Determination of natural Fe organic complexes in the surface waters of the Amundsen Sea

The distribution and biological availability of Fe is strongly controlled by its physical-chemical speciation within seawater, where colloids and Fe-organic complexes are dominant factors. In order to study the distribution and the biological availability of Fe the natural Fe organic complexes were determined in the surface waters of the Amundsen Sea (300 m).

Methods
Samples were collected using Go Flo bottles and filtered under ultra-clean conditions in flow benches (class 0). The concentration of iron binding ligands (organic compounds which strongly bind Fe) and their binding strength (conditional stability constant) were studied in 5 size classes here: unfiltered water, 0.2 μm filtered water, < 1000 kDa (Stereapore, Mitsubishi-rayon Co. Ltd, Nishioka and al., 2000, 2005), < 100 kDa and < 10 kDa ultra-filtrated water (Sartorius, Vivaflow 50, Schlosser and Croot, 2008). The left-over fraction from the ultra filtrations (retentates) were also analyzed for dFe and ligand characteristics to ensure a mass balance calculation and validate the ultrafiltration method. The dissolved iron concentrations in all the size fractions (and retentates) were measured (see dFe measurement section) using a chemo luminescence method (FIA) with acidified samples (pH 1.8). Total iron concentrations will be measured 6-12 months after the acidification of the unfiltered sample. Ligand characteristics were determined by using a complexing ligand titration with addition of iron (between 0 and 10 nM of Fe added) in buffered seawater (mixed NH3/NH4OH borate buffer, 5 mM). The competing ligand 'TAC' (2-(2-Thiazolylazo)-p-cresol) with a final concentration of 10 μM was used and the complex (TAC)2-Fe was measured after equilibration (> 15 h) by cathodic stripping voltammetry (CSV) (Croot and Johansson, 2000). The electrical signal recorded with this method (nA) was converted to a concentration of (TAC)2-Fe (nM). Subsequently, the ligand concentration and the binding strength were estimated using the non-linear regression of the Langmuir isotherm (Gerringa and al., 1995) and a newer "Leo" model currently built up (Gerringa et al, in prep). The voltammetric equipment consisted of a μAutolab potentiostat (Type II and III, Ecochemie, The Netherlands), a mercury drop electrode (model VA 663 from Metrohm). All equipment was protected against electrical noise by a current filter (Fortress 750, Best Power).

Sampling statistics
26 stations were sampled on this cruise. These included 14 profiles and 5 stations where different size fractions were analyzed after 4 filtrations with different filter sizes (0.2 μm cut-off, 1000 kDa, 100 kDa and 10 kDa). In addition, 11 Fe/Ligand experiments were analyzed. Special attention was given to determine the iron binding ligands before and after incubation with and without artificial ligands in these experiments, to look at the response of algae and the change of the ligand characteristics during the incubations.

Preliminary results
An average ligand concentration of 0.789 nEq was found on the NBP0901 cruise, varying from 0.2 and 1.6 nEq of Fe. Highest ligand concentrations were found at 10 m depth followed by a minimum at 25 m. Concentrations increased with depth to become rather constant at 200 and 300 m. Low Fe binding strength of the ligands was found at 10 and 25 m in the polynya suggesting freshly produced ligands by organisms (phytoplankton or bacteria) or by a change in the ligand content and characteristics due to the biologic activity (the pool of strong binding ligands may be removed or used). The ratio ligand/dissolved iron (Fig. 38) clearly shows differences between the surface water and deeper samples. Very high ratios (10) were found in the surface waters of the Pine Island polynya due to the low dissolved iron concentration and high ligand concentrations. In the deep water (200 and 300 m) of the polynya and the circumpolar deep water upwelling in front of the PIG the ligand/dissolved iron ratio was close to 1 indicating a saturation of the ligands by iron and the possibility for iron to be removed from the water column by precipitation.

