Silicon Isotopic Signatures Research Articles

Marine clay maturation induces systematic silicon isotope decrease in authigenic clays and pore fluids

Marine silicate alteration exerts a major influence on marine carbon and cation cycles, but has proven difficult to quantify. In this context, silicon isotopes of marine pore fluids became an important tracer. However, poorly constrained silicon isotope signatures of precipitates produced during silicate alteration (i.e. authigenic clays) remain a major source of uncertainty. Here we present in situ silicon isotope analyses of marine authigenic clays (intergrown iron-smectites and iron-glauconites) occurring within recent sediments from the Oregon margin, eastern North Pacific. We identify a trend to lower silicon isotopes (from −2.24‰ to −3.17‰), accompanied by decreasing aluminum/silicon ratios and increasing potassium oxide contents, which we interpret as an isotopic shift caused by progressive clay maturation via dissolution-reprecipitation reactions. Our modelling suggests that this clay maturation pathway, together with mixing of other fluid sources, may induce pore fluid silicon isotope shifts of up to −1.7‰, if sufficient newly precipitated clays are re-dissolved. This could potentially produce silicon isotopes values significantly lower than seawater and implies that conventional isotope-based approaches underestimate the prevalence of marine silicate alteration. Our findings highlight that clay maturation must be considered when interpreting silicon isotope signatures in terms of marine silicate alteration and upscaling to global element cycles.

Early differentiation processes on Mars inferred from silicon isotopes

Accretion of terrestrial planets involved partial or global melting events such as magma oceans or magma ponds. Mars experienced large-scale differentiation very early in its history, as shown by its 146Sm-142Nd and 182Hf-182W record. The broad variations in ε142Nd and ε142W of SNC meteorites highlight the presence of mantle sources that must have remained isolated, at least partly, after the crystallization of a global magma ocean. In this study, we have investigated whether the crystallization of the martian magma ocean could have generated mantle reservoirs characterized by different silicon isotope signatures, as the fractionation of Si isotopes between minerals and melts is known to depend on pressure. Thus, the goal of this study was to investigate whether there were any relationships between magma ocean crystallisation and possible variations in the Si isotope record of SNC meteorites. High resolution silicon isotope measurements were performed on twelve meteorites from the Shergottite, Nakhlite and Chassignite groups using a Neptune Plus MC-ICP-MS in dry plasma mode. The δ30Si values are in good agreement with previous studies but display a narrower range of variations with a mean value at −0.46 ‰ ± 0.07 (2SD). A magma ocean crystallization model shows that the range of δ30Si in SNCs is consistent with that generated by magma ocean crystallisation. In particular, there is a correlation between calculated 147Sm/144Nd for the moderately depleted mantle sources with δ30Si values; this correlation is consistent with the crystallization model if one includes trapped melt in the cumulates. In contrast, enriched shergottites displayed a very homogenous composition in Sm/Nd ratios, despite significant variability in δ30Si. This observation could be related to either fluid-rock interactions or redox effect during magma differentiation. Altogether, silicon isotope compositions of SNC provide new constraints about magma ocean crystallization processes in Mars.

