7 and 8). This is as expected, since the source function of 227 Ac is deep-sea sediments, and 227 Ac decreases in concentration as it is transported away from its source (Nozaki, 1984). Each profile was fit with a 1-D exponential function that describes the vertical transport of 227 Ac through the water column. The results from the 1-D exponential fit give vertical transport rates, named “apparent” vertical eddy diffusivity as a first approximation, which describes all processes that transfer tracers across density surfaces. These processes include vertical and horizontal mixing (Ku and Luo, 2009). If the source function is known, then the relative roles of horizontal and vertical transport can be examined. Furthermore, these mixing rates determined from the profiles of radionuclides can provide valuable context for boundary conditions in circulation models (Kuo and Veronis, 1970). 45 The profiles of 227 Ac throughout the 5 stations in the NEPB show a constant activity near the bottom 500 meters while decreasing exponentially above 500 meters toward the sea-surface. Vertical eddy diffusivities (K z ) coefficients derived from the full 227 Ac profiles range from 0.6 – 11.3 cm 2 s -1 . It has been shown that diapycnal mixing is linked to topography roughness and distance from the sea-floor (Hautala, 2018; Ledwell et al., 2000; Polzin et al., 1997). To examine how sea-floor topography controls the distribution of 227 Ac in the NEPB, these Kz’s derived from the 227 Ac profiles were plotted as a function of topography roughness (Figure 11). The topography roughness was calculated as the sample standard deviation of topography in the east-west direction around each station. The longitudinal direction was sampled around each station because hydrographic considerations suggest abyssal circulation moves in an eastward flow across the C-Disk-IV transect (Kawabe and Fujio, 2012). Figure 11 shows a positive correlation between K z and topography roughness for all 5 stations in the NEPB. There is an excellent correlation between K z and topography roughness, if we exclude station 3. Station 3 straddles the Mendocino Fracture Zone (MFZ) at 35˚N and 155˚W, and the nature of how roughness was calculated in the longitudinally direction may indicate a lower roughness than actually exist. Circulation has been shown to move along the MFZ and mixing may be enhanced within fracture zones (Decloedt and Luther, 2012; Kawabe and Fujio, 2012). Station 2 has the highest Kz value and its location is near the Musician Seamounts, which have rough topography (Hautala, 2018). Therefore, station 2 and 3 higher Kz values could suggest correlation with rough topography. Stations 4 and 5 have relatively smooth upsloping topography that extends toward the east. 46 Figure 11: Vertical Eddy Diffusivity (Kz) vs. Topography Roughness along the C-Disk-IV Transect. Kz’s were derived from full 227 Ac profile fits. The topography roughness was calculated by deriving the sample standard deviation of topography in the east-west direction around each station Figure 12: Upper Layer Vertical Eddy Diffusivity (K z ) vs. ∆density/∆depth K z ’s were derived from two-layer fit profiles. ∆density/∆depth is the slope in the upper layer in the Neutral Density vs. DAB plot (Fig. 8). 47 Ledwell et al., (2000) calculated K z values for the Brazil Basin from tracer released experiments (SF 6 ) and applied a 1-D model to evolution of the tracer after 26 months. They concluded that diapycnal mixing increases with depth and areas of around rough topography. Their Kz values ranged from 2-4 cm 2 /s 500 meters above rough topography to 10 cm 2 /s nearer the bottom. Due to the lack of 227 Ac sample points, high resolution Kz values could not be computed at specific depth ranges for this study. However, an eddy diffusion parameter was derived by fitting an exponential function to data from the upper layer, which was bounded by the break in slope on the neutral density vs. depth graph (Fig. 8). The two-layered structure of the 227 Ac profiles in figure 8 suggest higher K z values with increasing depth and this correlates well with decreased column stratification, which should larger be controlled by topography (Ledwell et al., 2000). The thickness of the deep mixed layer was also calculated by break in slope in the neutral density graphs and compared against depth and topography. Station 2 and 3 have the largest deep mixed layer thickness and this strongly correlates with increasing depth and topography. Lastly, an upper layer K z was also calculated in the two-layered structure and plotted against the upper layer slope of ∆density/∆depth from the neutral density vs. DAB graph (Fig. 12). There is an unresolved upper layer K z dependence on density stratification. Horizontal transport may be influencing the K z and further examination should be investigated in the future once more 227 Ac data are measured in the upper water column. 48 Figure 13: Illustration of Topography Throughout the NEPB. Open circles are stations along the C-Disk-IV transect. Red arrows are abyssal circulation pathways proposed by Kawabe and Fujio (2012). Hammond et al., (in prep) correlated the distribution of 227 Ac in the Deep South Pacific with density stratification. They determined that the vertical eddy diffusivities calculated from 227 Ac profiles increased with decreased stratification. This was most evident around the East- Pacific Rise (EPR), where rough topography is present. Figure 14 shows neutral density (kg/m 3 ) gradients along the C-Disk-IV transect. The neutral density stratification decreases significantly around stations 2 and 3. These two stations are areas of rough topography, where the Musician Seamounts and MFZ are present (Hautala, 2018). Furthermore, the high eddy diffusivities calculated around stations 4 and 5 correlate with an increase in column stratification toward the 49 high latitudes. Rough topography is not as significant between 40˚N and 50˚N as compared to the topography structures around stations 2 and 3. Figure 14: Neutral Density Gradients along the C-Disk-IV Transect. Red lines refer to C-Disk- IV stations where in-situ salinity, temperature, and pressure where measured. Neutral density was calculated using Ocean Data View. 50 For a radioactive tracer to provide the maximum time-resolution for transport processes, its source-function should be understood (Ku and Luo, 2009). Two approaches were undertaken to quantify the source function of 227 Ac from deep-sea sediments in the NEPB: direct measurement of 227 Ac via core incubation and indirect estimates based on isotope concentrations in sediments. The later approach used 210 Pb measurements in sediments to quantify bioturbation, which was applied to a modified transport equation developed by Nozaki et al., (1990). Both approaches are discussed below in more detail. Core incubation results are summarized in Table 6. Most station’s core incubation fluxes for 227 Ac are 3-10 times greater than the fluxes required to sustain 227 Ac observed in the water column. The errors associated with the core incubation are large (~30-40%), but they do not fall within uncertainty of the observed 227 Ac flux in the water column. This method might prove unreliable since such a large flux at every station is not present anywhere in the water column throughout the C-Disk-IV transect. The nature of how 227 Ac was sampled at the end of the core incubation period could produce values higher than actual flux. For example, all overlying water in the multi-core was siphoned out to the SWI, and some cores had an uneven SWI horizon. As seawater was siphoned from the multi-core, pore-water may have mixed into overlying water due to an uneven sediment horizon. Even small irregularities in the SWI horizon could advect excess 227 Ac that would not otherwise be supplied from 227 Ac diffusion during the incubation period. The source function of 227 Ac was also determined via Nozaki’s transport equation (Equation 5) based on the D b , 231 Pa activity, and fraction of 227 Ac released. The fraction of 227 Ac released in the pore water was assumed 70% (D. Hammond pers. comm.). The activity of 231 Pa was established by the profile of solid phase 227 Ac, assuming 227 Ac increases to maximum activity in the upper 5 cm. The maximum activity of 227 Ac in the upper sediments should reflect 51 the equilibrium of 231 Pa with 227 Ac. The most significant term in Equation 5 that controls the flux of 227 Ac across the SWI is the bioturbation rate, which was determined by the exponential fit to 210 Pb profiles (Nozaki et al., 1990). The bioturbation rate was calculated for each station by fitting equation 10 to the profile of excess 210 Pb in the upper few centimeters. Table 7 shows the results of the source function of 227 Ac via core incubation and Nozaki’s equation. The observed water column flux is also shown, which was derived from the total flux in the two-layered model (Table 3). Nozaki’s equation for the 227 Ac flux at stations 2 and 3 predicts the observed 227 Ac flux in the water column very well. Stations 1, 4, and 5 are a factor of 4 higher than the observed water column flux. This may suggest that stations 4 and 5 are not well suited for a simple 1-D model. The high latitudes along the C-Disk-IV transect are characterized by eastward circulation that travels along a westward dipping slope. The water that is observed above stations 4 and 5 could be reflecting an upper water column source from the west, which is lower in 227 Ac activity. For comparison at stations 2 and 3, the topography is shown to be rougher, but the depth remains constant throughout the area, so even if horizontal flow occurs, a 1-D model predicts Nozaki’s equation for the 227 Ac appropriately at these stations. At station 1, the 227 Ac flux based on Nozaki’s equation shows a flux that is 4x higher than the apparent water column flux, similar to stations 4 and 5. The complex topography around the Hawaiian ridge, or the westward circulation might cause a loss of 227 Ac in the water column above station 1. 