Drying, Igniting, and Weighing to Constant Mass

A chemist pulls a crucible from a 900 °C muffle furnace. She lets it cool in a desiccator for exactly 30 minutes. Then she places it on a microbalance that resolves 0.01 mg. The reading is 24.6688 g. She repeats the whole cycle — reheat, cool, weigh — and gets 24.6687 g. That agreement, to within 0.2–0.3 mg, is the operational definition of "constant mass." It is the quiet gatekeeper standing between a beautifully executed precipitation and a trustworthy number. Everything upstream — nucleation, digestion, filtration, washing, all covered under Precipitation Gravimetry — only pays off if the final solid can be converted into something specific. It must become a chemically defined, moisture-free, thermally stable compound of exactly known composition. This step matters because gravimetric analysis is, in the end, an exercise in mass bookkeeping. A chemist weighs a pure, stoichiometrically defined product. Then she works backward to the mass of analyte that produced it. If the "pure, stoichiometrically defined product" is secretly BaSO₄·xH₂O with an uncertain x, the arithmetic downstream is built on sand. The same is true if it is Fe₂O₃ still contaminated with occluded chloride. Or if it is a partially decomposed oxalate. Drying and igniting are the operations that force the precipitate to change. It must abandon its ambiguous, hydrated, ill-defined state. It must adopt a single reproducible chemical identity — one with a molar mass a chemist can look up with confidence. The big picture is thermal. As a wet precipitate is heated, it passes through a sequence of mass-loss events. These include free (occluded) water, adsorbed water, and water of crystallization. They also include decomposition of anions such as oxalate or carbonate. Finally, the precipitate arrives at a thermally stable oxide or salt. A thermogravimetric curve — mass vs. temperature — reveals distinct plateaus. Gravimetric method development consists of finding a temperature plateau that meets two conditions. It must be wide enough, and far enough above the decomposition edge. Then small furnace-temperature drift will not change the composition being weighed. "Constant mass" is the experimental proof that such a plateau has been reached. This article works through the thermal chemistry of common gravimetric precipitates. It covers the operational meaning and statistical logic of "constant mass." It also examines the physics of buoyancy and desiccator equilibration that can silently corrupt a mass reading. Three fully worked examples follow, spanning a routine ignition, an applied ore assay, and an exam-level buoyancy correction. It sets up, but does not compute, the stoichiometric bridge to analyte mass. That arithmetic belongs to Gravimetric Calculations and the Gravimetric Factor.

Key takeaways

Drying (110–160 °C) removes physically held water; igniting (500–1200 °C) removes water of crystallization and decomposes the anion into a stable oxide or salt.
"Constant mass" means two consecutive heat–cool–weigh cycles agree within a pre-set tolerance, typically ±0.2–0.3 mg for a macro precipitate.
The weighing form must have a fixed, known stoichiometry; the drying/ignition temperature must sit on a thermogravimetric plateau, not near a decomposition edge.
Desiccator cooling must be reproducible in both time and desiccant condition, because hygroscopic weighing forms re-adsorb atmospheric moisture measurably within minutes.
Air buoyancy on crucible and precipitate must be considered when the precipitate's density differs greatly from the calibration density of the balance, though for most gravimetric solids the correction is smaller than the reading uncertainty.
Failure to reach a genuine plateau — over-drying a hygroscopic hydrate, or under-igniting a carbonate — is a systematic error that no amount of replicate weighing can detect.
Key idea: a gravimetric result is only as good as the chemist's proof that the weighing form has stopped changing mass — and that proof must come from thermal chemistry, not from wishful stopping.

The core insight

Constant mass is not a weighing ritual — it is experimental evidence that the solid has reached a genuine, reproducible thermogravimetric plateau, so its molar mass is fixed and known.

