INDIVIDUAL FORAMINIFERAL ANALYSES: A REVIEW OF CURRENT AND EMERGING GEOCHEMICAL TECHNIQUES

Geochemical analysis of foraminiferal calcite has played a foundational role in paleoceanographic reconstructions, shaping our understanding of past climate conditions and ancient marine environments. Foraminiferal assemblages (e.g., transfer functions) and an understanding of foraminiferal paleoecology have long been used to inform our understanding of climate changes in the past (Cushman, 1932; Murray, 1897; Phleger, 1939; Vaughan, 1940). The use of foraminiferal geochemistry for paleoclimate reconstructions began shortly after the pioneering work of Harold Urey (1947), who postulated a correlation between the stable oxygen isotope (δ¹⁸O) composition of calcite and the water temperature under which it precipitated. This insight laid the groundwork for the subsequent development of foraminiferal calcite as an invaluable paleoclimate proxy. Seminal contributions firmly established the δ¹⁸O of foraminiferal calcite as a reliable indicator of calcification temperature, salinity, and ice volume (Emiliani, 1954; Emiliani, 1955; Emiliani and Epstein, 1953; Shackleton, 1967; Shackleton and Vincent, 1978). While stable isotope analyses were instrumental in the early stages of paleoceanographic reconstructions, it was the rapid exploration of foraminiferal trace element compositions that later garnered significant attention (Bender et al., 1975; Boyle and Keigwin, 1982; Chave, 1954; Delaney et al., 1985; Graham et al., 1982; Russell et al., 1996). These early studies revealed that a wealth of information about past ocean conditions, including temperature, nutrient concentrations, and the trace element chemistry of ancient water masses, can be gleaned from the trace element composition of foraminifera tests.

In the early days, foraminiferal geochemical analyses, whether for isotopes or trace elements, required substantial numbers of tests. The first stable isotope analyses required 5 mg of calcite (i.e., several hundred tests (Emiliani, 1954; Emiliani and Epstein, 1953)) whereas early trace element analyses required up to 100 tests for a single ‘bulk’ solution analysis (Bender et al., 1975; Boyle, 1981). In the decades that followed, advancements in analytical techniques enabled the analysis of much smaller samples (Martin et al., 2002; Ravelo and Hillaire-Marcel, 2007; Rosenthal et al., 1999; Yu et al., 2005). Today, it is possible to obtain stable isotope data on calcite samples weighing as little as 5 µg (e.g., Vonhof et al., 2020), allowing for individual foraminifera analyses (IFA) or in some cases the analysis of individual chambers (Meyer, 2023). In the case of trace elements, researchers are now able to analyze the trace element composition of not only individual specimens through solution-based methods (e.g., Marchitto, 2006), but also at sub-micron scale resolution, and gain insights into intratest trace element variability through various in situ techniques discussed herein, such as electron microprobe mapping (EMPA) and laser ablation inductively coupled plasma mass spectrometry (LA-ICP-MS), among others (Fig. 1).

The emergence of geochemical techniques allowing for reliable IFA offers several compelling advantages over traditional bulk analyses. Planktic foraminifera are relatively short lived (∼1 month) and thus interpretation of their individual test geochemistry provides a snapshot into paleoenvironmental conditions. The geochemistry of benthic foraminifera, which often have much longer life-spans (up to a year (Alve, 1999)), is similarly shaped by the environmental conditions and ecological niches of their preferred habitats either on or in the sediment (Rae et al., 2011). IFA provides the potential for higher temporally resolved investigations compared to traditional bulk sediment analysis (e.g., Glaubke et al., 2021). IFA also permits an assessment of population variance or variability (Figs. 1, 2), a metric that can be used to assess environmental change on shorter time scales, for example, annually or seasonally (e.g., Groeneveld et al., 2019; Rongstad et al., 2020). IFA techniques that can measure geochemical signatures in discrete domains or individual chambers of a given test allow for the investigation of vital effects and change in habitat conditions over a foraminifer’s lifetime (Figs. 1B–E). The tests sink post-mortem to the seafloor becoming part of the geologic record. Test chemistry can be compromised by scavenging processes associated with travel through the water column as well as diagenetic processes occurring on the seafloor (e.g., Poirier et al., 2021). Environmental signals interpreted from populations of individual test measurements of both benthic and planktic species can be further complicated by secondary processes, such as sediment mixing. While conducting and interpreting IFA can be difficult, the potential information gleaned from such detailed investigations can be invaluable to our understanding of past climates, ocean conditions, and biomineralization processes.

The investigation of micron-scale geochemistry within individual tests allows researchers to pursue new questions in paleoceanography, paleoecology, and biomineralization (Geerken et al., 2019; John et al., 2023; van Dijk et al., 2019). Moreover, paired isotope and trace element analyses can now be performed on the same individual test (e.g., Fritz-Endres et al., 2019; Kearns et al., 2023; Vetter et al., 2017; Wit et al., 2010), providing powerful new insight into environmental and biological controls on foraminiferal test composition. For example, trace elements can be analyzed via LA-ICP-MS followed by stable carbon (δ¹³C) and oxygen (δ¹⁸O) isotope analyses or nano-secondary ion mass spectrometry (NanoSIMS) imaging. These analytical approaches can also be paired with genetic analyses to reveal information about trace element and isotopic variability that may be driven by a foraminifers genotype or microbial associations (Ujiie et al., 2019; See also Lane et al., this volume). These advancements have enabled the acquisition of a wealth of data from single foraminifera, making them a vital, but hitherto underutilized tool in unraveling the complexities of Earth’s ocean and climate history.

This review aims to provide a comprehensive assessment of the available analytical tools and merits of IFA compared to traditional bulk analysis of pooled foraminiferal tests. Through a thorough examination of various analytical techniques, their strengths, weaknesses, applicability, and considerations, we will delve into the concept of choosing the right tool for the right question, that is, what information is needed and what tool, or tools, will provide this information. Further, this review seeks to offer insights into the advantages and limitations of IFA and provide guidance on when and how to use IFA for more robust and nuanced scientific investigations.

The tools and techniques are divided into two sections focused on: 1) stable oxygen and carbon isotope geochemistry and 2) trace element geochemistry. Each subsection introduces the history of the geochemical technique, gives general sample preparation details, and considerations and limitations. Note that in the interests of brevity, we do not provide details for all stable isotopes (e.g., δ¹¹B) and do not explore techniques that are applied to assess individual trace element incorporation systematics (e.g., x-ray absorption near-edge spectroscopy or “XANES”).

The first stable oxygen and carbon isotope measurements of foraminiferal tests were performed using dual-inlet gas-source isotope ratio mass spectrometry (GS-IRMS; Urey et al., 1951; (Emiliani and Epstein, 1953; Urey et al., 1951). In brief, this technique involves liberating CO₂ gas from foraminiferal carbonate under vacuum using phosphoric acid, and the drying and subsequent introduction of pure sample-derived CO₂ into the mass spectrometer to obtain isotopic ratios of carbon and oxygen. Seventy years on, GS-IRMS—including both traditional dual-inlet and continuous flow modalities—continues to play an important role in foraminiferal geochemistry and its applications. Over the years, advances in the efficiency of gas inlet systems and the sensitivity of ion sources now allow for modern GS-IRMS systems to routinely measure ∼10–25 µg of carbonate with high precision (<∼0.1‰, 1σ), allowing for analysis of individual foraminifera.