Linking silicon isotopic signatures with diatom communities

The use of silicon isotope ratios (expressed as δ30Si) as a paleolimnological proxy in lacustrine systems requires a better understanding of the role of lake processes in setting the δ30Si values of dissolved Si (dSi) in water and in diatom biogenic silica (bSi). We determined the δ30Si of modern dSi (δ30SidSi) and bSi (δ30SibSi) in three lakes in Lassen Volcanic National Park (LAVO), California (USA), and produced diatom assemblage compositional data from the modern system and from sediment core samples. In the modern systems, we observe the largest magnitude diatom Si isotope fractionations yet reported, at −3.4 and −3.9‰ for Fragilaria dominated samples. Using statistical approaches designed to condense multivariate ecological data, we can deconvolve assemblage-specific Si isotope fractionations from the combined diatom assemblage-δ30Si data. For example, samples dominated by generally deeper water euplanktic species have low δ30SibSi values (<−1.10‰). Conversely, samples dominated by shallow water planktic or benthic periphyton have higher δ30SibSi values (>−0.14‰). These data suggest that δ30Si records from LAVO lakes reflect species specific Si isotope fractionations and thus act as paleolimnological proxy for the aquatic-habitat of bSi production. Silicon isotope analysis should be coupled with diatom community composition data and other geochemical proxies for the most robust paleolimnological interpretations. We also construct a Si mass-balance for Manzanita Lake based on elemental fluxes. Despite a short residence time of ∼4 months, it is an efficient Si sink: about 30% of inflowing Si is retained in the lake sediments. An entirely independent Si isotope-based estimate agrees remarkably well. Burial fluxes of bSi derived from radiometrically dated sediment cores yield retention rates of about a factor of three higher, which might suggest groundwater is an important term in the lake Si budget.

Evaluating the geochemistry and paired silicon and oxygen isotope record of quartz in siliceous rocks from the ~3 Ga Buhwa Greenstone Belt, Zimbabwe, a critical link to deciphering the Mesoarchean silica cycle

Rocks that make up the Mesoarchean (~3Ga) Buhwa greenstone belt (BGB) of the Zimbabwe craton divide into three associations: shelf, basinal, and transitional. Chert and iron formation exist within the three associations, allowing the unique opportunity to compare textural and geochemical attributes of iron- and silica-rich rocks from distinct positions within a single basin. We analyzed major-, trace-, and rare-earth-elements, and silicon and oxygen isotopes of chert and iron formation from the shelf, basinal, and transitional associations within the BGB. Samples possess elemental signatures consistent with formation from mixed seawater and hydrothermal fluids (LaSN/LaSN* > 1; PrSN/YbSN < 1; Y/Ho = 28.2–44.4; EuSN/EuSN* = 1.8–2.8). Oxygen isotope values broadly overlap for samples from different associations (ranging from δ18OV-SMOW = 9.41 to 15.10‰ (2σ = 0.8‰)), with isotopic variation linked to the presence of oxide inclusions in some samples. Silicon isotope values by contrast, vary widely. Iron formations possess 30Si-depleted silicon isotope signatures (δ30SiNBS-28 = −2.5 to −0.5‰ (2σ = 0.05‰)), in line with previous studies. Basinal and transitional cherts possess 30Si-depleted silicon isotope signatures (δ30SiNBS-28 = −1.1 to −0.5‰ (2σ = 0.05‰)) that overlap with values measured from iron formation samples. Within the transitional association, cherts are not 30Si-enriched compared to iron formation, while chert in the basinal association is 30Si-enriched compared to iron formation. Depleted silicon isotope signatures could result from fractionation during adsorption, re-equilibration, post-depositional processes, or possibly microbial reduction processes. Silicon isotope heterogeneity recorded in chert preserved in different assemblages across the BGB may in part reflect a seawater signature. The range of silicon isotope values preserved within broadly co-eval rocks from the same assemblage may indicate that samples preserve local isotopic signatures of a heterogeneous reservoir. Future work focused on identifying basin-position dependent trends may inform interpretations of the compiled Precambrian silicon isotope record.

Early continental crust generated by reworking of basalts variably silicified by seawater

The Archaean continental crust comprises two major groups of silicon-rich granitoids: the tonalite–trondhjemite–granodiorite and granite–monzonite–syenite suites, which differ in their sodium-to-potassium ratios. How these felsic granitoids evolved from their mafic precursors remains elusive and the subject of great debate. Here, we present silicon isotopic constraints on the formation of representative trondhjemitic and granitic plutons from the Kaapvaal craton that range in age from 3.51–2.69 billion years ago. We identified very consistent silicon isotopic signatures, all uniformly 0.1–0.2‰ heavier than rocks of the modern continental crust. This unusual composition is explained by the melting of a mafic source that included significant proportions (15–35 wt%) of silicified basalts, which were common supracrustal rocks before 3 billion years ago. Before the melting event that formed the granitoid magmas at depth, portions of the mafic source rocks were enriched in silica by interaction with silica-saturated seawater. The addition of silica depresses the stability of amphibole at similar water activity, allowing trondhjemitic and granitic melt production at lower temperatures from protoliths with contrasting silica contents: 52–57 and ≥60 wt%, respectively. This explains why granitoids were able to form very early in Earth history but did not emerge in significant amounts on other rocky planets. Granitic continental crust in the Archaean formed from a basaltic source that was enriched in silica due to interaction with the early oceans before melting, according to silicon isotope analyses on rocks from the Kaapvaal craton.