Where this lives today

Gravimetric drying and ignition to constant mass, though sometimes viewed as a "classical" technique overtaken by instrumental methods, remains a working gold-standard procedure wherever absolute mass-based accuracy is required and cannot be sacrificed for speed. Pharmacopeial assays (USP, EP) still specify loss-on-drying and residue-on-ignition tests for raw pharmaceutical materials, using exactly this dry/cool/weigh-to-constant-mass logic to certify moisture content and inorganic ash content of active ingredients and excipients. Certified reference materials, used to validate every other analytical instrument in a laboratory (ICP-OES, XRF, ion chromatography), are themselves often assigned a certified composition using classical gravimetric ignition as the primary, traceable method — because gravimetry needs no external calibration curve, only a calibrated balance and known atomic weights, making it a "primary" method in metrological hierarchies. Materials science and cement chemistry use loss-on-ignition (LOI) testing, structurally identical to gravimetric constant-mass practice, to quantify volatile content (water, carbonate, organic matter) in clinker, clay, and soil samples — a control test run on every batch of Portland cement produced industrially. Geochemistry and soil science use ignition-based methods to determine organic matter content by mass loss between 105 °C (moisture) and 550–950 °C (organic carbon and carbonate combustion). In environmental and food analysis, "total solids" and "ash content" determinations are direct descendants of this same drying-to-constant-mass philosophy, applied to complex matrices rather than pure precipitates. Even in an era of automated thermogravimetric analyzers that record continuous mass-vs-temperature curves electronically, the conceptual core — find the stable plateau, prove stability by repeated measurement, trust only a mass that has stopped changing — remains unchanged from Berzelius's bench.

Ask the tutor: How do I choose between drying and igniting when I'm not sure whether my precipitate's water content is truly fixed or variable?

The physical picture

When a filtered precipitate leaves its filter paper or filtering crucible, it is a wet, ill-defined solid: part precipitate, part occluded mother liquor, part adsorbed water, sometimes part filter-paper cellulose. The purpose of the drying/ignition step is to strip away everything except a chemically defined weighing form — a compound whose formula, and hence molar mass, is known with the same certainty as a value from a table of atomic masses. Two thermal regimes are used, chosen according to what the precipitate is and what final form is wanted.

Drying (typically 100–160 °C, sometimes up to 250 °C) removes water that is only physically associated with the solid: surface-adsorbed water, and for some hydrates, loosely bound water of crystallization. Drying is used when the precipitate itself, once free of that water, is the desired weighing form — for example, AgCl dried at 110 °C, or the ammonium phosphomolybdate complex dried near 160 °C. Filter papers are not used for these; the precipitate is collected in a fritted (sintered-glass) crucible so that it can go straight into a drying oven — filter paper cannot survive.

Ignition (typically 500–1200 °C, in a muffle furnace) is more drastic. It is required when the natural precipitate form is a gelatinous hydroxide, an oxalate, a carbonate, or any compound whose anion must be thermally decomposed to leave a stable oxide. Classic cases: Fe(OH)₃·xH�2O ignited to Fe₂O₃ near 900–1000 °C; CaC₂O₄·H₂O ignited through CaCO₃ to CaO near 1000 °C (or held at a lower plateau to stop at CaC₂O₄ or CaCO₃, depending on protocol); Al(OH)₃ ignited to Al₂O₃. Ignition is also how the filter paper itself is removed — the precipitate is collected on ashless filter paper, folded and placed in a porcelain or platinum crucible, and the paper is charred and burned away before the furnace ramps to full ignition temperature, so the carbon leaves as CO₂ rather than being trapped as a black carbon residue inside the oxide. The reason a single temperature is not "good enough" without verification is that many of these solids pass through multiple mass-loss events as temperature rises, and the transitions between stable forms are not perfectly sharp. A thermogravimetric analysis (TGA) trace — mass fraction remaining, plotted against temperature — makes this vivid.

Figure

Idealized TGA curve for calcium oxalate monohydrate, CaC₂O₄·H₂O, heated at a controlled rate

The curve shows why a single arbitrary furnace setpoint is dangerous. Igniting at 500 °C might catch the sample mid-decomposition between CaC₂O₄ and CaCO₃ — a temperature where the composition is exquisitely sensitive to exact furnace time and temperature, and no two crucibles would agree. Igniting at 900–1000 °C sits on the broad CaO plateau, where a 50 °C furnace drift changes almost nothing. Choosing the ignition temperature is choosing a flat region of the mass-vs-temperature curve — and demonstrating "constant mass" experimentally confirms the chemist has actually landed on that plateau, rather than assuming it from a textbook temperature that may not match the actual furnace, crucible mass, or air circulation. A second physical subtlety is what happens after the furnace, during cooling. A red-hot crucible cannot go straight to a balance: convection currents from residual heat make the balance reading drift and unstable, and hot gas expansion buoys the pan unpredictably. The crucible is first cooled somewhat in air, then transferred to a desiccator holding a chemical desiccant (anhydrous CaCl₂, silica gel, or Drierite/Mg(ClO₄)₂) to exclude atmospheric moisture while it cools to room temperature. Cooling inside the desiccator, without a moisture-adsorbing solid still hot enough to re-adsorb water vapor as it approaches the last few degrees, must be timed reproducibly — typically 25–40 minutes — because different desiccants and different crucible masses equilibrate at different rates, and many oxide weighing forms (MgO, CaO, Al₂O₃) are themselves mildly hygroscopic and will slowly gain mass sitting in a poorly maintained desiccator.