52 Table 7: Excess 227 Ac Flux Determined from the Distribution of 227 Ac in the Water Column, Core Incubation, and Nozaki’s Transport Equation. The water column flux was estimated from the two-layered model. The bioturbation rate for Nozaki’s equation was estimated from curve fitting of excess 210 Pb in Fig. 9. The activity of 231 Pa was assumed as the maximum activity of 227 Ac in the upper sediments. Porosity, density of sediments, and fraction of 227 Ac released from 231 Pa was assumed 90%, 2.5 g/cm 3 , and 70%. The dimensionless partition coefficient, K, was assumed to be 10,000. Station Water Column Core Incubation A 231Pa D b Nozaki’s Eq. # dpm/m 2 -yr dpm/m 2 -yr dpm/g cm 2 /yr +/- dpm/m 2 -yr 1 21 ± 9 60 ± 24 2.5 0.005 0.002 79 ± 33 2 60 ± 25 352 ± 91 2.0 0.010 0.002 77 ± 12 3 52 ± 20 309 ± 103 2.5 0.005 0.001 78 ± 5 4 50 ± 19 125 ± 77 3.0 0.084 0.029 232 ± 78 5 17 ± 3 618 ± 160 1.0 0.022 0.004 78 ± 9 53 Conclusion The activity of 227 Ac increases with depth at all 5 C-Disk-IV stations in the Northeast Pacific. Vertical eddy diffusion rates derived from one-dimensional eddy diffusion-decay model increase with depth and areas around rough topography. Areas around the Musician Seamounts (Sta. 2) and the MFZ (Sta.3) show the highest vertical mixing rates, while stations 4 and 5, which are areas of westward dipping smooth topography, show lower mixing rates. Furthermore, mixing is enhanced in a benthic layer (~ 500 m thick) near the seafloor and may decrease exponentially above this layer. This is strongly correlated with water column stratification, which increases markedly above the mixed layer. The source function of 227 Ac was estimated from core incubations and bioturbation coefficients. Core incubation fluxes for 227 Ac are 3-10 times greater than the fluxes of 227 Ac required to support the decay observed in the water column. The errors associated with the core incubation were large (~30-40%), but they did not fall within uncertainty of the observed 227 Ac flux in the water column. This method might prove unreliable or modifications need to be implemented to determine the source function more accurately. For example, siphoning water from the core needs special attention to avoid siphoning 227 Ac from pore waters. Bioturbation coefficients and 231 Pa activities were also used to estimate the source function of 227 Ac through means of Nozaki’s equation. At Stations 2 and 3, this equation predicts fluxes comparable to the flux required by water column data. Stations 1, 4, and 5 do not predict the observed water column flux using Nozaki’s equation. The eastward flowing circulation in the high latitudes that would bring water toward the westward dipping topography may lower activity of 227 Ac in the water column at stations 4 and 5. At station 1, complex topography or westward circulation may also lower activity of 227 Ac in the water column. This 54 could suggest, if the source function and water column flux are correct, a 1-D model would not reflect the vertical distribution of 227 Ac very well and not predict a correct K z value for these regions. However, if depth and bioturbation are constant on large scales, then 227 Ac could provide a reasonable vertical eddy diffusivity, as shown by the basin stations, 2 and 3. Bioturbation processes and 231 Pa values in sediments will need to be further constrained to clarify the benthic flux. Understanding how non-local transport in sediment profiles influence the benthic flux of 227 Ac should be understood. 231 Pa values should be measured by another method and compared to the gamma counting measurements. Lastly, more measurements of 227 Ac are needed throughout this region to determine the transport and mixing pathways of 227 Ac and to conclude what model is best suited for these mixing regimes. 55 References Anderson, R.F., M. Bacon, P. Brewer. “Removal of 230Th and 231Pa from the open ocean.” Earth and Planetary Science Letters, vol. 62, (1982): 7–22. Bateman, H. “The solution of a system of differential equations occurring in the theory of radioactive transformations.” In Proc. Cambridge Philos. Soc, vol. 15, (1910): 423–427 Berelson, W.M., D.E. Hammond, and C. Fuller. "Radon-222 as a Tracer for Mixing in the Water Column and Benthic Exchange in the Southern California Borderland." Earth and Planetary Science Letters 61.1 (1982): 41-54. Berner, Robert A. Early Diagenesis: A Theoretical Approach. Princeton, NJ: Princeton UP, (1980). Boudreau, B. P. "Mathematics of Tracer Mixing in Sediments; II, Nonlocal Mixing and Biological Conveyor-belt Phenomena." American Journal of Science 286.3 (1986): 199- 238. Cochran, J. K., and S. Krishnaswami. "Radium, Thorium, Uranium, and 210Pb in Deep-sea Sediments and Sediment Pore Waters from the North Equatorial Pacific." American Journal of Science, (1980): 849-89. Decloedt, Thomas, and Douglas S. Luther. “Spatially heterogeneous diapycnal mixing in the abyssal ocean: A comparison of two parameterizations to observations.” Journal of Geophysical Research: Oceans, vol. 117, (2012). 56 Geibert, W., M. Rutgers Van Der Loeff, C. Hanfland, and H. Dauelsberg. "Actinium-227 as a Deep-sea Tracer: Sources, Distribution and Applications." Earth and Planetary Science Letters 198.1-2 (2002): 147-65. Geibert, Walter, M. Charette, G. Kim, W. S. Moore, J. Street, M. Young, and A. Paytan. "The Release of Dissolved Actinium to the Ocean: A Global Comparison of Different End- members." Marine Chemistry 109.3-4 (2008): 409-20. Chung, Y., and K. Kim. “Excess222Rn and the benthic boundary layer in the western and southern Indian Ocean.” Earth and Planetary Science Letters, vol. 49, no. 2, (1980): 351–359. Goldberg, Edward D., and Minoru Koide. "Geochronological Studies of Deep Sea Sediments by the Ionium/thorium Method." Geochimica Et Cosmochimica Acta 26.3 (1962): 417-50. Craig, H., Y. Chung, M. Fiadeiro. “A benthic front in the South Pacific.” Earth and Planetary Science Letters, vol. 16, no. 1, (1972): 50–65. Guinasso, N. L., and D. R. Schink. "Quantitative Estimates of Biological Mixing Rates in Abyssal Sediments." J. Geophys. Res. Journal of Geophysical Research 80.21 (1975): 3032-043. Hayes, C. T., Anderson, R. F., Jaccard, S. L., François, R., Fleisher, M. Q., Soon, M., & Gersonde, R. “A new perspective on boundary scavenging in the North Pacific Ocean”. Earth and Planetary Science Letters, 369–370, (2013): 86–97. Henderson, P. B., P. Morris, W. Moore, M. Charette. “Methodological advances for measuring low-Level radium isotopes in seawater.” Journal of Radioanalytical and Nuclear Chemistry, vol. 296, no. 1, (2012): 357–362. 57 Hammond, Douglas E., R. Marton, W. Berelson, T. Ku. “Radium 228 distribution and mixing in San Nicolas and San Pedro Basins, southern California Borderland.” Journal of Geophysical Research, vol. 95, no. C3, (1990): 3321. Hammond, D.E., M. Charette, W. Moore, P. Henderson, V. Sanial, L. Kipp, R. Anderson. “Actinium-227 as a tracer for solute transport in the deep ocean: Results from the Geotraces Pacific Transect” (2016), in prep. Hautala, Susan L. “The abyssal and deep circulation of the Northeast Pacific Basin.” Progress in Oceanography, vol. 160, 2018, pp. 68–82 Hou, Yi. “Estimating Benthic Si Flux in the North Pacific.” University of Southern California, Senior Thesis. (2017). Kawabe, Masaki, and Shinzou Fujio. “Effect of bottom slope in northeastern North Pacific on deep-Water upwelling and overturning circulation.” Journal of Oceanography, vol. 68, no. 2, (2012): 267–284. Kemnitz, Nathaniel. “Estimates of Bioturbation Rates Using Excess 210Pb in Eastern Tropical South Pacific.” University of Southern California, Senior Thesis. (2016). Ku, Teh-Lung, and Shangde Luo. “Chapter 9 Ocean Circulation/Mixing Studies with Decay- Series Isotopes.” Radioactivity in the Environment U-Th Series Nuclides in Aquatic Systems, (2008) 307–344. Lao, Yong, R. F. Anderson, W. S. Broecker, S. E. Trumbore, H. J. Hofmann, and W. Wolfli. "Transport and Burial Rates of 10Be and 231Pa in the Pacific Ocean during the Holocene Period." Earth and Planetary Science Letters 113.1-2 (1992): 173-89. 58 Ledwell, J. R., E. T. Montgon, K. L. Polzin, L. C. St. Laurent, R. W. Schmitt, J. M. Toole. “Evidence for enhanced mixing over rough topography in the abyssal ocean.” Nature, vol. 403, no. 6766, (2000): 179–182. Mashayek, A., H. Salehipo, D. Bouffard, C. P. Caulfield, R. Ferrari, M. Nikurash, W. R. Peltier, W. D. Smyth. “Efficiency of turbulent mixing in the abyssal ocean circulation.” Geophysical Research Letters, vol. 44, no. 12, (2017): 6296–6306. Moore, Willard S., and Ralph Arnold. “Measurement of 223Ra and 224Ra in coastal waters using a delayed coincidence counter.” Journal of Geophysical Research: Oceans, vol. 101, no. C1, (1996): 1321–1329. Moore, Willard S., and Pinghe Cai. “Calibration of RaDeCC systems for 223Ra measurements.” Marine Chemistry, vol. 156, (2013): 130–137. Munk, Walter, and Carl Wunsch. “Abyssal recipes II: energetics of tidal and wind mixing.” Deep Sea Research Part I: Oceanographic Research Papers, vol. 45, no. 12, (1998): 1977–2010. Nozaki, Yoshiyuki. "Excess 227Ac in Deep Ocean Water." Nature 310.5977 (1984): 486-88. Nozaki, Yoshiyuki, M. Yamada, H. Nikaido. “The marine geochemistry of actinium-227: Evidence for its migration through sediment pore water.” Geophysical Research Letters, vol. 17, no. 11, (1990): 1933–1936. Polzin, K. L. “Spatial Variability of Turbulent Mixing in the Abyssal Ocean.” Science, vol. 276, no. 5309, Apr. (1997): 93–96. Reid, D., D. R. Schink, R. M. Key. “Radium, thorium, and actinium extraction from seawater using an improved manganese-Oxidecoated fiber.” Deep Sea Research Part B. Oceanographic Literature Review, vol. 26, no. 12, (1979): 769. 59 Resing, Joseph A., Peter N. Sedwick, Christopher R. German, William J. Jenkins, James W. Moffett, Bettina M. Sohst, and Alessandro Tagliabue. "Basin-scale Transport of Hydrothermal Dissolved Metals across the South Pacific Ocean." Nature 523.7559 (2015): 200-03. Shaw, Timothy J, and Willard S Moore. “Analysis of 227Ac in seawater by delayed coincidence counting.” Marine Chemistry, vol. 78, no. 4, (2002): 197–203. Schlitzer, Reiner. “Interactive analysis and visualization of geoscience data with Ocean Data View.” Computers & Geosciences, vol. 28, no. 10, (2002): 1211–1218. Sun, Yin, and T. Torgersen. “The effects of water content and Mn-Fiber surface conditions on measurement by emanation.” Marine Chemistry, vol. 62, no. 3-4, (1998): 299–306. Talley, Lynne D., and Terrence M. Joyce. “The double silica maximum in the North Pacific.” Journal of Geophysical Research, vol. 97, no. C4, 1(992): 5465 Talley, Lynne D. “Descriptive Physical Oceanography.” 6th ed., Academic Press, (2011) Wolfe, Christa. “The accuracy of core incubations to determine benthic fluxes of 222-radon, 224-radium and 228-radium” University of Southern California, Master’s Thesis. (2011). 