Past and current geochemical conditions influence silicon isotope signatures of pedogenic clay minerals at the soil profile scale, Ethiopia

Soil processes strongly govern silicon (Si) mobility in terrestrial environments and its relation to other global biogeochemical cycles. The nature of inherited soil clay minerals can be highly diverse given the variability of their weathering environment. The influence of Si isotope fractionation factor in the initial geochemical conditions of clay precipitation can therefore be still expressed in inherited clay minerals in their new environment. We studied top- and subsoil of an Ethiopian Vertic Planosol derived from parent materials of similar sources of volcanism. The selected Planosol has an abrupt textural change at a depth of ~40 cm separating a bleached, silty (25 ± 1.8% clay) ash-derived soil horizon with a crumby structure from a heavy clayey (68 ± 3.4% clay) lacustrine-derived vertic soil horizon. The mineralogical assemblage of the clay fraction in top- and subsoil is characterized by similar proportions of 1:1 (kaolinite) and 2:1 (illite and smectite) layer-type clay minerals. This specific soil profile provides a unique opportunity to elucidate the influence of clay formation processes (inheritance versus neoformation) on the Si isotope signature of pedogenic clay minerals. Minerals of the clay fraction in the clayey vertic horizon are significantly enriched in light Si isotope (δ30Si = −1.41 ± 0.02‰) compared to the bleached, silty horizon (δ30Si = −0.69 ± 0.03‰). These results are corroborated by the preferential enrichment in Ge, relative to Si, in the clay fraction of the clayey subsoil compared to the silty topsoil (Ge/Si = 6.3 ± 0.14 and 4.0 ± 0.10 μmol mol−1, respectively). Our results demonstrate that geochemical conditions in lacustrine environment favor kinetically-driven Si isotope fractionation factor leading to lower Si isotope ratios in 2:1 clay minerals inherited in the new soil profile environment. The inherited soil textural conditions in the soil profile also contribute to on-going processes that result in larger Si isotope differences between the soil solution (CaCl2 extractable) and clay minerals. This implies that Si isotope signatures of clay minerals in the studied soil profile are influenced by a combination of inheritance processes in lacustrine environment and on-going neoformation processes in the soil profile. This finding has important implications for environmental studies using geochemical and Si stable isotope tracers to better understand current soil processes, to model elemental cycling in soil-plant systems and to quantify land-ocean element mass-balances.

Origin of silica in rice plants and contribution of diatom Earth fertilization: insights from isotopic Si mass balance in a paddy field