Figure

Anatomy of the drying/ignition/cool/weigh cycle

The core ideas

The central quantitative statement of "constant mass" is deceptively simple: two consecutive mass readings, taken after separate heat–cool cycles, must agree within an analytically justified tolerance.

|m_{n} - m_{n-1}| \le \delta_{\text{tol}}

where: — $m_n$ is the mass recorded after the $n$ -th ignition/cooling cycle (g) — $m_{n-1}$ is the mass from the previous cycle (g) — $\delta_{\text{tol}}$ is the tolerance, chosen so that it is small compared with the overall mass and with the required precision of the analysis (commonly 0.2–0.3 mg for a macro gravimetric precipitate weighed on an analytical balance good to 0.1 mg)

Reading. This is not a claim that the precipitate mass is now exactly time-invariant — it is a claim that any further mass change, over one more heat/cool cycle, is smaller than the smallest change the chemist's balance and protocol can reliably distinguish from noise. It is an operational, instrument-referenced criterion, not an absolute physical one. The choice of $\delta_{\text{tol}}$ is itself governed by the propagated effect on the final relative uncertainty of the analysis:

\frac{\delta_{\text{tol}}}{m_{\text{ppt}}} \ll \frac{s_{\%}}{100}

where: — $m_{\text{ppt}}$ is the total mass of the weighing form (g) — $s_{\%}$ is the target relative precision of the gravimetric result, in percent (commonly 0.1–0.2% for careful classical gravimetry)

Reading. If a chemist is aiming for 0.1% relative precision on a 0.400 g precipitate (i.e., ±0.0004 g absolute target uncertainty), a mass-constancy tolerance of 0.3 mg is already comparable to the whole error budget — meaning the tolerance itself, not just random weighing noise, becomes a significant contributor to overall uncertainty, and tightening it (or increasing precipitate mass) is often necessary. A second governing relationship concerns the plateau condition underlying why a given ignition temperature is trustworthy at all. If $m(T)$ is the thermogravimetric mass function, a valid weighing-form temperature $T^*$ must satisfy:

\left| \frac{dm}{dT} \right|_{T^*} \approx 0 \quad \text{over an interval } [T^* - \Delta T,\ T^* + \Delta T]

where: — $dm/dT$ is the local slope of the thermogravimetric curve (mass loss per degree) — $\Delta T$ is the expected furnace temperature variability in practice (often ±20–50 °C for a muffle furnace with a single thermocouple)

Reading. A furnace does not hold one exact temperature; it cycles around a setpoint. Any weighing form chosen must sit on a stretch of the mass curve flat enough that this normal thermal wobble does not change the mass — this is the physical reason ignition temperatures for oxides (900–1200 °C) are chosen so generously above their formation edge, rather than at the minimum temperature that works.

Derivation AWhy "two agreeing cycles," not one weighing, is required

Goal. Show why gravimetric protocols insist on repeated dry/ignite–cool–weigh cycles rather than accepting the first mass reading after nominal drying, and derive the logic connecting cycle-to-cycle agreement to confidence that the true equilibrium mass has been reached.

Step 1 — model the approach to equilibrium mass. Treat the precipitate's mass during a drying/ignition step as relaxing toward some limiting "true" weighing-form mass $m_\infty$ via a first-order-like process (moisture diffusing out, or a solid-state decomposition proceeding), so that after a heating period of duration $t$ :

m(t) = m_\infty + (m_0 - m_\infty)\, e^{-t/\tau}

Here $m_0$ is the mass at the start of that heating cycle and $\tau$ is an effective time constant governing how fast the sample approaches $m_\infty$ under the chosen furnace conditions.

Step 2 — express the mass remaining after one full standard heating time $t_1$ . If the standard protocol heats for a fixed time $t_1$ (say, 30–60 minutes), the mass measured is

m_1 = m_\infty + (m_0 - m_\infty)\, e^{-t_1/\tau}

which in general is not $m_\infty$ — there is always some residual offset unless $t_1 \gg \tau$ .

