Soil thermogravimetric analysis, abbreviated TGA, measures how a soil sample changes mass while it is heated through a controlled temperature programme. If you are trying to read a soil TGA curve, derivative thermogravimetry or DTG peak, loss-on-ignition interpretation, organic matter temperature interval or carbonate decomposition feature, start with the axes: sample mass or percent remaining against temperature. A loss of mass is an observation, not automatically the mass of organic carbon that disappeared.
The crucial reading skill is to separate what the balance measured from what caused each change. Water evaporation, organic compound reactions, mineral hydroxyl loss and carbonate decomposition can affect different or overlapping temperature ranges. Heating atmosphere, ramp rate, soil mineralogy and pretreatment influence those ranges. A DTG peak marks rapid mass change under the chosen programme; it is not a unique fingerprint of one biological carbon pool.
This guide explains soil TGA and DTG from first principles, including sample preparation, heating profiles, normalisation, derivative curves, thermolabile versus thermostable categories, carbonate signals, calibration, complementary measurements and realistic limits. Numerical examples are hypothetical, with explicit masses and units. The aim is to let readers translate a TGA figure accurately without turning a temperature interval into an unsupported claim about soil carbon permanence.
Quick answer: what is on a soil TGA plot?
The main TGA curve records sample mass as temperature increases. Depending on the graph, the y-axis may show milligrams remaining, mass percentage remaining or cumulative percent mass loss. The x-axis normally shows temperature in degrees Celsius, sometimes accompanied by time. A drop in the curve means mass left the pan or was transformed into volatile products under the instrument conditions.
A DTG curve expresses the derivative, the rate of mass change per degree or time, usually with a sign convention defined by the plotting software. A peak indicates a temperature range where mass changes quickly. It helps identify stages in a complex curve, but the peak’s chemical assignment needs independent knowledge and often complementary analyses.
The instrument is a balance inside a controlled furnace
Thermogravimetric instruments heat a small sample while monitoring its mass with a sensitive balance. A temperature programme specifies the starting temperature, heating rate, final temperature and possible holds. A controlled gas atmosphere may be air, oxygen, nitrogen or another specified environment, influencing combustion and decomposition reactions.
Read the programme before interpreting any peak. Organic matter can oxidise differently in air than it pyrolyses under inert gas, and minerals may lose water or decompose in either atmosphere. The same soil may produce different TGA patterns under different protocols. A curve is an experimental response to heating, not an unalterable signature that every sample has independent of method.
Soil is an intentionally complicated sample
Unlike a pure substance, soil contains particles, organic compounds, minerals, water, salts and potentially carbonates. Their transformations can overlap during heating. A mass change between 200 and 500 degrees Celsius might include contributions from several material classes rather than one identifiable molecule.
This complexity explains why TGA is useful for broad screening and comparison but risky when every peak is given a specific chemical identity without corroboration. A study should report soil type, parent material, pretreatment, carbonate presence and relevant reference methods. Do not interpret a peak’s cause from temperature alone when several plausible reactions overlap.
Sample pretreatment changes the opening part of the curve
A soil may be air dried, oven dried, sieved or ground before TGA. The starting water content and physical state affect early mass loss. Pre-drying at an elevated temperature can remove part of the water that an untreated sample would otherwise lose during the TGA run. Fine grinding can alter thermal contact and particle exposure.
Therefore, two studies with different pretreatment may report different percentages in the same temperature interval without reflecting a biological difference in the original soil. A careful results paragraph states whether percentages use the as-received, air-dry, oven-dry or preconditioned sample mass. That denominator is the first checkpoint before comparing anything about carbon.
Reading percent remaining versus percent lost
Suppose a fictional 100-milligram sample has 95 milligrams remaining at 150 degrees Celsius and 90 milligrams at 550 degrees. On a percent-remaining plot, the values are 100%, 95% and 90%. On a cumulative-loss plot, the corresponding values are zero, 5% and 10%. They describe the same mass history with opposite visual directions.
A reader who treats 90% remaining as 90% lost will invert the conclusion. Read the label, not the line shape alone. When comparing plots, consider whether an author normalised each curve to its own starting mass or plotted absolute milligrams, because samples of different original masses can show different absolute loss even at identical proportional behaviour.
Early mass loss often includes water
In many soil TGA protocols, a low-temperature interval includes evaporation of free or adsorbed water. The precise contribution depends on sample drying, clay surfaces, organic matter and instrument conditions. Some mineral or hydrated compound processes may overlap, so it is unsafe to call every low-temperature loss pure moisture without a method-based rationale.