Background and aims: The benefits of Si for crops is well evidenced but the biogeochemical cycle of Si in agriculture remains poorly documented. This study aims at identifying and quantifying the Si sources (primary and secondary soil minerals, amorphous silica, irrigation, Si-fertilizer) to rice plants. Method: Field experiments were carried out with and without application of diatomaceous earth (DE) under rice and bare conditions to determine the water and dissolved mass balance in paddy fields (Karnataka, Southern India). The fate of the Si brought by irrigation (DSi) (uptake by rice, uptake by diatoms, adsorption) was assessed through a solute mass balance combined with silicon isotopic signatures. Results: Above the ground-surface, about one third of the DSi flux brought by borewell irrigation (545 mmol Si.m−2) to bare plots and half of DSi in rice plots were removed from solution within minutes or hours following irrigation. Such rate is consistent with the rate of DSi adsorption onto Fe-oxyhydroxides but not with diatom blooms. In rice and rice + DE experiments, the isotopic fractionation factor (30e) between bore well and stagnant water compositions is close to −1 ‰, i.e. the isotopic fractionation factor known for rice, indicating that above-ground DSi removal would be dominated by plant uptake upon adsorption. Within the soil layer, pore water DSi decreases much faster in rice experiments than in bare ones, demonstrating the efficiency of DSi rice uptake upon adsorption. Total irrigation-DSi to plant-Si would then represent 24 to 36% in rice experiments (over 1460 ± 270 mmol Si m−2 in biomass) and 15 to 23% in rice + DE ones (over 2250 ± 180 mmol Si m−2). The δ30Si signature of whole plants was significantly different in the rice + DE plot analyzed, 0.99 ± 0.07 ‰, than in the rice one, 1.29 ± 0.07 ‰. According to these δ30Si signatures, the main Si source from the soil would be the amorphous silica pool (ASi). A slight contribution of DE to the rice plant could be detected from the Si isotopic signature of rice. Conclusions: The δ30Si signatures of the various soil-plant compartments, when associated to Si mass balance at soil-plant scale, constitute a reliable proxy of the Si sources in paddy fields. The solute Si balance is controlled by rice uptake in rice plots and by adsorption in bare ones. The main Si sources for the rice plants were soil ASi, irrigation Si and to a lesser extent Si fertilizer when it was applied.

Unveiling the Si cycle using isotopes in an iron-fertilized zone of the Southern Ocean: from mixed-layer supply to export

Abstract. A massive diatom bloom forms annually in the surface waters of the naturally iron-fertilized Kerguelen Plateau (Southern Ocean). In this study, silicon isotopic signatures (δ30Si) of silicic acid (DSi) and suspended biogenic silica (BSi) were investigated through the whole water column with unprecedented spatial resolution, during the KEOPS-2 experiment (spring 2011). We used δ30Si measurements to track the sources of silicon that fuelled the bloom, and investigated the seasonal evolution of the Si biogeochemical cycle in the iron-fertilized area. We compared the results from stations with various degrees of iron enrichment and bloom conditions to an HNLC reference station. Dissolved and particulate δ30Si signatures were highly variable in the upper 500 m, reflecting the effect of intense silicon utilization in spring, while they were quite homogeneous in deeper waters. The Si isotopic and mass balance identified a unique Winter Water (WW) Si source for the iron-fertilized area that originated from southeast of the Kerguelen Plateau and spread northward. When the WW reached a retroflection of the Polar Front (PF), the δ30Si composition of the silicic acid pool became progressively heavier. This would result from sequential diapycnal and isopycnal mixings between the initial WW and ML water masses, highlighting the strong circulation of surface waters that defined this zone. When comparing the results from the two KEOPS expeditions, the relationship between DSi depletion, BSi production, and their isotopic composition appears decoupled in the iron-fertilized area. This seasonal decoupling could help to explain the low apparent fractionation factor observed in the ML at the end of summer. Taking into account these considerations, we refined the seasonal net BSi production in the ML of the iron-fertilized area to 3.0 ± 0.3 mol Si m−2 yr−1, which was exclusively sustained by surface water phytoplankton populations. These insights confirm that the isotopic composition of dissolved and particulate silicon is a promising tool to improve our understanding of the Si biogeochemical cycle since the isotopic and mass balance allows resolution of processes in the Si cycle (i.e. uptake, dissolution, mixing).