60 Appendix A Optimizing Methods and Standardization for 227 Analysis using RaDeCC Overview Radium Delayed Coincidence Counters (RaDeCC) have proven to be a fast and reliable way to measure 223 Ra (t 1/2 = 11.4 days) and 224 Ra (t 1/2 =3.7 days) from seawater samples (Moore and Cai, 2013; Moore, 2008; Moore and Arnold, 1996). The radioactive ancestors of 223 Ra and 224 Ra can also be measured if secular equilibrium is assumed between the parent/daughter pairs (Moore and Arnold, 1996). For example, 227 Ac (t 1/2 = 21.8 years) is the long-lived grandparent of 223 Ra and has applications in deep-ocean diapycnal mixing. While protocols and standards have been developed for 223 Ra and 224 Ra calibration, 227 Ac calibration remains problematic (Scholten et al., 2010; Charette et al., 2012). This purpose of this supplement is to describe refinements of protocols and standardization for 227 Ac analysis by means of RaDeCC. The RaDeCC system The RaDeCC theory was developed by Giffin et al. (1963) and instrumentation was developed by Moore and Arnold (1996). Further refinements have been discussed by Moore (2008). The RaDeCC system measures 219 Rn (t 1/2 = 4 secs) and 220 Rn (t 1/2 = 56 secs) released by 223 Ra and 224 Ra adsorbed onto acrylic fibers impregnated with manganese dioxide (Mn-fibers). The system employs a delayed coincidence circuitry to distinguish 219 Rn and 220 Rn alpha decays 61 by detecting the timing of their polonium daughter decays ( 215 Po, t 1/2 =1.8 ms, and 216 Po, t 1/2 =145 ms). RaDeCC can thus distinguish the activities of 219 Rn and 220 Rn and hence, measure the activities of 223 Ra and 224 Ra that are adsorbed onto the Mn-fibers. The ancestors of these Ra isotopes, including 227 Ac, 228 Ra, and 228 Th can also be measured, if their shorter-lived daughters have grown into secular equilibrium. The RaDeCC system (Fig. A-1) recirculates a carrier gas, helium, through a closed loop which includes a pump, Lucas cell, flowmeter, and Mn-fiber holder. The helium is pumped through the flowmeter, the Mn-fibers, the Lucas cell, and then back to the pump. At USC, we have 6 RaDeCC systems, although only 5 are used consistently, as Channel 2 is often unreliable. Different channels employ slightly different designs, although the principles are basically the same. Each has a flowmeter, calibrated for the rate of helium that circulates through the system (L min -1 ). The flow rate is controlled by a valve which allows more or less helium through the system. Two variations for flow control have been employed, with one type using a valve that is simply opened or closed to vary the restriction for flow. The second type has a bypass loop straddling the pump, allowing flow to be short-circuited. When the valve is closed, all helium is diverted through the RaDeCC system, and when the valve is open, the short-circuit loop reduces flow through the Lucas cell portion of the system. The Lucas cell is a chamber that is coated with a ZnS phosphor. Alpha particles hitting the phosphor cause it to emit photons that are detected by the photomultiplier. When an alpha event is detected, the counting circuit opens a window lasting from 0.01 to 5.60 ms, termed the 219 window. Any counts detected during this time are recorded. A second window, the 220 window, then opens from 5.61 to 600 ms, and any counts detected are recorded. The timing is chosen to optimize detection of the short-lived 215 Po daughter of 219 Rn in the first window, and the longer-lived 216 Po daughter of 220 Rn in the second 62 window. Corrections must be made for the contributions of chance counts and decays of the 215 Po and 216 Po daughters that occur when the wrong window is open (Giffin et al., 1963). Details of our computation of these effects will be described in a subsequent article (Hammond et al., in prep.). Figure A-1: Schematic Diagram of the RaDeCC System. The efficiency of detecting the decays of Rn produced by the adsorbed 223 Ra and 224 Ra can be summarized: 𝐸 FMF =𝑅 O P Q P RSR 𝐸 9 (A-1) 63 where R i is the fraction of Rn that is released from the fiber into the He stream and carried into the detector before it can decay; V L /V SYS is the ratio of the Lucas cell volume to that of the total system; and E D is the efficiency of the Lucas cell and electronics for detecting the paired decays of Rn and its Po daughter Each of these factors will be discussed below. This equation assumes that circulation of gas in the system is sufficiently fast to keep its composition well mixed. In addition, when measuring the concentration of a dissolved Ra isotope, the efficiency of adsorption of Ra onto the fiber (E SORB ) during collection is also a critical factor. Subsequent sections of this appendix will consider the effects of E SORB , He flow rate, water content of fibers, and sample geometry on counting samples and standards, with the latter three parameters affecting R i , and the volume ratio. Manganese Fiber Geometry and Preparation Preparing efficient adsorption media for radium, thorium, and actinium extraction from seawater is essential for accurate analysis on RaDeCC systems. MnO 2 coated fibers have been shown to quantitatively absorb radium and actinium from seawater at flow rates of 1-1.5 L/min; thorium binds very strongly (~90%) at these flow rates as well (Reid et al., 1979). Traditionally, isotope extraction has been done using loosely clumped fibers housed in an acrylic cylinder (18 cm long, 3.5 cm ID), with PVC and quick connect fittings on each end (Moore, 2008). This fiber type and geometry will be termed Moore Fibers (MF). Henderson et al. (2013) have also shown that commercial grooved, acrylic water filters (produced by 3M, 5 µm pore size) can be cut to 10 cm length, impregnated with MnO 2 , sealed with a plastic plug on one end, and mounted with a compression spring inside a commercial acrylic filter holder (5”). This facilitates high fluid flow 64 through the sorption fiber. This design will be called Commercial Cartridge (CC). Samples collected in this study were in the CC geometry, although the standard that has been run regularly by our USC group (STD27) is in the MF geometry. Extensive experiments were done to compare the two fiber geometries, shown schematically with He flow paths in Fig. A-2. As part of this effort, multiple standards were prepared in each geometry (Table A-1) and the inter- calibration of two standard sources for 227 Ac have been compared. Cartridge Fibers (CC): The recipe of Henderson et al. (2013) was followed. First, cartridge acrylic fibers (3M catalogue #G80B81N, supplied cut to 4” by Brian Danner, Danner Associates, Rockland, MA) were soaked in DIW for 24 hours. Next, a bath of 0.5 M KMnO 4 was prepared that could hold 12 acrylic cartridges at one time. Cartridges were removed from the DIW and immediately placed in a 30 L bath of 0.5 M KMnO 4 , which was continually stirred at room temperature for 48 hours. Once the cartridges were removed, they were allowed to dry for 24 hours then washed with DIW, which removed excess KMnO 4 (~20-40 L). Cartridges were then allowed to sit for 48 hours to dry, flushed with copious quantities of DIW until effluent became clear, then placed in plastic bags. Each 30 L KMnO 4 bath solution was used for 2-3 batches (12 CC in each batch). Batches number was recorded and noted when used to examine any variations with sorption efficiency. Moore Fibers (MF): Moore fibers were prepared at USC for STDS K-8 and K-9, based on the method developed by Moore (1976). First, loose acrylic fibers were placed in a 0.5 M KMnO 4 solution heated to 70-80˚C. The acrylic fibers used were Acrilan B-16, 3 denier per filament, 2 inch cut length, made by Solutia (Colbert, 2004). The heated solution reacts with the acrylic to reduce the KMnO 4 to MnO 2 that is deposited on the fibers. After 10 minutes, the reaction is complete and the fibers are removed and rinsed in DIW. Each prepared solution 65 usually produces around 30 grams of Moore fibers (Colbert, 2004). For other MF standards prepared in this study (K-1, K-5 and K-7), MnO 2 coated fibers were purchased from Ralph Arnold of SCI (Scientific Computing, Inc., Columbia, SC). STD 27, used extensively for calibration, was prepared by Colbert (2004) using MF fibers prepared at USC. Figure A-2: Illustration of Moore Fibers and Commercial Cartridge Geometry. The Commercial Cartridge (CC) is shown schematically on the left and the Moore Fibers (MF) are on the right. Solid black represents the MnO 2 fibers housed within acrylic containers. The spring in CC geometry below the fiber forces the cartridge against seals at the top. Arrows show how He flows through these geometries. 66 Standard Preparation For standards prepared in the MF geometry, 20 g MnO 2 coated fibers (dry) were weighed out and placed in the acrylic containers. For the CC geometry, MnO 2 coated Commercial Cartridges were placed on a machined Delrin disc that sat on a spring (302SS, McMaster Carr #1986K31) in an acrylic filter holder designed for 5” fibers; the dry weight of coated acrylic fibers for each cartridge was approximately 60 grams. Acrylic containers were acid washed before Moore and cartridge fibers were placed inside. Fibers were washed thoroughly with DIW (~10 liters) after fibers were placed in containers. Ra-free seawater was prepared by passing it 3 times through a path with Moore fibers to remove radium and then filtered (Supor, 0.45 µm) to remove particulates. Aliquots (1.0 L) of this water were then spiked with known amounts of a solution containing 227 Ac and its daughters, briefly shaken, and immediately pumped (peristaltic pump with PVC tubing) through the fibers 5 times at a flow rate of ~0.5 L/min to ensure quantitative absorption of 227 Ac and its daughters from seawater (Reid et al., 1979). The 227 Ac spike was a calibrated solution purchased from Eckert and Zeigler and diluted by P. Henderson and M. Charette of WHOI in 1 N HNO 3 to a concentration of 5.9 dpm/g on April 26, 2017. 227 Ac spike aliquots of 1.2-1.3 ml were used, containing 1.2-1.3 meq of acid, less than the 1.9 meq of alkalinity that should be present in the surface seawater used. Standards were prepared for both MF and CC geometries. The seawater aliquots used were set aside for future analysis of 227 Ac, if needed. 67 RaDeCC Efficiency The calibration of RaDeCC for 227 Ac analysis is dependent on knowing the E TOT , as well as the binding efficiency of 227 Ac onto manganese fibers. The main factors that control E TOT for a given channel are the release of 219 Rn, which should depend on fiber moisture (Sun and Torgersen, 1998a), flowrate that influences Rn transfer from fibers to detector, Lucas cell detector efficiency, and system volume, which depends on sample geometry (Moore and Arnold, 1996; Moore, 2008; Colbert, 2004). These mechanisms are discussed below. The binding efficiency reflects the amount of 227 Ac that is bound onto Mn-fibers from the spiked standard solution. To test the binding efficiency of 227 Ac relative to its progeny, 5 K- series standards (Table A-1) were regularly counted on RaDeCC for up to 120 days after creation. Ideally, an 227 Ac spiked solution should be monitored for 120 days after creation as this allows equilibration of the progeny with the Ac parent. If absorption efficiencies of progeny and Ac differ, or the release efficiencies of progeny produced by recoil differs from those adsorbed from solution, the counting rate will change with time (Scholten et al., 2010). Only two of the K-series standards were monitored for 120 days due to time constraints. Another standard, STD27, was counted frequently on each channel, along with the K-series standards to monitor any drift in channel efficiency with time. STD27 was prepared by Steve Colbert in 2004 and has been counted on the 5 RaDeCC channels for many years. STD27 was prepared by soaking MF fibers in Ra-free seawater, spiked with dissolved EPA diluted pitchblende and monazite. 