Quantitative Characterization of Carbonate Mineralogy in Lake Yangzong Sediments Using XRF-Derived Calcium Signatures and Inorganic Carbon Measurements

Abstract

The carbonate content serves as a fundamental proxy in lacustrine sediments for reconstructing palaeoclimate and environmental changes. Although multiple analytical techniques exist for its quantification, systematic comparisons between different methodologies and the precise identification of carbonate mineralogy are still needed. In this study, a 1020 cm continuous sediment core (YZH-1) from Lake Yangzong in Yunnan Province was employed. Initially, the semi-quantitative calcium (Ca) concentration was obtained via X-ray fluorescence (XRF) core scanning. Subsequently, the total inorganic carbon (TIC) content was determined using both the loss on ignition (LOI) and gasometric (GM) methods to evaluate methodological discrepancies and potential biases. Furthermore, a quantitative regression model was developed to estimate carbonate abundance based on the relationship between XRF-derived Ca data and the analytically determined carbonate content. A comparative analysis revealed a strong positive correlation (r = 0.97) between LOI and GM measurements, though LOI-derived values are systematically elevated by 2.6% on average. This overestimation likely stems from the thermal decomposition of non-carbonate minerals during LOI analysis. Conversely, GM measurements exhibit a ~5% underestimation relative to certified reference materials, attributable to instrumental limitations such as gas leakage. Strong covariation (r = 0.92) between XRF-Ca intensities and the TIC content indicates that carbonate minerals in Lake Yangzong sediments predominantly consist of calcite. A transfer function was established to convert XRF-Ca scanning data into absolute Ca concentrations, leveraging the robust Ca-TIC relationship. The proposed quantification model demonstrates high reliability when applied to standardized XRF-Ca datasets, offering a practical tool for paleolimnological studies in similar geological settings.

1. Introduction

Endogenic carbonates constitute a prevalent mineral phase in lacustrine sedimentary systems, with their precipitation dynamics being governed by a complex interplay of environmental variables including climatic parameters, aqueous carbonate equilibria, and basin morphotectonic features. These authigenic minerals serve as sensitive paleoenvironmental archives, faithfully documenting hydrological fluctuations and climatic variations within lake catchments, thereby offering valuable insights for paleoclimatological [1,2], hydrological [3,4], sedimentological [5,6], and tectonic investigations. The quantification of the total inorganic carbon (TIC) in lacustrine sediments is conventionally performed using two principal analytical approaches: the gasometric method (GM) and loss on ignition (LOI) technique. The fundamental principle of GM, initially established by Bien [7], relies on acid-mediated carbonate decomposition with subsequent CO₂ evolution, where the liberated gas volume serves as the basis for TIC determination. This methodology offers distinct advantages including minimal sample requirements, operational simplicity, and rapid analytical throughput. When implemented with properly calibrated apparatus maintaining optimal hermetic integrity, the technique yields reproducible and accurate results. Nevertheless, measurement precision may be compromised by several confounding factors encompassing ambient temperature fluctuations, barometric pressure variations, equipment sealing efficiency, and operator-dependent variables, often introducing systematic biases [8,9,10]. Notably, the magnitude of analytical uncertainty exhibits an inverse correlation with TIC concentrations, becoming particularly pronounced in low-TIC samples [11].

The LOI method represents another prevalent approach for determining the total inorganic carbon content in sedimentary matrices. This method has irreplaceable importance in studying the carbon content in lake sediments, with value in multiple dimensions such as method characteristics, environmental indicators, and research applications, and has been widely used [12,13,14,15,16,17]. This technique offers distinct advantages including operational simplicity, cost-effectiveness, and high-throughput capability while maintaining satisfactory analytical reliability [18,19]. However, the method requires relatively large sample quantities and tends to yield slightly elevated values compared to actual TIC concentrations [20]. Heiri (2001) conducted experiments with a standardized lake sediment and revealed a strong initial weight loss at 950 °C, but the samples continued to lose weight at a very slow rate at an exposure duration of up to 64 h, although the differences became negligible. In addition, due to the incomplete combustion of volatile salts, the structural water of clay minerals or metal oxides, or organic matter in the sample at 550 °C, the residual organic matter and other substances are burned off when samples are burned at 950 °C, resulting in a higher measured TIC value [19]. While both the GM and LOI methods exhibit certain systematic errors, they generally provide reliable TIC quantification when properly implemented [21].