Step 3 — apply a second identical heating cycle starting from $m_1$ . Because the sample is now closer to $m_\infty$ than it was at the start of the first cycle (that is, $m_1$ replaces $m_0$ as the new starting mass), the same functional form gives

m_2 = m_\infty + (m_1 - m_\infty)\, e^{-t_1/\tau}

Step 4 — form the cycle-to-cycle difference. Subtracting,

m_1 - m_2 = (m_0 - m_\infty)\, e^{-t_1/\tau}\left(1 - e^{-t_1/\tau}\right)

This difference shrinks geometrically with each additional cycle (each cycle multiplies the residual offset from $m_\infty$ by the same factor $e^{-t_1/\tau}$ ). So if $|m_1 - m_2|$ is already smaller than the chosen tolerance $\delta_{\text{tol}}$ , it is overwhelmingly likely that $|m_2 - m_\infty|$ is smaller still — the sequence is converging, and the chemist can accept $m_2$ (or $m_3$ , if a third cycle is run to confirm) as an excellent proxy for the true, fully-equilibrated weighing-form mass $m_\infty$ , without ever measuring $m_\infty$ directly.

Step 5 — state the resulting protocol rule. Because a single mass reading, however carefully taken, gives no information about how close $m_1$ is to $m_\infty$ (that requires knowing $\tau$ , which is unknown and precipitate-specific), the only experimentally accessible surrogate for "has equilibrium been reached" is the rate of change between successive readings. This is precisely the constant-mass criterion: repeat until consecutive readings agree within tolerance, then stop.

Go deeper · Derivation B · 1 minReconciling anhydrous vs. hydrate weighing forms — the stoichiometric water-loss check

Derivation BReconciling anhydrous vs. hydrate weighing forms — the stoichiometric water-loss check

Goal. For a precipitate that could plausibly be weighed either as a hydrate (e.g., MgNH₄PO₄·6H₂O, "MAP") or as its ignited anhydrous oxide/pyrophosphate form (Mg₂P₂O₇), derive the relationship a chemist uses to confirm — from measured mass loss alone, without independent structural proof — that ignition has driven the reaction to true completion at the expected stoichiometric endpoint.

Step 1 — write the ignition reaction. Magnesium ammonium phosphate hexahydrate decomposes on strong ignition (≈1000–1100 °C) according to

2\,\text{MgNH}_4\text{PO}_4\cdot 6\text{H}_2\text{O}_{(s)} \xrightarrow{\Delta} \text{Mg}_2\text{P}_2\text{O}_{7(s)} + 2\,\text{NH}_{3(g)} + 13\,\text{H}_2\text{O}_{(g)}

Two moles of hydrate combine (losing ammonia and all water, and formally condensing two phosphate units) to give one mole of magnesium pyrophosphate.

Step 2 — compute the expected fractional mass loss. Molar mass of MgNH₄PO₄·6H₂O is $M_1 = 245.41\ \text{g/mol}$ (Mg 24.305, N 14.007, H₄ 4.032, P 30.974, O₄ 64.00, then 6 H₂O = 108.10 — summing gives 245.41 g/mol). Molar mass of Mg₂P₂O₇ is $M_2 = 222.55\ \text{g/mol}$ (2 Mg = 48.61, 2 P = 61.95, 7 O = 112.00 → 222.56 g/mol; using standard atomic weights gives 222.55). For 2 mol of hydrate ( $2M_1 = 490.82$ g) forming 1 mol of pyrophosphate ( $M_2 = 222.55$ g), the expected fractional mass retained is

f_{\text{retained}} = \frac{M_2}{2M_1} = \frac{222.55}{490.82} = 0.45340

so the expected fractional mass lost on ignition is

f_{\text{lost}} = 1 - f_{\text{retained}} = 0.5466

Step 3 — compare to the experimentally measured mass loss. If a chemist weighs the hydrated precipitate ( $m_{\text{hydrate}}$ ) before ignition and the ignited pyrophosphate ( $m_{\text{ignited}}$ ) after reaching constant mass, the observed fractional loss is

f_{\text{obs}} = \frac{m_{\text{hydrate}} - m_{\text{ignited}}}{m_{\text{hydrate}}}

Step 4 — use agreement (or disagreement) as an independent completeness check. If $f_{\text{obs}}$ matches $f_{\text{lost}} = 0.5466$ within the propagated weighing uncertainty, this is strong evidence that ignition proceeded exactly to Mg₂P₂O₇ and not to some intermediate (partially dehydrated, or partially decomposed) composition — because any incomplete or over-driven reaction would produce a different characteristic fractional loss. This is why gravimetric method-validation protocols often report both the pre-ignition and post-ignition constant masses: constant mass alone proves stability at the endpoint, but the fractional-loss check proves the endpoint is chemically the expected one, guarding against a stable-but-wrong intermediate plateau.

