Introduction

Ultrafiltration (UF) and microfiltration (MF) are the two most widely deployed low-pressure membrane technologies in water and wastewater treatment, serving applications ranging from drinking water production and seawater reverse osmosis (SWRO) pretreatment to industrial process water, membrane bioreactors, and tertiary wastewater polishing. While both technologies use porous membranes as absolute physical barriers to separate suspended and colloidal material from water, the differences in their pore size, retention characteristics, operating pressure, and fouling behaviour create distinct application envelopes that engineers and plant operators must understand to make informed technology selections.

This article provides a technically detailed comparison of UF and MF membrane technologies — examining pore structure, rejection mechanisms, module formats, flux behaviour, and application suitability — to guide engineers and decision-makers in selecting the optimal membrane technology for their specific water quality and treatment objectives.

Pore Size and Classification

The fundamental distinction between UF and MF lies in pore size, which directly determines the size range of particles and molecules each technology can reject. Jaffrin (2015) provides the standard classification:

• Microfiltration (MF): Mean pore diameter ranging from 0.05 µm to 5 µm (50–5,000 nm). Transmembrane pressure (TMP) typically 0.1–3 bar. MF membranes reject suspended particles, bacteria, protozoan cysts (Giardia at 8–15 µm, Cryptosporidium at 3–7 µm), and most colloidal material. However, dissolved organic molecules, viruses (typically 0.02–0.3 µm), and truly dissolved species pass through MF pores.
• Ultrafiltration (UF): Mean pore diameter ranging from 2 nm to 50 nm (0.002–0.05 µm). UF membranes are characterised by their molecular weight cut-off (MWCO), typically ranging from 2,000 to 1,000,000 Daltons. TMP operation is higher than MF at 1–8 bar. UF rejects not only particles and bacteria but also macromolecules, proteins, colloids, and viruses — anything larger than the nominal pore size. Dissolved salts, low-molecular-weight organics, and truly soluble species pass through.

This pore size difference translates into a fundamentally different separation capability. MF provides clarification — it produces water free of particulate turbidity. UF provides both clarification and disinfection-grade pathogen removal, meeting the U.S. EPA’s Long Term 2 Enhanced Surface Water Treatment Rule (LT2ESWTR) requirements for Cryptosporidium removal credits without chemical disinfection.

Rejection Mechanisms and Water Quality

Both UF and MF membranes operate primarily by size exclusion (sieving) — particles larger than the largest membrane pores are physically retained on the feed side while water and smaller constituents pass through. However, the mechanisms by which contaminants smaller than the nominal pore size are retained differ between the technologies and have important practical implications.

MF membranes typically feature a symmetrical or slightly asymmetric pore structure with a relatively narrow pore size distribution. The rejection of particles smaller than the nominal pore size depends on the formation of a dynamic filtration layer — a cake of accumulated particles on the membrane surface that progressively reduces the effective pore size. During the initial minutes of a filtration cycle (before the cake layer establishes), smaller particles and some bacteria may pass through. This transient period of reduced rejection, known as filtrate turbidity spiking at the start of each cycle, is a recognised operational characteristic of MF systems.

UF membranes, with their significantly smaller and more uniform pores, rely less on cake layer formation for pathogen rejection. The UF pore size is inherently smaller than bacteria (typically 0.5–5 µm) and most viruses (20–300 nm), providing an intrinsic removal barrier. The superior and more consistent pathogen rejection of UF is reflected in regulatory frameworks that assign higher log removal credits to UF than to MF for Cryptosporidium (typically 4-log for UF vs. 2–3 log for MF) and viruses. For applications where pathogen barrier integrity is paramount — drinking water production from surface sources, water reuse, or pharmaceutical process water — UF provides a demonstrably higher safety margin.

The Jaffrin textbook notes that because membrane pores are never perfectly uniform, there can be a distribution where some pores are larger than the nominal cut-off. This means that in MF, a fraction of particles near the nominal pore size may pass through larger pores. In UF, the extremely small pore dimensions make this effect less significant, but the concept of molecular weight cut-off — defined as the molecular weight at which 90% rejection is achieved — acknowledges that rejection is statistical rather than absolute at the molecular scale.