Calcium (Ca) ranks among the most abundant elements in lacustrine sediments and serves as a primary constituent of carbonate minerals, with its concentration bearing significant climatological and hydrological implications [22,23,24]. The advent of high-resolution X-ray fluorescence (XRF) core scanning technology has facilitated the application of elemental geochemistry in paleoclimate reconstruction across diverse temporal and spatial scales [25,26,27,28]. The XRF core scanning method offers distinct advantages including high analytical resolution, rapid measurement capability, and cost-effectiveness. However, several sediment characteristics may potentially compromise scanning accuracy, including the water content, grain size distribution, porosity, fracture density, and core surface topography. Consequently, XRF-derived elemental data represent semi-quantitative measurements that require systematic evaluation and calibration [29,30]. Recent methodological validation studies have demonstrated the reliability of XRF core scanning technology through multiple verification approaches, confirming its effectiveness in characterizing geochemical element variations along sediment cores [16,23,31,32,33,34]. A comparative study by Lei et al. (2011) on sediments from Zigetang Lake (Tibetan Plateau) established a robust conversion function between XRF Ca intensity and the absolute Ca concentration (determined by ICP-OES) using linear regression analysis [35]. Subsequent investigations of Lake Fuxian sediments [36] and peat deposits from Xingyi [37] revealed strong correlations between XRF-derived Ca intensities and the inorganic carbon content, indicating a predominant co-occurrence of Ca and carbonate as CaCO₃ in both lacustrine and swamp facies within the karst regions of Southwestern China. These findings provide theoretical justification for employing the TIC content to quantitatively calibrate XRF Ca intensity measurements.

Given these considerations, this study focuses on Lake Yangzong, a typical plateau lake in Yunnan Province, as the research subject. A comprehensive analysis was conducted on a 10.2 m sediment core retrieved from the lake, employing high-resolution XRF continuous scanning to determine the relative concentrations of calcium. For TIC quantification, parallel measurements were performed using both the GM method and LOI technique. The carbonate content data obtained from these two analytical approaches were systematically compared and cross-validated. Subsequently, a correlation analysis was performed between the corrected carbonate content and XRF-derived Ca intensity measurements. Based on these comparative analyses, we developed a novel quantitative calibration method for converting XRF continuous scanning data on Ca intensity to the absolute Ca content using TIC measurements as reference values, specifically tailored for Lake Yangzong sediments.

2. Geographical Setting of Study Site

Lake Yangzong (24°51′–24°58′ N, 102°58′–103°01′ E; elevation ~1770 m a.s.l.) is a rift-basin lake situated in the central Yunnan Plateau of Southwestern China (Figure 1). With a surface area of approximately 31 km² and a catchment area of about 192 km², the lake exhibits an average water depth of 20 m, reaching maximum depths of 30 m [38,39]. The lake occupies a pull-apart basin formed by subsidence along the western branch of the middle Xiaojiang Fault system. The surrounding mountainous terrain consists primarily of Permian shale, limestone, and basalt formations, while Quaternary deposits of gravel, sand, and clay dominate the piedmont plains [40].

Climatologically, the Lake Yangzong region is influenced by the South Asian Monsoon (SAM), exhibiting characteristic seasonal patterns with distinct wet summers and dry winters. Meteorological records from the nearby Kunming station (1951–2012 CE; located ~32 km northwest of the lake at ~1800 m a.s.l.) indicate a mean annual temperature of 15.1° C, with monthly averages ranging from 8.3 °C in January to 20 °C in July. The mean annual precipitation totals 987 mm, with approximately 80% occurring during the monsoon season (May to October). Hydrologically, Lake Yangzong represents a semi-closed, warm monomictic system characterized by winter mixing and seasonal stratification. The lake receives water inputs primarily from atmospheric precipitation, the Yangzong River (the major inflow located in the south), and several smaller seasonal streams along the Xiaojiang Fault system. The primary outflow occurs through the Tangchi Channel, an artificial drainage structure constructed in 1388 CE in the northeastern section of the lake [41].

3. Materials and Methods

3.1. Coring

During a field campaign in July 2013, sediment coring operations were conducted in the central basin of Lake Yangzong (water depth: 23 m; coordinates: 24°54′42″ N, 103°0′7″ E) using an Austrian UWITEC coring platform (UWITEC GmbH, Mondsee, Austria) equipped with a 3 m long, 60 mm diameter piston corer. Through successive overlapping drilling operations, nine deep sediment cores were successfully retrieved. From these, a continuous 10.2 m composite sediment profile (designated as YZH-1; Figure 1) was constructed through careful depth matching and elemental scanning correlation. The intact sediment cores were maintained in their original PVC liners during transport and storage. In the laboratory, each core was longitudinally sectioned into two halves. One half was archived at 4 °C for permanent preservation, while the other half was systematically subsampled at 0.5 cm intervals (0–120 cm depth) and 1 cm intervals (121–1020 cm depth) following XRF core scanning, preparing the samples for subsequent multi-proxy analytical procedures.