Worked examples

Example 1 — Foundational: recognizing constant mass from a raw weighing log (BaSO₄). A precipitate of BaSO₄, collected on ashless filter paper, is charred and ignited at 800 °C in a porcelain crucible. The crucible-plus-precipitate mass is recorded after each of four successive 30-minute ignition/cool cycles:

Cycle	Crucible + BaSO₄ (g)
1	26.8843
2	26.8809
3	26.8804
4	26.8803

The empty (tared) crucible mass, determined beforehand by the same protocol, is 25.9982 g. Step 1 — apply the constant-mass criterion. The tolerance for a macro BaSO₄ precipitate on a balance readable to 0.1 mg is conventionally taken as 0.3 mg (0.0003 g).

|m_3 - m_2| = |26.8804 - 26.8809| = 0.0005\ \text{g (0.5 mg)} \;\; \text{— exceeds tolerance}

|m_4 - m_3| = |26.8803 - 26.8804| = 0.0001\ \text{g (0.1 mg)} \;\; \text{— within tolerance}

Step 2 — accept the correct cycle. Cycles 3–4 agree within tolerance; cycles 2–3 do not. The chemist accepts $m_4 = 26.8803$ g as the constant-mass value — cycle 4 is the first one that satisfies the criterion against its predecessor, and, because the trend (26.8843 → 26.8809 → 26.8804 → 26.8803) is monotonically decreasing and shrinking in step size, this is consistent with the exponential-approach model derived above, not with random noise. Step 3 — compute the mass of BaSO₄.

m_{\text{BaSO}_4} = 26.8803 - 25.9982 = 0.8821\ \text{g}

Limiting-case check. If cycle 2 had been mistakenly accepted as final (a common student error — stopping too early because "it looks stable"), the reported precipitate mass would be $26.8809 - 25.9982 = 0.8827$ g, a difference of 0.6 mg, or about 0.07% relative — small here, but for a tighter-precision method this exact kind of premature stopping is a classic hidden systematic error, since it is not caught by ordinary replicate weighing (the balance itself reads perfectly reproducibly at each individual weighing).

Example 2 — Applied: iron ore assay via Fe₂O₃, including the ignition rationale. An analyst dissolves 0.5124 g of an iron ore sample, oxidizes all iron to Fe³⁺, and precipitates it as hydrous iron(III) oxide, Fe(OH)₃·xH₂O (a gelatinous, poorly crystalline, highly hydrated precipitate — notoriously variable in water content, x is not a fixed small integer). The precipitate is filtered on ashless paper, and the paper is charred and ignited at 900–950 °C. Step 1 — explain why ignition (not drying) is mandatory here. Fe(OH)₃·xH₂O cannot serve as a weighing form: its water content is not stoichiometrically fixed (x varies with precipitation conditions, aging, and drying humidity), so its molar mass is undefined. Only conversion to a single stable, well-defined stoichiometric oxide removes this ambiguity:

2\,\text{Fe(OH)}_3\cdot x\text{H}_2\text{O}_{(s)} \xrightarrow{\Delta,\ 900\text{–}950\,°C} \text{Fe}_2\text{O}_{3(s)} + (3+2x)\,\text{H}_2\text{O}_{(g)}

Fe₂O₃'s composition is fixed by simple valence chemistry (Fe³⁺, O²⁻) and does not depend on how much water the original gel happened to trap. Step 2 — the constant-mass log. Successive ignition cycles give crucible + Fe₂O�3 masses of 29.7761 g, 29.7714 g, 29.7702 g, 29.7701 g (tare crucible mass 29.2969 g). Consecutive differences: 4.7 mg, 1.2 mg, 0.1 mg. With a 0.3 mg tolerance, the last pair (29.7702 → 29.7701, 0.1 mg) is the first to qualify; the analyst accepts $m_{\text{Fe}_2\text{O}_3} = 29.7701 - 29.2969 = 0.4732$ g. Step 3 — sanity-check against the plateau argument. The 900–950 °C ignition window is chosen because Fe₂O₃ is thermally stable over a wide range (it does not decompose further, nor does it re-hydrate at these temperatures) — so even if the actual furnace fluctuates ±30 °C around the 925 °C setpoint, the mass would not change. Contrast this with attempting to isolate an intermediate iron oxyhydroxide (FeOOH) — such a phase exists only over a narrow temperature band and would be a poor, furnace-sensitive weighing form. Step 4 — foreshadow the next arithmetic step without performing it. The mass fraction of iron in the original ore sample follows from the known stoichiometric ratio of Fe to Fe₂O₃ (2 mol Fe per mol Fe₂O₃) — the gravimetric factor that converts $m_{\text{Fe}_2\text{O}_3}$ into $m_{\text{Fe}}$ and then %Fe in the ore. That conversion, and its formal name, is the subject of Gravimetric Calculations and the Gravimetric Factor; here the essential point is only that 0.4732 g is now a trustworthy number, because it was obtained at constant mass on a genuine thermogravimetric plateau.