Killingley et al. (1981) first reported stable isotopic IFA-δ¹⁸O and -δ¹³C measurements in single tests of large-sized (>840 µm) Orbulina universa, Globigerinoides conglobatus, and Globorotalia tumida from Ontong-Java Plateau sediments. They posited that the large stable isotopic variability observed between individual tests of foraminifera found within the same stratigraphic sample must arise due to either bioturbation, metabolic/vital effects, depth-habitat variability, or fractionation driven by environmental (i.e., thermal) and oceanic variability across seasonal to longer timescales. More recent work has uncovered how other post-depositional processes can influence single-test stable isotope distributions, such as transport via turbidity currents (Fritz-Endres et al., 2019) and selective diagenetic alteration related to seafloor methane seepage (Clemens et al., 2023). Nevertheless, the pioneering measurements by Killingley et al. (1981) and their inferences spurred studies that statistically scrutinized IFA geochemical population variance and debated their implications for paleoceanographic reconstructions (Boyle, 1984; Schiffelbein, 1986; Schiffelbein and Hills, 1984). These studies, in turn, paved the way for proxy system models and statistical inferences employed in more recent studies (Bienzobas Montávez et al., 2024; Glaubke et al., 2021; Khider et al., 2011; Leduc et al., 2009; Lougheed and Metcalfe, 2022; Thirumalai and Clemens, 2020; Thirumalai et al., 2013; Tindall et al., 2017; White et al., 2018) of utilizing the ‘time slice’ approach (Thirumalai and Maupin, 2023), where many specimens are analyzed from a single sample or samples that span a short duration of time, to discern variability between two different time intervals (Fig. 2).

Studies focusing on the stable isotopic variability of individual tests of foraminifera have been used to address a plethora of paleoceanographic problems; examples include delineating the impact of sediment reworking on paleoceanographic and paleoecologic inferences (Billups and Spero, 1996; Hupp et al., 2022; Stott and Tang, 1996); reassessing foraminiferal ontogenetic and (paleo)ecological trends (Billups and Spero, 1995; Houston et al., 1999; Spero and Lea, 1993); and reconstructing terrestrial paleohydrology (Tang & Stott, 1993), including so-called “meltwater spikes” (Killingley et al., 1981; Spero and Williams, 1990). More recent studies have focused on settings where upper-ocean temperature variability is the main driver of IFA-δ¹⁸O variability to capture climatic processes with a frequency much higher than the sedimentation rate (e.g., years and decades rather than 1000s of years), such as ENSO (Khider et al., 2011; Leduc et al., 2009), the Indian Ocean Dipole (IOD; (Thirumalai et al., 2019), and seasonal upwelling (Naidu et al., 2019). Future approaches could combine stable isotope IFA with amino acid racemization (AAR) of specimens from the same interval. AAR is a relatively new tool for estimating the ages of foraminifera that are too old for radiocarbon dating and for detecting sediment/fossil reworking (see Murray-Wallace et al., this issue).

Common statistical approaches to scrutinize and compare stable isotopic (and other geochemical) IFA measurements include visualizing frequency distributions via histograms, contrasting box and whisker (and swarm) plots, and quantile-quantile (i.e., “Q-Q”) plots (Fig. 2). Typically, IFA measurements grouped from within a stratigraphic horizon of a marine sediment core are compared to those from another sedimentary sample (oftentimes, the “core-top” or the most modern portion of the marine sediment core is chosen). One of the advantages of utilizing IFA to contrast seasonal-to-interannual variability across different mean climatic states is that average (or median) IFA values may be removed from each dataset to focus on differences in variance (Schmitt et al., 2019; Thirumalai et al., 2019). Accordingly, box plots of anomalies (with median values removed) have been used to visualize differences in variability, where box size can be indicative of the interquartile range of sample data—a metric representing the middle 50% of the samples (Fig. 2B). Q-Q plots (Fig. 2C) visualize contrasts between the tails of the two distributions and have been applied to discerning changes in seasonality and interannual modes of variability such as the El Niño/Southern Oscillation across different periods (Ford et al., 2015; Glaubke et al., 2024; Thirumalai et al., 2019).

Similar to the limitations of multi-test foraminiferal records of stable isotopes conducted without paired TE measurements, IFA distributions of stable isotopic measurements can be challenging to interpret. For example, multivariate in situ and post-depositional controls of foraminiferal IFA-δ¹⁸O and -δ¹³C can have confounding effects on the resulting distributions (e.g., Clemens et al., 2023; Khider et al., 2011; Leduc et al., 2009). Fortunately, via several techniques detailed in this paper, it is now possible to estimate seasonal-scale variability in the oxygen isotope composition of seawater (δ¹⁸O_sw) by utilizing paired Mg/Ca-δ¹⁸O measurements on individual foraminiferal tests (Vetter et al., 2017; Wit et al., 2010). Such applications are still in their preliminary stages; future work could focus on rigorous methodological calibrations to extract high-fidelity reconstructions of temperature and δ¹⁸O_sw distributions from paired IFA data. The δ¹³C of individual benthic foraminifera are commonly assessed in paleoceanographic reconstructions. In contrast, the planktic IFA-δ¹³C data are underutilized in paleoceanographic reconstructions due to species-specific offsets from the δ¹³C of dissolved inorganic carbon in seawater, making δ¹³C data (whether bulk or IFA) difficult to interpret. We suggest that the coupling of IFA-δ¹³C and biogeochemically-sensitive TE data has the potential to open up new avenues into understanding foraminiferal paleoecology and its implications for paleoceanography in various settings, and increase the utility of IFA-δ¹³C data (e.g., Groeneveld et al., 2019; Jonkers et al., 2022). As researchers continue to reduce sample size requirements (currently a minimum ∼5 μg of calcite is required for GS-IRMS), new avenues of research could include single chamber analyses to explore stable isotope variability driven by ontogeny as well as single chambers grown in controlled conditions.

Secondary ion mass spectrometry (SIMS) is a relatively recent addition to the toolkit of analytical techniques that can be used for individual foraminifera analyses (see Kozdon, this issue, for a more detailed review). SIMS analysis allows for measurement of stable isotope composition on micron-scale domains of a foraminiferal test prepared in cross section. Currently, stable carbon isotopes are measured on spots or ‘pits’ as small as ∼9 to 10 μm, whereas stable oxygen isotopes can be measured on spots ranging from ∼3 to 10 μm, with a compromise on decreasing analytical precision at progressively smaller spot sizes. During SIMS analysis, a primary ion beam (i.e., Cs⁺ for δ¹⁸O and δ¹³C) is focused upon the cross-sectional surface of a sample material. Secondary ions are sputtered from the sample surface and accelerated into the mass spectrometer where the secondary ion beam is doubly-focused before being measured by a collection of ion detectors (Valley et al., 2009). Unlike GS-IRMS where both carbon and oxygen isotopes are measured from the CO₂ gas produced by the reaction of phosphoric acid and carbonate, only a single isotope system (e.g., δ¹³C or δ¹⁸O) is measured with each SIMS analysis. With respect to foraminifera, SIMS has been used to measure stable carbon, oxygen, sulfur (δ³⁴S; Borrelli et al., 2020), lithium (δ⁷Li; Vigier et al., 2015), and boron (δ¹¹B; Rollion-Bard and Erez, 2010) isotopes in modern and ancient foraminifera.

