Frequently Asked Questions About Pure Water/Ultrapure Water Systems

Frequently Asked Questions About Pure Water/Ultrapure Water Systems

This document primarily addresses and discusses issues encountered in the daily design, construction, and operation of pure water/ultrapure water systems. It therefore provides highly practical guidance for resolving common challenges—a crucial point!

1. Q: How is the nitrogen gas requirement for ultrapure water nitrogen-sealed tanks calculated?
A1: When no data is available, nitrogen consumption for ultrapure tanks is estimated based on water usage, though actual gas consumption is lower than this value. Two scenarios exist: constant-level control and variable-level control.
A2: Maximize gas supply to account for extreme conditions. Assume tank volume V. At high liquid level, water production stops while water points operate at full load until low liquid level is reached, assuming maximum water flow rate Q1. Then, water usage points cease, and water production operates at maximum capacity, assuming maximum production flow rate Q2. This scenario represents maximum gas consumption, with one cycle duration T = V/Q1 + V/Q2. The instantaneous gas supply flow rate only needs to reach Q1. Typically, an air storage tank is installed, and the average gas supply can reach V/T, where the average gas supply is V/T multiplied by the safety factor.

2. For ultrapure water systems, how should material grades be selected for metal tanks and vessels during initial metal ion leaching in downstream treatment stages like polishing mixed beds and related equipment?
A: Downstream treatment typically involves TOC removal and polishing mixed beds. TOC units use 316L stainless steel, while polishing mixed beds employ FRP lined with PVDF.

3. When producing ultrapure water, does the Cold Frost moderator favor mixed beds or EDI?
A: This depends on the industry. The solar industry predominantly selects EDI treatment, while the thermal engineering industry generally opts for mixed-bed treatment.
A2: Personally, I favor EDI because it represents a cleaner process.
A3: Currently, larger-scale solar production still predominantly uses mixed-bed systems for pure water, primarily due to lower initial equipment investment costs, making it more acceptable to owners. The same applies to pure water for semiconductor production. For ultrapure water in 8-inch and larger semiconductor wafer production, extremely stringent requirements exist for metal ions, silicon, TOC, dissolved oxygen, and particulates. Therefore, process determination and final equipment/piping selection are critical. Equipment linings and piping are almost universally PVDF or even PTFE!

4. For polishing mixed beds, which offers better cost advantage: FRP lined with PVDF or stainless steel? Also, given the high pressure at the terminal end, is FRP suitable for withstanding such pressure? If stainless steel is used, how would initial leaching of Zn, Ca, Na, etc., be managed?
A: FRP lined with PVDF is slightly more expensive than stainless steel. Stainless steel requires polishing-grade treatment, with all processing steps completed before shipment to ensure no metal ion leaching during use.

5. What resin should be used in laboratory ultrapure water systems to achieve ultrapure water? How much water can one liter of resin process?
A: Ultrapure water quality depends on your source water; there's no universal standard for defining “ultrapure.” Additionally, the processing capacity per liter of resin varies based on the resin type. No resin in the world is that miraculous: turning tap water into 18.2 MΩ ultrapure water. .....

6. At what terminal water quality level should terminal UF be used instead of terminal filtration? Or what are the advantages of terminal UF over filter cartridges?
A: Though terminal UF is very expensive, it's primarily used to remove particles. When extremely stringent particle requirements are needed, terminal UF becomes essential because filters simply cannot meet such high standards. Based on my experience, when 0.1μm particles must be below 10 pcs/L, terminal UF must be designed.

7. Are there specific membrane material requirements for terminal UF? Can standard UF membranes used in pretreatment suffice? I see many systems using Asahi Kasei membranes.
A: Terminal UF typically uses polysulfone (PSF) membranes. Pretreatment UF membranes have too high a molecular weight cutoff and are unsuitable for terminal processing. Asahi Kasei offers UF membranes for both pretreatment and terminal applications.
A1: Most utilize PVDF material, as standard ultrafiltration membranes fail to meet requirements. This stems from two primary reasons: ① Material composition is inadequate ② Cutoff molecular weight is insufficient [6000-10000 Dalton]. Asahi Kasei's membranes perform well and are widely adopted in most ultrapure water systems [18.2 MΩ·cm].

8. How is CO₂ generated after DI water passes through TOC-UV treated? Can polishing resin remove CO₂?
A: Resin has no removal effect on gases. However, CO₂ in water forms an equilibrium with CO₃²⁻ and HCO₃⁻, meaning it converts into ions. These ions can then be removed by polishing resin.
9. How is constant liquid level control achieved for the pure water tank? How is water fed into the pure water tank?
A: Constant liquid level requires a proportional valve. Install a pressure sensor on the pure water tank to control the valve opening via pressure readings.

10. Is the security filter typically sized to match the reverse osmosis membrane dimensions? For example, if I select a 4-inch x 40-inch reverse osmosis membrane, what size security filter should I choose?
A: The size of the security filter is not directly related to the RO membrane size but depends solely on the water flow entering the RO system. Determine the required filter cartridges based on your specific water flow rate.