Example 3 — Exam/competition level: air-buoyancy correction on a constant-mass reading. A student determines the constant mass of a silica (SiO₂) precipitate, obtained by igniting a gelatinous silicic acid precipitate at 1100 °C, as 0.5217 g on a single-pan analytical balance that was calibrated using stainless-steel standard weights of density $\rho_{\text{wt}} = 8.00\ \text{g/cm}^3$ . The balance was calibrated (i.e., "zeroed against buoyancy") assuming ambient air density $\rho_{\text{air}} = 0.0012\ \text{g/cm}^3$ . The ignited SiO₂ has density $\rho_{\text{SiO}_2} = 2.20\ \text{g/cm}^3$ (amorphous/cristobalite mixture typical of ignited silicic acid). Find the true (vacuum) mass of the SiO₂, and comment on whether the correction matters at the precision of this analysis. Step 1 — recall the buoyancy correction formula. A balance reading $m_{\text{obs}}$ (calibrated against weights of density $\rho_{\text{wt}}$ in air of density $\rho_{\text{air}}$ ) relates to the true mass $m_{\text{true}}$ of a sample of density $\rho_{\text{sample}}$ by

m_{\text{true}} = m_{\text{obs}}\left[1 + \rho_{\text{air}}\left(\frac{1}{\rho_{\text{sample}}} - \frac{1}{\rho_{\text{wt}}}\right)\right]

Step 2 — plug in numbers.

\frac{1}{\rho_{\text{sample}}} = \frac{1}{2.20} = 0.4545\ \text{cm}^3/\text{g}, \qquad \frac{1}{\rho_{\text{wt}}} = \frac{1}{8.00} = 0.1250\ \text{cm}^3/\text{g}

\frac{1}{\rho_{\text{sample}}} - \frac{1}{\rho_{\text{wt}}} = 0.4545 - 0.1250 = 0.3295\ \text{cm}^3/\text{g}

m_{\text{true}} = 0.5217\ \text{g} \times \left[1 + (0.0012)(0.3295)\right] = 0.5217 \times (1 + 0.0003954)

m_{\text{true}} = 0.5217 + 0.5217\times 0.0003954 = 0.5217 + 0.0002063\ \text{g} = 0.52191\ \text{g}

Step 3 — interpret the size of the correction. The correction is +0.0002 g, i.e. +0.2 mg, on a 0.5217 g sample — a relative correction of only +0.040%. On a balance that reads to 0.0001 g (0.1 mg), this correction is about twice the smallest readable digit — non-negligible at the very edge of instrument resolution, but smaller than the 0.3 mg constant-mass tolerance from Examples 1–2, meaning that in typical macro-gravimetric practice this buoyancy correction is usually folded into overall method uncertainty rather than applied explicitly. Step 4 — limiting-case check. If the sample density equaled the calibration weight density exactly ( $\rho_{\text{sample}} = \rho_{\text{wt}} = 8.00\ \text{g/cm}^3$ ), the bracket term vanishes and $m_{\text{true}} = m_{\text{obs}}$ exactly — no correction, as expected, since a balance calibrated against weights of a given density introduces zero net buoyancy bias for a sample of that same density. The correction here is driven entirely by how much lower SiO₂'s density (2.20 g/cm³) is compared to the steel calibration standard (8.00 g/cm³) — light, voluminous solids like ignited silica or gelatinous oxides display the largest buoyancy corrections among common gravimetric weighing forms, precisely because they displace much more air per gram than a dense metal.