A 2025 soil thermogravimetric mass-loss study observed a major low-temperature water-related feature in a set of soils, while higher-temperature behaviours reflected both organic and mineral properties. The study’s ranges illustrate its experimental system, not universal cutoffs to be imposed on every parent material.
Mid-temperature losses need mixed chemical interpretation
Soil organic matter can undergo oxidation, volatilisation and decomposition over broad overlapping temperature intervals. Some soil minerals also lose structural water over temperature ranges where organic compounds react. A cumulative mid-temperature mass loss therefore need not equal soil organic carbon content.
If a paper uses a chosen interval as a proxy for organic matter, check how it validated that proxy against independently measured soil organic carbon. A calibrated correlation can be useful within its tested sample classes; it should not be presented as a universal identity. Different clay and carbonate compositions can change the relationship significantly.
High-temperature features can include carbonates
Carbonate minerals can release carbon dioxide at elevated temperatures under appropriate conditions. A soil containing significant carbonate may therefore show high-temperature mass loss not attributable to organic matter. The location and size of the feature depend on mineral chemistry, sample size, heating rate and atmosphere.
The 2025 research on TGA and soil mineral stabilisation properties documents a high-temperature feature associated with carbonate-rich sample groups, around its reported interval. Readers should identify carbonate presence before using a high-temperature loss to infer ‘recalcitrant organic carbon’. It may largely represent inorganic mineral decomposition.
Derivative thermogravimetry reveals hidden shoulders
In a cumulative TGA curve, several overlapping losses can look like one broad downward step. A DTG plot can show local maxima or shoulders where the mass-loss rate changes most strongly. This derivative representation can help researchers select intervals or recognise that a broad pattern contains multiple processes.
But differentiation amplifies noise, and smoothing choices may alter apparent peak positions or the number of small features. A high DTG peak is a rapid rate of mass change with temperature, not necessarily a larger total amount lost. A narrow sharp event can have a tall derivative peak but smaller integrated mass than a broad low peak.
Peak height is not the same as peak area
Imagine two DTG features: one tall and narrow, another shorter but wide across more degrees. The integrated mass loss is related to the area under the signed derivative across a temperature interval, not simply the maximum peak height. A taller DTG feature may correspond to less total lost mass if it occupies a small interval.
An informative article shows the mass remaining before and after each feature, or integrates the derivative appropriately. Do not rank carbon fractions merely by derivative peak height. The difference is the thermal equivalent of confusing a momentary flow rate with the total volume transferred over time.
Heating rate shifts apparent peak positions
If the furnace temperature rises faster, reactions with finite kinetics may reach their most rapid mass-loss stage at different measured temperatures. Temperature gradients within larger or poorly conducting samples can also influence curves. This means peak position is not an immutable molecular thermometer.
When comparing soils, align heating rate, sample mass, pan geometry and atmosphere. If one publication uses a rapid ramp and another a slow ramp, identical chemical components can yield differently shaped curves. A peak at 400 degrees in one method should not automatically be equated to the same material as a peak at 400 degrees in another.
Oxidising and inert atmospheres answer different questions
Air or oxygen can support oxidation of organic material, while inert atmospheres may favour thermal decomposition without the same oxidation reactions. Mineral transformations and gas release can also depend on atmosphere. Some methods use multiple heating segments or a switch in gas conditions to separate classes of processes.
An atmosphere change can alter the interpretation of mass loss at a given temperature. Do not compare a nitrogen-pyrolysis TGA curve with an air-combustion curve as though they shared a fixed temperature-to-carbon fraction map. A strong caption includes atmosphere and ramp conditions with the axes.
DTG sign conventions can reverse a visual conclusion
The derivative of mass with temperature is usually negative during mass loss. Some software plots the negative derivative so loss events appear as positive peaks. Other software displays the raw negative derivative below zero. Without checking the plotted sign, readers may incorrectly describe a loss event as mass gain.
Look for a label such as derivative mass loss or minus dM/dT. Use the units and axis convention in your explanation. If a figure contains both TGA and DTG lines, check which colour corresponds to which axis before reading peak heights. Mixed dual-axis figures are particularly vulnerable to accidental misdescription.
Worked example: separating successive mass-loss intervals
Suppose a hypothetical soil sample begins at 100 milligrams, falls to 95 at 150 degrees, 90 at 550 and 86 at 850. Relative to the original mass, the losses are 5 milligrams in the first interval, 5 milligrams in the second and 4 milligrams in the third. They correspond to 5%, 5% and 4% of starting mass.