68 The 219-channel efficiencies vs. time are tabulated in Table A-2. The 219 efficiency was determined by measuring the corrected counts per minute (cpm) in the 219-channel and dividing by the activity of spike solution used to prepare the standard (Table A-1). Figures A-3 and A-4 show all 5 RaDeCC detectors 219-channel efficiency with time for STD27, STDK-2, and STDK- 6. STD27 and STDK-2 each have constant efficiencies during the 120-day counting period, indicating the counting system efficiency did not change and sorption of Ac and its progeny were similar. Furthermore, STDK-6 has no obvious change in efficiency during the 60-day counting period and is similar in efficiency with STDK-2. Three MF K-series standards, STDK-1, STDK-5, and STDK-7, all show an obvious decline in 219-channel efficiency with time, particularly between 10 – 40 days (Figure A-5). This decrease in efficiency for the MF standards could be related to the inability of the MF fibers used (supplied by SCI, Inc.) to adsorb Ac quantitatively. Alternatively, it may reflect a decreased ability of these fibers to release 219 Rn generated from the 227 Ac that was originally sorbed. Each standard was fit with a Bateman equation to determine the initial and final activities of Ra and Ac. As a constraint, 227 Th was assumed to be initially equal to 223 Ra; this assumption produced a fit that was visually better than if the initial 227 Th was left as a free parameter. Counts were done on multiple channels, with the efficiency for each channel determined from STD27 used to compute the activity. Figure A-5 shows dpm (disintegration per minute) vs. days after creation. Each MF standard’s initial activity (m3 in the fit) is the initial value for 223 Ra. The final activity (m1 in the fit) is the effective activity for 227 Ac sorbed: all of these show a drop of about 20-50%. There are a few possible causes to explain why the MF standards decline in 219-channel efficiency with time and the CC standards do not. First, the CC standards were recently (~2-5 months) created and their cartridges were damp for many 69 weeks before standard preparation. The fibers for MF standards K1, K-5 and K-7, on the other hand, were created many (~5) years earlier, were stored dry for a long period of time, and did not appear as dark as the CC fibers. The newly created CC standards might adsorb Ac and Th more strongly than aged fibers. To test this hypothesis of “new” and “aged” MnO 2 fibers binding Th and Ac more strongly, new MF fibers were prepared using the hot bath technique described above, and used to prepare STDK-8 and STDK-9. Table A-2 shows the first 3 weeks of counting STDK-8 and STDK-9. There is no obvious drop in efficiency during this time. All other MF STDK-series show an obvious drop in the first 3 weeks. This could suggest that these “new” MF fibers absorb Th more strongly than “aged” fibers. More time is needed to observe the absorption of Ac (~120 days) and release efficiency of 219 Rn. The system efficiency also reflects how well 219 Rn is transferred from the Mn-fibers into the Lucas cell where it is detected. One factor in eq. A-1 is (V L /V Sys ). Based on the volume differences for the MF and CC geometries, the CC efficiency should be 0.84 times the MF efficiency. It is apparent that this factor does not fully explain the lower ratio of standards K2 and K6 in comparison to K-8 and K-9, as all were prepared from the same spike. Another cause for the difference in these two geometries is the effectiveness of 219 Rn release from the two geometries, which may be affected by He flowrate. To evaluate the effect of flowrate, each geometry was run on each channel at varying flowrates. Different systems have different flowmeters, and the optimum flow on each meter had been chosen based on experiments done by Colbert (2004), using MF geometry. Figure A-6 shows STDK-series and STD27 219-channel efficiency with changing flow rate on all 5 detectors, plotted relative to the optimum flow that has been in use for many years. Both CC and MF geometries were measured. When standards K-1 and K-5 were used, counts were done over less than a week, after the activity of these had 70 largely stabilized, so this should not be a factor in the patterns observed. The MF geometry standards (STD27, K-1 and K-7) appear to be nearly on a plateau at the optimum flow rate on each channel, with the possible exception of K-1 on channel 5. Linear regressions have been fit to the data from ~20% below optimum, up to ~50% above optimum. Slopes are modest (Table A-3) for this geometry, generally less than ~5% of the efficiency at the optimum with a 10% flow rate change. A few detectors show elevated slopes for the MF geometry, especially the MF STDK-series. This could be related to points chosen below the optimal flow rate, which are sensitive to lower flow rates and show a 30% drop between the optimal flow rate 219-channel efficiency. On the other hand, standards in the CC geometry continue to increase steadily above the optimal flow rate, typically ~5-10% with a 10% increase in flow, about twice the slopes for the MF geometries on each channel. It is likely that He has preferential flow paths through some parts of the grooved commercial cartridges, so some of the 219 Rn produced in CC geometries may not be as efficiently transported, in comparison to the behavior of the MF geometry. A plateau may be reached at higher flow rates for the CC geometry (Fig. A-6). A drawback of increasing flow rates beyond the optimum used is that the water carried from the fiber to the detector will increase, reducing the length of reliable counting time for a run. Standard efficiencies below the optimal flow rates are even more sensitive to flow rate. In part, this is related to the short half-life of 219 Rn and long residence time within the system (Colbert, 2004; Moore, pers. comm.). 219 Rn will thus decay before it can be reach the Lucas cell at lower flow rates. It is clear that once optimal flow rates are selected, some care must be taken to set them properly, and the counting efficiency should be determined with a standard geometry like that of samples. Furthermore, when comparing activities determined in different geometries on multiple detector, the geometry corrections may vary if flow rates are not identical. 71 The final set of tests on counting efficiency was to determine sensitivity to fiber water content. The emanation efficiency of 219 Rn can be considered as the number of radon atoms released per number of radon atoms generated. The moisture content of the Mn-fibers has been shown to be an important factor for Rn emanation from Mn-fibers (Sakoda et al., 2011; Sun and Torgersen, 1998a). The interpretation for this phenomenon is that the stopping power of water is much greater than air (or helium), and when 219 Rn recoils from the Mn-fibers into the void space, it can be stopped in the water before it penetrates into the adjacent fibers, where it then diffuses into the stream of helium (Sakoda et al., 2011). Thus, the water content on the fibers is important in controlling how much radon enters into the stream of helium from the MnO 2 surface. Too much water will reduce the ability of Rn to diffuse from the water into the He stream before it can decay, and too little will allow it to penetrate more deeply into adjacent fiber material. Figure A-7 shows the 219-channel efficiency with changing water content for two CC geometry samples. A natural seawater sample from the South Pacific (SO245, collected with a pump flow of ~7 L/min) and STDK-2 (prepared at ~0.5 L/min) were used for this experiment. The water content is the ratio of water (g) to dry MnO2-coated fibers (g). There is little change in the 219-channel efficiency with % water change over the range investigated for the CC geometries. CC cartridges run in this experiment seem far less sensitive to moisture content than MF geometries (Sun and Torgersen, 1998a). This could be related to the small pore sizes of CC fibers and water surface tension that leads to more uniform water distribution, compared to the loose MF fibers. 72 Summary of Experiments To determine a satisfactory release and calibration factor for the RaDeCC system, channels 3 and 4 were rejected due to poor reproducible results for STD27 in the last 7 months. Inconsistent flow rates, large errors in standard measurements, and detector age contributed to rejection of data produced from these channels. As a result of the experiments done here, it appears that calibration of STD27 to the MF STDK-series are 10% different for its initial 223 Ra activity (Table A-4). STDK-1, STDK-8, and STDK-9 are 10% higher and STDK-5 and STDK-7 are 10% lower in 223 Ra activity compared to STD27. STDK-7 low activity might be related to the procedure of soaking the fibers overnight in the Ac solution, compared to the other standards that had Ac solution pumped through them. The 227 Ac calibration of STD27 to the MF STDK-series are a bit more complicated. STDK-1, STDK-5, and STDK-7 all drop significantly in activity by the first month after creation (Table A-3). This suggest that Th and Ac did not adsorb onto these “aged” fibers (supplied by SCI, Inc). The loss of Ac for these standards is 20-50% (Figure A-5). The “new” MF fibers, STDK-8 and STDK-9, show a steady dpm for the first 3 weeks after creation. Both “new” MF standards are 10% higher in their initial 223 Ra activity compared to STD27. This could suggest that STD27 has lost activity in the radium parents or the “new” MF fibers release radium more efficiently. STDK-1 also shows a 10% increase in its initial 223 Ra activity compared to STD27. This could be related to age of the fibers, the preparation of the fibers, or the time between preparing and spiking the fibers. More time and measurement must be allowed for STDK-8 and STDK-9 to resolve this problem of fiber preparation. 