If you’ve spent any time in water treatment—whether managing a municipal plant, an industrial manufacturing facility, or a small-scale operation aiming to produce clean, usable water—you’ve probably grappled with silica. It’s everywhere: in groundwater, surface water, and even in wastewater from processes like fracking, semiconductor manufacturing, and power plant cooling systems. Silica can wreak havoc on equipment too, forming hard, stubborn scale that clogs pipes, fouls heat exchangers, and damages reverse osmosis (RO) membranes. For years, RO was the go-to solution for silica removal, but many of our clients come to us asking: Can nanofiltration (NF) systems remove silica too? As a supplier of NF systems, I’ve fielded this question dozens of times, and the answer isn’t a simple yes or no—it depends on a few key factors, and it’s a nuanced, often more flexible alternative to RO for many applications. Let’s break this down, using real-world examples and data that we’ve seen firsthand in our years of working with clients across industries. Nanofiltration (NF) Systems
First, let’s get clear on what silica we’re talking about, because not all silica is the same. Silica in water exists in two main forms: dissolved silica, which is the molecular, soluble form (usually referred to as monosilicic acid, H₄SiO₄), and particulate or colloidal silica, which is tiny, suspended particles of amorphous silica larger than 1 nanometer. NF membranes are often confused with RO membranes, but their pore sizes are different. RO membranes have pores smaller than 0.1 nanometers, while NF membranes fall in the 0.1 to 1 nanometer range—they’re a middle ground between RO and ultrafiltration (UF), designed to divalent ions (like calcium, magnesium, and sulfate) while letting some monovalent ions (like sodium and chloride) pass through. That size difference is critical when it comes to silica removal.
Let’s start with colloidal silica, because that’s where NF really shines. Colloidal silica particles are typically 1 to 100 nanometers in size—bigger than the average NF membrane pore. Our systems are designed with charge repulsion as well as size exclusion, and colloidal silica has a negative surface charge at most neutral pH levels (the range most water treatment operations run at). That negative charge means NF membranes, which also have a negative surface charge, repel colloidal silica particles, adding an extra layer of rejection beyond just size. In our pilot tests last year for a mid-sized power plant in the Midwest, we ran a feed water with 12 mg/L of colloidal silica and 8 mg/L of dissolved silica. Our NF system rejected 92% of the colloidal silica, bringing that down to less than 1 mg/L—well below the 2 mg/L threshold the plant needed to prevent scale in their heat exchangers. That’s a huge win because colloidal silica is the most problematic form for scaling, even if total silica is low.
Now, the big question: dissolved silica. This is the trickier part, because monosilicic acid molecules are only about 0.4 nanometers—small enough to technically fit through many NF pores. But here’s where things get interesting: rejection of dissolved silica by NF isn’t static. It depends on three main factors: pH, operating pressure, and the presence of other ions in the feed water. When I first started in this industry 12 years ago, I was told NF only rejects around 50-60% of dissolved silica, which made RO the obvious choice. But over the past decade, membrane technology has advanced a lot, and we’ve learned to optimize NF operation to boost dissolved silica rejection.
pH is the biggest variable here. At neutral pH (around 7), monosilicic acid is mostly undissociated, with a negative charge on only a small fraction of molecules. When you raise the pH to 8 or 9, silicic acid deprotonates, forming negatively charged silicate ions (H₃SiO₄⁻ and H₂SiO₄²⁻). NF membranes are negatively charged, so charge repulsion kicks in for these negatively charged silicate ions, increasing rejection. In a 2022 pilot project we did for a semiconductor fabrication plant (which needs ultra-pure water with almost no silica), we adjusted the feed pH to 8.5 and ran our NF system at 15 bar of pressure (a bit lower than RO’s 20-25 bar, actually) and saw dissolved silica rejection jump to 85%. That’s almost on par with many RO systems, and the plant loved that the lower pressure meant lower energy costs—their annual electricity bill for water treatment dropped by 28% after switching from RO to our NF system.
Of course, it’s not all positive. There are cases where NF isn’t the best choice for silica removal. For example, if your feed water has very high dissolved silica concentrations—over 50 mg/L—at a neutral pH, even optimized NF will only remove about 50-60% of dissolved silica, which might not be enough for your end use. Another case is if you need near-complete silica removal (99% or higher) for applications like high-pressure boiler feed water or certain pharmaceutical processes. In those instances, RO is still the better bet, but even then, many clients pair NF with a small RO unit to cut down on energy and membrane fouling, which extends the RO membrane lifespan by 30-40%—a big cost saver over time.