The total cumulative loss is 14%, leaving 86%. These arithmetic fractions describe mass changes, not validated chemical allocations. If the first 5 milligrams is mostly moisture, it should not be counted as organic carbon. If the final 4 milligrams includes carbonate CO2, attributing it all to refractory organic matter would be equally mistaken.
Why different percentage denominators matter
If a paper reports loss as a fraction of the original 100 milligrams, the second 5-milligram interval is 5%. If it reports loss relative to the mass remaining after initial drying, the same second interval is 5 divided by 95, or approximately 5.26%. Both can be internally correct, but they do not mean the same percentage.
This denominator issue matters when converting an as-received TGA sample to a dry-matter reference. Record which starting mass each reported interval uses, especially if some stages are normalised separately. Mixing percentages from different denominators can cause a reconstructed cumulative loss to exceed or understate the actual mass change.
TGA can complement independent soil carbon measurements
A soil carbon analyser can provide an organic or total carbon measurement through a distinct method, while TGA shows how the sample responds to heating. Comparing TGA intervals against measured organic carbon, nitrogen, carbonate or clay can reveal useful empirical relationships in a tested dataset.
An original 2004 study, Rapid Screening of Soil Properties Using Thermogravimetry, examined correlations between thermal losses and soil properties across tested mineral soils. Such validation matters because the interpretation is not automatic for organic horizons or all soil types. A reader should distinguish calibrated prediction from direct thermogravimetric measurement.
Thermolabile does not mean short lived in the field
Thermolabile and thermostable describe responses to heating within a method and chosen temperature intervals. They are not exact synonyms for rapidly versus slowly decomposed in an ecosystem. A compound or association that decomposes at a relatively low laboratory temperature might still persist under field conditions because of physical protection or limited microbial access.
Conversely, a high-temperature thermal event does not guarantee that a carbon pool will remain in soil for centuries. Soil carbon persistence depends on environment, biology, minerals, aggregation and transport. A TGA stability index may be useful as an operational descriptor, but it is not an independent measurement of carbon residence time.
What a thermal stability index can summarise
Some investigators use ratios of mass losses from selected temperature intervals to describe relative thermal behaviour. A thermogravimetric index may compare a lower-temperature loss with a higher-temperature loss or place both within a normalised scheme. Its value depends entirely on which bands are defined and the direction of the ratio.
An index increasing in one paper may indicate a different property from an identically named index defined with the inverse formula elsewhere. The 2025 thermogravimetric soil organic matter study relates specific interval losses to soil mineral properties using a stated stability index. Read the formula before interpreting a trend. A numeric index is a compressed description of the heating response; without its components and validation, it is not a direct age or microbial degradation constant.
Clay minerals may lose structural water
Some soil minerals contain hydroxyl groups or other water associated with their crystal structures. Heating can drive dehydroxylation and related transformations that release water at temperatures overlapping organic matter decomposition. The resulting mass loss can appear in an interval that an oversimplified chart labels ‘organic carbon’.
This is why the soil’s mineralogical composition matters. A clay-rich soil and a quartz-rich sandy soil with equal independently measured organic carbon can produce different TGA curves. Mineral controls, X-ray diffraction or other chemical measurements may help identify important transformations. A reader should avoid allocating all mass loss from a mid-temperature window to organic material merely because the paper uses a convenient colour code.
Carbonate decomposition has stoichiometric meaning
When calcium carbonate decomposes under suitable thermal conditions, a simplified reaction is CaCO3 forming CaO plus CO2. The molar mass of CaCO3 is about 100.09 grams per mole, while CO2 is about 44.01 grams per mole. If a measured high-temperature mass loss is securely attributable only to this carbonate reaction, the corresponding carbonate amount can be estimated stoichiometrically.
However, real soil may contain different carbonates, overlapping losses and additional mineral processes. A carbonate-related high-temperature DTG peak supports investigating inorganic carbon; it is not sufficient by itself to guarantee pure calcium carbonate. Any calculated carbonate content should state the reaction assignment and corroborating evidence.
Worked carbonate calculation with the correct denominator
Suppose, purely illustratively, 4 milligrams of sample mass were lost as CO2 from decomposition of CaCO3 alone. The equivalent CaCO3 mass would be 4 times 100.09 divided by 44.01, approximately 9.10 milligrams. The carbon in that released CO2 would be 4 times 12.01 divided by 44.01, approximately 1.09 milligrams. These are not the same quantities.