11. What is the relationship between water flow rate and the number of filter cartridges? How is the filtration rate determined?
A: The water production capacity per cartridge varies by manufacturer. Some produce 2 t/h per cartridge, others 1.5 t/h, and some 1 t/h.
A: Common cartridges are 3“ in diameter, with lengths ranging from 10” to 40“. For a 10” 5-micron cartridge, the recommended design flow rate should not exceed 1 m³/h.

12. Have you implemented RO concentrate water reuse systems? (Recovery rate: 75%)
A: In our projects, clients typically discharge RO concentrate water. Some collect it for toilet flushing, but we haven't done actual treatment systems. Designing RO concentrate reuse follows the same principles as standard RO systems, though parameters differ slightly.

13. Question: Have you designed a two-stage RO system without an interstage pump? What type of RO membrane should be selected?
A: I haven't designed a two-stage RO system without an interstage pump. However, I can confirm that the membrane selection for such a system is identical to a conventional two-stage RO system. The main differences lie in stricter control requirements and increased operational risks.

14. Where is the most economical and practical location for recirculating water from each unit in an ultrapure water production process?
A: Stepwise recirculation is the most economical and practical approach. This involves routing EDI concentrate to the first-stage RO product tank, while second-stage RO concentrate is recirculated to the ultrafiltration tank or feedwater tank.

15. For common ultra-pure water instruments like TOC analyzers, SiO₂ analyzers, particle counters, and dissolved oxygen meters, which brands do you typically select?
A: We exclusively use imported instruments for these applications. We're not familiar with domestic brands and wouldn't consider them. Take conductivity meters as an example: domestic brands like Shangtai become inaccurate for measuring water above 10 MΩ·cm. For such applications, GF or THORTON instruments are essential.

16. Are there pressure requirements for ammonia used in nitrogen purging? If so, what is the specified pressure?
A: The minimum nitrogen pressure depends on the micro-pressure regulator selected. For example, Fisher regulators typically require a minimum of 0.2 MPa. However, standard nitrogen supply pressures are generally above 0.5 MPa, so this is not an issue.

17. If scale inhibitors replace water softeners in the process, how should the volume of the inhibitor tank be determined? Additionally, how is the volume of the regeneration salt tank configured for the softener determined?
A: The dosing tank volume should accommodate 3 days' worth of chemical dosage. The regeneration salt tank specification should be determined based on the volume of the regeneration salt solution.
A1: ① The scale inhibitor tank capacity should account for the number of days the configured chemical solution will last. I typically design for 1-3 days per solution batch. For the feedwater end, a recommended scale inhibitor dosage concentration is 3-5 ppm. ② The design principle for the regenerant salt tank of the softener is the same. Each full load should support 1-2 regenerations. For resin regeneration, calculate based on 100g pure NaCl per liter of resin. The saturated salt solution concentration is approximately 25%. Considering lower winter temperatures, a slightly lower concentration can be used. This allows calculation of the required volume.

18. How is the media packing height typically calculated for multimedia filters, activated carbon filters, softeners, and mixed-bed units? Is it the same for carbon steel and FRP tanks?
A: While the theory is consistent, specific data varies slightly between manufacturers. Generally, the heights for MMF and ACF are relatively fixed, but SF and MB heights vary based on resin volume. Since FRP tanks have fixed heights (e.g., 72-inch models are 1.8 meters tall), while carbon steel dimensions can be customized, some differences arise.

19. What water quality can be achieved with just pretreatment + two-stage reverse osmosis? What about pretreatment + two-stage reverse osmosis + EDI + polishing mixed bed? And what if it's two-stage reverse osmosis + EDI + polishing mixed bed + terminal ultrafiltration?
A: ① Depending on local feedwater quality, dual-stage RO output may range from 0.2-1 MΩ·cm, potentially reaching 1 MΩ·cm under optimal conditions. ② EDI output guarantees over 16 MΩ·cm; adding a polishing mixed bed extends this to over 17.5 MΩ·cm. ③ Adding terminal ultrafiltration (UF) at the end does not enhance water quality, as UF only removes particles.
20. Is there a significant difference in water production between 0.1μm and 0.22μm filter cartridges (of the same length)?
A: Since the pore size difference is minimal, the flow rate difference is not substantial. More significant factors include: ① Cartridge length and dimensions; ② Cartridge material. For example, a PES cartridge of the same size has a much higher flow rate than a PP cartridge.

21. What are the differences between these two processes? (Should the heat exchanger be placed before or after the carbon filter?) ① Sand filter + Carbon filter + Water tank + Heat exchanger + RO ② Sand filter + Water tank + Heat exchanger + Carbon filter + RO Also: For two-stage RO with alkaline addition, should the alkali be added to the pre-tank or directly into the RO inlet pipe?
A: We commonly implement “sand filter + carbon filter + storage tank + heat exchanger + RO.” Alkaline dosing for the secondary RO is added to the RO inlet pipe.
A1: The approach here is largely similar. (1) Adding at the inlet can slightly improve chemical utilization efficiency due to higher temperatures. If placing it earlier, I recommend adding it directly before the sand filter. (2) If the dosing point is too close to the secondary high-pressure pump, add it to the storage tank; if it's far from the pump, add it directly to the inlet pipe. The distance threshold is 40 times the pipe diameter. Exceeding this distance requires the second approach; otherwise, use the first.

