How Immiscible Liquid Phases Separate Inside Tower Distributors

Liquid distributors are normally designed on the assumption that the entering liquid behaves as one uniform phase. That assumption fails when the feed contains two immiscible liquids, such as hydrocarbon and water, organic solvent and aqueous solution, or a heavy acid phase mixed with a lighter organic phase.

Even if the two phases arrive as a temporary dispersion, they may separate inside the distributor header or trough. Once separation occurs, outlets no longer receive the same composition. Some discharge mainly the light phase, while others discharge a disproportionate amount of the heavy phase.

The distributor may therefore achieve excellent total-flow uniformity while producing severe composition maldistribution.

Why Phase Separation Occurs

Two immiscible liquids tend to separate because of density difference and interfacial tension. The rate depends on droplet size, viscosity, residence time, turbulence, and the geometry of the distributor.

A feed pipe may keep the phases dispersed because velocity and fittings create turbulence. When the mixture enters a large header or trough, velocity falls sharply. Droplets collide and coalesce, and gravity moves the heavier phase downward.

Common conditions that promote separation include:

Large distributor hold-up volume.

Long residence time before discharge.

Low liquid rate during turndown.

Large density difference between phases.

Low continuous-phase viscosity.

Weak inlet mixing.

Dead zones at header ends.

Unequal branch elevations.

Intermittent or slugged two-phase feed.

The phase arrangement inside the distributor depends on geometry. In a horizontal header, the heavy phase collects at the bottom, where bottom-facing holes may preferentially discharge it. In an open trough, the light phase may occupy the upper layer while the heavy phase feeds lower orifices first.

Why Total Flow Measurements Can Be Misleading

A bucket test measures volume or mass from each outlet. It usually does not measure phase composition.

Two outlets can discharge the same total mass but radically different liquid mixtures. One may contain mostly water while another contains mostly hydrocarbon. If the packing bed is intended to contact vapor with both components, local mass transfer will be inconsistent.

Composition maldistribution can cause:

Reduced absorption or stripping efficiency.

Local corrosion where the aqueous or acidic phase concentrates.

Poor catalyst or reactive-packing wetting.

Channeling caused by different viscosities and surface tensions.

Local temperature variation during exothermic reactions.

Unstable interface behavior in downstream equipment.

Incorrect tower material-balance interpretation.

In reactive absorption, the consequence can be especially serious. If the active reagent is concentrated in one phase, regions receiving the other phase may have almost no usable reaction capacity.

Distributor Geometry and Phase Preference

Outlet elevation is critical. Bottom outlets naturally access the liquid present at the bottom of a header. Side outlets may draw from different vertical layers depending on liquid level. Standpipes can preferentially withdraw the upper phase while leaving the heavy phase behind.

Branch connections also matter. A branch taken from the top of a main header may receive more light phase. A bottom branch may receive more heavy phase. Symmetrical piping does not guarantee symmetrical composition.

Pressure-drop devices can remix the phases temporarily, but remixing immediately upstream of a large, calm trough may provide little benefit because the phases separate again before discharge.

The inlet device should therefore be considered part of the distributor. Its purpose may need to include composition homogenization, not merely reduction of inlet momentum.

Essential Design Judgments

The first decision is whether the phases should remain mixed. In some processes, uniform phase composition across the bed is essential. In others, one phase is unwanted contamination and should be separated upstream rather than distributed.

If both phases must enter the bed uniformly, the designer must estimate the separation timescale and compare it with distributor residence time. A large distributor that appears hydraulically stable may actually be harmful because it provides more time for phase stratification.

The second decision concerns discharge velocity. Higher outlet velocity can reduce residence time and may entrain both phases, but it can also create excessive jet momentum, poor spreading, or erosion.

The third decision is whether a single distributor is appropriate. Separate distribution systems may be more reliable when phase ratios vary widely or when the phases have very different physical properties.

The fourth decision is turndown. A distributor that keeps phases mixed at design rate may stratify during low-rate operation because header velocity and turbulence decrease.

Engineering Solutions

Possible design approaches include:

Installing a static mixer immediately upstream of the distributor.

Reducing header volume and residence time.

Using multiple feed points rather than one large inlet.

Avoiding long dead-end headers.

Orienting branches to sample the full pipe cross-section.

Using recirculation where the process permits it.

Providing controlled turbulence inside the distributor.

Separately metering and distributing each phase.

Selecting outlet geometry that does not withdraw only one vertical layer.

A static mixer is not automatically sufficient. The distance between the mixer and the final outlets must be short enough to prevent re-separation. The pressure drop must also be acceptable, and the mixer must be compatible with fouling and solids.

When separate distribution is selected, designers must consider whether the two outlet patterns interact properly on the packing surface. Simply installing two independent grids does not guarantee microscale mixing.

Testing Requirements

A water-only factory test cannot verify performance for an immiscible two-liquid service. A representative test should use safe liquids with similar density ratio, viscosity ratio, and interfacial behavior where practical.

The test should evaluate:

Total flow from each outlet.

Phase fraction from each outlet.

Startup behavior before steady mixing develops.

Low-rate and maximum-rate conditions.

Response to changing phase ratio.

Drainage and retained heavy-phase inventory.

Restart behavior after settling.

Sampling repeatability.

Outlet samples should be collected simultaneously or over the same time interval. Sequential sampling can produce false conclusions if the inlet composition fluctuates.

Inspection and Operating Checks

Before startup, verify header slope, branch elevations, mixer orientation, drain locations, and absence of unintended pockets. Confirm that vents do not allow vapor accumulation to displace one phase.

During operation, compare temperature, corrosion, and composition profiles across the tower. Unexplained differences between sampling points may indicate phase segregation rather than ordinary liquid-flow maldistribution.

During shutdown, inspect low points for trapped heavy liquid, deposits, or corrosion. A distributor that retains one phase can create both process contamination and maintenance hazards.