Chemical Composition Analysis of Borosilicate Glass 3.3

Borosilicate glass 3.3 contains 80–81% silicon dioxide, 12–13% boric oxide, and controlled sodium, aluminium and alkaline-earth oxides. Verify it by XRF, ICP-OES and wet chemistry, after HF digestion or borate fusion. This confirms compliance with the 3.3 × 10⁻⁶ K⁻¹ expansion grade for laboratory glassware, pharmaceutical packaging and hermetic seals.

Principle & mechanism

The behavior of borosilicate glass 3.3 follows from its oxide architecture. Silicon dioxide builds the continuous silicate network and confers chemical resistance. Boric oxide lowers the expansion coefficient and improves thermal-shock endurance. Sodium oxide acts as a network modifier. It cuts melting temperature but degrades hydrolytic durability when excessive. Aluminium and alkaline-earth oxides stabilize the network and suppress devitrification. The 3.3 designation names the expansion coefficient, not a compositional figure, so conformity must be proven analytically. Two physical principles dominate the measurement. In X-ray fluorescence, primary X-rays eject inner-shell electrons. The emitted characteristic fluorescence is converted to oxide concentrations through calibration and matrix correction. In ICP-OES, an acid digest is nebulized into an argon plasma and quantified against element-specific emission wavelengths. Classical titrimetric and gravimetric procedures operate on comparable digests. They serve as referee checks when results are disputed.

Sample

Preparation — HF Digestion & Borate Fusion
Both preparation routes start from a representative, finely ground glass powder. For HF digestion, the powder is treated with hydrofluoric acid, usually with nitric acid, in closed PTFE vessels. Microwave or hotplate heating continues until the silicate lattice decomposes. Excess fluoride is then bound by adding boric acid or mannitol. This also limits volatile boron loss as BF₃. The clear digest is diluted for plasma analysis. Certified reference materials of comparable matrix should run in parallel. This keeps recovery and blank levels traceable. For XRF, borate fusion is preferred: the powder is blended with lithium tetraborate or a tetraborate–metaborate mixture, fused above the melting point and cast as a homogeneous bead, often with a non-wetting agent. Fusion removes particle-size and mineralogical effects. Bead mass-to-flux ratio and ignition loss must be recorded so concentrations are normalized correctly.

Elemental

Quantification — XRF, ICP-OES & Wet Chemistry
Each technique covers a defined part of the oxide panel. XRF on the fusion bead quantifies silicon, aluminium, sodium, calcium, magnesium, iron and titanium within one measurement cycle. Boron sits at too low an X-ray energy for reliable direct determination. B₂O₃ is therefore obtained by ICP-OES, by acid–base titration or by difference. ICP-OES on the HF digest determines silicon, boron, sodium, potassium, aluminium and the trace oxides. Internal standardization with scandium, yttrium or rhodium solutions corrects drift. Wet chemistry retains referee value. Gravimetric silicon dioxide after HF volatilization and mannitol-assisted titration of boric oxide anchor results in contractual disputes. This dual-technique logic mirrors Ca/P Ratio Analysis of Hydroxyapatite per ISO 13779, where ICP-OES and XRF data are cross-checked rather than treated as interchangeable. Calibration standards, wavelength selection and spectral-interference checks must be documented. This ensures the reported oxide sum closes against the specification.

Sensitivity,

Linear Range & Repeatability
Performance should be demonstrated per instrument and scope, not assumed. ICP-OES delivers detection limits at the microgram-per-litre level in the digest. Its linear range spans several orders of magnitude. Internal standardization tightens repeatability. The usable figure for the glass depends on the dilution factor applied. XRF reaches detection limits in the low milligram-per-kilogram range for minor elements. It sustains a wide working range for major oxides through the fusion calibration. Repeatability is verified by duplicate beads, duplicate digests and periodic re-measurement of a control glass. Agreement between the two preparation routes is a stronger quality statement than any single numerical claim. Control charts, inter-method reconciliation and periodic proficiency checks should run as routine. Acceptance limits for bias and scatter belong in the written test plan, not in post hoc judgement after results appear.

Co-Testing

Hydrolytic Resistance & Thermal Expansion
Composition results gain meaning when read beside the two defining performance tests. Hydrolytic resistance is assessed on glass grains boiled in water. The released alkali is titrated to assign a hydrolytic class. A harsher autoclave vapor test probes the same surface-durability balance. It follows the same titration logic already used in wet chemistry. Thermal expansion is measured by dilatometry. The mean coefficient over 20–300 °C should reproduce the nominal 3.3 × 10⁻⁶ K⁻¹. The three results are causally linked. Alkali drift shifts the hydrolytic class, while the silica–borate balance moves the expansion coefficient. This linkage is decisive in failure work. Hermetic glass-to-metal seals in relays, connectors and sealed cells rely on matched expansion. Connector and Relay Failure Analysis Testing therefore pairs dilatometry with composition data to separate material causes from assembly causes.

Application

Scenarios & Acceptance Criteria
The analysis serves incoming inspection of tubing and formed ware, process control in melting plants, verification after raw-material or supplier changes, and pharmacopoeial classification of type I packaging glass. Acceptance is expressed as oxide ranges. Silicon dioxide and boric oxide must fall inside the specified nominal bands. Sodium oxide, alkaline-earth oxides and trace oxides must stay within their stated maxima. The analytical total should close against the specification, with boric oxide reconciled by the independent method. These bands are far narrower than those in routine Flat glass testing, where soda-lime compositions tolerate much higher alkali levels. Performance acceptance runs in parallel. The grain and vapor hydrolytic classes must match the declared class. The dilatometric coefficient must stay within the stated tolerance around 3.3 × 10⁻⁶ K⁻¹. Where sealed energy-storage cells are involved, the same data set feeds Battery Failure Analysis Testing when leakage or durability loss is investigated. For pharmaceutical use, organic surface residues such as phthalates are assessed separately. GC-based procedures of the kind applied in diethyl phthalate detection are used.