海水中铁（Fe）的分布与生物可利用性，主要受其物理化学形态的强烈调控，其中胶体与铁-有机络合物是主导控制因素。为研究铁的分布与生物可利用性，本研究于阿蒙森海（Amundsen Sea）300 m层的表层水体中测定了天然铁有机络合物。

## 方法
样品采用Go Flo采水器采集，并于0级超净工作台内完成过滤操作。本研究针对5个粒径分级开展铁结合配体（强结合铁的有机化合物）浓度及其结合强度（条件稳定常数）的分析，分级设置如下：未过滤水样、0.2 μm过滤水样、<1000 kDa级水样（Stereapore膜，三菱丽阳有限公司（Mitsubishi-rayon Co. Ltd），引用Nishioka等，2000、2005）、<100 kDa及<10 kDa超滤水样（赛多利斯（Sartorius）Vivaflow 50超滤装置，引用Schlosser和Croot，2008）。同时对超滤过程中的截留液（retentates）进行溶解态铁（dFe，dissolved Fe）与配体特征分析，以验证质量平衡计算结果及超滤方法的可靠性。

所有粒径分级水样及截留液中的溶解态铁浓度，均采用酸化至pH 1.8的样品，通过化学发光法（FIA）测定（详见溶解态铁测定部分）。未过滤水样酸化后6~12个月，将测定总铁浓度。

配体特征通过络合配体滴定法测定：在以5 mM氨-氯化铵硼酸盐缓冲体系缓冲的海水中，添加0~10 nM的铁进行滴定。本实验采用终浓度为10 μM的竞争配体TAC（2-(2-噻唑偶氮)-对甲酚，2-(2-Thiazolylazo)-p-cresol），在平衡时长>15 h后，通过阴极溶出伏安法（CSV，cathodic stripping voltammetry）测定(TAC)₂-Fe络合物浓度（引用Croot和Johansson，2000）。该方法记录的电信号（单位为纳安，nA）被转换为(TAC)₂-Fe的浓度（单位为纳摩尔，nM）。随后，通过朗缪尔等温线（Langmuir isotherm，引用Gerringa等，1995）的非线性回归分析，以及当前正在构建的新型‘Leo’模型（Gerringa等，待发表），估算配体浓度与结合强度。

伏安分析设备包括μAutolab电位仪（Ⅱ型与Ⅲ型，荷兰Ecochemie公司）以及瑞士万通（Metrohm）VA 663型滴汞电极。所有设备均通过电流滤波器（Fortress 750，Best Power公司）屏蔽电气噪声干扰。

## 采样统计
本航次共布设26个采样站位，其中包含14条断面剖面，另有5个站位分别采用4种不同孔径的滤膜（0.2 μm截留孔径、1000 kDa、100 kDa及10 kDa）进行分级过滤并分析各粒径级样品。此外，共完成11组铁/配体实验并开展分析。本实验重点测定了添加与不添加人工配体的培养实验前后的铁结合配体含量，以探究藻类响应及培养过程中配体特征的变化规律。

## 初步结果
在NBP0901航次中，配体平均浓度为0.789 纳当量（nEq），铁结合配体浓度介于0.2~1.6 nEq之间。配体浓度在10 m深度达到峰值，25 m深度降至最低；随后浓度随深度增加而升高，在200 m与300 m深度趋于稳定。

冰间湖区域10 m与25 m深度的配体铁结合强度较低，这表明配体可能由生物（浮游植物或细菌）新生成，或是因生物活动导致配体库与特征发生改变（强结合配体库可能被消耗或移除）。配体/溶解态铁比值（图38）清晰反映出表层水体与深层水样之间的显著差异。松岛冰间湖（Pine Island polynya）表层水体中该比值高达10，这源于其较低的溶解态铁浓度与较高的配体浓度。在冰间湖深层水体（200 m与300 m）以及松岛冰川（Pine Island Glacier，PIG）前缘上升的环极深层水中，配体/溶解态铁比值接近1，表明配体已被铁饱和，铁可能通过沉淀作用从水体中移除。