Changes in diatom productivity and upwelling intensity off Peru since the Last Glacial Maximum: Response to basin-scale atmospheric and oceanic forcing

New records of stable silicon isotope signatures (δ30Si) together with concentrations of biogenic opal and organic carbon from the central (9° S) and northern (5° S) Peruvian margin reveal changes in diatom productivity and nutrient utilization during the past 20,000 years. The findings are based on a new approach using the difference between the δ30Si signatures of small (11-32μm) and large (>150μm) diatom fractions (Δ30Sicoscino-bSi) in combination with the variance in diatom assemblages for reconstruction of past upwelling intensity. Combination of our records with two previously published records from the southern upwelling area off Peru (12-15° S) shows a general decoupling of the environmental conditions at the central and southern shelf mainly caused by a northward shift of the main upwelling cell from its modern position (12-15° S) towards 9° S during Termination 1. At this time only moderate upwelling intensity and productivity levels prevailed between 9° S and 12° S interpreted by a more northerly position of Southern Westerly Winds and the South Pacific Subtropical High. Furthermore, a marked decrease in productivity at 12-15° S during Heinrich Stadial 1 coincided with enhanced biogenic opal production in the Eastern Equatorial Pacific, which was induced by a southward shift of the Intertropical Convergence zone and enhanced northeasterly trade winds. Modern conditions were only established at the onset of the Holocene. Past changes in preformed δ30Si signatures of subsurface waters reaching the Peruvian Upwelling System did not significantly affect the preserved δ30Si signatures.

Stable silicon isotope signatures of marine pore waters – Biogenic opal dissolution versus authigenic clay mineral formation

Dissolved silicon isotope compositions have been analysed for the first time in pore waters (δ30SiPW) of three short sediment cores from the Peruvian margin upwelling region with distinctly different biogenic opal content in order to investigate silicon isotope fractionation behaviour during early diagenetic turnover of biogenic opal in marine sediments. The δ30SiPW varies between +1.1‰ and +1.9‰ with the highest values occurring in the uppermost part close to the sediment–water interface. These values are of the same order or higher than the δ30Si of the biogenic opal extracted from the same sediments (+0.3‰ to +1.2‰) and of the overlying bottom waters (+1.1‰ to +1.5‰). Together with dissolved silicic acid concentrations well below biogenic opal saturation, our collective observations are consistent with the formation of authigenic alumino-silicates from the dissolving biogenic opal. Using a numerical transport-reaction model we find that approximately 24% of the dissolving biogenic opal is re-precipitated in the sediments in the form of these authigenic phases at a relatively low precipitation rate of 56μmolSicm−2yr−1. The fractionation factor between the precipitates and the pore waters is estimated at −2.0‰. Dissolved and solid cation concentrations further indicate that off Peru, where biogenic opal concentrations in the sediments are high, the availability of reactive terrigenous material is the limiting factor for the formation of authigenic alumino-silicate phases.

Effect of diagenetic phase transformation on the silicon isotope composition of opaline sinter deposits of Geysir, Iceland

Detectible δ30Si variations in present-day chemical silica deposits have stimulated the application of silicon isotopes to infer environmental conditions from ancient equivalents. Interpretations of δ30Si signatures remain problematic in view of potential post-depositional changes, of which magnitudes and underlying mechanisms are largely unknown. A critical issue in the interpretation of isotope data from cherts concerns the extent to which early-diagenetic processes modify original δ30Si signatures. Here, we report δ30Si variations in opal-A, opal-A/CT and opal-CT from fossil sinter deposits in an active discharge apron in the Geysir geothermal field, Iceland. Opal-A samples show an average δ30Si of −0.7±0.2‰, while opal-CT samples are isotopically lighter with an average δ30Si of −2.0±0.4‰, implying a sizable shift of approximately 1.3‰ between the different phases. This shift can be explained by repetitive dissolution/re-precipitation processes, diffusive transport or temperature differences during phase transitions. On average, the fossil opal-A tends to be less negative in δ30Si than amorphous silica that recently precipitated from the hydrothermal water. The difference can be attributed to primary variability in isotopic fractionation that accompanies precipitation out of spring water at the surface, or to a post-depositional release of surface 28Si at the onset of diagenetic formation. Our results corroborate the perception that original silicon isotope signatures of silica, acquired during chemical precipitation from a saturated solution, may not be preserved in the geological record, and that post-depositional changes must be taken into account when interpreting data from ancient chert deposits.