Whereas GS-IRMS has been the more traditional method of measuring foraminiferal stable isotope compositions, SIMS provides a unique approach to addressing specific questions in paleoceanography and paleoecology. For example, investigations into diagenetic influences on stable isotope compositions have been largely focused upon the δ¹⁸O of foraminiferal tests, with few studies (e.g., Vergnaud-Grazzini, 1979) investigating potential impacts on δ¹³C composition. Use of SIMS to measure the in situ δ¹³C compositions of foraminifera associated with time periods marked by carbon isotope excursions (CIE; e.g., the Paleocene Eocene Thermal Maximum or “PETM”) has shown that the δ¹³C of fossil foraminifera is not immune to diagenetic influences. By analyzing micron-scale domains throughout the test wall of fossil foraminifera, researchers have been able to identify variability within the test wall indicative of diagenetic overprinting caused by rapid carbonate chemistry fluctuations associated with short-lived hyperthermal events like the PETM (Hupp et al., 2023; Kozdon et al., 2018). Exclusion of SIMS analyses from the diagenetic portions of the test wall has resulted in identification of much larger CIE magnitudes when compared to GS-IRMS records where both primary biogenic and diagenetic calcites contribute to the isotopic signature of the CO₂ gas that is analyzed via dissolution of whole tests (Hupp et al., 2023; Kozdon et al., 2018; Zhang et al., 2020). In addition to using SIMS analysis of foraminifera to reconstruct CIE magnitudes during ancient hyperthermal events, researchers have used SIMS δ¹³C analysis to identify diagenetic overprinting on foraminiferal tests associated with proximity to methane seeps (Panieri et al., 2017) and to investigate carbon isotope incorporation into cultured foraminifera (Vetter et al., 2014).

A similar approach has been taken to characterize diagenetic overprinting and vital effects that influence the oxygen isotope composition of foraminiferal tests (Rollion-Bard et al., 2008; Kozdon et al., 2009a; Kozdon et al., 2013; Wycech et al., 2018a; Livsey et al., 2020). For example, several researchers have used SIMS δ¹⁸O analysis to constrain differences in oxygen isotope composition between lamellar and gametogenic crust calcites in modern species of planktic foraminifera, including Neogloboquadrina pachyderma (Kozdon et al., 2009b; Livsey et al., 2020) and Trilobatus sacculifer (Wycech et al., 2018a), providing critical insights into a vital effect that should be considered when evaluating paleorecords.

Perhaps one of the most powerful applications of SIMS is the potential for SIMS analysis to be paired with other in situ techniques. Combined trace element geochemistry, measured by techniques such as EMPA or LA-ICP-MS, with stable isotope compositions measured via SIMS have provided important insights into paleoceanographic changes throughout the Cenozoic (Kozdon et al., 2011) as well as foraminiferal biomineralization processes (e.g., Vetter et al., 2013, 2014; Livsey et al., 2020). For example, several researchers have paired in situ measurements of Mg/Ca with SIMS-derived δ¹⁸O data in planktic foraminifera to investigate hydrologic changes in the Paleogene (Kozdon et al., 2013, 2020) and Pliocene (Wycech et al., 2022).

Lastly, much work has been done to constrain the similarities and differences in data derived from GS-IRMS versus SIMS. Wycech et al. (2018b) and Balestra et al. (2020) have found a consistent offset whereby foraminifera analyzed by SIMS yielded δ¹⁸O values that were on average ∼0.9‰ lower than the same foraminiferal calcite being measured by GS-IRMS. Recent work by Wycech et al. (2022) also found an inter-instrument δ¹³C offset of 0.6-0.8‰ in untreated and cleaned core top O. universa tests, but no detectable offset in cultured or vacuum roasted core top tests. The cause of these observed offsets is unresolved and additional work is needed to constrain potential inter-instrumental differences.

SIMS analysis of foraminifera requires that the tests and grains of a standard reference material (e.g., UWC-3 calcite grains) be placed in a mounting medium, typically epoxy, and ground to cross-section (Figs. 3A–B). Epoxy mounts are subsequently cleaned and gold coated prior to analysis. Analytical time for a single spot analysis typically ranges from ∼5 to 8 minutes depending on the instrument settings. Ideally, individual foraminifera will be oriented in the same position (e.g., in umbilical view) when creating the sample mount to allow for consistency in domains available to measure during SIMS analysis. Foraminifera have a wide variety of test structures which present unique challenges for sample preparation, so it is highly recommended that a user contact the SIMS laboratory for instructions prior to scheduling analysis.

Diligent quality control is required after analysis to characterize the analyzed domain geochemically and visually (Fig. 3C). Quality control metrics should be implemented for both the data and pit appearance (sensuWycech et al., 2022). Such metrics for the data may include acceptable ranges for ¹³C¹H/¹³C (or ¹⁸O¹H/¹⁸O) ratios or ion yields for the sample versus the standard to assess potential contamination from organic matter or epoxy. Meanwhile, each SIMS pit should be imaged in high magnification under a scanning electron microscope to ensure the pit did not cross-cut cracks or sputter into voids or test-adjacent epoxy (e.g., Fig. 3C). After quality control, the data should be processed with consideration to the overall project question. For example, the best option in paleoceanographic studies may be to average SIMS measurements from each test and then average the test values for each sample to create a time series of SIMS-based measurements (e.g., Wycech et al., 2022; Hupp et al., 2023). In other cases, the research question may benefit from no averaging of SIMS measurements to understand intratest variability (Vetter et al., 2013a; Vetter et al.; Wycech et al., 2022).

Special considerations are required in every step of SIMS analysis of foraminifera. For example, during sample preparation it is important to understand how the targeted domain appears in cross-section and if it is sufficiently large enough to be separately measured from other domains via a 3 to 10-μm SIMS spot. If foraminifera from multiple samples are being analyzed, preservation of the targeted domain should be reviewed for each sample, ideally via scanning electron microscope (SEM), as post-depositional processes such as dissolution or diagenetic calcite addition may destroy or obfuscate the targeted domain in certain intervals (Hupp et al., 2023; Wycech et al., 2018a). SIMS analyses inherently have poorer precision than GS-IRMS (<0.2‰) such that the in situ technique is only usefully applied to questions where the expected isotope variability exceeds the instrumental precision, which varies based on the laboratory, pit size, and instrument conditions. For example, reported δ¹⁸O measurements of foraminifera using 10-μm pit sizes have recorded average analytical precision as low as 0.3‰ (Kozdon et al., 2020; Wycech et al., 2018b), however the average analytical precision increases to 0.7‰ when using 3-μm pit sizes (e.g., Kozdon et al., 2020). Furthermore, the average analytical precision reported for δ¹³C of foraminifera has been 0.7‰, which may be too large to confidently reconstruct isotopic δ¹³C changes if the expected magnitude is relatively low (e.g., ∼1‰).

Beyond analytical caveats, the research question must consider which signal is being captured by the targeted domain. For example, the chambers of foraminifera can have distinct isotopic values due to ontogenetic factors, such as water depth migration and juvenile-stage disequilibrium (Bemis et al., 1998; Birch et al., 2013; Oppo and Fairbanks, 1989; Takagi et al., 2016). Other processes, such as gametogenesis and diagenesis, add distinct calcite with a more positive δ¹⁸O onto the tests (Wycech et al., 2018a, b). For this reason, SIMS analyses should strategically target only the domain of interest, whether that be only one chamber (e.g., the ultimate or penultimate chamber) of multiple tests, the interior of the test wall in species that add thick gametogenic calcite, or domains that are known to be less affected by diagenetic calcite. In cases where multiple tests are analyzed to reconstruct isotopic time series, one must consider how many analyses per test and how many tests per sample are required to be representative of the population. Foraminiferal populations that experience a large degree of seasonality, for example, may not be well-suited for SIMS analysis, as the number of analyses (i.e., the number of specimens per species per sample) required may exceed what can be accomplished under funding or time constraints.

