1. Why Porous Titanium Is the Default for Aggressive Chemical Service

Commercially-pure titanium (GR1, GR2) passivates spontaneously in any environment that contains dissolved oxygen or water. The passive film is TiO₂, 2 – 7 nm thick, and re-forms within milliseconds if it is damaged by abrasion or a chemical transient. The same film that makes titanium biocompatible in implants is what makes a sintered titanium filter cartridge survive in wet chlorine gas, ferric chloride etch baths, and hot nitric acid — the three streams that eat 316L in months. Palladium-stabilised grades (GR7, GR11) extend the passive window into dilute reducing acids; GR12 (Ti-0.3Mo-0.8Ni) brings crevice-corrosion resistance on hot halide brines without the cost of a nickel-base alloy.

The mechanical and thermal envelope of porous titanium is narrower than that of 316L — usable temperature caps out around 280 °C in oxidising service and 320 °C in inert, against 600 °C+ for 316L — but almost every corrosion failure we see in porous metal filtration is in the room-temperature to 90 °C range, where titanium is unrivalled. Where the temperature exceeds titanium’s envelope, the standard practice is to drop back to alloy-C / Hastelloy for the housing and keep the porous element in titanium up to about 90 °C. For the broader porous-metal decision, our Titanium vs Stainless Steel Sintered Filters reference frames the question for non-chemical buyers.

2. The Corrosion Modes That Actually Drive Material Choice

Three modes — not one — account for nearly every premature failure of porous metal filters in chemical service. Pick the wrong mode, pick the wrong alloy.

2.1 Pitting and Crevice Corrosion in Halide Streams

316L pits in chloride-bearing streams above about 60 °C and in the presence of an oxidiser — FeCl₃, CuCl₂, wet chlorine, hypochlorite, even dissolved oxygen in hot brine. Pitting on a porous 316L cartridge looks like pin-prick perforations on the OD, with the rest of the wall structurally intact; differential pressure drops to near zero, and the cartridge is scrap. Titanium is immune to pitting in any halide solution short of wet fluoride, because the passive film is stable across the entire halide-pH-temperature envelope. The crevice between a porous cartridge and its support ring is the most common pitting site on 316L; on titanium, the crevice is passivated as soon as the ring is wetted.

2.2 Hydrogen Embrittlement on Acid + Cathodic Couples

Titanium absorbs hydrogen in strongly reducing acid (HCl above about 5 %, H₂SO₄ above 10 %, anhydrous methanol-HCl blends) and at any cathodic site, including the inside of a vessel where the housing is steel and the porous element is titanium. Hydrogen charging causes hydride formation, embrittlement, and cracking in welded zones. The standard mitigation is a positive electrochemical potential (Pt or Ru-Ir anode in line), keeping the wall above 80 °C, and avoiding fresh acid + galvanic couples. Specifiers often switch to Hastelloy C-276 for the worst-reducing streams rather than fight the hydrogen charging; we walk through that decision in §3.

2.3 General and Crevice Corrosion in Hot Brine

Seawater and chlor-alkali brine above about 90 °C attack 316L in the crevice between the porous element and its seat; titanium survives, but the cheaper GR12 (Ti-0.3Mo-0.8Ni) is even better in the 90 – 200 °C hot-brine window because the small nickel and molybdenum addition widens the passive range. For desalination RO guard filters where the temperature sits at ambient, plain GR2 is the right call.

3. Chemical Compatibility Matrix and the Ti-vs-SS316L-vs-Hastelloy Decision

The matrix below summarises the working envelope of a sintered GR1 / GR2 titanium filter cartridge against the eight streams that drive most chemical-industry RFQs. Ratings assume continuous duty, 1 µm nominal grade, ΔP ≤ 1.0 MPa, and the temperatures listed. The right column shows the next alloy up the cost curve when titanium is borderline. For the broader porous-metal selection question, our Titanium vs Stainless Steel Sintered Filters reference frames the same decision in non-chemical language.

The decision rule is short. If the stream is oxidising and chloride-bearing, pick titanium. If the stream is reducing and chloride-bearing, pick Hastelloy. If the temperature is below 60 °C, the stream is dilute, and there is no biological or Ni-sensitive downstream, 316L is the cheapest call and is often right. The failure mode that we still see too often is a 316L cartridge in chlor-alkali brine at 70 °C; the cartridge passes incoming inspection and fails in six to nine months, leaving iron hydroxide on every downstream surface. Switching to a sintered GR2 titanium filter eliminates the bleed and gives an order-of-magnitude life gain.

4. Porosity, Surface Engineering, and What Actually Affects Service Life

Once the chemistry is fixed, the next three knobs are nominal pore size, surface finish, and wall thickness. None of them changes the material, but together they shift the bubble-point integrity, the regeneration cycle count, and the life-time ΔP window by an order of magnitude.

4.1 Nominal Pore Size and Bubble-Point Integrity

Porous titanium is rated by nominal retention, not absolute. A 1 µm candle typically retains > 99 % of 1 µm particles in a single pass and 95 % at 0.5 µm. The bubble-point integrity test — pressurising an isopropanol-wetted candle from the inside and recording the pressure at which the first steady bubble stream appears — is the cleanest field check. For a 1 µm GR1 candle the minimum is 0.18 MPa; below that the candle has cracked, glazed, or wetted out. For 5 µm candles the minimum is 0.10 MPa. These are the numbers we run on every lot before release, and they are the numbers the customer can repeat in incoming inspection with a simple N₂ regulator and a beaker of IPA.