73 To determine the release factor for the CC geometry, flow rates and moisture content were evaluated against 219-channel efficiency. Table A-3 summarizes the flow experiment. The ratios of CC STDK-series and STD27 were used to determine the release factor. Excluding the MF STDK-series due to significant loss of 227 Ac and using only STD27, the release factor for the CC geometry is 69 ± 0.06%. This factored multiplied by the volume factor (0.84) describes the first 2 terms on the RHS of Eq. A-1. The moisture content for this geometry has little effect (Figure A-7). Therefore, 69% of the 219 Rn generated in the CC geometry will be transferred into the Lucas cell where it is detected. The absolute efficiency for CC geometry STDK-series is summarized in Table A-5. The ratios of STDK-2/ STDK-6 and STD27 are shown to the right. Furthermore, ratios of both CC STDK-series and STDK-8 are shown as well. The release factor for CC STDK-series and STD27 is 0.74 ± 0.04%. STDK-8 release factor is 10% lower (0.67 ± 0.06%). This release factor for CC STDK and STDK-8 is in agreement with the above-mentioned flow rates (69%). Therefore, a calibration factor of 10% must be added to the release factor for STD27 219- channel efficiency to find a satisfactory efficiency for the CC geometry. The total efficiencies for the CC and MF geometry are summarized in Table A-6. The CC geometry was determined from STDK-2 and STDK-6 and the MF geometry was determined from STDK-8 and STDK-9. 74 Conclusion STDK-2 and STDK-6 absolute efficiency for the 219-channels should be satisfactory for the standardization of 223 Ra and 227 Ac measurements in the CC geometry (Table A-6). Both standards show no obvious change in activity with time, with STDK-2 showing similar activities 180 days since creation (created July 6, 2017). STD27 absolute efficiency for the 219-channel in the MF geometry is 10% lower for 223 Ra calibration. STDK-8 and STDK-9 are 10% more efficient for detecting 223 Ra than STD27. 227 Ac calibration in this geometry still needs further investigation. Fiber preparation needs to be studied and explored for this geometry. 75 76 77 Figure A-3: STD27 219-Channel Efficiency vs. Time. RaDeCC detectors 1, 3, 4, 5, and 7 are shown above. These represent counts 3-5 hours long, except for channel 4 (3-4 hours due to more rapid water build-up). Error bars represent ±1 s based on counting statistics. STD27 was prepared by S. Colbert (2004) with fibers soaked in Ra-free seawater, spiked with dissolved EPA-diluted pitchblende and monazite. STD27 also contains dpm 232 Th and its progeny in secular equilibrium. STD27 has no obvious change in efficiency with time. 78 79 80 Figure A-4: STDK-2 and STDK-6 219-Channel Efficiency vs. Time. RaDeCC detectors 1, 3, 4, 5, and 7 are shown. STDK-2 and STDK-6 are CC geometry with a nominal activity of 7.9 dpm. Both K-series standards have no obvious change in efficiency with time. 81 82 Figure A-5: Bateman Equation Fits for STDK-1, STDK-5, and STDK-7 Activities. All three standards were run on all RaDeCC detectors. The Bateman curve was fitted to each MF K-series standard due to significant decrease in activity with time, assuming initial 223 Ra and 227 Th were in secular equilibrium. The Red dashed line on y-axis refers to the activity of 227 Ac spike that was passed through the fibers. Sample activity was calculated using the efficiency for STD27 on each detector. 83 84 85 Figure A-6: 219-Channel Efficiency vs. Flow Rate. RaDeCC detectors 1, 3, 4, 5, and 7 for MF and CC geometries are shown. Flow rate was plotted relative to flow meter set point normally used for each detector. Each standard was fit with a linear function (y=mx+b, with x=flow used minus normal flow set) and the resulting y-intercept was used to calculate the 219-channel efficiency in Table A-3. Data points within the bold dash lines were used to fit the linear function, ranging from 20% below normal flow to 50% above normal flow (arbitrary choices). 86 Figure A-7: 219-Channel Efficiency vs. % Water Content. A natural seawater sample from the South Pacific and STDK-2 were used for this experiment for detectors 1 and 7. The 219-channel efficiency for the SO245 sample was determined by choosing activity of the sample determined around 50% water content. The errors shown are counting statistics. 87 Table A-1: Standards prepared. Activities for K-series standards are in dpm unsupported 227 Ac on 4/27/2017. Geometries are either Moore Fiber (MF) or Commercial Cartridge (CC). STD27 was prepared by S. Colbert (2004) with fibers soaked in Ra-free seawater, spiked with dissolved, EPA diluted pitchblende and monazite. This standard also contains 29.76 dpm 232 Th and its progeny in secular equilibrium. Standard ID Geometry Activity (dpm) Creation Date STDK-1 MF 7.89 Jul 6, 2017 STDK-2 CC 7.91 Jul 6, 2017 STDK-3 CC 7.89 Jul 6, 2017 STDK-4 CC 7.89 Jul 6, 2017 STDK-5 MF 7.89 Nov 16, 2017 STDK-6 CC 7.89 Nov 16, 2017 STDK-7 MF 7.28 Dec 12, 2017 STDK-8 STDK-9 STD27 MF MF MF 7.28 7.28 2.56 Jan 18, 2017 Jan 18, 2017 Jun 10, 2004 88 Table A-2: 219-Channel Efficiency for RaDeCC Detectors for Standards Used. The 219-channel efficiency was determined by the ratio of the raw counts per minute (cpm) in the 219-channel and the known activity of 227 Ac that was bound to the Mn-fibers. STDK-1, STDK-2, STDK-5, STDK-6, STDK-7, STDK-8, STDK-9, and STD27 have nominal activities of 7.9, 7.9, 7.9, 7.9, 7.3, 7.3, 7.3, and 2.6 dpm. Detector 1 219 Eff. +/- date Detector 3 219 Eff. +/- date STDK-1 0.300 0.011 7/6/2017 STDK-1 0.285 0.011 7/10/2017 MF 0.315 0.016 7/11/2017 MF 0.293 0.011 7/13/2017 0.295 0.013 7/27/2017 0.299 0.012 7/27/2017 0.234 0.009 9/15/2017 0.273 0.011 9/15/2017 0.253 0.012 9/21/2017 0.297 0.013 10/5/2017 0.265 0.014 11/19/2017 0.255 0.015 10/10/2017 0.188 0.012 11/27/2017 0.265 0.012 11/15/2017 0.212 0.015 12/18/2017 0.233 0.012 11/29/2017 0.262 0.011 12/4/2017 0.233 0.011 12/6/2017 0.230 0.015 12/15/2017 0.221 0.018 12/20/2017 0.205 0.012 1/24/2018 STDK-2 0.158 0.009 9/20/2017 STDK-2 0.160 0.008 10/2/2017 CC 0.168 0.009 10/9/2017 CC 0.166 0.008 10/4/2017 0.179 0.009 10/11/2017 0.167 0.010 10/6/2017 0.172 0.013 10/23/2017 0.174 0.007 10/8/2017 0.166 0.010 11/3/2017 0.152 0.007 10/20/2017 0.161 0.010 11/17/2017 0.161 0.010 10/30/2017 0.179 0.010 11/19/2017 0.176 0.013 11/8/2017 0.167 0.015 12/11/2017 0.174 0.011 11/17/2017 0.186 0.014 1/26/2018 0.168 0.012 11/19/2017 0.169 0.010 11/24/2017 0.156 0.009 12/9/2017 0.189 0.011 12/15/2017 0.163 0.011 1/11/2018 STDK-5 0.257 0.013 11/18/2017 STDK-5 0.283 0.015 11/17/2017 MF 0.262 0.013 11/20/2017 MF 0.230 0.013 11/20/2017 0.194 0.012 11/24/2017 0.239 0.008 11/21/2017 0.219 0.008 11/28/2017 0.229 0.013 11/28/2017 0.185 0.009 12/4/2017 0.190 0.010 12/1/2017 0.185 0.012 12/15/2017 0.133 0.010 12/8/2017 0.141 0.011 1/20/2018 0.220 0.017 12/13/2017 0.178 0.014 1/13/2018 STDK-6 0.186 0.009 11/17/2017 STDK-6 0.170 0.010 11/17/2017 CC 0.189 0.012 11/20/2017 CC 0.172 0.008 11/20/2017 0.144 0.009 11/21/2017 0.167 0.011 11/27/2017 0.159 0.009 12/1/2017 0.177 0.012 11/29/2017 0.167 0.012 12/2/2017 0.177 0.012 11/30/2017 0.154 0.011 12/9/2017 0.161 0.009 12/11/2017 0.144 0.011 12/13/2017 0.173 0.009 12/12/2017 0.166 0.010 1/11/2018 STDK-7 0.278 0.021 12/15/2017 STDK-7 0.311 0.015 12/13/2017 MF 0.200 0.010 1/23/2018 MF 0.293 0.017 12/18/2017 0.230 0.010 1/24/2018 89 STDK-8 0.319 0.016 1/19/2018 STDK-8 0.381 0.017 1/19/2018 MF 0.348 0.017 1/20/2018 MF 0.381 0.014 1/22/2018 0.309 0.016 1/29/2018 0.328 0.011 1/26/2018 0.298 0.014 2/5/2018 0.361 0.018 1/29/2018 0.339 0.010 1/30/2018 0.351 0.012 1/31/2018 STDK-9 0.353 0.018 1/22/2018 STDK-9 0.410 0.021 1/20/2018 MF 0.312 0.015 1/24/2018 MF 0.346 0.016 1/22/2018 0.270 0.013 1/29/2018 0.341 0.016 1/25/2018 0.321 0.012 1/30/2018 0.328 0.016 1/29/2018 0.340 0.017 1/31/2018 0.314 0.012 2/5/2018 STD27 0.336 0.027 8/31/17 STD27 0.245 0.025 7/28/17 MF 0.238 0.028 9/18/17 MF 0.332 0.031 9/14/17 0.272 0.026 9/27/17 0.323 0.029 9/30/17 0.302 0.023 10/2/17 0.284 0.030 10/12/17 0.266 0.027 10/6/17 0.249 0.031 10/23/17 0.286 0.022 10/11/17 0.268 0.016 11/2/17 0.305 0.022 10/20/17 0.326 0.021 11/10/17 0.295 0.024 10/27/17 0.229 0.024 11/18/17 0.311 0.023 11/6/17 0.250 0.021 11/22/17 0.262 0.029 11/13/17 0.236 0.030 11/28/17 0.283 0.024 11/17/17 0.223 0.024 12/11/17 0.288 0.040 11/28/17 0.284 0.031 1/22/18 0.299 0.030 12/6/17 0.349 0.026 2/1/18 0.258 0.028 12/12/17 0.298 0.041 1/8/18 0.287 0.029 2/2/18 Detector 4 219 Eff. +/- date Detector 5 219 Eff. +/- date STDK-1 0.251 0.018 7/7/17 STDK-1 0.441 0.025 7/7/17 MF 0.240 0.017 8/1/17 MF 0.439 0.020 7/9/17 0.290 0.012 8/31/17 0.347 0.015 7/18/17 0.231 0.015 10/7/17 0.368 0.020 7/19/17 0.194 0.017 11/10/17 0.375 0.012 7/31/17 0.213 0.014 12/18/17 0.362 0.016 9/1/17 0.306 0.011 10/8/17 0.289 0.011 11/14/17 0.333 0.016 11/18/17 0.308 0.017 12/12/17 0.279 0.019 12/17/17 0.284 0.015 12/19/17 0.275 0.012 1/2/18 90 STDK-2 0.170 0.018 7/13/17 STDK-2 0.231 0.009 7/11/17 CC 0.160 0.010 7/20/17 CC 0.228 0.014 9/13/17 0.174 0.011 7/24/17 0.240 0.013 9/15/17 0.169 0.017 9/13/17 0.269 0.011 10/10/17 0.179 0.011 10/11/17 0.225 0.011 10/20/17 0.154 0.015 10/31/17 0.251 0.012 10/25/17 0.157 0.011 11/13/17 0.236 0.010 11/10/17 0.180 0.013 11/20/17 0.268 0.013 11/17/17 0.176 0.015 11/21/17 0.245 0.014 12/11/17 0.178 0.012 11/27/17 0.219 0.014 1/9/18 0.200 0.010 12/6/17 0.179 0.011 1/5/18 STDK-5 0.211 0.011 11/17/2017 STDK-5 0.347 0.012 11/16/2017 MF 0.172 0.011 11/21/2017 MF 0.340 0.013 11/20/2017 0.177 0.012 12/2/2017 0.271 0.013 11/29/2017 0.147 0.012 12/12/2017 0.240 0.006 11/29/2017 0.198 0.016 12/13/2017 0.263 0.017 11/30/2017 0.185 0.009 12/15/2017 0.227 0.011 12/1/2017 0.188 0.014 1/8/2018 0.193 0.011 12/11/2017 0.191 0.012 1/8/2018 0.189 0.014 1/22/2018 STDK-6 0.193 0.011 11/18/2017 STDK-6 0.259 0.014 11/17/2017 CC 0.157 0.009 11/24/2017 CC 0.217 0.011 11/19/2017 0.180 0.012 11/28/2017 0.220 0.017 11/28/2017 0.170 0.015 12/6/2017 0.217 0.010 12/2/2017 0.183 0.013 12/11/2017 0.202 0.011 12/8/2017 0.175 0.014 12/12/2017 0.238 0.014 12/15/2017 0.173 0.008 1/12/2018 STDK-7 0.235 0.014 12/17/2017 STDK-7 0.320 0.013 12/15/2017 MF 0.261 0.020 12/19/2017 MF 0.314 0.017 12/18/2017 0.167 0.009 12/30/2017 0.350 0.017 12/20/2017 0.342 0.016 1/12/2018 STDK-8 0.346 0.017 1/22/2018 STDK-8 0.408 0.016 1/18/2018 MF 0.304 0.016 1/24/2018 MF 0.447 0.016 1/25/2018 0.281 0.016 1/29/2018 0.445 0.017 1/26/2018 0.404 0.016 2/4/2018 STDK-9 0.326 0.020 1/19/2018 STDK-9 0.481 0.020 1/19/2018 MF 0.340 0.012 1/20/2018 MF 0.446 0.023 1/26/2018 0.311 0.017 1/26/2018 0.443 0.017 1/31/2018 0.339 0.019 1/30/2018 0.365 0.019 2/2/2018 0.417 0.015 2/5/2018 91 STD27 0.275 0.033 6/19/17 STD27 0.391 0.027 7/7/17 MF 0.264 0.034 6/30/17 MF 0.335 0.038 8/1/17 0.228 0.033 7/8/17 0.430 0.038 9/11/17 0.202 0.030 7/10/17 0.435 0.033 9/15/17 0.280 0.035 7/11/17 0.394 0.029 9/20/17 0.256 0.024 7/31/17 0.414 0.030 9/26/17 0.260 0.033 9/1/17 0.428 0.035 9/29/17 0.257 0.029 9/7/17 0.355 0.022 10/16/17 0.288 0.039 10/3/17 0.398 0.025 10/22/17 0.267 0.036 10/5/2017 0.329 0.031 11/6/17 0.265 0.036 10/9/17 0.372 0.030 11/23/17 0.263 0.035 10/15/17 0.395 0.028 11/27/17 0.228 0.034 11/4/17 0.331 0.021 12/6/17 0.205 0.022 11/14/17 0.335 0.035 12/7/17 0.263 0.026 11/19/17 0.343 0.039 12/13/17 0.220 0.029 11/29/17 0.395 0.037 1/12/18 0.309 0.033 12/1/17 0.342 0.024 1/22/18 0.235 0.020 12/8/17 0.177 0.024 12/20/17 0.189 0.030 1/18/18 0.231 0.020 1/31/18 Detector 7 219 Eff. +/- date STDK-1 0.423 0.021 7/8/17 MF 0.369 0.015 9/13/17 0.382 0.020 9/19/17 0.354 0.012 9/30/17 0.360 0.014 10/9/17 0.375 0.014 