Ask the laboratory which fraction it measures
Request silica with units and method, identifying reactive, dissolved or total silica. Do not automatically treat the difference between two tests as an exact colloidal fraction without reviewing preparation and analytical recovery. Methods can respond differently to polymerised species and sample handling.
Include pH, temperature, iron, aluminium, hardness and alkalinity, together with the remaining projection inputs. Coagulant carryover and pH changes can affect silicate deposits. The assessment needs the chemistry actually reaching RO after the installed pretreatment rather than only the composition of an untreated source-water sample.
| Result | Required question | Implication |
|---|---|---|
| Reactive silica | Which method and SiO₂ basis? | Projection for measured fraction |
| Total silica | Which preparation and recovery? | Assess non-reactive fractions |
| Iron and aluminium | Any carryover at RO inlet? | Silicate and deposit risk |
Separate particle removal from solubility control
Suitable UF or other filtration can retain particular colloids, depending on size, stability and conditions. Conventional UF should not be described as a dissolved-silica removal process. Measure both relevant fractions in feed and filtrate when testing an upstream barrier, rather than relying on a single total result.
A safety cartridge may be insufficient for submicron particles and is not the primary process for a continuous colloidal load. Review coagulation and separation where appropriate while preventing metals or chemical carryover from creating a different fouling problem inside the RO plant.
Project concentrate under limiting conditions
As a first approximation without silica passage or precipitation, feed containing 20 mg/L as SiO₂ at 75% recovery would give 20/(1−0.75) = 80 mg/L in the concentrate. The calculation does not establish that 80 mg/L is acceptable: pH, temperature and composition remain essential.
The FilmTec manual distinguishes silica from metal-silicate fouling and requires assessment of concentrate chemistry. A proposed antiscalant needs an efficacy envelope applicable to that case. Compatibility with the membrane and effectiveness against a particular deposit are separate questions that require their own supporting evidence.
Design a trial that distinguishes the mechanisms
Collect simultaneous samples before and after pretreatment and from RO streams. Compare balances allowing for analytical uncertainty, pressure trends and normalised performance. A reduction in reactive silica may indicate a change in chemical form rather than actual physical removal from the water.
Where deposits exist, preserve samples for characterisation before selecting cleaning chemistry. Polymerised silica can be difficult to remove; do not improvise aggressive treatment by analogy with calcium carbonate. The quotation should define target fractions, recovery, pretreatment and the reassessment trigger when feedwater conditions change.
Common questions
Does UF protect against every silica form?
No. It may separate particular colloids, while dissolved silica normally passes conventional UF. Identify which fraction limits the actual installation.
Does low turbidity exclude silica problems?
No. Dissolved silica can concentrate without initially increasing turbidity. Colloids may also be inadequately represented by a single turbidity reading.