Page 430 -

MEMBRANE PROCESSES 13.27

Recovery Considerations. As recovery increases, the following factors must be consid-
ered in designing an RO or NF system.
Scaling. The concentration factor and potential for scaling increase as recovery in-
creases. The source feedwater composition must be evaluated to estimate maximum op-
erating recovery and the necessary pretreatment requirements (for example, pH adjust-
ment or scale inhibitor addition). The concentrations of solutes are greater near the
membrane surface than in the bulk stream due to concentration polarization.
Hydraulics. Optimal performance requires adhering to minimum concentrate and
maximum feed flow conditions for membranes. Feed flow to the first element in a pres-
sure vessel and concentrate flow from the last element in a pressure vessel must satisfy
the manufacturer's stated requirements.
System design must provide adequate membrane concentrate flow. Concentrate stag-
ing of membranes and pressure vessels is typically used for recoveries greater than 50%
to 60% (see Membrane Module Arrays and Staging). Some small systems are designed
with concentrate recycle to produce flows above the minimum specified by the membrane
manufacturer.
Source Water Use. The required volume of source feedwater necessary to produce
the same volume of permeate decreases as the recovery rate increases. Maximizing re-
covery rates minimizes both the source water requirement and the volume of concentrate
generated.
Permeate Water Quality. Feed-concentrate average salinity increases as recovery in-
creases. Because the flow of solutes through the membrane is a direct function of their
concentration in the feed concentrate stream, permeate quality decreases as recovery
increases.

Solute Rejection and Solute Passage. The removal, rejection, or passage of solutes in
a membrane system requires consideration of several variables.
Manufacturer's Specifications. RO and NF membranes are rated for nominal and
minimum rejections based on a specific test condition. Each RO membrane manufacturer
typically provides both design and minimum specifications relating to percent rejection
for sodium chloride (NaC1). With NF modules, specifications are also given for selected
divalent salts, for example, magnesium sulfate (MgSO4), and possibly organics in terms
of general molecular weight cutoff in daltons.
For example, low-pressure spiral-wound RO membrane elements are commonly rated
to have 96% to 99% salt rejection at 150 to 225 psig (1,035 to 1,550 kPa) feed pressure;
25 ° C feed temperature; 8% to 15% recovery; 1,500 to 2,000 mg/L NaC1 feed; and a pH
of 5.7 to 7.0.
Seawater spiral-wound RO elements are typically rated at 99% to 99.7% salt rejection
at 800 psig (5,520-kPa) feed pressure; 25 ° C feed temperature; 7% to 10% recovery;
32,000 to 35,000 mg/L NaC1 feed; and pH of 5.7 to 7.0.
NF membranes used for membrane softening, THMFP, and color removal typically
have 95% to 98% MgSO4 rejection with a 1,000 to 2,000 mg/L MgSO4 feed solution (or
greater than 70% salt rejection with 1,500 mg/L NaC1 feed) at 70 to 100 psig (483 to 689
kPa) feed pressure; 25 ° C feed temperature; 10% to 15% recovery; pH of 6.5 to 7.0; and
a molecular weight cutoff in the 150- to 400-dalton range.
Inorganic versus Organic Solutes. Both RO and NF membranes reject ionic and many
nonvolatile organic solutes to a high degree. Composite membranes typically reject
organic compounds better than do cellulose acetate or polyamide hollow fine-fiber
membranes.
In general, volatile organic compounds are poorly rejected by all membrane types (less
than 50%), although certain composite formulations have considerably higher rates. NF
membranes reject multivalent ionic and many nonvolatile organic solutes to a high de-