22. Assuming the reverse osmosis feedwater pH is 7.5, what is the product water pH? If a decarbonizer is used post-RO, by how much can the pH increase?
A: After reverse osmosis, the pH typically decreases by 1-1.5, resulting in a product water pH of 6-6.5. If a decarbonizer is installed downstream, the pH can reach 6.5-7. Note that pH is influenced by numerous factors; the above parameters are for reference only.

23. Why is the regeneration flow rate for mixed-bed regeneration 10 m/h, while that for separate anion/cation bed regeneration is 5 m/h—a twofold difference? Rain2007: The anion bed regeneration flow rate should be set lower, while the cation bed can be higher. I believe the anion bed regeneration rate could be set below 5.
A: This is because during mixed-bed regeneration, if acid and base are regenerated simultaneously, a flow rate of 10 is required. If acid and base are regenerated separately, only a flow rate of 5 is needed. Since the regeneration of the cation and anion beds uses only one type of regenerant, the flow rate is 5. In summary: for simultaneous acid-base regeneration, use a flow rate of 10; for separate acid-base regeneration, use 5. Additionally, the flow rate selection primarily relates to the regeneration pump's flow capacity. For separate regeneration at 5, calculate the pump flow based on this rate. For simultaneous regeneration, calculate the pump flow based on a rate of 10.

24. Can 1stRO + EDI reliably achieve 10M? Since I'm not very familiar with EDI, could you recommend a cost-effective EDI module model?
A: 1stRO plus EDI is not recommended, but if operated well, achieving 10M should be feasible. Regarding EDI recommendations, there are many brands, but IONPURE is among the best.

25. What should be considered when designing an electronic ultrapure water purification plant?
A: If you're referring solely to the pure water workshop, there are no special requirements. Before starting a new project, ensure the workshop design is complete, including aspects like interface connections and drainage.

26. The stainless steel tank has a breather. How is this breather selected? If it depends on the tank size, how exactly is it chosen based on the tank dimensions?
A: The flow rate is a key specification for air vents. Vents vary in size based on tank capacity. Consult the vent manufacturer for specific sizing details.

27. Our activated carbon tower requires 6-7 cleaning cycles per replacement. Are there methods or processes to reduce cleaning frequency?
A: Initial cleaning cycles for activated carbon filters inherently take time. There are no shortcuts for this process.

28. Today I worked on a project using a degassing membrane unit, primarily for removing carbon dioxide and oxygen. However, the documentation states: “Employ a single-stage RO + degassing membrane to remove CO2 instead of single-stage RO + alkali addition + double-stage RO.” I'm a bit confused—isn't the degassing membrane equivalent to the alkali addition unit? Why was the double-stage RO omitted?  
A: This likely refers to the pre-treatment section for EDI! Actually, if the original text is phrased this way, it could be misleading and should be corrected! Typically, the EDI pretreatment stage uses a 2st RO with alkaline addition in between to remove CO2, ensuring low CO2 content entering the EDI. However, some processes opt for a single RO stage to save costs. In such cases, adding a degassing membrane is recommended to remove CO2 from the water before it enters the EDI. But the degassing membrane cannot replace the function of the second RO stage.

29. How do you determine the hardness of tap water in a specific location? It seems to vary significantly across regions. I couldn't find much online—most results were about drinking water quality...
A: Water quality data varies considerably by location, so testing locally is best. If that's not feasible, consult regional hydrological data. Crucially, accumulate data from various projects to build a reference database.

30. I'm currently working on a nanofiltration unit and found references describing nanofiltration as low-pressure reverse osmosis. So the transmembrane pressure difference required for nanofiltration must be lower than reverse osmosis! But the literature states nanofiltration's transmembrane pressure difference ranges from 0.5 MPa to 2.0 MPa—that's an incredibly wide range! Plus, the maximum value is significantly higher than reverse osmosis's transmembrane pressure difference. I don't understand why? I asked in a group chat, and someone also answered that it depends on the specific water quality. But surely there must be a general ballpark figure, like in reverse osmosis where the first-stage high-pressure pump typically operates at 130-150 m and the second-stage at around 120 m.
A: Claiming NF membrane pressure drop ranges from 0.5-2.0MPa is impossible. In reality, NF membrane pressure drop is only around 0.1-0.2MPa, similar to reverse osmosis values.

31. For post-treatment of wastewater after deep processing, what size should the raw water tank be for a daily flow of 4000 tons? How should the filter be selected?
A: The raw water tank design should account for half a day's flow volume. Regarding the second question, your description is unclear, so I cannot provide an answer.