11/17/17 0.343 0.012 11/21/17 0.373 0.016 12/12/17 0.320 0.013 12/13/17 0.277 0.013 1/8/18 0.247 0.015 1/13/18 STDK-2 0.244 0.013 7/19/17 CC 0.247 0.013 9/15/17 0.245 0.008 9/29/17 0.257 0.011 10/26/17 0.253 0.011 11/1/17 0.277 0.014 11/14/17 0.292 0.017 11/18/17 0.283 0.013 11/28/17 0.260 0.012 12/12/17 0.245 0.014 12/16/17 0.283 0.011 12/17/17 0.274 0.008 1/2/18 92 STDK-5 0.395 0.013 11/17/2017 MF 0.307 0.012 11/19/2017 0.291 0.013 11/27/2017 0.278 0.010 12/6/2017 0.182 0.013 12/7/2017 0.195 0.010 12/9/2017 0.205 0.009 12/11/2017 0.297 0.015 12/12/2017 0.283 0.011 12/27/2017 STDK-6 0.254 0.010 11/16/2017 CC 0.243 0.013 11/20/2017 0.257 0.015 11/29/2017 0.232 0.013 12/1/2017 0.248 0.011 12/6/2017 0.260 0.018 12/7/2017 0.270 0.014 12/15/2017 0.254 0.018 12/27/2017 STDK-7 0.394 0.020 12/13/2017 MF 0.374 0.011 12/26/2017 0.285 0.015 1/22/2018 STDK-8 0.464 0.019 1/20/2018 MF 0.429 0.019 1/22/2018 0.434 0.022 1/24/2018 0.419 0.017 1/29/2018 0.356 0.017 1/30/2018 0.461 0.019 1/31/2018 0.382 0.020 2/2/2018 0.366 0.011 2/5/2018 STDK-9 0.478 0.024 1/18/2018 MF 0.464 0.020 1/22/2018 0.361 0.014 1/24/2018 0.401 0.017 1/29/2018 0.370 0.018 2/1/2018 0.378 0.018 2/4/2018 STD27 0.400 0.035 7/13/17 MF 0.424 0.038 9/1/17 0.469 0.023 9/12/17 0.422 0.028 9/22/17 0.458 0.035 10/4/17 0.437 0.031 10/10/17 0.362 0.028 11/5/17 0.384 0.031 11/24/17 0.454 0.038 12/1/17 0.387 0.027 12/6/17 0.438 0.032 12/26/17 0.393 0.040 1/19/18 0.298 0.042 1/26/18 0.338 0.029 1/26/18 0.348 0.039 1/31/18 93 Table A-3: 219-Channel Efficiency for Standards Used as a Function of Flow Rate. Flow rate was normalized to average flow rate run for each detector. Each standard was fit with a linear function (y=mx+b) and the resulting y-intercept was used for the final 219-channel efficiency. The ratios of each standard are shown to the far right. Channel # Standard ID Geometry Slope +/- intercept +/- Channel 1 STDK-6 CC 0.009 0.002 0.154 0.006 STDK-7 MF 0.009 0.002 0.198 0.006 STD27 MF -0.008 0.003 0.289 0.008 Channel 3 STDK-2 CC 0.012 0.004 0.130 0.009 STDK-1 MF 0.009 0.005 0.200 0.010 STD27 MF -0.002 0.002 0.285 0.006 Channel 4 STDK-2 CC 0.008 0.006 0.170 0.007 STDK-7 MF 0.006 0.002 0.195 0.003 STD27 MF -0.001 0.007 0.242 0.008 Channel 5 STDK-2 CC 0.010 0.005 0.189 0.008 STDK-1 MF 0.011 0.005 0.232 0.015 STD27 MF 0.013 0.007 0.300 0.013 Channel 7 STDK-2 CC 0.020 0.010 0.211 0.013 STDK-1 MF 0.004 0.007 0.262 0.012 STD27 MF -0.002 0.009 0.369 0.015 94 Table A-4: MF Standard Efficiencies for RaDeCC Detectors. The STDK-series 219-channel efficiency for the first 5-20 days after creation are shown. STD27 219-channel efficiency is the last 7-month average. Ratios are STDK-series and STD27 (K/27). Error in the ratio is error propagation from the standard error. Weighted average is from channels 1, 5, and 7 only. Sample ID Channel # E219 Average Std Error Sample # (N) K/27 sig dpm sig Spike dpm STD27 1 0.287 0.006 16 Last 7 3 0.277 0.012 13 month 4 0.246 0.007 21 average 5 0.378 0.009 17 7 0.401 0.010 15 STDK-1 1 0.307 0.007 2 1.07 0.03 8.46 0.26 7.89 First 5 3 0.289 0.004 2 1.04 0.05 8.24 0.37 7.89 day 4 0.251 0.018 1 1.02 0.08 8.06 0.64 7.89 average 5 0.440 0.001 2 1.17 0.03 9.20 0.22 7.89 7 0.423 0.021 1 1.06 0.06 8.33 0.46 7.89 Weighted 1.11 0.02 Average STDK-5 1 0.260 0.002 2 0.91 0.02 7.15 0.16 7.89 First 5 3 0.283 0.015 3 1.02 0.07 8.07 0.55 7.89 day 4 0.191 0.020 2 0.78 0.08 6.14 0.66 7.89 average 5 0.347 0.012 1 0.92 0.04 7.24 0.31 7.89 7 0.395 0.013 1 0.99 0.04 7.77 0.31 7.89 Weighted 0.93 0.02 Average STDK-7 1 0.272 0.021 1 0.95 0.08 6.91 0.55 7.28 First 5 3 0.302 0.009 2 1.09 0.06 7.95 0.42 7.28 day 4 0.248 0.013 2 1.01 0.06 7.34 0.45 7.28 average 5 0.320 0.013 1 0.85 0.04 6.17 0.29 7.28 7 0.394 0.020 1 0.98 0.06 7.15 0.41 7.28 Weighted 0.93 0.03 Average STDK-8 1 0.318 0.011 4 1.11 0.04 8.09 0.32 7.28 First 20 3 0.357 0.009 6 1.29 0.06 9.39 0.47 7.28 day 4 0.310 0.019 3 1.26 0.09 9.19 0.63 7.28 average 5 0.426 0.012 4 1.13 0.04 8.21 0.30 7.28 7 0.414 0.015 8 1.03 0.04 7.52 0.32 7.28 Weighted 1.08 0.01 Average STDK-9 1 0.319 0.014 5 1.11 0.05 8.11 0.40 7.28 First 20 3 0.332 0.007 4 1.20 0.06 8.74 0.42 7.28 day 4 0.329 0.007 4 1.34 0.05 9.74 0.36 7.28 average 5 0.430 0.019 5 1.14 0.06 8.29 0.42 7.28 7 0.408 0.021 6 1.02 0.06 7.42 0.41 7.28 Weighted 1.09 0.01 Average 95 Table A-5: CC Standard Efficiencies for RaDeCC Detectors. STDK-2 and STDK-6 219-channel efficiency for all RaDeCC detectors are shown. STDK-2 and STDK-6 219-channel efficiency is the average from July - January, 2018 and November - January, 2018. The ratio of the STDK- series and STD27 is the geometry factor between CC and MF. Sample ID Channel # E219 Average Std Error Sample # (N) K/27 +/- K/K-8 +/- STDK-2 1 0.171 0.003 9 0.596 0.016 0.525 0.024 3 0.167 0.003 13 0.604 0.028 0.469 0.025 4 0.173 0.004 12 0.704 0.026 0.558 0.040 5 0.241 0.005 10 0.639 0.021 0.594 0.046 7 0.263 0.005 12 0.657 0.020 0.626 0.030 STDK-6 1 0.164 0.006 8 0.571 0.024 0.503 0.026 3 0.171 0.017 7 0.618 0.067 0.480 0.049 4 0.176 0.004 7 0.716 0.028 0.567 0.037 5 0.225 0.008 6 0.597 0.026 0.555 0.044 7 0.252 0.004 8 0.630 0.018 0.600 0.024 Table A-6: CC and MF geometry 219-channel efficiencies for USC RaDeCC detectors 1, 3, 4, 5, and 7. CC geometry 219-channel efficiency was determined from the average 219-channel efficiency of STDK-2 and STDK-6. The MF geometry 219-channel efficiency was determined from the average 219-channel efficiency of STDK-8 and STDK-9. The CC geometry is satisfactory for 227 Ac standardization. The MF geometry is satisfactory only for 223 Ra standardization. Sample ID Channel # 219-channel Efficiency Error CC Geometry 1 0.167 0.035 3 0.169 0.055 4 0.175 0.055 5 0.233 0.034 7 0.258 0.038 MF Geometry 1 0.319 0.018 3 0.345 0.011 4 0.320 0.021 5 0.425 0.040 7 0.416 0.038 96 References Charette, M.A., H. Dulaiova, M.E. Gonneea, P.B. Henderson, W.S. Moore, J.C. Scholten, and M.K. Pham. “GEOTRACES radium isotopes interlaboratory comparison experiment.” Limnology and Oceanography: Methods, 10, 2012, 451-463. Colbert, Steven, L. “Radium isotopes in San Pedro Bay, California: Constraint on inputs and use of nearshore distribution to compute horizontal eddy diffusion rates.” University of Southern California, PhD Dissertation (2004). Giffin, C., A. Kaufman, W.S. Broecker. “Delayed coincidence counter for the assay of actinon and thoron.” J. Geophys. Res. 68, (1963): 1749–1757 Henderson, P. B., P. Morris, W. Moore, M. Charette. “Methodological advances for measuring low-Level radium isotopes in seawater.” Journal of Radioanalytical and Nuclear Chemistry, vol. 296, no. 1, (2012): 357–362. Moore, Willard S., and Ralph Arnold. “Measurement of 223Ra and 224Ra in coastal waters using a delayed coincidence counter.” Journal of Geophysical Research: Oceans, vol. 101, no. C1, (1996): 1321–1329. Moore, Willard S. “Fifteen years experience in measuring 224Ra and 223Ra by delayed- Coincidence counting.” Marine Chemistry, vol. 109, no. 3-4, (2008): 188–197. Moore, Willard S., and Pinghe Cai. “Calibration of RaDeCC systems for 223Ra measurements.” Marine Chemistry, vol. 156, (2013): 130–137. Moore, Willard S. “Sampling 228Ra in the deep ocean.” Deep Sea Research, vol. 24, no. 4, (1976): 207. 97 Reid, D., D. R. Schink, R. M. Key. “Radium, thorium, and actinium extraction from seawater using an improved manganese-Oxidecoated fiber.” Deep Sea Research Part B. Oceanographic Literature Review, vol. 26, no. 12, (1979): 769. Sakoda, Akihiro, Y. Ishimori, K. Yamaoka. “A comprehensive review of radon emanation measurements for mineral, rock, soil, mill tailing and fly ash.” Applied Radiation and Isotopes, vol. 69, no. 10, (2011): 1422–1435 Scholten, Jan C., M. K. Pham, O. Blinova, M. A. Charette, H. Dulaiova, M. Eriksson. “Preparation of Mn-Fiber standards for the efficiency calibration of the delayed coincidence counting system (RaDeCC).” Marine Chemistry, vol. 121, no. 1-4, (2010): 206–214. 98 Appendix B Rapid measurement of 227 Ac Using a High Purity Germanium Well-Type Detector Actinium-227 (t 1/2 = 22 y) is a member of the 235 U decay series, which is directly produced from 231 Pa (t 1/2 = 32,800 y) by alpha decay. Natural occurrences of 227 Ac activity are extremely low, since the 235 U decay series is only 0.72% of the 238 U series. In deep-sea marine environments, 227 Ac can reach activities around 3.0 dpm/g or slightly higher depending on the 231 Pa that is present in deep-marine sediments. Measurements of this rare radioisotope are difficult. Recent techniques include: ion exchange with alpha spectroscopy analysis of 227 Th its daughters, delayed coincidence counting of its 219 Rn great grand-daughter, liquid scintillation counting, and gamma ray spectroscopy (Alharbi et al, 2016). All these methods, except for gamma ray spectroscopy, require difficult analytical separation of 227 Ac. The current effort here will employ a gamma spectrometry method that is non-destructive and simple to measure using a High Purity Germanium Well-Type Detector. The USC radioisotope lab has two high HPGe well detectors (HPGe ORTEC, 120 cm 3 active volume). Detector 1 (MCB2) has a Full Width Half Maximum resolution of 1.23 eV at 122 keV and 2.10 keV at 1330 keV. Detector 2 (MCB1) has 1.47 eV at 122 keV and 2.06 keV at 1330 keV. Both detectors have a useful energy range between 30.0 – 2000 keV. However, below 100 keV, some self-absorption may occur, and the sample matrix needs to be considered when measuring different environmental samples. Detector efficiencies were determined from counting the activities of known standards obtained from EPA (SRM-1 diluted pitchblende and SRM-2 diluted monzonite) and Eckert & 99 Zeigler (WHOI E&Z Actinium-227 CRM). Standards were 3.0 cm high geometry in polyethylene tubes and corrections were made to sample results to account for the lower sample heights that were used. Samples were counted for 2-4 days, and Table B-1 shows the energies used in both detectors for the determination of 227 Ac in deep-sea marine sediments. Additional information is listed for each isotope in Table B-1, which are things to consider when measuring each energy region. Self-absorption and coincidence factors should cancel out in the standard and the sample because the pitchblende and sample are both mixed sediments. Actinium-227 beta minus decays to 227 Th, with very weak gamma emissions. Therefore, direct measurement of 227 Ac via gamma ray spectroscopy is challenging. However, the 227 Ac granddaughter, 223 Ra, has a relatively strong gamma at 270 keV that can be quantified if enough time is given to measurement. Table B-1 shows numerous gamma lines of the 235 U series that can be measured on a HPGe well-detector. This work will focus on 2 energy regions, 270 keV and 338 keV. These energy regions were chosen due to high branching ratio intensity of 223 Ra and non-interfering gammas that cannot be corrected easily. 