Although SIMS and GS-IRMS measurements may be completed on tests from the same samples, comparison between the values should consider the fact that inter-instrument offsets have been observed for both δ¹⁸O and δ¹³C in foraminifera (Wycech et al., 2018a, 2022; Balestra et al., 2020). Acknowledgement of such offsets highlights the need for a fine-tuned research question when analyzing foraminifera by SIMS. For example, current δ¹⁸O-temperature calibrations are based on GS-IRMS measurements such that assumptions of comparability are required when applying SIMS δ¹⁸O measurements to these relationships. A biogenic carbonate standard has not yet been developed for SIMS measurements and thus studies rely on in-house calcite (e.g., marble) standards (e.g., Wycech et al., 2018a; Balestra et al., 2020). Due to this limitation and the unique signals captured by SIMS, the technique is most suited to address research questions that draw on the fine-scale spatial resolution that SIMS affords. Areas of growth for SIMS methodology should therefore focus on development of a biogenic carbonate SIMS standard and further investigation into the cause and consistency of observed offsets between SIMS and GS-IRMS methods. Future directions for SIMS studies could include investigations focused on better understanding intratest isotopic variability due to ecological, ontogenetic, or biomineralization processes. For example, from a paleoceanographic perspective, SIMS allows for distinction of primary and diagenetic isotopic signatures which could provide greater insight into the magnitude of isotopic change associated with abrupt climate change events like Eocene hyperthermals and other events marked by carbon or oxygen isotope excursions. Lastly, development of in situ IFA techniques in general widens the possibilities for paired analysis of stable isotope and trace element geochemistry in micron-scale domains, furthering our ability to reconstruct past ocean-climate change at ever finer resolution.

There are a multitude of techniques that can be used to constrain trace element (TE) geochemistry of foraminiferal calcite. Here, we focus on a subset of techniques that can yield individual test average TE/Ca data and intratest TE/Ca image ‘maps’ of the 2D distribution of trace elements within a test wall. We begin by discussing the techniques that are commonly applied for investigating individual test and individual chamber TE/Ca data (i.e., LA-ICP-MS) and then discuss techniques that require embedding the tests in epoxy resin and analyzing the tests in cross-section, including electron microprobe analysis (EMPA, though sometimes also referred to as electron probe microanalysis or EPMA), nano-secondary ion mass spectrometry (NanoSIMS), and Time-of-Flight LA-ICP-MS (TOF-LA-ICP-MS).

Laser ablation inductively coupled plasma mass spectrometry (LA-ICP-MS) utilizes a high energy, rapidly pulsed laser (nanosecond to femtosecond pulse duration ‘fired’ several times per second) to ideally ablate a very thin layer from the surface of the sample for transport in a gas stream to an ICP-MS instrument. First developed in the mid-1980s (Gray, 1985), LA-ICP-MS has since found widespread application in Earth, environmental, and biomedical sciences, chiefly because the technique offers a balance between spatial resolution, analytical rapidity, detection limits, and cost that is unparalleled by most, if not all, alternatives (Sylvester and Jackson, 2016). The technique can quantify trace elements at <10% level accuracy and precision at ppb-level and even ppt-level concentrations. In the case of line scans, LA-ICP-MS can resolve chemical heterogeneity on a lateral scale of a few to 10s of microns and for spot analyses, can yield highly spatially resolved (sub-micron) data, depending on the analytical methods (e.g., Evans and Müller, 2018; Heinrich et al., 2003; Müller and Fietzke, 2016; Sinclair et al., 1998; Vander Putten et al., 1999). Applied to foraminifera, hundreds of specimens/chambers can be analyzed per day, with the resulting data allowing information regarding the chemical heterogeneity in tests to be leveraged while also determining average chamber or whole-test values (e.g., Fig. 4).

Since its inception, the trend in laser ablation instrumentation in the Earth sciences has broadly been towards lower wavelength lasers, with deep-UV (193 nm) lasers now dominant over 213 nm and 266 nm solid-state (Nd:YAG) systems. The reason for this is that shorter wavelengths result in improved laser-sample coupling for many materials (minerals) of interest, including calcite (Park and Haglund, 1997). Specifically, greater absorbance (laser-sample coupling) reduces laser-induced heating and mechanical fracturing (Müller and Fietzke, 2016), resulting in smaller particles and a narrower particle size distribution (Hathorne et al., 2008). In turn, this minimizes inter-element fractionation during sampling and transport to the mass spectrometer (Alexander et al., 1998; Jochum et al., 2012), resulting in more accurate and more precise data, particularly when a non-matrix matched standardization procedure is used, as is common when analyzing CaCO₃ samples (Eggins et al., 2003; Evans et al., 2015b; Fehrenbacher et al., 2015; Reichart et al., 2003). While LA instruments are typically characterized by a relatively narrow depth of focus, such that the technique is best applied to samples with limited topography, differences in test diameters or chamber wall thicknesses can be accommodated without substantial analytical ramifications. As such, foraminifera <1 mm can be analyzed with minimal sample preparation (Fig. 4A).

Foraminifera sample preparation for LA-ICP-MS is potentially less laborious than that required for solution-based ICP-MS or analytical techniques that require highly polished sample surfaces (e.g., SIMS), but care should nonetheless be taken to ensure adequate removal of contaminant phases, especially adhered clay particles (e.g., Creech et al., 2010; Evans et al., 2015a; Inglis et al., 2023). As is the case with other geochemical techniques, preparing the samples for analysis is study/sample-type dependent: fossil tests may need rinses only to remove adhered clays (which are often not fully removed from the inside of the test) or may require aggressive cleaning to remove contaminant phases (e.g., metal oxides; Barker et al., 2003), whereas cultured specimens are typically only oxidatively cleaned to remove remnant organic matter (e.g., Schmidt et al., 2022; Smith et al., 2020; van Dijk et al., 2017). There is perhaps an under-recognized trade-off between the convenience of analyzing tests intact, and the ease with which contaminant phases may be removed when tests are crushed and chemically cleaned in ‘bulk’ prior to measurement via solution ICP-MS. While poorly preserved or contaminated portions of the test can often be excluded during post-LA-ICP-MS analysis, a benefit of spatially resolved data obtained from intact tests (Creech et al., 2010; Vetter et al., 2013b), it is likely that at least some proportion of these contaminant phases would not be present if the same tests were crushed before cleaning (although see Marr et al., (2013) in which it was first demonstrated that statistically identical data can be derived from LA and solution approaches following careful processing). In addition, it should be considered that foraminiferal tests are biomineral-organic composites that potentially contain phases other than calcite (e.g., fluid inclusions, metastable precursor phases; Arns et al., 2022; Branson et al., 2016; Jacob et al., 2017), such that LA may sample such phases that are more readily removed by (aggressive) cleaning procedures in which tests are first crushed, particularly in the case of pristine samples (Gray et al., 2023), given that these phases may be readily lost post-mortem (Mezger et al., 2018). Nonetheless, good agreement between solution and LA-ICP-MS Mg/Ca and Sr/Ca data has been demonstrated (Evans et al., 2016; Fehrenbacher et al., 2020), whereas substantial offsets in certain samples related to possible secondary phases have been observed in the case of Na/Ca (Gray et al., 2023).

