Pilot-Scale Resin Ion Exchange Column Train (Lead-Lag with Regeneration)

Ion exchange is easy to demonstrate and hard to design, because almost every number that matters only appears after the resin has been run to exhaustion two or three times on the real liquor. The lead–lag train exists to do that properly: two columns in series during service, with a third either regenerating or standing by, so the circuit never stops and the resin is always used to its full capacity rather than to the point where the operator got nervous.

Why lead–lag rather than a single column

In a single-column circuit the run has to stop the moment the effluent reaches the breakpoint, which means the bottom of the bed is still loaded and the top is saturated — a poor use of resin and an interrupted flowsheet. In a lead–lag arrangement the lead column is run past its own breakthrough, because the lag column behind it is still clean and catches the leakage. When the lag column begins to show the ion, the lead is taken off-line for regeneration, the lag is promoted to lead, and the freshly regenerated column takes the lag position. Continuous service, full resin utilisation, and a guard bed that absorbs the upsets.

The consequence for design is that each column has to be sized on the breakthrough curve of the one behind it, not on its own. A train sized on single-column data will either regenerate too early or leak.

Bed depth and flow rate are not independent

Everything in ion exchange is expressed in bed volumes per hour, and the reason is that the kinetics are governed by contact time in the bed rather than by absolute flow. Published pilot guidance puts the minimum resin volume at about 375 mL and the minimum classified bed depth at 760 mm, with a service flow of 2 to 50 BV/h — 8 to 20 BV/h being the usual band, and 1 to 10 BV/h where selectivity is high and the ion is valuable. A bed shallower than about 760 mm has too few transfer units to reach the leakage specification no matter how slowly it is run, and it channels.

Regeneration is slower by design: 2 BV/h is typical, with a contact time of at least 30 minutes, a slow displacement of 1 to 2 BV to push the regenerant through without mixing it, and a fast rinse of 2 to 10 BV. Pushing regenerant faster saves time and wastes chemical, because the exchange that strips the resin is itself a mass-transfer step and it does not get faster just because the pump does.

What the pilot has to produce

• The breakthrough curve, in bed volumes to a defined leakage (usually 5% of feed assay), on the actual liquor at the actual temperature.
• Loading capacity in grams of metal per litre of resin, measured, not taken from the resin datasheet — the datasheet figure is for a pure solution.
• Regeneration efficiency: how much of the loaded metal comes off per bed volume of regenerant, and whether the capacity is restored to the same value on the third cycle as on the first.
• Pressure drop and backwash expansion at the intended flow, because a bed that cannot be backwashed 30–50% will eventually plug with fines.
• Swell between the loaded and regenerated forms. A resin that swells 20% will crack a column designed without the allowance.

The operating discipline that decides the result

Never let the bed run dry, and never introduce air. Bubbles are almost impossible to remove once they are in a resin bed, and a bed with bubbles channels for the rest of its life. Run every test to exhaustion rather than stopping halfway, because ion exchange is reversible and a paused column simply re-equilibrates with the liquor it is sitting in — which looks like premature exhaustion on the next run and is not. And run three consecutive cycles before changing any condition: the first cycle on fresh resin is rarely representative of the second or third.