Hydraulic Performance and Flux

The hydraulic permeability of a membrane — the flux (L/m²·h, or LMH) produced per unit of transmembrane pressure — is inversely related to pore size. MF membranes, with their larger pores, exhibit hydraulic permeabilities typically 3–10 times higher than UF membranes treating the same feed water. This translates into higher sustainable fluxes: MF systems typically operate at 60–120 LMH for drinking water applications, while UF systems operate at 40–80 LMH under comparable conditions.

However, this apparent advantage of MF is moderated by fouling behaviour. The larger pores of MF membranes are more susceptible to internal pore fouling — where particles enter the pore structure and become trapped, causing progressive and often irreversible permeability loss. UF membranes, with their tighter pore structure, are more resistant to internal fouling but are more susceptible to surface fouling (cake layer formation) by organic macromolecules, particularly biopolymers such as polysaccharides and proteins. The net effect is that the long-term sustainable flux difference between UF and MF is often less than the clean-water permeability difference would suggest.

Jaffrin quantifies membrane resistance (R_m) as the inverse of hydraulic permeability times fluid viscosity: R_m = (µ·L_p)⁻¹. This relationship emphasises that flux is temperature-dependent through viscosity, and that design flux must be corrected for the minimum operating temperature. A system designed for 25°C may experience a 30–40% flux reduction at 5°C if adequate membrane area is not provided — a critical consideration for plants in temperate or high-altitude locations.

Module Formats: Hollow-Fibre vs. Spiral-Wound

Both UF and MF membranes are available in multiple module configurations, but the format that dominates each market segment reflects the different operational requirements of the two technologies.

Hollow-fibre modules dominate both UF and MF in the water treatment market. These modules consist of thousands of individual fibres — self-supporting tubular membranes typically 0.5–2.0 mm in diameter — potted at one or both ends. Filtration can be configured as outside-in (feed on the shell side, permeate extracted from the fibre lumen) or inside-out. Outside-in configurations, preferred for most drinking water UF/MF applications, can tolerate higher feed solids loading because the large shell-side volume accommodates solids accumulation. Inside-out configurations, where feed enters the fibre lumen, require more stringent pre-screening but facilitate more effective hydraulic cleaning since the flow path is well-defined.

Pressurised hollow-fibre systems enclose the fibres in a pressure vessel (housing) and drive filtration with a feed pump. Submerged (immersed) systems place the fibre bundles directly in an open tank and draw permeate by vacuum. Submerged systems generally have lower capital cost per unit area but operate at lower fluxes and are preferred for larger installations where tankage cost is offset by membrane cost savings.

Spiral-wound modules — the dominant format for reverse osmosis and nanofiltration — have also been adapted for UF, primarily in industrial and RO pretreatment applications. A spiral-wound element comprises membrane sheets separated by permeate spacer mesh, wound around a central permeate collection tube. Feed flows axially through the feed spacer while permeate spirals inward to the central tube. Spiral-wound UF elements offer higher packing density (membrane area per unit volume) than hollow-fibre, but are less tolerant of suspended solids and require more rigorous pretreatment. The DHP training material notes that format selection must account for feed type, feed quality, and flowrate, with hollow-fibre generally preferred for surface water treatment and spiral-wound for groundwater or RO pretreatment where solids loading is lower.

Temperature Correction

Temperature has a pronounced effect on membrane flux because water viscosity decreases with increasing temperature. The DHP design guidelines emphasise that flux values must be normalised to a standard temperature (typically 20°C or 25°C) using the viscosity ratio correction. Tabatabai (2014) presents the temperature correction based on Darcy’s law:

J_nor = J_t × exp[2148/(t + 273.15) − 2148/(25 + 273.15)]

where J_nor is the flux normalised to 25°C, and t is the actual feed temperature in °C. This correction is essential for accurately comparing flux data across seasons and for sizing membrane plants to meet demand at the minimum design temperature. Failure to apply temperature correction during design can result in plants that fail to meet production targets during winter months.

Application Selection: When to Use UF vs. MF

The decision to specify UF or MF for a given application must weigh multiple factors beyond pore size alone. The following framework provides practical guidance:

Drinking Water Treatment

UF is increasingly the default choice for surface water treatment where pathogen removal credits are required and where variable raw water quality — seasonal turbidity, algal blooms, or elevated natural organic matter (NOM) — demands consistent barrier performance. UF delivers turbidity consistently below 0.05 NTU and provides 4-log virus removal, allowing plants to meet disinfection requirements with reduced chemical dosage. MF may be acceptable for groundwater treatment under the direct influence of surface water where raw water quality is relatively stable and the primary objective is turbidity and bacterial removal, but the regulatory trend favours UF for new surface water plants.