During sample analysis, termed ‘ablation’, a fluence is selected such that (typically) 50-200 nm of material is removed from the sample surface by each laser pulse (when using a ns pulse length laser) and is transported to an ICP instrument of choice, usually in a He or He+Ar environment. The amount of ablated material/pulse is tunable by changing the laser energy (e.g., at higher energies, more material is removed per pulse). Given the very small amount of material removed and the development of two-volume or small-volume ablation cells, which provide fast sample washout times (Arrowsmith and Hughes, 1988; Eggins et al., 2003; Müller et al., 2009), LA-ICP-MS offers spatially resolved data at a resolution approaching that of NanoSIMS in the vertical direction (Figs. 4D–E) when utilizing ‘spot’ (depth profiling) analysis. However, given a typical beam diameter of ∼10–100 µm, resolving spatial heterogeneity at a sub-µm scale is only possible in samples that are (broadly) homogeneous perpendicular to the laser beam, which fortuitously includes many species of foraminifera (Creech et al., 2010; Eggins et al., 2003; Evans et al., 2015a; Fehrenbacher et al., 2017; Jochum et al., 2019; Spero et al., 2015).

A time-consuming aspect of LA-ICP-MS to consider is that, ideally, each individual specimen is tracked. That is, each individual foraminifer is given a sample ID, possibly individually imaged (light microscopy or SEM; e.g., Fig. 4) and weighed, and then individually removed from the sample slide after analysis to prepare the sample for follow-up analyses, such as SIMS or GS-IRMS (although less common, this could equally be performed in the case of solution measurements). Additionally, if one assumes a laser session of 12 hours, with an analysis time of up to 2 minutes per spot analysis plus the time required for the analysis of standards; then ∼75–150 tests can be analyzed per day (depending on the species and number of spots per specimen), equivalent to 3–6 solution-based bulk samples. Indeed, this decrease in sample throughput may make LA-ICP-MS a more suitable technique for ‘time-slice’ reconstructions (e.g., specific intervals such as the Holocene, deglacial, LGM), paired with solution-based analyses for generating the time-resolved data spanning the time-slices of interest.

Finally, we offer additional aspects of this technique for consideration: not all TE/Ca ratios can readily be measured on available LA-ICP-MS systems and moreover, a limited number of analytes can be measured at a time depending on the ICP (Q-ICP-MS vs. sector field, for example). This is because it is often desirable to have a relatively fast ‘sweep time’ [i.e., the total amount of time the mass spectrometer takes to measure count rates for every selected analyte (≈element) once] low to keep the time-resolution of the analysis highly resolved. While this is a consideration for any ICP-MS based technique, there is a particular need to maintain analytical rapidity in the case of LA-ICP-MS as this is a key control on achievable spatial resolution. As such, this may limit the number of analytes depending on the necessary dwell times, broadly a function of, for example, the concentration of the analyte in the samples as well as transport efficiency and ease of ionization.

Within the context of the above analytical benefits, the simple sample preparation, and the spatially resolved nature of LA-ICP-MS analysis, data from this technique has been leveraged in four main ways:

1. To determine inter-chamber variability in proxy trace element ratios, especially Mg/Ca, as a means of utilizing the time-resolved data to understand environmental variability on seasonal and interannual timescales. For example, chamber-by-chamber measurements of foraminifera have been used to reconstruct seasonality and interannual surface ocean temperature variability on a range of timescale, from modern to Eocene (Evans et al., 2013; Groeneveld et al., 2019; Guo et al., 2019; Schmitt et al., 2019; Wit et al., 2010).

2. To utilize spatial chemical heterogeneity to inform our understanding of foraminifera biomineralization and/or ecology, particularly in species that build layered tests with strongly contrasting trace element ratios (Allen et al., 2011; Eggins et al., 2004; Fehrenbacher et al., 2017; Hathorne et al., 2009; Hori et al., 2018; Kearns et al., 2023; Richey et al., 2022; Spero et al., 2015).

3. To exclude portions of fossil tests that are compromised by diagenetic overgrowths, dissolution, or other forms of contamination, as described above (Fig. 4D), particularly in cases where this would be difficult or impossible to remove or account for via solution ICP-MS cleaning procedures (Creech et al., 2010; Evans et al., 2015a; 2021; Sadekov et al., 2008; Van Raden et al., 2011).

4. To derive geochemical datasets from foraminifera grown partially in laboratory culture experiments. Given that many species have not been observed to reproduce in culture (most planktics) or are long-lived such that it is often not feasible to rely on reproduction as a means of deriving test calcite grown under experimental conditions, many experimental studies have derived a methodology to account for chambers formed prior to specimen collection. A common way of doing so is to use spatially resolved LA-ICP-MS analysis to ensure that only chambers precipitated under experimental conditions are used in, for example, laboratory culture calibrations of geochemical proxies (Fehrenbacher et al., 2018; Geerken et al., 2018; van Dijk et al., 2017) or to understand biomineralization processes (Evans and Müller, 2018; Levi et al., 2019), often coupled with a visual (Dueñas-Bohórquez et al., 2011; Hauzer et al., 2018; Mojtahid et al., 2023) or chemical/isotopic label (Evans et al., 2015a, 2016; Fehrenbacher et al., 2017) added to, or prior to, the experimental seawater to ensure unambiguous identification of culture-grown material (Fig. 4F).

LA-ICP-MS has been used to analyze foraminifera for more than two decades, with rapid growth in the popularity of the technique over the last ∼10 years. Current and future research will continue to shed light on the ontogenetic and biomineralization processes that modulate geochemical heterogeneity in widely used paleo-proxies (e.g., Mg/Ca), while laser ablation will remain a key tool of choice when contaminant phases cannot be separated from more pristine portions of the test, necessitating a spatially resolved approach. Laser ablation analyses will be of particular importance as new proxies are developed and investigated further (e.g., controls on sodium incorporation (Gray et al., 2023; Watkins et al., 2021) and barium variability in non-spinose foraminifera (Fritz-Endres et al., 2022; Richey et al., 2022). The pioneering research of Meilland et al. (2023) and Davis et al. (2020) demonstrated that planktic foraminifera can reproduce within laboratory cultures; this opens the gate for using LA-ICP-MS to investigate geochemical variability amongst specimens grown fully in culture, a feat once thought prohibitively difficult.

Recent developments that are likely to become more popular in the future include coupling the laser system to a multi-collector ICP-MS (LA-MC-ICP-MS), which permits the analysis of isotope systems, particularly δ¹¹B (Evans et al., 2021; Raitzsch et al., 2020; Standish et al., 2019), and has already been used to produce δ¹¹B data from samples prohibitively small for traditional solution MC-ICP-MS (e.g., Babila et al., 2022). Future possibilities include ‘split-stream’ analysis in which the gas stream from the laser is split between two mass spectrometers with, for example, one analyzing TE/Ca ratios (e.g., on a Q-ICP-MS), and the other isotopes (e.g., δ¹¹B, on a MC-ICP-MS). This approach has been used in other disciplines, for example in petrochronology, to simultaneously determine TE concentrations and U-Th-Pb ages (Kylander-Clark et al., 2013) or in other carbonates to simultaneously measure TE concentrations and ⁸⁷Sr/⁸⁶Sr (Prohaska et al., 2016), and may find particular utility in foraminifera given that spatially co-located δ¹¹B and TE analysis is desirable in both the fields of biomineralization and paleoclimate reconstruction, but must currently be separated. In addition, laser induced breakdown spectroscopy (LIBS) instruments are now commercially available and may be retrofitted to standard LA instruments in some cases. This technique, which uses the emission lines of elements, or those resulting from molecular recombination in the ablation plume (e.g., Gaft et al., 2014), opens up the possibility of quantifying elements that are traditionally very difficult to measure via ICP-MS, particularly the halogens, which may open up new research directions in the fields of biomineralization and paleo-chemical oceanography (Ram and Erez, 2023).