For a 100-milligram original soil sample, this idealised contribution corresponds to roughly 9.10% CaCO3 by mass and 1.09% inorganic carbon by mass. It would be incorrect to call the 4-milligram CO2 loss 4 milligrams of carbon or 4% calcium carbonate. The chemical mass conversion is a separate step from the observed thermal loss.
Overlapping carbonate and organic reactions prevent a unique shortcut
If two transformations release gas across overlapping temperatures, a broad mass-loss interval cannot be uniquely assigned using TGA alone. For example, the high-temperature region may reflect carbonate decomposition together with mineral hydroxyl changes or later organic reactions, depending on sample and atmosphere. A mathematical partition requires additional information.
Complementary gas analysis, mineralogy or independent inorganic carbon testing may help constrain the components. Without such evidence, use wording such as ‘mass loss in a region consistent with carbonate decomposition’ rather than declaring an exact carbonate mass. A precise temperature boundary drawn on the plot does not automatically separate chemically overlapping reactions.
Loss on ignition uses an operational boundary
Loss on ignition, abbreviated LOI, commonly heats a sample to a chosen temperature for a specified time and calculates mass lost. It can be an inexpensive proxy for organic matter under appropriately calibrated conditions, but the result can include water, mineral and carbonate contributions. A TGA curve is more detailed because it follows mass continuously or in small steps during a heating programme.
An LOI percentage and a TGA interval percentage can differ because of pre-drying, heating atmosphere, duration and sample composition. Neither should be relabelled as organic carbon without an applicable validation. The companion soil organic carbon guide explains why carbon mass, organic matter mass and soil stock are distinct concepts.
Organic matter mass is not organic carbon mass
Organic matter includes elements such as hydrogen, oxygen and nitrogen in addition to carbon. As compounds react during heating, the gases leaving the balance can include more than carbon dioxide. Consequently mass lost during oxidation cannot be assumed to equal the original mass of organic carbon.
A particular empirical conversion from organic matter to organic carbon may be used in some soil contexts, but it is not a universal chemical identity. The most reliable approach is to report independently measured carbon and use TGA for thermal patterns or calibrated predictors. A reader should not turn 6% TGA loss into 6% organic carbon without explaining the chemical assumptions.
Thermal curves are not the same as DSC heat-flow curves
Differential scanning calorimetry, abbreviated DSC, observes thermal energy flow associated with events under a heating programme. TGA observes sample mass change. A reaction can release or absorb heat while also changing mass, but the numerical axes represent different properties. Some combined instruments measure both and produce plots that resemble one another.
The soil microcalorimetry guide addresses heat flow in another measurement context. The units should make the distinction obvious: milligrams or percent mass for TGA, and instrument-defined heat-flow units for DSC or calorimetry. Confusing these produces false conversions from temperature peaks to biological energy generation.
A derivative peak cannot be used as an energy peak
A tall DTG signal means rapid change in mass per temperature or per time. It does not directly tell you whether the underlying reaction was endothermic or exothermic. Heat flow requires an appropriate calorimetric measurement or thermodynamic model. Likewise, a DSC feature need not coincide with a mass change if a process involves a phase transition without volatile release.
When a published figure overlays TGA and a heat-flow signal, read each axis and sign convention separately. Do not use the same verb for the curves. ‘The sample lost mass rapidly’ and ‘the process released heat’ are different observations that may be connected only after further interpretation.
How to reconstruct a cumulative-loss curve
Suppose a sample loses 3% of original mass in the first interval, 7% in the second and 2% in the third, all stated on the same initial-mass basis. Total cumulative loss is 12%, leaving 88% remaining. If the paper plots mass remaining, subtract cumulative loss from 100 rather than stacking the interval percentages upward.
The calculation changes if each percentage is instead relative to the mass entering that interval. For example, losing 7% of the 97% that remained after the first interval removes 6.79% of the original mass, not 7%. This illustrates why authors should always state their denominator and why readers should resist adding percentages from differently normalised reports.
A mass balance exposes missing stages
A complete TGA report should make initial sample mass, final residue and total loss consistent. If a paper claims interval losses summing to 18% but the curve shows only 12% total loss, something about normalisation, boundaries, baseline drift or data reporting needs investigation. The difference is not merely a cosmetic issue.
