Electronic-grade methyl tert-butyl ether (MTBE, CAS 1634-04-4) is defined by SEMI C37 and is specified for photoresist stripper duty where trace metal contamination must remain in the sub-ppb range. In front-end semiconductor processing, MTBE is not used as a bulk solvent but as a controlled organic component in resist stripping and cleaning formulations. The material’s solvent profile includes a boiling point of 55.2 °C, a flash point of -28 °C, a density of 0.740 g/mL at 20 °C, and a surface tension near 19.4 mN/m. These properties allow rapid wetting of resist surfaces and efficient evaporation after spin rinsing, while also imposing explosion-proof handling and low-temperature storage requirements. Electronic-grade lots are typically certified to an assay of ≥ 99.8% by gas chromatography using ASTM D5441, a water content below 100 ppm by Karl Fischer titration per ASTM E1064, and individual trace metal concentrations at or below 0.1 ppb to 1 ppb depending on the element. The sub-ppb metals criterion is tied to gate dielectric and interconnect contamination budgets, not to solvent purity in the conventional industrial sense.

What Limits Sub-ppb Metal Control in MTBE Stripper Service?

The trace-metal table in SEMI C37 applies to the final packaged solvent, but the manufacturing path from etherification over acidic ion-exchange resin to distillation, storage, and packaging determines whether sub-ppb limits are sustainable. Iron and chromium can migrate from stainless steel distillation reboilers and transfer lines; sodium and calcium can derive from water used in final washing; aluminum can be introduced from zeolite or alumina drying beds. Electronic-grade producers therefore route distilled MTBE through fluoropolymer-lined or electro-polished stainless steel equipment, followed by sub-micron filtration and packaging in high-purity fluoropolymer or glass containers. Inductively coupled plasma mass spectrometry (ICP-MS) is the standard technique for verifying sub-ppb metal concentrations in an organic matrix. The analytical procedure generally involves controlled evaporation of a 50 g to 100 g sample, acid digestion, and matrix-matched calibration with NIST-traceable standards. Incoming inspection in semiconductor fabs typically sets action limits at 0.1 ppb for sodium, potassium, calcium, iron, copper, zinc, and aluminum, with a total mobile-ion budget below 1 ppb because these elements are known to degrade oxide reliability and alter metal-semiconductor contact behavior. Published round-robin data for this specific SEMI C37 configuration is limited, but the analytical approach is consistent with ultra-trace organic solvent analysis in the semiconductor industry.

Photoresist stripper duty subjects MTBE to conditions that differ from analytical or laboratory use. In immersion batch benches and single-wafer spray processors, MTBE is commonly blended with polar aprotic solvents, glycol ethers, and alkanolamines to remove positive-tone novolak, polyhydroxystyrene, and chemically amplified resist films after etch or ion implantation. The solvent’s low surface tension and low viscosity improve penetration into undercut resist geometries and high-aspect-ratio lines, while its aprotic ether structure promotes swelling of nonpolar resist matrices. However, MTBE alone does not remove carbonized or heavily silylated resist surfaces. Process experience on manufacturing tools shows that stripper bath life is limited primarily by dissolved resist solids and water ingress rather than by MTBE conversion. Once water exceeds approximately 0.5 wt%, the blend can lose aggressiveness and promote corrosion of exposed aluminum or copper interconnects. Filtration through 0.05 μm to 0.1 μm PTFE cartridges removes particulate resist debris but not soluble metal ions; therefore, sub-ppb metal control must be maintained by raw-material purity and point-of-use plumbing.

When Batch Immersion and Single-Wafer Spray Tools Use MTBE Blends

Equipment selection changes the risk profile for MTBE-containing stripper formulations. In batch immersion tools, the solvent is heated to 50 °C to 70 °C in a quartz or PTFE bath, and the high vapor pressure of MTBE requires local exhaust and nitrogen blanketing to maintain the headspace below flammable limits. In single-wafer spray tools, the solvent is dispensed through PFA tubing and a point-of-use filter, then spun off at high rotational speed. The flash point of -28 °C and lower flammability limit of 1.6 vol% in air require that spray chambers be purged with nitrogen before processing and that tool interlocks monitor oxygen concentration. Thermal degradation or peroxide buildup can produce acidic byproducts that raise the solvent’s metal extraction capacity from stainless steel or immersed components. For that reason, many fabs specify MTBE with low peroxide content and avoid over-drying with active alumina because acidic alumina can catalyze tert-butyl ether cleavage. The batch-to-batch variance in electronic-grade MTBE is most often observed in water content and trace levels of isobutylene, methanol, and tert-butyl alcohol, which are monitored by GC per ASTM D5441 and by Karl Fischer titration per ASTM E1064.