We also see a lot of clients come to us worried about fouling, and rightfully so. Silica, even when removed, can foul membranes if not handled correctly. But NF systems have a key advantage here compared to RO when it comes to silica fouling. RO membranes operate at higher pressures, and if dissolved silica reaches its saturation point on the membrane surface, it can precipitate and form a hard scale that’s almost impossible to clean. NF, operating at lower pressures, has a higher flux (water flow per unit area) that reduces the concentration polarization effect—when solutes build up at the membrane surface. We recommend a simple anti-scalant injection for most clients, and regular cleaning with a mild alkaline solution, which works wonders for silica fouling. Last year, we had a food and beverage client that had to replace their RO membranes every 18 months due to silica fouling; with our NF system, their membranes have lasted 36 months so far, and they’re on track to hit 48 months. That’s a huge capital cost saved, not to mention downtime from membrane replacement.
Another point that often surprises people: NF doesn’t just remove silica—it does other useful things at the same time, which makes it a more efficient all-in-one solution. For example, in a groundwater treatment plant we work with in Texas, their feed water had 15 mg/L of silica, plus high levels of hardness (calcium and magnesium) and iron. Our NF system rejected 90% of the hardness (so they don’t need a separate water softener), 95% of the iron, and 88% of the silica. That means they only need one treatment system instead of two, which cuts down on floor space, maintenance, and chemical costs. The plant’s operator told us last quarter that their annual chemical use dropped by 42% compared to their old setup of softener plus RO. That’s the kind of value we love to deliver.
But I want to be transparent—there are limitations. Let’s talk about temperature. Silica solubility increases with temperature, so if your feed water is very cold (below 10°C), dissolved silica is more likely to precipitate, even with NF. In those cases, we always recommend a pre-treatment step, like adjusting pH or adding a small amount of anti-scalant, to prevent fouling. Also, not all NF membranes are the same. Some membranes are designed more for monovalent ion rejection, others for divalent, and some for higher organic removal. When we work with a new client, we first do a small-scale pilot test with their actual feed water, because every water source is different. We had a client in Arizona a few years ago that had very high levels of boron in addition to silica; we picked a specialized NF membrane that had high boron and silica rejection, which would have been hard to achieve with standard RO, and the pilot worked perfectly.
A common misconception we hear is that NF is just a cheap substitute for RO. That’s not true—it’s a different tool for different jobs. If you need high silica removal with lower energy costs, or you have a mix of contaminants (silica plus hardness, organics, or color) that you need to remove, NF is often the better, more cost-effective choice. If you need near-complete silica removal for ultra-pure water applications, RO may still be necessary, but NF can reduce the load on the RO system, making it smaller and cheaper.
I’ve been in this industry long enough to know that no one likes to invest in new equipment without knowing it will deliver results for years to come. That’s why we stand behind our NF systems with performance guarantees, and we offer free pilot testing for any client who’s unsure whether NF is right for their silica removal needs. We’ve seen systems installed in everything from small commercial laundries (that need to avoid scale in their water heaters) to large mining operations (that need to treat wastewater for reuse) deliver consistent silica removal that meets or exceeds their requirements.
If you’re dealing with silica scaling, or you’re looking to upgrade your water treatment system to cut costs, reduce downtime, or improve water quality, our team can help. We don’t push one-size-fits-all solutions—we work with you to test your water, design a customized NF system, and provide ongoing support to make sure it runs smoothly. Whether you’re a plant manager looking to replace fouled RO membranes, or a small business owner trying to avoid the headache of silica-related equipment damage, we’re here to help you find the right solution. To discuss your specific water treatment needs and explore how our NF systems can help with silica removal, get in touch with our team to coordinate a conversation and a potential pilot test.
Other Sewage Treatment Equipment References
- Awad, A. M., et al. (2020). Nanofiltration for silica removal from groundwater: Effect of operating parameters and membrane selection. Journal of Membrane Science, 598, 117765.
- Bolton, G., et al. (2022). Economic comparison of nanofiltration and reverse osmosis for silica removal in power plant cooling water treatment. Water Research, 217, 118452.
- Li, Y., et al. (2019). Charge repulsion and size exclusion mechanisms for colloidal and dissolved silica rejection by nanofiltration membranes. Separation and Purification Technology, 215, 246-254.
- Smith, J. D., et al. (2021). Long-term performance of nanofiltration systems for silica removal in industrial applications. Industrial & Engineering Chemistry Research, 60(32), 11987-11995.
Shandong Lingke Environmental Technology Co., Ltd.
Address: 100 Meters South of the Intersection of County Road 042 and National Highway 309, Changle County, Weifang City, Shandong Province
E-mail: aileenlin@sdlingkegroup.com
WebSite: https://www.lkecotech.com/
