Mtbe Supplier For Electronic Grade Sub Ppb Metals And Semi C37 Evidence
Evaluation of a methyl tert-butyl ether supplier for electronic grade: sub-ppb metals and SEMI C37 evidence begins with the recognition that MTBE (CAS 1634-04-4) is a volatile peroxide-forming solvent with a boiling point of 55.2 °C and a flash point of -28 °C. In semiconductor manufacturing, the solvent is used in photoresist thinning, edge bead removal, and solvent-based cleaning where residual metal ions after evaporation create direct yield loss mechanisms. A certificate of analysis reporting 99.9% assay is insufficient; qualification requires lot-specific trace-metal data at sub-ppb levels and a SEMI C37 evidence package that connects the analytical result to the container, the packaging line, and the preserved sample.
The purification route for electronic-grade MTBE typically starts with C4 raffinate-derived solvent that is washed to remove water-soluble acids, then distilled in a dedicated stainless steel or glass-lined column. Because MTBE forms a minimum-boiling azeotrope with water, simple distillation leaves low-level water and polar impurities. The separation of MTBE from water is a process conflict because the minimum-boiling azeotrope contains approximately 3.5 wt% water at atmospheric pressure and boils near 52.6 °C; the purified solvent leaving a conventional distillation column still retains water at azeotropic composition. To reach water at or below 50 ppm, the distillation overhead is passed through a 4A molecular sieve bed with bed regeneration at 250 °C under nitrogen, or a pressure-swing adsorption loop. Published data for the specific water breakthrough curve in electronic-grade MTBE drying is limited; equipment sizing is usually confirmed by moisture analyzer challenge testing rather than literature correlations.
What Does Sub-Ppb Metal Certification Actually Require?
Sub-ppb means individual metal concentrations below 1 µg/kg, but for front-end-of-line processes the practical requirement is a method reporting limit at or below 0.05 ppb for alkali metals and transition metals. Sodium, potassium, calcium, iron, copper, and zinc are frequently specified at ≤0.1 ppb each; aluminum, chromium, nickel, manganese, tin, and lead may be specified at ≤0.05 ppb because of their impact on gate oxide integrity and carrier lifetime. Verification is performed by inductively coupled plasma mass spectrometry (ICP-MS). Sodium and potassium require cold plasma or triple-quadrupole operation to reduce argon-based interferences at mass 23 and mass 39. A supplier's SEMI C37 evidence package includes the raw ICP-MS data, method blank, spike recovery, and calibration traceability to NIST SRM 3100 series standards.
| Parameter | Method | Electronic-grade control level |
|---|---|---|
| Assay as MTBE | GC-FID, capillary column with internal standard | ≥99.9 area% |
| Water | ASTM D6304-16, Karl Fischer coulometry | ≤50 ppm |
| Peroxides as H₂O₂ | ASTM E298-17a, iodometric titration | ≤1 ppm |
| Nonvolatile residue | ASTM D1353-13 | ≤2 ppm |
| Na, K, Ca, Fe, Cu, Zn | ICP-MS, cold plasma | ≤0.1 ppb each |
| Li, Mg, Al, Cr, Mn, Ni, Co, Pb, Sn | ICP-MS, reaction/collision cell | ≤0.05 ppb each |
| Particles ≥0.2 µm | Laser particle counter, ISO 21501-1 calibration | Per SEMI C37 particle ceiling |
The reliability of sub-ppb metal data is controlled by the sampling path more than by the analytical instrument. A 1 L PFA bottle rinsed with the same solvent lot and filled under nitrogen contributes less than 0.005 ppb of common metals if the bottle has been leached in 10% nitric acid and dried in an ISO 14644-1:2015 Class 4 environment. Stainless steel fittings are excluded from electronic-grade loading loops unless they are 316L electropolished and passivated, and even then they are used only where mechanical strength is required. The supplier should provide the loading manifold bill of materials and cleaning log because cross-contamination from a previous solvent transfer is a common root cause of lot failure that does not appear in the bulk tank certificate.
On a 300 mm wafer track, MTBE is dispensed through a high-purity pump with shot-to-shot variation below 1.5%. The solvent's surface tension of approximately 19 mN/m at 20 °C allows wetting of high-aspect-ratio structures, but the same wetting behavior carries dissolved metal ions into recessed areas where evaporation leaves nonvolatile residues. Post-dispense exhaust must maintain a face velocity above 0.3 m/s across the cup to prevent solvent vapor from condensing on chamber walls and redepositing metal salts. Wafer inspection after MTBE contact is typically performed on a Surfscan SP5 or equivalent; lots are held or rejected when particle adders for ≥0.09 µm exceed 100 per wafer.
