Take 40 nm as the dividing‑line. For mature process nodes 40 nm and above, batch‑tank cleaners are preferred. For advanced nodes below 40 nm, single‑wafer cleaners are mandatory. Wafer size determines equipment platform; throughput and budget are secondary factors. Once process node and wafer size are confirmed, most selection disputes can be resolved.

Core Comparison Table

Why 40 nm is the dividing‑line

The 40 nm threshold comes from the physical limit of particle control. Batch‑tank cleaning relies on megasonic cleaning for mass particle removal. When device feature size shrinks, fluid boundary‑layer interference occurs among densely‑arranged wafers, creating cleaning dead zones and particle re‑adhesion. At the 28 nm node, a single 15 nm particle can break down gate oxide and cause full‑wafer scrap. Single‑wafer processing treats each wafer independently with dedicated drain, fundamentally eliminating cross‑contamination. In 2025, single‑wafer cleaners reached 68 % penetration in 300 mm advanced fabs.

Reverse application scenarios

For power devices, analog chips, MEMS, SiC/GaN compound semiconductors and wafer reclaim lines, batch‑tank cleaners remain the optimal choice. Take Valin CSE batch‑tank megasonic cleaner as example: it supports silicon, GaAs, SiC and other substrates; 50‑wafers per batch; particle performance: particles ≥0.2 μm less than 10 per wafer. Widely deployed for LED epitaxy and MEMS production. Deploying single‑wafer cleaners for mature‑node processes leads to unnecessary capital expenditure.

5‑Dimensional Evaluation for Procurement Engineers

1. Process Node Matching: Key metrics: Particle Removal Efficiency (PRE), Defect Reduction Capability (DRC). Advanced tools require PRE/DRC ≥99.99 %.

2. Throughput Calculation: Single‑wafer throughput = Number of chambers × (1 / per‑wafer processing time) × equipment utilization. Typical 12‑chamber ~ 20‑chamber 12‑inch tools deliver 150‑300 wafers per hour.

3. Process Compatibility: RCA standard clean (SPM / DHF / SC‑1 / SC‑2) as baseline. For special processes: high‑temperature SPM, IPA‑Marangoni drying.

4. Supplier Assessment: Spare‑part response target ≤24 hours. Domestic reference: Valin CSE has delivered >3000 wet‑processing sets; its supporting Chemical Delivery System (CDS) has over 200 installations in 14‑28 nm fabs.

5. Total Cost of Ownership (TCO): Capital procurement accounts for only 40‑50 % of TCO. Chemicals, spare parts and maintenance constitute the majority. Reserve 15‑20 % extra budget for supporting utilities including CDS and UPW (Ultra‑Pure Water).

Decision‑Making Workflow (5‑question check list)

1. Node < 40 nm? → Select single‑wafer cleaner.

2. 8‑inch fab? → Prioritize batch‑tank; add single‑wafer for special layers.

3. 12‑inch fab for 90‑130 nm? → Evaluate hybrid solution: batch‑tank rough clean + single‑wafer fine clean.

4. Budget < 10 million RMB? → Batch‑tank or entry‑level domestic single‑wafer unit.

5. Need flexible capacity expansion? → Prefer dual‑platform supporting both 8‑inch & 12‑inch wafers.

Three Common Misconceptions

1. Blind preference for single‑wafer tools: For mature nodes, single‑wafer accuracy brings no yield improvement, while capital expense (Capex) surges 3‑5×.

2. Focus only on unit price, ignore chemical consumption: Single‑wafer chemical consumption is 3‑5× higher than batch‑tank. Request vendor data for chemical‑saving designs such as fixed‑discharge & fixed‑replenishment.

3. Skip process validation: Even identical models may show 2‑3 percentage‑point PRE deviation in different fab environments. Perform small‑lot validation with your actual product. Real test data outweigh vendor presentations.

Frequently‑Asked‑Questions (FAQ)

Q1: Can batch‑tank be used for 40 nm?

A: Permitted only for non‑critical layers. Critical layers such as gate and contact hole must adopt single‑wafer cleaning.

Q2: What is the performance gap between domestic and imported batch‑tank cleaners?

A: Domestic tools can mostly replace imports for conventional RCA cleaning for ≥90 nm processes. Gaps exist in long‑term stability of high‑grade quartz tanks and maturity of special modules such as Marangoni drying. Domestic batch‑tank tools deliver high cost‑performance for LED, MEMS and power‑device manufacturing.

Q3: How to verify vendor process capability?

A: Request three sets of documents: mass‑production DRC data from third‑party fabs; small‑lot PRE report using your similar product; ≥12‑month Equipment Overall Efficiency (OEE) track record (OEE ≥85 %).

Q4: Can legacy 8‑inch production lines be upgraded?

A: Retrofitting batch‑tank into single‑wafer configuration costs 60‑70 % of new‑tool purchase. If remaining fab lifetime >5 years: direct replacement is advised. If remaining lifetime <3 years: keep existing batch‑tank equipment for better economy.

Process node defines equipment type; wafer size defines platform; throughput defines chamber count; budget defines brand. Always run real‑product process validation and make decisions based on measured data.