3.2. XRF Core Scanning

The XRF core scanning analysis was performed at the Key Laboratory of Plateau Lakes Ecology & Global Change, Yunnan Normal University, utilizing an Avaatech XRF Core Scanner (Avaatech B.V., Amsterdam, The Netherlands) [42]. Prior to scanning, all core surfaces were carefully smoothed and protected with a 4 μm thick Ultralene film to both minimize measurement artifacts caused by surface irregularities and prevent potential contamination of the XRF probe. For calcium detection, scanning parameters were set as follows: spatial resolution of 0.5 cm (0–120 cm depth) or 1 cm (121–1020 cm depth), with an analyzed area of 5 mm (width) × 10 mm (length). The instrument was operated at 10 kV voltage and 1 mA current with an integration time of 30 s per measurement. The resulting elemental data, expressed as counts per second (cps), provide semi-quantitative information that reliably reflects relative variations in chemical composition rather than absolute elemental concentrations.

3.3. Gasometric Method

The sediment samples were ground to approximately 100 mesh, and 0.6~0.7 g of each sample was weighed after drying at 60 °C for 24 h and then treated with 10 mL of dilute hydrochloric acid in a closed-system reaction vessel, with the evolved CO₂ gas being quantitatively collected and measured volumetrically. To ensure data quality, we implemented a rigorous quality control protocol: one parallel sample was added to every five samples, and one standard sample was added to every ten samples to test the accuracy of the experimental process and the reliability of the experimental data [7,20,43]. The standard sample was sodium carbonate reagent with a purity of 99.9%. Repeated experiments yielded an average sodium carbonate content of 95%, indicating a reproducibility error of ±5%. The analyses were carried out in the Key Laboratory of Plateau Lakes Ecology & Global Change, Yunnan Normal University. The total inorganic carbon (TIC) content was calculated using Equation (1).

W_GM = (a·P·V·M)/(R·T·m)

(1)

In the given formula, W_GM represents the percentage content of TIC. The parameters P and T correspond the atmospheric pressure and absolute temperature recorded during the experimental procedure, respectively. The correction coefficient a was determined through a sequential analysis of 10–15 certified standard samples. This coefficient is derived from the ratio between the measured gas volume under actual laboratory conditions and the theoretical gas volume expected from the complete reaction of the standard sample under ideal conditions, serving to compensate for potential systematic errors in the measurement system. The variable V denotes the experimentally measured volume of evolved CO₂ gas. The constants R, M, and m represent the universal gas constant (8.314 J·mol⁻¹·K⁻¹), the atomic mass of carbon (12 g·mol⁻¹), and the precisely measured sample mass, respectively. All sediment samples from core YZH-1 were subjected to TIC quantification using this gasometric methodology.

3.4. Loss on Ignition Method

The LOI method is a conventional approach for quantifying both the organic matter content (LOI at 550 °C, hereafter LOI_550°C) and total inorganic carbon (TIC) content (LOI_950°C) in sedimentary samples [19]. In this study, approximately 3.00 g of homogenized sample material was initially processed using an agate mortar and pestle to ensure particle size consistency. Prior to analysis, samples underwent sequential thermal treatment as follows: (1) drying at 105 °C for 2 h to remove residual moisture, (2) combustion at 550 °C for 4 h to oxidize organic matter, and (3) calcination at 950 °C for 2 h to decompose carbonate minerals. The analyses were conducted at the Institute for Ecological Research and Pollution Control of Plateau Lakes of Yunnan University. The TIC content was subsequently calculated using Equation (2).

W_LOI = M₁·(weight at 550 °C − weight at 950 °C)/(weight at 105 °C·M₂)

(2)

Within the given formulation, W_LOI represents the percentage concentration of total inorganic carbon (TIC). The parameters M₁ and M₂ correspond to the molar masses of carbon (12 g/mol) and carbon dioxide (44 g/mol), respectively. For this investigation, a comprehensive set of 363 sediment samples underwent TIC quantification employing the LOI methodology.