The riverine silicon isotope composition of the Amazon Basin

We present here the first large-scale study of riverine silicon isotope signatures in the Amazon Basin. The Amazon and five of its main tributaries were studied at different seasons of the annual hydrological cycle. The δ30Si signature of the dissolved silicon (DSi) exported to the estuary (weighted for DSi flux) for the period considered is estimated at +0.92‰. A river cross-section shows the homogeneity of the Amazon River regarding DSi concentration and isotope ratio. The biogenic silica (BSi) concentration measured in surface water from all rivers is generally small compared to the DSi reservoir but large variations exist between rivers. Very low isotope signatures were measured in the upper Rio Negro (δ30Si=+0.05±0.06‰), which we explain both by an equilibrium between clay formation and dissolution and by gibbsite formation. The Si isotope fractionation in the Andean tributaries and the Amazon main stem can be explained by clay formation and follow either a Rayleigh or a batch equilibrium fractionation model. Our results also suggest that the formation of 2:1 clays induces a fractionation factor similar to that of kaolinite formation.

Silicon pool dynamics and biogenic silica export in the Southern Ocean inferred from Si-isotopes

Abstract. Silicon isotopic signatures (δ30Si) of water column silicic acid (Si(OH)4) were measured in the Southern Ocean, along a meridional transect from South Africa (Subtropical Zone) down to 57° S (northern Weddell Gyre). This provides the first reported data of a summer transect across the whole Antarctic Circumpolar Current (ACC). δ30Si variations are large in the upper 1000 m, reflecting the effect of the silica pump superimposed upon meridional water transfer across the ACC: the transport of Antarctic surface waters northward by a net Ekman drift and their convergence and mixing with warmer upper-ocean Si-depleted waters to the north. Using Si isotopic signatures, we determine different mixing interfaces: the Antarctic Surface Water (AASW), the Antarctic Intermediate Water (AAIW), and thermoclines in the low latitude areas. The residual silicic acid concentrations of end-members control the δ30Si alteration of the mixing products and with the exception of AASW, all mixing interfaces have a highly Si-depleted mixed layer end-member. These processes deplete the silicic acid AASW concentration northward, across the different interfaces, without significantly changing the AASW δ30Si composition. By comparing our new results with a previous study in the Australian sector we show that during the circumpolar transport of the ACC eastward, the δ30Si composition of the silicic acid pools is getting slightly, but significantly lighter from the Atlantic to the Australian sectors. This results either from the dissolution of biogenic silica in the deeper layers and/or from an isopycnal mixing with the deep water masses in the different oceanic basins: North Atlantic Deep Water in the Atlantic, and Indian Ocean deep water in the Indo-Australian sector. This isotopic trend is further transmitted to the subsurface waters, representing mixing interfaces between the surface and deeper layers. Through the use of δ30Si constraints, net biogenic silica production (representative of annual export), at the Greenwich Meridian is estimated to be 5.2 &amp;amp;pm; 1.3 and 1.1 &amp;amp;pm; 0.3 mol Si m−2 for the Antarctic Zone and Polar Front Zone, respectively. This is in good agreement with previous estimations. Furthermore, summertime Si-supply into the mixed layer of both zones, via vertical mixing, is estimated to be 1.6 &amp;amp;pm; 0.4 and 0.1 &amp;amp;pm; 0.5 mol Si m−2, respectively.

Silicon isotopes in allophane as a proxy for mineral formation in volcanic soils

Weathering of basaltic ash in volcanic areas produces andosols, rich in allophane and ferrihydrite. Since the rate of mineral formation is very useful in climate and geochemical modelling, this study investigates Si isotope compositions of allophane as a proxy for mineral formation. Allophane formed in contrasting conditions in five Icelandic soil profiles displays silicon isotope signatures lighter than the basalt in less weathered soils (−0.64 ± 0.15‰), and heavier in more weathered organic-rich soils (+0.23 ± 0.10‰). The fate of the dissolved Si in those volcanic soils strongly depends on Al availability. In organic-rich soils, most of Al is humus-complexed, and the results support that Si precipitates as opaline silica by super-saturation, leaving an isotopically heavier dissolved Si pool to form allophane with uncomplexed Al. This study highlights that Si isotopes can be useful to record successive soil processes involved in mineral formation, which is potentially useful in environmental paleo-reconstruction.