Limiting cases, scope & common traps

Situation	What happens	Correct handling
Precipitate is a fixed, well-known low-temperature hydrate (e.g., some sulfates)	Drying alone (100–160 °C) gives a stable, correctly stoichiometric weighing form	Use drying, not ignition; verify the plateau is not a partial-dehydration edge
Precipitate is gelatinous, variable-hydration (Fe(OH)₃·xH₂O, Al(OH)₃·xH₂O, silicic acid)	Water content is not fixed; only ignition to the anhydrous oxide gives fixed stoichiometry	Ignite to a well-known, broad high-temperature oxide plateau (Fe₂O₃, Al₂O₃, SiO₂)
Ignition stopped mid-decomposition (e.g., CaC₂O₄ → CaCO₃ transition zone)	Mass is highly sensitive to exact time/temperature; not reproducible between crucibles	Push ignition well past the transition onto the next stable plateau, or hold deliberately at a validated lower plateau with tighter temperature control
Hygroscopic weighing form (CaO, MgO, some Al₂O₃) cools in a poor desiccator	Sample slowly re-adsorbs atmospheric water during cooling/weighing, mass creeps upward	Use fresh, effective desiccant; weigh promptly; check for monotonic mass gain (opposite sign from normal ignition-drift) as a diagnostic
Filter paper not fully charred before high-temperature ignition	Residual carbon reduces oxide (e.g., converts some Fe₂O₃ to Fe or FeO) or gets trapped as black specks, corrupting stoichiometry	Char slowly at low flame/low furnace temperature with adequate air access before ramping to full ignition temperature
Only one heat/cool/weigh cycle performed	No evidence equilibrium mass was reached; possible large systematic error from premature stopping	Always perform at least two, ideally three, cycles and require consecutive agreement within tolerance
Balance reading unstable immediately after removing crucible from desiccator	Residual convection currents (crucible not fully at ambient temperature)	Allow full cooling time; verify with a blank crucible timing test

The chemist behind it

Analytical gravimetry as a rigorously quantitative discipline owes an enormous debt to the Swedish chemist Jöns Jacob Berzelius, whose systematic gravimetric determinations of atomic weights in the early nineteenth century depended precisely on the drying/igniting/constant-mass discipline this article describes. Berzelius performed thousands of careful precipitations — sulfates, oxalates, phosphates — and, crucially, insisted on reproducible ignition and repeated weighing before accepting a mass as definitive, understanding intuitively (well before formal statistical treatments of measurement existed) that a single weighing proved nothing about whether decomposition or dehydration was complete. His atomic weight tables, astonishingly accurate for the instrumentation of his era, were possible largely because he refused to trust a precipitate's mass until it stopped changing between successive treatments — the same operational discipline codified today as "dry/ignite to constant mass."

Origins · 1 minJöns Jacob Berzelius and how this was discovered

Jöns Jacob Berzelius

1779–1848 · Swedish chemist; pioneer of quantitative gravimetric analysis and systematic atomic weight determination

Berzelius combined obsessive attention to precipitate purity, drying, and reproducible weighing with a broader theoretical program — he coined much of modern chemical notation and formalized the law of definite proportions through gravimetric evidence. His laboratory practice, demanding that a precipitate's mass be shown stable before it was trusted, anticipated by a century the formal statistical logic now taught as "constant mass" verification.

History

- 1803–1818 — Berzelius performs extensive gravimetric determinations of elemental combining weights, refining precipitation, washing, and ignition technique.

- 1828 — Berzelius publishes a comprehensive atomic weight table, largely gravimetrically derived, that remains close to modern values for many elements.

- 1861 — Robert Bunsen's development of the laboratory burner enables more controlled, reproducible ignition temperatures in routine analytical practice.

- 1887 — Introduction of platinum crucibles and improved muffle furnace design standardizes high-temperature ignition in analytical laboratories.

- 1915–1920s — Systematic thermogravimetric studies (Honda and others) formalize the mass-vs-temperature curve concept underlying modern ignition-temperature selection.

Why can't we just dry every precipitate at 110 °C instead of igniting some at 900–1000 °C?

Drying at 110 °C removes only physically adsorbed and loosely bound water; it does not decompose anions like oxalate, carbonate, or hydroxide, nor does it drive off strongly bound water of crystallization in gelatinous hydrous oxides. For precipitates whose natural filtered form already has a fixed, known composition (e.g., AgCl, some anhydrous sulfates), 110 °C drying is entirely sufficient and is in fact preferred, since it is gentler and faster. But precipitates such as Fe(OH)₃·xH₂O, Al(OH)₃·xH₂O, or CaC₂O₄·H₂O have compositions that are not fixed at 110 °C — their water content varies with precipitation history — so drying alone leaves an undefined molar mass. Only high-temperature ignition, decomposing the anion or hydroxide completely into a single well-defined oxide, removes this ambiguity. The choice between drying and ignition is therefore dictated by the chemistry of the specific precipitate, not by convenience.