4. Results

4.1. XRF Scanning Intensity of Ca Elements

The X-ray fluorescence core scanning results from the YZH-1 sediment core reveal significant variations in Ca intensity, ranging from 0.10 × 10⁶ to 0.97 × 10⁶ counts per second (cps), with a mean value of 0.62 × 10⁶ cps. The vertical Ca distribution profile (Figure 2a) exhibits four statistically distinct stratigraphic intervals: Phase I (1020–763 cm depth) displays the lowest mean Ca intensity (0.37 × 10⁶ cps) within the core sequence. This interval maintains relatively stable values, characterized by an initial increase, followed by a decrease and then a subsequent increase. Phase II (763–494 cm) initiates with a rapid Ca enrichment event, evolving through three well-defined evolutionary stages that culminate in peak intensity (0.97 × 10⁶ cps) at the phase boundary (494 cm). Phase III (494–83 cm) sustains elevated Ca concentrations with an average value of 0.76 × 10⁶ cps), modulated by four pronounced cyclical variations that likely reflect periodic environmental changes. Phase IV (83–0 cm) documents a dramatic depletion in the Ca content, decreasing precipitously from 0.90 × 10⁶ to 0.10 × 10⁶ cps, representing the minimum observed values in the entire core stratigraphy.

4.2. TIC Contents

The total inorganic carbon (TIC) content of the YZH-1 core, assessed using both the GM method and the LOI method, ranges from 0.14% to 8.47% and from 1.34% to 8.30%, respectively, with average values of 5.47% and 5.61%. Measurements obtained via the LOI method are generally slightly higher than those derived from the GM method. Both methodologies exhibit similar trends and can be categorized into four distinct stages (Figure 2b): Stage I (1020–845 cm) displays the lowest and most stable TIC concentrations in the core sequence, with the GM and LOI measurements averaging 2.6% and 3.5%, respectively. During Stage II (845–626 cm), both methodologies document a progressive enrichment in the carbonate content, with values increasing from 3.1% (GM) and 4.5% (LOI) to 5.6% (GM) and 7.1% (LOI). The subsequent Stage III (626–86 cm) maintains the highest TIC levels throughout the core, characterized by fluctuating values averaging 7.0% (GM) and 7.5% (LOI). Finally, Stage IV (86–0 cm) exhibits an abrupt depletion in inorganic carbon, with both techniques recording near-zero values—the minimum observed in the entire borehole profile.

5. Discussion

5.1. Comparison and Correction of TIC Results Between GM Method and LOI Method

A robust positive correlation (r = 0.97, p < 0.001) exists between the TIC contents determined using the GM and LOI methods, as evidenced by linear regression: TIC_LOI = 0.85 × TIC_GM + 1.40 (Figure 3a). This strong agreement validates the methodological reliability of both approaches [9,21]. As previously discussed, systematic errors affect both methodologies. The GM method typically yields underestimations due to inherent measurement limitations [8,9,10,11], while the LOI results tend to overestimate the TIC content. Notably, even after calibration, persistent discrepancies suggest LOI-derived values may incorporate systematic overestimation. Several factors contribute to LOI measurement variability, including the duration of air exposure before post-heating weighing, the crucible position in the muffle furnace, and sample volume. Additionally, LOI measurements can be influenced by the combustion of residual organic matter and the decomposition of mineral crystalline water, often resulting in overestimation [19]. Conversely, the GM method quantifies the carbonate content through acid-induced CO₂ evolution and volumetric measurement. When proper instrumental calibration and airtight conditions are maintained, this approach provides highly reproducible determinations despite its tendency toward slight underestimation [2]. In the YZH-1 core, 86% of GM measurements were lower than the corresponding LOI values, with the discrepancy magnitude being inversely proportional to the TIC content (Figure 3a). This relationship confirms that the sample TIC concentration is a critical factor influencing measurement accuracy [20]. Consequently, this study employs calibrated GM results as the reference TIC values for all subsequent analyses as this approach minimizes organic matter interference, reduces temperature-dependent artifacts, and demonstrates superior reproducibility at low TIC concentrations.

5.2. Comparison and Quantitative Conversion of Ca Element XRF Scanning Intensity and TIC Content

A comparative analysis of the calibrated TIC content and Ca element XRF scanning intensity in core YZH-1 demonstrates remarkable consistency in their stratigraphic variation patterns (Figure 2). Statistical evaluation reveals a robust positive correlation (r = 0.92, p < 0.001, Figure 3b) between these parameters, as evidenced by the linear regression equation:

TIC_GM (%) = 8.96 × Ca-_XRF/10⁶ − 0.08

(3)

These findings indicate that calcium and inorganic carbon in Lake Yangzong sediments predominantly occur as calcium carbonate (CaCO₃). The lake’s watershed exposes primarily Permian, Carboniferous, and Devonian carbonate formations, whose weathering releases Ca²⁺ and HCO₃- ions into the aquatic system. Subsequent authigenic carbonate precipitation under lacustrine conditions logically favors CaCO₃ formation. Supporting evidence comes from adjacent Lake Fuxian, where Li et al. [36] demonstrated strong covariance between the TIC content, Ca XRF intensity, and calcite XRD peak intensity. This tripartite correlation confirms calcite (CaCO₃) as the dominant carbonate mineral phase, with other carbonate species occurring in trace amounts below detection limits.