Edge bead removal requires a solvent that dissolves uncured photoresist at the wafer edge without lifting the bulk resist. MTBE's Hansen solubility parameters place it within the solubility envelope of many phenolic and acrylic resists, but the same solvency means that dwell time must be limited to 2–5 s at ambient temperature to avoid undercutting. A dispense nozzle positioned at 1.0 mm from the edge with a 0.2 mm orifice produces a solvent stream that removes the bead without generating liquid aerosol. Metal contamination in this operation is critical because the edge is a high-voltage and leakage region; a single 0.1 ppb sodium spike can shift flatband voltage on test structures.
When Peroxide Formation Limits Shelf Life of High-Purity Solvent
Because MTBE contains a tertiary carbon adjacent to an ether oxygen, exposure to air generates peroxides through an autocatalytic radical pathway. Electronic-grade specification limits peroxides to ≤1 ppm as H₂O₂ by iodometric titration; however, the induction period for peroxide accumulation in vented storage can be as short as 30 days. Storage under nitrogen with headspace oxygen below 1 vol% is required to maintain peroxide values below specification over a 24-month shelf life. The container must be grounded during transfer because MTBE's low electrical conductivity permits static accumulation; a 0.45 µm fluoropolymer filter in the dispense line prevents particle introduction during decanting. No inhibitor such as BHT is added to electronic-grade MTBE because inhibitor residues and their reaction products exceed the nonvolatile residue specification.
Elastomer compatibility is limited to PTFE, PFA, FEP, and Kalrez; Viton, EPDM, nitrile, and neoprene swell or extract organic and metallic impurities into the solvent. Transfer pumps with aluminum bronze or brass wetted parts are not used for electronic-grade service because redox-active metals accelerate peroxide decomposition and contribute ion contamination. If released, vapor can travel to ignition sources because the lower explosive limit is 1.6 vol% and the upper explosive limit is 8.4 vol%.
A supplier's lot-specific SEMI C37 evidence package is reviewed before the solvent is released to manufacturing. The package should include the certificate of analysis, ICP-MS trace-metal data for at least 12 elements, water and peroxide results, nonvolatile residue, particle counts, and a packaging statement identifying the cleaned fluoropolymer-lined shipping container. Retained samples stored at 5 °C under nitrogen for 24 months support root-cause investigations if a lot is linked to wafer metal contamination. Published data for point-of-use stability of MTBE in photoresist thinning formulations is limited; therefore, each fab must verify that the solvent meets the same metal limits after 30 days of recirculation in the dispense system.
SEMI C37 Evidence Package Expectations Without Ambiguity
| Evidence item | Required content | Acceptance criterion |
|---|---|---|
| SEMI C37 compliance statement | Specification revision, lot number, retest date | Current revision specified in purchase order |
| ICP-MS metal analysis | Each required element, method detection limit, blank, spike recovery | MDL ≤ 1/3 of specified limit; spike recovery 80–120% |
| Water | Karl Fischer coulometric result | ≤50 ppm |
| Peroxides | Iodometric titration result | ≤1 ppm as H₂O₂ |
| Nonvolatile residue | ASTM D1353-13 result | ≤2 ppm |
| Particle count | Laser particle counter data | Per SEMI C37 particle ceiling |
| Packaging log | Fluoropolymer liner leach data | Cation/anion values below specified limits |
| Change notification | Feedstock, column turnaround, sieve replacement | Requalification after any change |
Residual tert-butyl alcohol and methanol are the principal oxygenate impurities. Because tert-butyl alcohol has a boiling point of 82.4 °C and methanol 64.7 °C, they are removed by distillation; however, polar oxygenate impurities can preferentially adsorb onto molecular sieves and then desorb during regeneration if the regeneration cycle is not complete. Therefore, the electronic-grade supplier must specify the regeneration endpoint as a dew point or moisture analyzer reading rather than time alone. A supplier that cannot provide lot-specific evidence for metal content, water, peroxide, and particles should not be qualified for front-end-of-line solvent use, regardless of bulk assay.