How is the constant-mass tolerance (like 0.2–0.3 mg) actually chosen, rather than just picked arbitrarily?

The tolerance should be set so that it is small compared to the overall target uncertainty of the gravimetric result, and comparable to (not much larger than) the balance's own reading precision. If the balance reads reliably to ±0.1 mg and the analyst is targeting 0.1–0.2% relative precision on a roughly 0.4–0.5 g precipitate, an absolute uncertainty budget near 0.4–1 mg is implied; a constant-mass tolerance of 0.2–0.3 mg keeps the "have we truly stopped losing mass" uncertainty as a modest, not dominant, contributor to that total budget. Tightening the tolerance further (say to 0.05 mg) usually buys little, since ordinary environmental noise (air currents, static, minor temperature drift of the balance) often sits near that level anyway.

Does the order — dry first, then ignite — matter, or can ignition happen directly from the wet filtered solid?

Order matters for practical and safety reasons. A precipitate that is still visibly wet, placed directly into a hot furnace, can spatter, bump, or even explode gently as trapped water flash-vaporizes, potentially ejecting fine precipitate particles out of the crucible and causing a low, non-recoverable result. Standard practice is to air-dry or gently oven-dry (or, for a filter-paper precipitate, to slowly char the paper at low flame) to drive off bulk water gradually before ramping to full ignition temperature. This staged heating protects both the precipitate's integrity and the crucible/filter paper combination from thermal shock.

If ignition drives off gases like CO₂, NH₃, or H₂O, do we need to correct the crucible's own mass for anything?

The crucible itself is brought to its own constant mass (tared) using the identical drying/ignition/cooling protocol as the loaded crucible, before the precipitate is ever added — so any small, reproducible mass changes intrinsic to the crucible material (loss of adsorbed moisture from the porcelain or platinum surface, for instance) are already captured in the tare and cancel out when the empty crucible mass is subtracted from the loaded, ignited crucible mass. This is why the tare crucible must be ignited under conditions matching the sample ignition — a mismatched tare protocol (e.g., tared at 110 °C but sample ignited at 900 °C) introduces a systematic bias.

Why does a desiccator use a chemical desiccant rather than just letting the crucible cool in open air?

Open-air cooling exposes the hot crucible and its ignited weighing form to ambient humidity for the many minutes it takes to reach room temperature, during which many oxide and salt weighing forms — even non-obviously hygroscopic ones — can adsorb a measurable film of atmospheric water onto their high-surface-area, freshly ignited surface. A desiccator, sealed and charged with a strong desiccant (silica gel, anhydrous CaCl₂, or Mg(ClO₄)₂/Drierite), maintains a very low internal humidity so the sample cools without regaining mass. The desiccant itself must be periodically regenerated or replaced, since a saturated desiccant no longer protects the sample and is a common hidden source of the "mass slowly creeping up" symptom.

Is the buoyancy correction from Example 3 ever actually applied in routine gravimetric work, or is it mostly a teaching exercise?

For most classical macro-gravimetric determinations, the buoyancy correction (typically a few tenths of a milligram on samples of several hundred milligrams to a gram) is smaller than, or comparable to, the constant-mass tolerance and ordinary weighing noise, so it is frequently absorbed into the overall uncertainty rather than applied as an explicit correction. It becomes practically important in high-precision metrological work — certifying reference materials, determining atomic weights, or any situation demanding uncertainties well below 0.01% — where sample density differs substantially from the balance's calibration-weight density and every contribution to the error budget must be individually accounted for.

What's the relationship between "constant mass" here and the later gravimetric-factor calculation?

Constant mass answers the question "is this a trustworthy, well-defined mass?" The gravimetric factor answers a completely different question: "given that trustworthy mass of the weighing form, how much analyte does it represent?" The two are sequential and independent — a beautifully constant, perfectly reproducible mass is still useless for quantitation if the chemist does not know, or gets wrong, the stoichiometric ratio between the weighing form and the original analyte. That stoichiometric bridge, and its formal definition, is the proper subject of Gravimetric Calculations and the Gravimetric Factor.