223 Ra at 270 keV was chosen due to its high branching ratio and sharing of the region with its daughter, 219 Rn. Together, these isotopes will add to the counting statistics for measurement of 223 Ra. 219 Rn can be assumed to be in equilibrium with 223 Ra since its half-life is only 4 seconds. Furthermore, radon leakage can be ignored due to its short half-life. Caution must be used when measuring 223 Ra at 270 keV due to the interference of 228 Ac at 270.2 keV (BR% = 3.4). This energy region can be corrected for 228 Ac if the activity of 228 Ac is known. 228 Ac gives off two reliable gammas at 338.3 and 911.2 keV (BR% = 11.3 and 26.6). Detector efficiency drops exponentially with energy from 120 keV, so the 338 keV was chosen over the 911 keV to give better statistics for the activity of 228 Ac. Another problem that arises 100 with measurement of 228 Ac at 338.3 keV is the interference from 223 Ra at 338.3 keV (BR% = 2.8). Therefore, there are two energy regions that have the same two isotopes interfering with each other. Below is a system of equations that will solve for 223 Ra at 270 keV. The count rate (CR) is calculated by the following relationships for 228 Ac and 223 Ra at 270 and 338 keV: CR 223Ra-270keV = A 223Ra x x 270keV x BR 223Ra-270 keV (B-1) CR 223Ra-338keV = A 223Ra x x 338keV x BR 223Ra-338 keV (B-2) CR 228Ac-270keV = A 228Ac x x 270keV x BR 228Ac-270 keV (B-3) CR 228Ac-338keV = A 228Ac x x 338keV x BR 228Ac-338 keV (B-4) where: BR is the branching ratio, A is the activity, and x is the counting efficiency for a specific energy region. Total efficiencies and branching ratios used in these equations can be found in Table B-1. For a sample containing 223 Ra and 228 Ac, both plus 219 Rn will share the 270 keV energy peak: x total-270keV = x 270keV (BR 223Ra-270 keV + (g)BR 228Ac-270 keV + BR 219Rn-270 keV ) = CR 270keV /A 235U_pitchblende (B-5) Where (g) = A 235U /A 232Th in pitchblende standard. Based on the calculated counting efficiency for the 270 keV energy region (x total-270keV ), using the respective BR of 223 Ra, 228 Ac, and 219 Rn 101 at 270 keV and using the activity ratio of 235 U/ 232 Th of 3.53 in the EPA pitchblende standard in use, x 270keV can be found using the following equation: x total-270keV = x 270keV (0.137 + 0.034/3.53 + 0.108 ) (B-6) x total-270keV = x 270keV (0.255) (B-7) Using the following relationship to represent the CR at 270 keV and 338 keV for 223 Ra and 228 Ac: CR 223Ra-270keV = ∑CR 270keV - CR 228Ac-270keV (B-8) CR 223Ra-338keV = ∑CR 338keV - CR 228Ac-338keV (B-9) CR 223Ra-270keV can be solved if we set up a system of equations for B-1, B-2, B-3, B-4, B-8, B-9. Plugging in B-3 and B-4 into B-8 and B-9, and setting up a system of 4 equations we can arrive at corrected factors for the total count rates ∑CR 270keV and ∑CR 338keV . Below is a Matlab script that solves these systems of 4 equations: 102 Matlab Script: syms e270 e338 A228 A223 CR270 CR338 x y BR1 BR2 BR3 BR4 %x = CR 223Ra @ 270 keV %y = CR 223Ra @ 338 keV e270=0.222;%absolute eff. of the 270 keV = eff. total = (BR223Ra + BR219Rn + BR228Ac/3.53) e338=0.104; %relative eff. of the 338 keV = eff. total = (BR228Ac + BR223Ra/0.38) BR1=9.7e-3; %(g)BR of 228Ac @ 270 keV (0.0343/3.53), 3.53 is ratio of activity of 235U/232Th in pitchblende BR2=0.032; %BR of 228Ac @ 338 keV (0.113/3.53), 3.53 is ratio of activity of 235U/232Th in pitchblende BR3=0.235; %combine both 223Ra and 219Rn BR @ 270 keV since total eff. is derived this way BR4=0.0279; %BR of 223Ra @ 338keV eqn1 = x == CR270 - (e270)*(A228)*(BR1); eqn2 = y == CR338 - (e338)*(A228)*(BR2); eqn3 = x == (e270)*(A223)*(BR3); eqn4 = y == (e338)*(A223)*(BR4); sol = solve([eqn1, eqn2, eqn3, eqn4], [y, x, A228, A223]); xSol = sol.x; ySol = sol.y; A223Sol = sol.A223; A228Sol = sol.A228; xx=vpa(xSol,3) xx=1.04*CR270 - 0.671*CR338 There are two corrected factors for the determination of CR 223Ra-270keV. The correction factor for 223 Ra at 270 keV is higher than 1.0 because the correction factor from 228 Ac at 338 keV has the influence of 223 Ra from this energy region. 103 Table B-1: HPGe Total Efficiencies and Branching Ratios. Total efficiencies were calculated from known standards measured on the well-style gamma-detectors MCB1 and MCB2. The total efficiency shown below is measured on MCB2. Monazite and Actinium standard efficiencies were not used in the above calculations. 227 Th was not used for the determination of 227 Ac due to 214 Pb interference at 259.3 keV. The activity of 232 Th and 235 U in pitchblende standard is 15 and 52.8 dpm. All isotopes in the 232 Th and 235 U chains are believed to be in equilibrium. isotope energy keV total efficiency BR % Efficiency Source Comments 223 Ra 269.5 0.0567 13.7 Pitchblende Standard Interference from 228 Ac at 270.2 keV (3.43%) and 219 Rn at 271.2 (10.8%) 219 Rn 271.2 0.0567 10.8 Pitchblende Standard Interference from 223 Ra at 269.5 keV (12.7%) and 228 Ac at 270.2 keV (3.43%) 228 Ac 270.2 0.0567 3.4 Pitchblende Standard Interference from 228 Ac and 219 Rn 228 Ac 338.3 0.0185 11.3 Monazite Standard Interference from 223 Ra at 338.3 keV (2.79%) 223 Ra 338.3 0.0062 2.8 Pitchblende Standard Interference from 228 Ac at 338.3 keV (3.43%) 227 Th 256.3 0.0177 7.0 Actinium Standard 214 Pb interferes at 259.2 keV (0.55%) 104 Figure B-1: Schematic Illustration for the Determination of 223 Ra at 270 keV 105 References Alharbi, Sami H., and Riaz A. Akber. “Broad-Energy germanium detector for routine and rapid analysis of naturally occurring radioactive materials.” Journal of Radioanalytical and Nuclear Chemistry, vol. 311, no. 1, Aug. (2016): 59–75. Siegel, P. B. “Gamma spectroscopy of environmental samples.” American Journal of Physics, vol. 81, no. 5, (2013): 381–388. 106 Appendix C Commercial Cartridge Absorption Efficiency and Counting Protocol Dissolved phase 227 Ac was collected via a dual-flowpath in-situ pump (McLane WRT- LV), which was processed through two grooved acrylic Commercial Cartridges (CC) impregnated with MnO 2 that sat in series. Having two cartridges organized in series allows for absorption efficiencies of radium, actinium, and thorium from seawater. These cartridges have shown to quantitatively absorbed radium and actinium at flow rates below 1.0 L/min (Reid et al., 1979). Flow rates during the C-Disk-IV pump casts varied from 5-7 L/min. Large volumes of seawater (900-1500 L) were pumped during a 4-hour pumping period, which determined the flow rate. Absorption efficiencies for 227 Ac was established by having two CC organized in series. The following equations were used to determine the absorption efficiency of 227 Ac: 𝑒 =1− V $ (C-1) where A and B are the activities of 227 Ac absorbed onto cartridges A and B. Cartridge A is placed before cartridge B in series. After the activities were calculated for both A and B cartridges, a linear fit was obtained by plotting cartridge activity B vs. A. Figure C1 shows the results of the fit. The slope of the linear fit is the B/A term in equation 1. All station’s A and B cartridges are present in the fit. The average absorption efficiency was 54%. This is a little lower than expected (Hammond et al., in prep). 107 Figure C-1: Linear Fit for Commercial Cartridges B vs. A. The slope of the linear fit is the B/A term in equation C1. The linear line was forced through zero. All station’s A and B cartridges in the C-Disk-IV cruise are present in the graph. Sample activity was calculated as followed: 𝐴 %WXX = $YV AZ[ \]RR (C-2) where A corr is the corrected activity for the sample, f miss is the fraction of 227 Ac that was missed by both cartridges, and A and B are cartridges A and B defined in E-B1. The f miss term was calculated as followed: 108 𝑓 _O`` =(1−𝑒)(1−𝑒) (C-3) where e is the absorption efficiency defined in E-B1 and found through the linear fit in figure B1. The average fraction missed was 21% for all stations. This was used for all station’s activity correction. The value for e varied at a few stations, especially for deeper casts. However, the 21% average f miss corresponds reasonably well with upper water column 231 Pa activities (Table 1). The discrepancies in e between stations and depths are attributed to faulty sealing around the first cartridge, allowing a large fraction of flow to bypass it (Hammond et al., in prep). Finally, the concentration at each station’s depth was calculated by dividing the A corr by the volume of seawater that passed through the two cartridges. Counting Protocols The 227 Ac activity of each cartridge was measured on RaDeCC for a period that averaged between 180 – 300 minutes. Each cartridge was measured at least three times on different detector systems. Errors associated with cartridge activities are derived from the standard deviation of replicate counts. Detector efficiencies for 227 Ac are described in Appendix A. For most detectors, the efficiency ranged between 16 – 25% for commercial cartridges. The resulting concentrations for 227 Ac in the upper water column agree well with 231 Pa measured from Hayes et al., (2013). 109 References Hayes, C. T., Anderson, R. F., Jaccard, S. L., François, R., Fleisher, M. Q., Soon, M., & Gersonde, R. “A new perspective on boundary scavenging in the North Pacific Ocean”. Earth and Planetary Science Letters, 369–370, (2013): 86–97. Reid, D., D. R. Schink, R. M. Key. “Radium, thorium, and actinium extraction from seawater using an improved manganese-Oxidecoated fiber.” Deep Sea Research Part B. Oceanographic Literature Review, vol. 26, no. 12, (1979): 769. 110 Appendix D Flux of a radionuclide during a core incubation The following derivation has been done by D. Hammond (pers. Comm). If a benthic flux occurs into well-mixed water of thickness h 1 sitting above a sediment core at a constant flux J dC/dt = J/h 1 - lC (D-1) where C is concentration of excess radionuclide and t 1 is the time of the elapsed interval. The solution to this equation is lC 1 = (J/h 1 )(1 - exp(-lt 1 )) + lC 0 exp(-lt 1 ) (D-2) Incubation during a second interval for t 2 (with a different height h 2 ) leads to lC 2 = (J/h 2 )(1 - exp(-lt 2 )) + lC 1 (exp(-lt 2 )) (D-3) Subsequent intervals lead to the solution lC n = J(a n ) + lC 0 exp(-lSt) (D-4) where lC n is the concentration after the nth interval and St is total time elapsed prior to the n-1 incubation interval. The parameter a is given by a 1 = (1 - exp(-lt 1 ))/h 1 (D-5) and subsequent values for a i will be given by: a i = (1 - exp(-lt i ))/h i + (a i-1 )exp(-lt i ) (D-6) Computation of the flux based on a sample after draw n is then: J = (lC n - lC 0 exp(-lSt))/ a n (D-7) If water from multiple cores is combined into a single sample, the value for a n should be weighted by the volume used from each core. If lC n is in dpm/m3 and an is in 1/m, the units of flux J would be atoms/m2-min. Multiplying by the decay constant of the isotope in 1/y converts units to dpm/m2-y.