5.3. Quantitative Calculation of Ca Element Based on XRF Scanning Intensity

Following the identification of CaCO₃ as the dominant carbonate mineral phase in Lake Yangzong sediments, we derived the absolute calcium content from gasometrically determined TIC measurements using Equation (4). Building upon this foundation, we developed a robust conversion algorithm (Equation (5)) through a correlation analysis between the calculated Ca concentrations and XRF scanning intensities, as mathematically described by Equations (3) and (4).

Ca_GM (mg/g) = 10 × TIC_GM (%) × 40/12

(4)

Ca_FV (mg/g) = 298.7 × Ca_XRF/10⁶ − 2.67

(5)

In the above formulations, the coefficients 12 and 40 correspond to the atomic masses of carbon (C) and calcium, respectively, while Ca_FV represents the calculated absolute calcium content derived from the fitting procedure. Equation (5) establishes the quantitative conversion algorithm for determining the absolute Ca concentrations from XRF core scanning data in Lake Yangzong sediments. The validation of this conversion algorithm demonstrated excellent agreement between the calculated and measured absolute Ca contents, as evidenced by a strong positive correlation (r= 0.92, p < 0.001) shown in Figure 4a and Figure 5a.

By applying this conversion function to the Ca element XRF scanning intensity data from the Lake Fuxian sediment core, the fitted Ca values were calculated. These fitted values show a highly significant positive correlation (r = 0.97; Figure 4b) with the measured values, following a linear relationship.

The implementation of this conversion algorithm to the calcium XRF intensity dataset from the Lake Fuxian sediment core [36] yielded calculated Ca concentrations that demonstrated an exceptionally strong linear correlation (r = 0.97, p < 0.001; Figure 4b) with direct measurement values, as described by the following regression equation:

Ca_FV (mg/g) = 1.08 × Ca_GM (mg/g) − 0.98

(6)

The stratigraphic profiles of the calculated and measured Ca concentrations exhibit near-perfect congruence throughout the sediment core (Figure 5b), unequivocally validating the reliability of the conversion algorithm. Therefore, based on the established predominance of CaCO₃ in the lacustrine sediments, the calibrated XRF calcium scanning intensities coupled with the conversion algorithm developed in this study enable the rapid and precise determination of the absolute CaCO₃ content throughout sediment cores. This methodology provides a robust and efficient technical solution for a high-resolution quantitative analysis of carbonates in lacustrine core sequences.

6. Conclusions

This study presents a comprehensive analysis of a 1020 cm continuous sediment core (YZH-1) retrieved from Lake Yangzong, Yunnan Province, China. Initially, high-resolution X-ray fluorescence core scanning was conducted to establish calcium intensity profiles throughout the sediment sequence. Subsequently, parallel determinations of the carbonate content were performed using both the gasometric method and loss on ignition technique, with particular emphasis on methodological discrepancies and their underlying controlling factors in lacustrine sediment analysis. Additionally, calibrated carbonate content data were correlated with Ca scanning intensities to elucidate sediment carbonate mineralogy. This integration enabled the development of a robust quantitative model for the carbonate mineral content based on XRF-derived Ca data and direct carbonate measurements.

The analytical results reveal a strong positive correlation (r = 0.97, p < 0.001) between the carbonate content determinations obtained via the LOI and GM methods, with the LOI-derived values systematically exceeding the GM measurements by an average of 2.6%. LOI techniques may overestimate the carbonate content due to the concurrent decomposition of non-carbonate mineral phases, while GM measurements typically underestimate true values by approximately 5%, as verified through certified reference material analyses, primarily attributable to instrumental limitations including system airtightness. Both methodologies exhibit excellent concordance with depth-resolved Ca XRF scanning profiles throughout the core sequence.

Based on rigorous method validation, we recommend employing systematically calibrated GM results as the reference carbonate content. The observed strong covariance between XRF Ca intensities and corrected TIC values (r = 0.92, p < 0.001) provides compelling evidence that carbonates in Lake Yangzong sediments occur predominantly as CaCO₃. This study establishes a robust conversion algorithm that quantitatively relates XRF Ca scanning intensities to the absolute Ca content, expressed as Ca (mg/g) = 298.7 × Ca_XRF/10⁶ − 2.67. Given the documented predominance of calcite in southwest Chinese lacustrine systems, this algorithm demonstrates high reliability when applied to standardized and calibrated XRF datasets, showing significant potential for regional paleolimnological studies.