Seawater Reverse Osmosis Pretreatment

Both UF and MF are deployed as pretreatment for SWRO, and both produce permeate with SDI (Silt Density Index) consistently below 3, and typically below 2, which is the target for reliable RO operation. The Tabatabai (2014) research confirms that UF/MF pretreatment produces RO feed water quality that is not affected by variations in raw water quality — a decisive advantage over conventional granular media filtration during algal bloom events. MF is often selected for this application on the basis of lower capital cost and higher flux, but the Altmann et al. (2023) ceramic UF study noted that polymeric MF/UF plants in the Gulf of Oman experienced 30–40% capacity reduction during harmful algal bloom (HAB) seasons, indicating that membrane chemistry and operating strategy may be more important than the UF/MF distinction for challenging seawater applications.

Wastewater Treatment and Reuse

In membrane bioreactors (MBRs), both UF and MF membranes are used, with the choice often driven by the membrane supplier’s standard product rather than a technical distinction. The Metcalf (2017) comparison of a Toray submerged MF MBR and a Norit sidestream UF MBR at Darvill found that both technologies produced excellent effluent quality that exceeded the conventional activated sludge process for all measured parameters. The UF system achieved higher flux (37.5 vs. 17 LMH) but this was primarily attributable to the sidestream configuration and higher operating pressure rather than pore size alone.

Industrial Process Water

UF is generally preferred where the objective extends beyond particle removal to include macromolecule rejection — for example, in the production of high-purity water for electronics manufacturing, pharmaceutical water-for-injection systems, or boiler feed water where silica and colloidal iron removal are critical. MF is suitable for cooling tower blowdown treatment, produced water filtration in oil and gas, and food and beverage applications where the target contaminants are primarily suspended solids and bacteria.

Design and Operational Optimisation

The DHP training programme identifies several key considerations for MF/UF system design and optimisation that apply regardless of the pore size classification:

• Flux selection: Design flux must balance capital cost (membrane area) against operational cost (fouling rate, chemical cleaning frequency, and energy). Conservative design at 60–80% of the critical flux — the flux above which fouling becomes irreversible — maximises membrane life and minimises chemical consumption.
• Recovery: The fraction of feed water converted to permeate. Typical recoveries range from 90–95% for drinking water applications, with the concentrate (reject) stream representing 5–10% of feed flow. Higher recovery reduces water losses but increases feed-side concentration of foulants, accelerating fouling.
• Backwash strategy: Periodic reversal of permeate flow to dislodge accumulated foulants. Backwash frequency (typically every 20–60 minutes), duration (30–90 seconds), and flux (1.5–3× filtration flux) must be optimised for the specific feed water and membrane type. UF systems generally require more frequent or aggressive backwashing than MF due to the greater accumulation of organic foulants on the tighter membrane surface.
• Chemically enhanced backwash (CEB): Periodic addition of chemicals (NaOCl, acid, or caustic) to the backwash water to remove organic and inorganic foulants that physical backwashing cannot address. CEB frequency is a primary determinant of plant operating cost and chemical consumption.

Conclusion

The choice between ultrafiltration and microfiltration is not a simple matter of “tighter is better.” While UF provides inherently superior pathogen rejection and greater resilience to raw water quality variation, MF offers higher flux and lower capital cost for applications where the treatment objectives are limited to turbidity and bacterial removal. The optimal selection depends on a systematic evaluation of raw water characteristics, treatment objectives, regulatory requirements, and lifecycle cost analysis. The trend in drinking water and advanced treatment applications is unmistakably toward UF as membrane costs converge and regulatory expectations for pathogen removal increase, but MF retains a significant role in industrial applications and where cost sensitivity is the dominant consideration.

For water treatment projects across Indonesia requiring expert membrane technology selection, system design, and operational support, TIWA offers independent, application-focused guidance drawing on extensive experience with both UF and MF membrane systems in Southeast Asian water conditions.