Checking closure helps locate errors before assigning biochemical meaning. Some intervals overlap or exclude baseline corrections, while derivative peaks may not account for all gradual losses. The correct approach is to reconcile the measurement definitions rather than forcing totals to agree by adding an arbitrary adjustment.
Why a small sample may not represent a field
TGA often analyses relatively small portions of a prepared soil sample. A mineral or organic fragment can influence the local composition, and incomplete homogenisation may increase variability. Replicate furnace runs address measurement repeatability, while independently collected soil cores address spatial variability; the two are not equivalent.
A study comparing management treatments should report independent sampling units and how the subsample was chosen. A smooth representative TGA curve is not proof that an entire landscape shares its profile. Where stones, roots or carbonate nodules are present, sample preparation can change what material reaches the pan. Record those decisions before comparing groups.
Instrument baseline and buoyancy corrections matter
Heating a furnace changes gas density and can influence the apparent force measured by a sensitive balance. Instrument drift, baseline variation, pan materials and sample positioning can affect the trace. Blank runs, calibration standards and corrections are used according to the instrument method to separate these contributions from sample reactions.
A DTG curve can magnify small baseline artefacts because taking a derivative emphasises changes. If a tiny peak appears near a calibration transition or noise region, do not automatically assign it to a rare soil organic compound. A good analyst examines replicate traces and correction procedures before reporting a new chemical feature.
Why derivative smoothing can create or conceal features
Derivative curves are often smoothed to reduce noise. Different smoothing windows can merge close peaks, shift apparent maxima or remove small shoulders. This does not mean smoothing is inherently wrong; it means reported fine structure depends partly on data processing.
A reader should compare the underlying mass-loss curve with the derivative where available. If an article claims four distinct organic carbon fractions solely because software located four derivative maxima, ask whether the peaks were reproducible and independently linked to chemistry. A numerical feature is not automatically a separate ecological pool.
Mineralogy determines which temperature bands are plausible
Quartz-rich sands, clay-rich soils, carbonate-bearing materials and iron-rich horizons can display different thermal behaviours. Water retention, mineral transformations and carbonate content influence the shape even when organic carbon contents overlap. A single temperature partition applied to all soils can therefore create systematic bias.
The 2004 screening study validated useful regressions in its tested mineral soils but did not establish a universal conversion applicable to every organic horizon. The 2025 temperature-loss analysis likewise examines mineral and organic relationships. Readers should check whether the reported model covers the mineralogy of the new samples to which it is applied.
Comparing treatments without confusing cause and correlation
Suppose grassland samples show a larger mid-temperature loss than cultivated samples. The pattern could reflect differences in organic matter, mineral composition, moisture, management history or several interacting factors. TGA can detect a thermal response; assigning management as the unique cause requires suitable experimental design and environmental context.
A cross-sectional correlation between a mass-loss interval and soil carbon can support prediction within tested conditions. It does not prove the material lost in that interval was exclusively carbon or that changing the land use alone caused the signal. Clear educational writing distinguishes association, chemical assignment and causal explanation.
A note on thermodynamic stability and ecological persistence
A sample requiring higher temperatures to lose mass may be described as more thermally stable under that programme. That statement does not necessarily predict its rate of microbial decomposition in ambient soil. Field persistence can arise from restricted access, oxygen limitation, mineral associations and other factors that a dry heating furnace does not reproduce.
The user reading a soil carbon paper should ask what concept the authors actually operationalised. A temperature-based stability index and a radiocarbon-based historical signature are complementary forms of evidence, not interchangeable measurements of a single ‘carbon age’. Read the soil radiocarbon guide to compare thermal response with isotope histories.
Connection to mineral-associated soil carbon
Soil carbon can interact with mineral surfaces, and the proportions of mineral and organic components may influence TGA signals. This creates an opportunity for comparing thermal profiles with measured clay content, oxide composition and carbon pools, but interpretation requires independent measurements and a well-defined sample population.
The POM and MAOM guide explains operational separation of organic matter pools by size and density. TGA of recovered fractions could show different heating responses, yet the fraction name does not automatically explain every DTG peak. Combining two methods strengthens a hypothesis only when neither is treated as a perfect source identifier.
Check which temperature axis and ramp were used
A TGA figure may display temperature against mass, time against mass or a programmed ramp with holds. If the x-axis is time, a sharp transition may coincide with a hold at fixed temperature rather than a narrow intrinsic decomposition temperature. If it is temperature, the heating rate affects the conversion between the two representations.