To embed or not to embed a foraminifer in epoxy begs the question: What knowledge are you seeking? Why would a researcher embed foraminifera tests in epoxy (a time consuming and irreversible process) and generate moderately-to-highly time-consuming images of a limited number of trace elements? Techniques to analyze foraminifera in cross section are typically conducted to explore 2D intratest TE variability. Indeed, there are faster tools (e.g., LA-ICP-MS) that can yield information about intratest TE variability within foraminifera tests, which offer an advantage of producing a more extensive array of TE information, including a greater number of analytes/isotopes compared to 2D-imaging instruments. However, high-resolution 2D images obtained from foraminifera tests polished through a cross-section characterize the spatial distribution of TEs within the test, and this is only possible with TE mapping. Trace element mapping can be achieved using many techniques including laser ablation trace element mapping (Fig. 5A), ToF-LA-ICP-MS (Fig. 5B), electron microprobe analysis (Fig. 5C), NanoSIMS (Fig. 5D), or ToF-SIMS (not discussed in this review, see Bonnin et al. (2019) for further details on this technique).

All the techniques discussed below involve first embedding a foraminifera test in epoxy resin and polishing the test to form a cross-section through the test wall. Typically, tests are cleaned prior to embedding, but the extent of cleaning is study dependent. Fossil specimens are often only rinsed in methanol and Milli-Q water to remove clays (e.g., Fehrenbacher and Martin, 2014; John et al., 2023; Sadekov et al., 2005), however additional oxidative and reductive steps can be included dependent upon test preservation/contamination. For example, studies have explored the effectiveness of cleaning on trace element distributions, especially regarding understanding variability of redox-sensitive elements (Glock et al., 2012; Pena et al., 2008; Pena et al., 2005). Specimens grown in culture are typically only oxidatively cleaned to remove remnant organic matter (e.g., Davis et al., 2017; Geerken et al., 2018; van Dijk et al., 2019).

Electron microprobe analysis (EMPA), often called electron probe microanalysis (EPMA), is a technique used to determine the elemental composition of a sample. It involves bombarding a sample with a focused electron beam, which induces the emission of characteristic X-rays from the elements present. By measuring the energy and intensity of these X-rays, the composition of the sample can be quantified. Electron microprobe analysis provides high spatial resolution, on the order of micron to sub-micron scales (Fig. 1C). While it is technically possible to analyze elements from beryllium to uranium, most trace elements in foraminiferal calcite are below the limit of detection (LOD), the lowest concentration of an element that can be quantified above background. As applied to foraminiferal calcite, studies using EMPA analyses typically focus on ‘major’ trace elements that are above the LOD, such as magnesium (Balestra et al., 2021; John et al., 2023), or to identify contaminant phases like iron or manganese (e.g., Pena et al., 2008; Pena et al., 2005). Additionally, the ability of the instrument to target elements of interest may be limited due to the capabilities of the instrument. Tungsten filament electron microprobe instruments operate at lower voltages and have lower spatial resolution. Field-emission electron microprobe instruments use a field emission source, which can emit a higher density of electrons compared to tungsten filament microprobe analyzers. As a result, field-emission microprobes offer superior imaging due to a more tightly focused beam and improved sensitivity in detecting trace elements within a sample (i.e., better LOD).

Early researchers used EMPA spot analysis to analyze Mg/Ca variability in planktic foraminifera. In the case of spot analysis, the electron beam is focused to 10–50 µm in diameter, and the test is analyzed in ‘spots’ that are spatially separated. Results from these early papers found that there was significant intratest Mg/Ca variability, but the range was generally in agreement with growth conditions or diagenetic alteration (e.g., Brown and Elderfield, 1996; Lipps and Ribbe, 1967; Nurnberg et al., 1996). Results from Brown & Elderfield (1996) are particularly notable as this seminal study demonstrated that foraminiferal crust calcite had a different composition compared to early ontogenetic calcite, motivating future studies to use this tool to further understand foraminiferal biomineralization (see de Nooijer et al., 2014, Section 6).

EMPA imaging of the TE distribution in foraminiferal calcite in 2D evolved from spot-based analyses. Here, the beam is focused and then scanned or ‘rastered’ on a region of interest, thus providing details about the elemental composition over a larger spatial area (typically 50–100 µm²; Fig. 5C). Imaging resolution as small as 0.3 µm can be achieved with a highly focused electron beam, but the actual resolution is lower (on the order of 1–2 µm) due to the excitation volume within the sample that emits X-rays. Analyses take between 1 to 18 hours depending on the size of the region of interest and resolution of the image. The first foraminifera image map focused on quantifying Mg variability in the species Orbulina universa (Eggins et al., 2004). Results confirmed that Mg was not homogeneously distributed within the test wall but rather consists of high and low Mg/Ca intercalated layers or ‘Mg-bands’ (see fig. 4 in Eggins et al. 2004). A follow-up study found that intratest Mg-banding was common in many planktic foraminifera species (Sadekov et al., 2005). These results spurred a number of studies to explore the cause for the high intratest-Mg variability, investigating questions such as: Is the Mg-banding diurnally modulated? Is banding due to migration in the water column? Does this variability complicate the use of Mg/Ca as a paleothermometer? Subsequent studies determined the temperature range implied by the Mg/Ca-banding is far too large to be linked to migration in the water column because the implied temperature range (>14°C) exceeds the depth range most species are acclimated to (Eggins et al., 2004). Culture experiments later demonstrated that these layers are, indeed, diurnally modulated (Spero et al., 2015; Fehrenbacher et al., 2017) and form at constant calcification temperature (Spero et al., 2015; Davis et al., 2017; Fehrenbacher et al., 2017). The mechanism responsible for such banding, however, remains elusive.

Researchers have also used EMPA mapping to identify contaminant phases on foraminiferal calcite that can complicate paleoceanographic interpretations (Pena et al., 2005, 2008), assess intratest variability driven by dissolution (e.g., Fehrenbacher & Martin, 2014) or ontogeny (Davis et al., 2017; Hathorne et al., 2009; Steinhardt et al., 2015), and assess water column controls on test geochemistry (Jonkers et al., 2016). More recent studies have used EMPA to assess the effect of preservation on trace element geochemistry (Staudigel et al., 2022; John et al., 2023).

Although EMPA is a useful tool for generating a 2D map of the distribution of trace elements in foraminiferal calcite, it has notable limitations. First, it can measure only 4–5 analytes at a time, depending on the instrument, with one detector typically dedicated to Ca. Second, as noted above, there are a limited number of elements that can be analyzed on an EMPA because the concentrations of many TEs in foraminiferal calcite fall below the LOD, typically 5 to 300 ppm depending on the analyte or instrument/electron source (tungsten or field-emission). Future directions for EMPA analysis include increasing the use of field emissions instruments, which have a better LOD and permit the analysis of elements that are in low concentrations in foraminiferal calcite, such as sulfur and sodium (van Dijk et al., 2019), or exploring the distribution of elements incorporated during culture in seawater with elevated TEs. Furthermore, EMPA could be used as a survey tool for identifying specimens that are ideal candidates for more expensive analysis, such as NanoSIMS, or for putting LA-ICP-MS profiles into context (e.g., Davis et al., 2017).

Nano-secondary ion mass spectrometry (NanoSIMS) is a technique for chemical microanalysis that provides spatially resolved information about element, molecule, and isotope composition of a sample (Herrmann et al., 2007). The principle of this technique is similar to stationary or “static” SIMS, described further above. The sample is bombarded with a highly energetic ion beam, which results in an ionization of elements at the upper sample surface and a consequent liberation of secondary ions (Herrmann et al., 2007). These secondary ions are filtered within a mass-filter according to their mass-to-charge ratio and subsequently quantified by a detector. In contrast to static SIMS, in which the sample is moved below a stationary ion beam, the ion beam is focused to a spatial resolution down to 50 nm and then moved dynamically on the surface of the sample to generate a molecule, element, or isotope image of the sample surface (Herrmann et al., 2007). Depending on the analyzed elements, either a Cs⁺ or O^– primary ion beam can be used to facilitate negative or positive ion formation, respectively. The number of ion species that can be analyzed simultaneously with this technique, depends strongly on the instrument (e.g., number of channels in the sector-field) but is usually around seven (Mueller et al., 2012).