MTBE-containing strippers are used in metal lift-off, dry-film removal, and post-ash residue dissolution. In lift-off processing, the solvent must undercut the sacrificial photoresist pattern without swelling the underlying metal or dielectric feature. Swelling can produce metal flagging or pattern collapse. MTBE’s low polarity and high vapor pressure provide fast drying, but this also means that dwell time on the wafer is short in spray tools; therefore, process engineers compensate with temperature and flow control. In post-ash residue removal, MTBE alone is not effective for inorganic halogenated residues, and it is blended with polar solvents and amines to remove organometallic byproducts. The solvent’s sub-ppb metals certification is most valuable in front-end processes where the wafer surface is bare silicon, silicon oxide, or silicon nitride and any metallic contamination becomes electrically active. In back-end processes, the same purity requirement is relaxed but not eliminated because metal residues on copper/low-k interconnects can increase leakage current and reduce time-dependent dielectric breakdown lifetimes.

Peroxide Control, Storage Stability, and Incompatibility Boundaries

MTBE undergoes slow autoxidation to form methyl tert-butyl peroxide and di-tert-butyl peroxide when exposed to air, heat, or ultraviolet light. Semiconductor-grade MTBE is generally not inhibited with phenolic antioxidants because residual inhibitor contributes nonvolatile residue and may alter photoresist stripping selectivity. Instead, storage under nitrogen with headspace oxygen below 0.5% by volume and temperature maintained below 25 °C suppresses peroxide formation. Peroxide test kits and iodometric titration are used at receipt and before use; action limits commonly range from 10 mg/kg to 50 mg/kg, depending on the tool and process. The solvent must not be combined with strong oxidizing acids, peroxide generators, or unpassivated reactive metal surfaces. Because MTBE has a relatively high vapor pressure of approximately 27 kPa at 20 °C, closed containers are required to prevent evaporation and moisture uptake. The solubility of water in MTBE is approximately 4.8 g per 100 g at 20 °C; water ingress above specification can be detected quickly by Karl Fischer titration and should trigger solvent replacement rather than on-site distillation, because fab-level distillation cannot reliably restore sub-ppb metal levels.

Incoming quality control for photoresist stripper duty frequently combines gas chromatographic assay with ICP-MS metal analysis and particle counting. A typical sampling protocol draws solvent from the container under nitrogen, measures water immediately, and retains a sub-sample for metal testing. Direct organic injection into an ICP-MS equipped with oxygen addition and a platinum-skimmer cone can be used for soluble metals, but the more conservative evaporation-acid digestion approach provides lower detection limits for sodium, potassium, calcium, iron, copper, zinc, aluminum, chromium, nickel, lead, and titanium. Method detection limits are typically between 0.005 ppb and 0.05 ppb in the original solvent when using class 100 cleanroom sample preparation. SEMI C37 does not eliminate the need for fab-level incoming inspection because contamination can occur during container filling, shipping, and dispensing. Particle specifications such as ≤ 25 particles/mL at 0.5 μm and larger are verified by optical particle counters, and any statistical outlier requires the lot to be re-qualified before release to the wet bench.

The following table summarizes representative electronic-grade MTBE certification parameters for photoresist stripper duty; the controlling limits are those in the current edition of SEMI C37.

Trace Metal Removal Cannot Be Reestablished by Point-of-Use Filtration Alone

Although the solvent may leave the manufacturer at 0.1 ppb metal levels, point-of-use contamination can occur in transfer lines, filters, and dispensing equipment. Stainless steel wetted parts are common sources of iron, chromium, and nickel; thus, electronic-grade MTBE distribution systems use high-purity fluoropolymers such as PFA and PTFE, with 0.05 μm polishing filters at the tool inlet. Filtration alone does not remove dissolved metal cations; it removes particulate metal oxides and abraded polymer particles. Dissolved metal control requires pre-validated containers, sealed couplers, and inert gas blanketing. Facilities that run MTBE in photoresist stripper recirculation loops monitor metal accumulation by sampling at multiple points, including the bulk reservoir, the filter housing, and the wafer contact area. When copper, zinc, or iron concentrations exceed 0.5 ppb, the bath is often drained or treated with ion-exchange resin beds because sub-ppb metal excursions on post-chemical mechanical planarization wafers can alter oxide breakdown voltage and metal semiconductor work-function characteristics.

In high-volume manufacturing, MTBE-based stripper formulations are monitored for etch rate on blanket resist wafers, residue count after spin rinsing, and metal contamination on silicon monitor wafers. The process control limits are typically set by the fab’s yield model rather than by the solvent supplier. For example, a post-strip metal specification might require less than 1×10^10 atoms/cm² of mobile ions on the wafer surface, verified by vapor phase decomposition ICP-MS or total reflection X-ray fluorescence. MTBE contributes to this surface cleanliness only when its own metal content is held below the incoming specification and when the stripping tool’s wetted surfaces are compatible. Polished stainless steel wetted parts can release iron and chromium at levels that exceed sub-ppb solvent limits; therefore, PTFE, PFA, and CTFE wetted components are preferred. The operational boundary for MTBE in photoresist stripper duty is defined by flammability, peroxide formation, and metal extraction from bath hardware, not by solvency alone. These factors determine the maximum bath turnover interval, the nitrogen purge rate, and the point-of-use filtration configuration.