4.2 Surface Finish and Cake Release

As-sintered titanium has a Ra of about 3.2 µm; pickled / passivated surface is similar; electropolished surface can drop below 0.4 µm. Smoother surface means lower cake adhesion and faster regeneration — particularly for spent-catalyst recovery, where Pd/C and Pt/C stick to a rough surface and require aggressive caustic soak to remove. The trade-off is cost: electropolishing adds two to three weeks to lead time and a 15 – 25 % price premium. For most chlor-alkali and pulp-and-paper duty, the as-sintered finish is fine.

4.3 Wall Thickness and Mechanical Headroom

Standard 3 mm wall handles 2.5 MPa differential and back-pulse at 0.6 MPa. Thicker walls (4 – 4.5 mm) are specified when the customer wants a margin against water hammer or transient over-pressure, or when the candle is unusually long (60″ and above). Thinner walls (2 mm) are used in compact OEM skids where the working ΔP is below 0.6 MPa and the candle is 20″ or shorter. The general rule: keep wall temperature above 80 °C if the stream is reducing-acid, to keep hydrogen out of the lattice. The reverse is true in pure water and oxidising service, where cold wall actually passivates faster.

5. Four Field Cases — What the Call Actually Looks Like

Below are four duty positions where a sintered titanium filter is the answer, drawn from real shipping logs. Each line item is sized for a single 7-candle housing; for higher flow rates we parallel housings, not lengthen the candles.

5.1 Chlor-Alkali Brine Polishing (membrane-grade NaCl feed)

A 250 t/d chlor-alkali plant was getting 3-month cartridge life on 316L, with iron-hydroxide bleed into the membrane cells. We replaced the 10″ × 5 µm 316L elements with 30″ × 5 µm GR2 sintered titanium elements and a periodic in-place 5 % HCl pickle. The first 18 months of service ran without a single cartridge change; iron bleed dropped below the ICP-OES detection limit; cell membrane life extended from 18 to 36 months. The economic case was carried by the membrane life, not the filter savings.

5.2 Spent Pd/C Catalyst Recovery in API Hydrogenation

A CDMO running Pd/C-catalysed hydrogenation in 200-L batches was losing 3 % of catalyst per cycle to filtration cake loss. We specified 1 µm GR1 candle filters in a 4-candle housing, with back-pulse every 30 min during the reaction. Catalyst recovery went from 97 % to 99.4 %, the MTC for residual Ni/Fe/Co in the isolated API dropped below 5 ppm, and the catalyst budget dropped by 38 % over the next four quarters. The key win was trace-metal control, which the customer’s regulatory team had been chasing for two years.

5.3 Pharmaceutical Intermediate Polishing in Wet Methanol-HCl

A custom synthesis on an active pharmaceutical ingredient used a 5 % HCl-in-methanol stream for a crystallisation step, and the 316L polishing filter in the mother-liquor line was failing every 6 weeks. We specified 2 µm GR7 (Pd-stabilised) titanium, and added a 2 % water makeup to keep the stream out of the anhydrous-HCl window where hydrogen charging is fastest. The first set of cartridges ran 14 months with no ΔP excursion; the customer’s annual spend on filter consumables dropped by 60 %.

5.4 Bright Electroplating Drag-out Recovery (Pd / Rh drag-out)

A PCB plating line was running 60 g/L Pd drag-out from a drag-out rinse, and the existing 1 µm PP melt-blown cartridge was shedding fibres into the precious-metal recovery tank. We replaced it with a 0.5 µm GR1 titanium candle, and added a weekly 10 % oxalic acid soak. Pd recovery went from 92 % to 99.1 %; the rinse tank stays visually clear; the precious-metal refund from the refiner went up measurably within the first month. The shutdown exposure for change-out dropped from 4 hours per week to 30 minutes per month.

6. Cleaning, Regeneration, and End-of-Life Decisions

A sintered titanium filter is a consumable, but a long-life one. Three regeneration paths cover most field fouling, and the end-of-life criterion is well-defined and measurable. For the broader porous-metal regeneration discussion, including back-pulse rig design and on-site steam sanitisation, see our How to Clean and Regenerate Sintered Metal Filters reference.

6.1 The Three Regeneration Paths

Back-pulse with filtrate: 0.4 – 0.6 MPa reverse pulse every 30 – 60 min during normal duty. Recovers ~80 % of ΔP in well-characterised catalyst service; the cheapest path and the one that should be designed in from the start. Hot caustic / detergent soak: 5 % NaOH at 70 °C for 1 – 4 h to hydrolyse organic foulant films; rinse with DI water until conductivity is below inlet. Acid pickling: 10 % oxalic or 15 % citric at 60 °C for 2 – 6 h to dissolve oxide scale and metal hydroxide deposits; rinse, then re-passivate in 30 % HNO₃. Pickling is the most aggressive and is the path that ends the life of a poorly-vented cartridge in two or three cycles.

6.2 End-of-Life Indicators

Replace the candle when the bubble-point integrity test drops more than 25 % from the MTC baseline, or when ΔP at rated flow exceeds 0.8 MPa and back-pulse no longer recovers. Expected service life: 4 – 6 years in chlor-alkali brine, 18 – 36 months in spent-catalyst recovery, 3 – 5 years in seawater pre-filtration, 2 – 4 years in pharmaceutical polish. The high end of the range comes from installations that monitor integrity monthly and pre-empt the failure rather than waiting for ΔP to climb.

7. Frequently Asked Questions