Contrasting silicon isotope signatures in rivers from the Congo Basin and the specific behaviour of organic‐rich waters

We investigate the dissolved δ30Si of the Congo River, the world's second largest riverine source of Si to the ocean. Small tributaries rich in dissolved organic carbon running through wetlands (“Black Rivers”) exhibit the lowest δ30Si ever measured in running surface waters (+0.02 ± 0.15‰), whilst the main branch and largest tributaries have higher values (+0.98 ± 0.13‰), well within the average of what has been measured so far. Our data suggest that the contribution of Black Rivers to the total discharge of the basin is 22 ± 10% and that δ30Si is mostly controlled by weathering intensity rather than fluxes. We propose both a mass and Si‐isotopic balance model, which suggest that the distribution of Si in the particulate and/or dissolved components in Congo rivers results mainly from mixing between two types of weathering regimes: one where clays are formed and remain stable, and one where they are dissolved under the action of organic matter.

Silicon isotopes in spring Southern Ocean diatoms: Large zonal changes despite homogeneity among size fractions

We determine Southern Ocean diatom silicon isotopic signatures and compare them with the previously published data for dissolved silicic acid from the same locations. Five stations distributed along the WOCE SR-3 transect (Australian Sector of the Southern Ocean) in different biogeochemical provinces are presented: Polar Front and Inter-Polar Front Zones (PFZ–IPFZ), Southern Antarctic Zone (AZ-S), Seasonal Ice Zone (SIZ). Total (> 0.4 μm), medium-sized (20–70 μm), and large diatoms (> 70 μm) were sampled at 2–4 depths in the upper 150 m. Silicon isotopic compositions of biogenic silica (diatoms) and seawater were then measured by MC-ICP-MS, in dry plasma mode using external Mg doping. Results are expressed as δ 29Si relative to the NBS28 standard. The isotopic composition of diatoms ( δ 29Si BSi) is generally homogeneous in the mixed layer and does not exhibit a systematic isotopic fractionation linked to a size effect. δ 29Si BSi are always lighter than the ambient dissolved silicic acid signatures ( δ 29Si DSi), reflecting the preferential uptake of light isotopes by diatoms. A trend of lighter isotopic signatures southward is observed both in diatoms and seawater samples but the δ 29Si BSi latitudinal gradient is much steeper. A diatom signature as low as − 0.26‰ in the southernmost SIZ station strongly contrasts with the + 0.65‰ signature measured on PFZ diatoms. The difference between the ambient dissolved silicic acid and diatom isotopic signatures, Δ 29Si, strongly increases southward: from 0.4 in the PFZ up to 1.08‰ in the SIZ. This points toward occurrence of mixing events in the PFZ–IPFZ with diatoms not being under equilibrium with their surrounding water and/or, possible variation of the diatom–seawater equilibrium fractionation factor, 29 ε. Apart from mixing, we found that the other parameters likely responsible of such variation are temperature, dissolved Si contents and, Si specific uptake and dissolution rates although at this stage none of these could be clearly recognized as the leading cause. Thorough examination of these parameters through in vitro experiments reflecting the extreme Southern Ocean conditions is needed to determine whether the observed latitudinal variation of Δ 29Si reflects real variable fractionation or results from non-equilibrium or different time-scales recorded between dissolved and biogenic Si isotopic signatures. Our results also call for the development of more realistic models for describing short-term isotopic composition changes due to e.g. Si consumption, export and resupply via mixing. Finally, by comparing δ 29Si BSi within and below the mixed layer, we could identify a two-step history of the PFZ–IPFZ bloom in contrast to the recently started diatom bloom in the SIZ.