Within foraminiferal research, this technique has found applications to study processes ranging from biomineralization (Fig. 5D) to intracellular bio(geo)chemistry and metabolism (LeKieffre et al., 2018). The technique is rarely used for quantitative analysis but can produce quantitative results comparable to bulk ICP-MS or static SIMS with an acceptable accuracy (Glock et al., 2019). For the microanalysis of the (calcareous) test walls, specimens are typically embedded in a resin followed by an exposition of a cross-section through the walls by grinding and polishing as with EPMA and SIMS. To study metabolic and biogeochemical processes of living foraminifera, specimens are typically fixed alive, decalcified and subsequently cut into thin sections for a coupled transmission electron microscopy (TEM)/NanoSIMS approach. This approach has a slight disadvantage, since it is not possible yet to analyze the soluble content within vacuoles or the cytoplasm, although cryo-SEM/EDS has had some success in this realm (Khalifa et al., 2016; Okada et al., 2024). The preparation method for TEM imaging involves a dehydration step that decreases the water content within the cell, which results in the loss of soluble content within the cytoplasm and vacuoles (Ayache et al., 2010). This can be overcome by using cryo-fixation instead (Newcomb et al., 2012), but due to the lack of a NanoSIMS that could be operated under cryogenic temperatures, this coupled cryo-TEM/NanoSIMS approach is not yet feasible. Recent advances in the development of a cryo-NanoSIMS may change this in the near future (Meibom et al., 2023).

Studies focused on biomineralization and test geochemistry of calcareous foraminifera that utilize NanoSIMS are relatively broad, including both cultured or natural-grown specimens. For example, Fehrenbacher et al. (2017) cultured Neogloboquadrina dutertrei under day/night cycles using labeled ⁸⁷Sr within the culture seawater and demonstrated that this species has a diurnal growth pattern, similar to Orbulina universa, and grows nearly all its high-Mg bands after the final chamber forms. In addition, several studies that utilized NanoSIMS revealed that many chemical elements within foraminiferal calcite are associated with organic heterogeneities in the foraminiferal test walls in rotaliid as well as in miliolid foraminifera (Geerken et al., 2019; Glock et al., 2019; Kunioka et al., 2006; Roepert et al., 2020). The chemical elements that so far have been shown to be associated with the organic content in the test walls include Mg and Sr (Kunioka et al., 2006), I (Glock et al., 2019), S (Geerken et al., 2019), and F and Cl (Roepert et al., 2020). Furthermore, Mn is weakly associated with organic heterogeneities (van Dijk et al., 2019). The study by van Dijk et al. (2019) is also a good example of how a complex approach using various microanalytical techniques, such as EMPA, micro-X-ray fluorescence (μXRF), and micro-X-ray absorption near-edge structure (μXANES), can be used to decipher biomineralization pathways. For example, those authors demonstrated that different processes affect Mn and Mg uptake into foraminifera and that photosynthetic symbiont activity can influence Mn incorporation. Finally, NanoSIMS has facilitated measurements of calcite precipitation rates in great detail (Geerken et al., 2022) and the study of rapid diagenetic exchange of isotopes in foraminiferal calcite with the surrounding water (Cisneros-Lazaro et al., 2022).

The relatively new coupled TEM/NanoSIMS approach brought several advances regarding the understanding of physiological and metabolic processes in foraminifera. For example, it was shown that a species of the Ammonia tepida morphogroup substantially reduces its heterotrophic metabolism when experiencing prolonged exposure to anoxia (LeKieffre et al., 2017). In addition, several previously unknown processes related to foraminiferal nitrogen metabolism have been revealed, which provided new indications of the role of these organisms in the marine nitrogen cycle (Bird et al., 2020; LeKieffre et al., 2018). Another interesting insight from this technique is the degradation of cytoplasm of an aphotic kleptoplastic foraminifera after light exposure (Jauffrais et al., 2019). The diversity of studies on foraminifera involving NanoSIMS measurements indicates how versatile this tool can be despite its relatively high costs, labor intensive sample preparations, and low performance for making quantitative measurements.

The majority of mass spectrometers operate in such a way as to separate ions on the basis of their mass-to-charge ratio (m/z), with most using a quadrupole or sector-field ICP-MS to measure the chemical make-up of foraminifera (e.g., Rosenthal et al., 1999; Yu et al., 2005). In this case, the ions are generated in an Ar plasma ion-source, focused through an ion lens system, and introduced into a quadrupole mass filter, which for a given setting only lets ions pass with a specific m/z. Because most ions generated by an Ar-plasma are +1 charged, the quadrupole can sequentially select for a given element, excluding other ions and thereby allowing the concentrations of Mg, Ca, and other elements to be determined. Shortly before the invention of the quadrupole mass filter in 1953, the first time-of-flight (TOF) mass spectrometer was developed (See Campana, 1987). These mass spectrometers operate on a different principle to quadrupole MS. In TOF-MS, discrete packages of ions are accelerated by an electric field of known strength, giving all the ions of the same charge in the package the same kinetic energy. The ions are then introduced into a relatively long flight tube and, because the speed of each ion depends on its m/z, the heavier ions travel more slowly down the tube than the lighter ions. Arrival time at a single detector is thus dependent on m/z. In contrast to quadrupole MS where ion collection is sequential, in TOF-MS all the ions in the package are measured quasi-simultaneously allowing detection of ions from across the mass spectrum.

The first TOF using an ICP ion source was developed in 1994 (Myers et al., 1994) but suffered from a number of analytical issues, such as poor sensitivity, that prevented its wide use. Modern TOF-ICP-MS instruments, such as the icpTOF (TOFWERK AG) or the Vitesse (Nu Instruments), have overcome many of these challenges and now offer a viable alternative to Q-ICP-MS, particularly when it comes to analysis by laser ablation. This is because TOF-ICP-MS can provide near full spectra (Na to U, or Li to Mn) collecting 10’s of mass spectra every millisecond. LA-TOF-ICP-MS coupled to the latest low-dispersion laser ablation cells that have washout times as low as 1 millisecond (e.g., the COBALT cell from Teledyne; Van Malderen et al., 2020) can therefore rapidly generate high-spatial resolution 2D near full-mass spectrum images free of the imaging artifacts that plague images from LA-Q-ICP-MS (van Elteren et al., 2018). Such imaging artifacts arise in LA-Q-ICP-MS because the sequential detection of ions requires slow-laser scanning speeds, introducing spot-scale mixing that blurs the resulting image obscuring fine details (van Elteren et al., 2018). The latest LA-TOF-ICP-MS can operate with high laser repetition rates (100’s Hz) and rapid rates of stage translation (100’s microns per second), resulting in each laser pulse on the sample being fully separate from the proceeding one, leading to crisp, artifact-free images more akin to Electron Microprobe than LA-Q-ICP-MS (Van Malderen et al., 2020). Application of LA-TOF-ICP-MS is growing (e.g., Savard et al., 2023), though to our knowledge it has only been applied to foraminifera in one study that demonstrated limits of quantification for marine carbonates with a 1-μm laser spot of ∼0.5 mg/g with uncertainties of ± 7%. By using a low-dispersion laser ablation cell and fast laser repetition rate (400 Hz), Standish et al. (in review) generated a 1-μm-per-pixel 0.4 × 0.5 mm² map of an entire Globigerinoides ruber test sectioned and mounted in resin in around 1 hour (including necessary standards). Intratest variability was identified in a number of elements across the mass range (e.g., Ce/Ca), with clear banding in Mg/Ca, Na/Ca, and Sr/Ca amongst others (Fig. 5B). This rapidity of mapping by LA-TOF-ICP-MS allows for population-level investigations of intratest heterogeneity, while the low quantification limit allows a broader metallomics-based view of test heterogeneity. An additional potential major strength of TOF-MS in the field of paleoceanography is that all ions are collected whether you are interested in them or not. As LA-TOF-ICP-MS based imaging grows therefore large datasets containing >50 elements could routinely be collected on foraminifera that, if properly archived, could be later “mined” as new proxies are developed and interpretations of existing proxies are refined.