Asset Metadata

Creator Kemnitz, Nathaniel James (author)

Core Title Actinium-227 as a tracer for mixing in the Deep Northeast Pacific

Contributor Electronically uploaded by the author (provenance)

School College of Letters, Arts and Sciences

Degree Master of Science

Degree Program Geological Sciences

Publication Date 04/10/2018

Defense Date 04/09/2018

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag ²²⁷Ac,actinium-227,C-Disk-IV,diapycnal mixing,mixing,Northeast Pacific,OAI-PMH Harvest

Language English

Advisor Berelson, William (committee member), Hammond, Douglas (committee member), Moffett, James (committee member)

Creator Email kemnitz@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c89-8673

Unique identifier UC11671203

Identifier etd-KemnitzNat-6192.pdf (filename),usctheses-c89-8673 (legacy record id)

Legacy Identifier etd-KemnitzNat-6192.pdf

Dmrecord 8673

Document Type Thesis

Rights Kemnitz, Nathaniel James

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA

Abstract (if available)

Abstract ²²⁷Ac (t₁/₂ = 22 y) is produced by decay of ²³¹Pa (t₁/₂ = 32,800 y). ²³¹Pa is scavenged in the water column by falling particulates that carry it to sediments, where it decays to its more soluble daughter. A fraction of this ²²⁷Ac diffuses out of deep-sea sediments and is transported vertically and horizontally as it decays in the water column, providing a distribution that can be used to infer transport rates for other solutes in the deep ocean. ²²⁷Ac was measured throughout the water column and sediments at 5 stations along a cruise track from Hawaii to Alaska in the Northeast Pacific Basin. Sampling took place in August 2017 onboard the R/V Kilo Moana (C- Disk-IV cruise). The supported activity of ²²⁷Ac in the water column was determined from previous measurements of its ²³¹Pa parent made by others who worked near this region. ❧ The activity of excess ²²⁷Ac increases with depth at all 5 stations in the Northeast Pacific. Apparent vertical eddy diffusivities (Kz) derived from a one-dimensional eddy diffusion-decay model increase with depth and in areas near rough topography. Areas around the Musician Seamounts and the Mendocino Fracture Zone (MFZ) show the highest values of Kz, while areas in the higher latitudes, where the seafloor is characterized by westward dipping smooth topography, show lower values of Kz. Furthermore, mixing is enhanced in a benthic layer (∼ 500 m thick) near the seafloor and may decrease exponentially above this layer. This is strongly correlated with column stratification, which increases markedly above the benthic mixed layer. ❧ Two independent approaches were undertaken to quantify the source function of ²²⁷Ac from deep-sea sediments in the Northeast Pacific: direct measurement of ²²⁷Ac fluxes via core incubation and indirect estimates based on gamma counting ²¹⁰Pb and ²²⁷Ac in sediments. Core incubation fluxes for ²²⁷Ac are 3-10 times greater than the fluxes of ²²⁷Ac required to support the decay observed in the water column, but may be affected by handling artifacts. The latter approach, which uses ²¹⁰Pb and ²²⁷Ac measurements in sediments, provides a reasonable match for the benthic flux required by a 1-D model to match the water column data observed at the stations 2 and 3, which occupy a more open basin, characterized by constant depths throughout a large area. The sediment data at the higher latitude stations 4 and 5, and station 1 near the Hawaiian ridge predicts a flux that is 3-4 times greater than required by the ²²⁷Ac inventory in the overlying water column. Horizontal transport of water far from the bottom into a dipping slope could reduce the water column inventory of ²²⁷Ac below what is predicted by the benthic flux. This would suggest that a 1-D model may not predict a correct Kz value for these regions. However, if depth and bioturbation are constant along the flow path, then ²²⁷Ac should provide a well-constrained vertical eddy diffusivity and water column inventory, as shown by the basin stations, 2 and 3.