Write the units after every calculated derivative. A rate expressed in milligrams per degree Celsius differs from milligrams per minute. If the heating programme changes mid-run, conversion between these derivatives may require segment-specific rates. Do not report a peak from a time plot as a temperature derivative without checking the programme.
The same curve can support several reporting questions
A soil TGA dataset might be used to compare water-related loss among samples, study carbonate-related features, screen organic matter properties or examine thermal profiles under different management. These are related but distinct questions. Each needs a defined interval, appropriate calibration and a stated chemical interpretation.
An article should not switch mid-paragraph from ‘fraction of mass lost’ to ‘fraction of soil carbon stored’ without a new measurement or model. The best presentation separates raw curves, calculated interval losses, statistically fitted predictors and subsequent environmental conclusions. That structure helps non-specialists follow how a physical balance becomes an ecological statement.
Independent exercise: is the carbonate interpretation justified?
A sample loses 4 milligrams above 650 degrees Celsius and a student assumes it is all carbon dioxide from CaCO3. Without mineralogical evidence, that chemical assignment is only a hypothesis. If the assignment were independently supported, the illustrative calcium carbonate mass would be approximately 9.10 milligrams and the carbon mass about 1.09 milligrams, not 4 milligrams of carbon.
The student must distinguish the conditional stoichiometric calculation from observed mass loss. This is a broadly transferable lesson: a mathematically correct conversion can rest on an unjustified identification of the thing being converted. First validate the source reaction, then perform the numerical calculation.
Independent exercise: peak height versus integrated loss
Two fictional DTG peaks occur in a derivative graph. Peak A is tall but spans a narrow temperature interval, while peak B is half as tall but extends across five times as much temperature. Is A necessarily associated with greater lost mass? No. Under simple rectangular approximations, B would integrate to a larger area despite its lower maximum height.
Real peaks are not rectangular, but the lesson holds: integration across the temperature range determines total mass change, subject to baseline and overlap choices. The exercise corrects the habit of equating a visually dramatic peak with the dominant soil carbon pool.
The most common mistakes and their repairs
The first mistake is equating all TGA mass loss with organic carbon; repair it by considering water, mineral and carbonate processes. The second is confusing percent remaining with percent lost; repair it by reading the y-axis and computing the complement. The third is treating DTG peak height as total mass loss; repair it by integrating or comparing before-and-after masses.
Another mistake is interpreting a high-temperature peak as ancient or biologically resistant carbon; repair it by checking mineralogy and understanding the heating context. A fifth is comparing curves with different atmospheres and ramps as if their temperature bands were identical; repair it by aligning methods. A sixth is forgetting starting-mass denominators; repair it by reconstructing a complete mass balance.
Frequently asked questions
Is TGA the same as measuring soil organic carbon? No, it measures mass changes during heating. Does every DTG peak identify a unique compound? No, reactions can overlap. Is the low-temperature loss always water? Often important, but not automatically exclusive. Can carbonates produce high-temperature losses? Yes, under appropriate conditions. Is a high-temperature mass-loss interval proof of long carbon residence time? No.
Does a tall DTG peak always represent the most mass? No. Can heating rate move peak temperatures? Yes. Is DSC identical to TGA? No, it tracks thermal energy flow rather than mass. Can TGA screen soil mineral properties? It can within validated methods, not by universal conversion. Must soil preparation be stated? Yes, because drying and grinding affect the curve.
Original research and connected study routes
The 2025 open-access study of soil thermogravimetric mass-loss assignments links observed thermal behaviour to organic matter, extractable fractions and mineral properties in a defined soil set. Siewert’s 2004 soil-property screening study reports independent calibration of thermal-loss profiles against conventional properties for mineral soils. Both show why chemical validation matters before labelling peaks.
Useful comparisons include soil organic carbon and stocks, POM and MAOM size–density separation, soil radiocarbon ages, ATR-FTIR and DRIFT soil spectra and cross-linguistic transfer. The reading route keeps measured mass, inferred chemistry and soil function distinct.
The practical conclusion
A soil TGA curve tells us how mass changed during a known heating programme. DTG shows where that mass changed most rapidly. Those are powerful, repeatable measurements, but assigning losses to water, organic matter, mineral reactions or carbonate decomposition requires context and sometimes independent chemical tests.
A confident reader checks the axes, atmosphere, ramp, baseline, starting-mass denominator and sample mineralogy before attaching ecological meaning to a peak. This prevents the most consequential mistake in soil thermal analysis: presenting a temperature-defined mass loss as a directly measured pool of persistent organic carbon.