The techniques described above provide a wealth of detailed, spatially resolved geochemical information that can provide broad insights into paleoenvironments, population dynamics, biomineralization processes, vital effects, and diagenetic processes. IFA data can be highly informative but are complex by nature. To be applied effectively, it is critical to match the scale of the technique to the nature of the question and to be prepared to deal with the data that these techniques produce. The arduous process of collecting these data is only half the work. In this section, we review some of the considerations a user should be aware of when choosing an IFA technique and highlight some of the key challenges when working with IFA data and combining these data across scales and techniques.

The best choice of IFA technique depends strongly on the nature and scale of the question posed (Fig. 1). Broadly speaking, lower resolution techniques are better suited to reconstructing environmental conditions (as sample throughput is higher), whereas higher resolution techniques are more suitable for examining the processes of biomineralization and/or diagenesis. The choice of a single IFA technique for a specific question is relatively straightforward. However, one of the key advantages of IFA techniques is the ability to analyze multiple geochemical systems within the same materials. These types of paired analyses provide unparalleled information for understanding the incorporation and preservation of geochemical signals in foraminifera but bring additional layers of complexity to the analytical and interpretive process.

Appropriately combining IFA techniques requires an understanding of the specific analytical details of each technique, which determine the order in which they must be applied to the samples. For example, to examine intra-population variability in Mg/Ca and δ¹⁸O, it is ideal to conduct these analyses on the same exact test material. This could be accomplished by first measuring Mg/Ca via LA-ICP-MS spot analyses and stable isotopes using GS-IRMS. This latter technique involves the dissolution of the test in phosphoric acid, so it must be conducted after the LA-ICP-MS analyses (or chambers must be fragmented). Another approach to measuring both these geochemical systems could be to use SIMS for δ¹⁸O and EMPA or NanoSIMS for Mg/Ca. These techniques all require embedding the sample in resin and polishing the test to expose a cross-section of the sample (i.e., they are highly compatible in terms of sample preparation). However, both sample damage considerations and analytical costs might dictate the order of analysis. Broadly, EPMA < NanoSIMS < SIMS in terms of both cost/availability and the extent of sample damage, so it would be prudent to conduct analyses in this order, first identifying regions of interest in EPMA, then examining them in more detail by NanoSIMS, before using these compositional data to choose specific areas for SIMS analyses of stable isotopes. It is also important to realize that it is impossible to extract a sample from the resin mount required for analyses conducted under high vacuum, so it is recommended that all other geochemical characterization(s) must be undertaken before these analyses are carried out.

Another critical consideration when applying IFA techniques is how to combine data from different spatial scales of analysis. Of the techniques considered above, GS-IRMS requires the dissolution of entire chambers of tests, LA-ICP-MS examines 10s-of-micron subregions of individual chambers, SIMS analyses examine 3–10 μm spots within a polished cross-section, EPMA has an excitation volume of 0.5–2 μm, and NanoSIMS can achieve 50-nm beam diameters. All foraminifera are chemically and structurally heterogeneous across all these length-scales, and the analyses of these features with different spatial resolutions can dramatically alter the patterns observed (Fig. 1). Even so, these techniques can be combined by comparing data across spatial scales with extreme care, extensive use of careful correlative imaging to ensure the location of analyses is well constrained, and down-sampling data from higher resolution techniques before comparison to lower-resolution measurements.

Another important factor in combining these data is recognizing that the same geochemical feature can look very different when analyzed at different spatial scales (Figs. 1B–E). For example, consider GS-IRMS and SIMS measurements of stable isotopes. Gas source-IRMS will provide a single value for an entire foraminiferal test or chamber (Figs. 1B–C), whereas multiple SIMS spot measurements from within that test will reveal a wide variability of values (Fig. 1E) which, if the material has been sufficiently sampled, should have a mean value close to the GS-IRMS measurement. It is also important to note that the variability within the SIMS measurements will be a direct function of the length-scale of the variability within the sample, and the size of the SIMS analysis spot. If there are numerous sharp transitions between distinct compositions within the test (as is well-established for Mg/Ca, for example, Eggins et al., 2004), larger spot sizes will lead to more ‘smoothing’ during sampling and reduce the observed variability, and vice versa (Figs. 1D–E). These issues are particularly noticeable within the trace elements, numerous of which are known to exhibit substantial systematic banding within the test wall (e.g., Kunioka et al., 2006).

There is a fundamental incompatibility between spatially resolved IFA data collected at different resolutions from a heterogeneous sample. This complicates the comparison of data collected by different techniques and with different spatial resolutions. Ultimately, the only way to robustly compare data with different spatial resolutions is to down-sample the higher resolution technique so that it matches the lower-resolution technique at the precise locations that the lower-resolution measurements were made. Ideally, this down-sampling should be physically based and take account of the characteristic energy distribution of the analysis beam. For example, one could down-sample a grid of NanoSIMS analyses by convolution with a gaussian kernel that approximates the size and shape of a SIMS analysis beam. A further consequence of this is that it is not possible to meaningfully compare individual spot measurements with substantially different spot sizes—the smaller spot should be rastered over the sample to allow subsequent down-sampling to the resolution of the larger beam.

The aim of this review is to highlight how combining IFA data across techniques and special and temporal resolutions can provide invaluable insights but is not a trivial task. Doing this well requires extensive data analysis, and the fine detail of this analysis could have a profound effect on the eventual conclusions of the comparison. All data processing and analysis should therefore be carefully documented and recorded at all stages, including the techniques applied and software used. For example, LA-ICP-MS data can be analyzed using LA-Tools (Branson et al., 2019), in MS Excel (Jochum, 2007; Öğretmen et al., 2022), or Iolite (e.g., Sadekov et al., 2019). NanoSIMS data can be analyzed, for example, using Fiji (which requires the plugin OpenMIMS) or via Look@NanoSIMS, a free software provided by Cameca Instruments. It is critical that anyone attempting these types of analysis adhere strictly to the principles of FAIR data (Wilkinson et al., 2016), provide raw data alongside their publications, when possible, and consider sharing additional information and/or code required to reproduce their analysis from those data (e.g., Searle-Barnes et al., 2023).

While this review is focused on the inorganic geochemistry of foraminifera, new methods for constraining the genotype of different foraminiferal taxa as well as potential links between foraminifera microbiome associations and geochemistry are being developed with workflows for pairing inorganic and organic techniques underway. For example, it is now possible to obtain an individual foraminifer’s genotype, microbiome associations, morphology (via microCT scanning or SEM), and test geochemistry (e.g., Ujiie et al., 2019; Lane et al., this issue). In summary, IFA techniques, both individually and when paired with each other, provide a plethora of approaches and future research directions for foraminiferal researchers that can be pursued with the understanding of identifying the right tool for the right question.