Lily pipe placement: the geometry of whole-tank CO₂ distribution

The difference between a well-placed and a poorly-placed lily pipe can be the difference between consistent plant growth across an entire aquascape and a tank where one corner always lags. Understanding why requires a small amount of fluid dynamics — but the practical rules that follow are simple.

The Coanda effect

When a fluid jet flows close to and parallel to a solid surface, it tends to attach to that surface and follow its contour rather than expanding freely into open space. This is the Coanda effect — a well-documented phenomenon in fluid mechanics, observed across liquids and gases alike.

In a planted aquarium, a lily pipe outlet positioned close to the back glass and aimed along that wall tends to produce a jet that hugs the glass and travels further than one aimed into open water. The practical guidance widely used in the hobby — 2–4 cm from the glass — is a reasonable working distance: close enough to encourage wall attachment, far enough that glass lily pipes don't vibrate or rattle. The exact threshold varies with flow rate and outlet diameter, but the principle holds across common filter sizes.

The back wall is where you want the initial jet to travel, because it forms the first leg of the whole-tank circulation cell — the gyre.

Building a gyre

A gyre is a large-scale, coherent circulation cell. A correctly placed lily pipe creates one that works in two planes simultaneously. Viewed from above (top-down), the jet sweeps along the back wall, wraps around the far end wall, travels back along the front glass, and rises up the near end wall to return — a horizontal oval covering the full tank footprint. Viewed from the side, the same jet drives a vertical loop: surface flow toward the far end, descent at the far end wall, return along the substrate, and rise near the outlet. The animated panels in the next section show both of these views. One continuous three-dimensional loop — the same water column visiting every corner.

This matters for CO₂ distribution because CO₂ is consumed at the leaf surface, not in the open water column. A single pass of CO₂-rich water is not enough. What you need is continuous circulation that constantly refreshes the boundary layer immediately around each plant — the thin film of CO₂-depleted water that clings to every leaf surface; see the physics of the CO₂ boundary layer for why this matters as much as injection rate. A healthy gyre does this simultaneously across the whole tank.

Animated flow: what each placement actually does

The four panels below show how water circulates — or fails to — under different outlet positions and angles. Particles represent water movement. Watch how a back-corner outlet fills the entire tank volume, while a centre or surface-aimed outlet leaves large areas completely stagnant.

The five placement rules

These rules follow directly from the fluid dynamics above. They apply equally to ADA glass lily pipes, ceramic versions, and standard plastic filter return outlets.

1. Back corner, not centre. A centre-back position splits the flow into two weak half-tank gyres with dead zones in the near corners and a collision zone along the front glass. A back corner creates one large gyre covering the whole tank.
2. 2–4 cm from the back glass. A practical working distance for wall attachment — close enough to encourage the jet to hug the glass, far enough that glass lily pipes don't vibrate against the wall. Higher flow rates may tolerate slightly more distance; very low flow rates benefit from being closer.
3. Tip 1–2 cm below the waterline. Deep enough to create surface movement without breaking the surface. If the outlet tip is at or above the waterline, it creates surface turbulence and direct CO₂ loss.
4. Aimed parallel to the back wall, not into it. The outlet should fire along the glass, not at a right angle to it. A slight downward angle of 5–10° helps the jet reach the far end before it rises.
5. Fire toward the longer dimension. In most tanks this means left-to-right (or its mirror). The longer the jet travels before hitting the far end wall and descending, the more of the tank volume the gyre encompasses. In near-cube tanks the difference matters less.

What to look for at the surface

A correctly placed lily pipe creates a thin, fast-moving film of water across the surface — visible as a gentle shimmer when you look across the waterline from the side. This surface movement provides adequate gas exchange to maintain dissolved oxygen and allows CO₂ to off-gas at night when injection has stopped. (In a CO₂-injected tank during injection hours, dissolved CO₂ is well above atmospheric equilibrium — the tank is a net CO₂ emitter, not absorber. The benefit of surface movement during the photoperiod is primarily O₂ outgassing and replenishment, not CO₂ absorption. Avoiding excessive surface agitation during injection hours limits unnecessary CO₂ loss; gentle surface movement that maintains O₂ without vigorous turbulence is the target.)

What you do not want is visible surface rippling — waves, splashing, or broken water. This indicates the outlet is too close to the surface, angled upward, or running at a flow rate too high for the outlet size. Each accelerates CO₂ off-gassing significantly.

If you are logging CO₂ stability with a pH monitor, poor lily pipe placement often shows up as an inconsistent mid-session curve — CO₂ peaks and then drops irregularly rather than holding steady. This frequently indicates localised turbulence at the outlet causing CO₂ loss faster than the diffuser replaces it. Repositioning the outlet often smooths the curve significantly. See the guide to CO₂ stability for how to read and act on these curves, and the pH monitor guide for how to log them.

Which way should the outlet point?

This is the question that matters, and it is worth being precise about it — because the common advice to "keep the outlet at the back" describes where the pipe hangs, when the thing that actually drives the gyre is which way it fires. Those are two different variables, and confusing them causes real mistakes.

The rule that governs circulation is simple: fire the outlet parallel to a long wall, along its full length. That parallel relationship is what produces Coanda attachment and drives the whole-tank loop. A jet that runs the length of the back glass does this; so does a jet that runs the length of the front glass — hydraulically, a front outlet firing parallel to the front pane drives a complete gyre that is simply the mirror image of the recommended one (along the front glass, around the far end wall, back along the rear glass, and up to return). Its flow coverage is essentially identical. So if you have seen that setup running well, it is not a fluke; the physics is symmetric.

What breaks the gyre is firing perpendicular — across the short axis, into the opposing pane. And this is the mistake that hides behind a position-based rule, because it happens at the back just as easily as at the front. An outlet mounted correctly at the back but pointed forward at the front glass hits it in 30–40 cm, without ever running alongside a wall long enough to develop coherent circulation, then deflects into chaotic local turbulence rather than a whole-tank gyre. A front outlet pointed back at the rear glass fails in exactly the same way, for exactly the same reason. Position is not what saves you here; direction is.

So why is "at the back" still the standard advice? Because of a visual argument, not a flow one. The point of a glass lily pipe is that it is less intrusive than a standard plastic filter return; placing it in the foreground, between the viewer and the aquascape, eliminates that advantage entirely, and a front outlet also drops its surface film and any ripple onto the very pane you look through. Back placement hides the hardware behind the plant canopy, out of the primary sightline. If a particular hardscape genuinely forces the glassware to an end wall or the front, that is fine for circulation provided the jet still runs parallel to a long wall — you are trading looks, not flow.

Same side or opposite — where does the inlet go?

The inlet generally works best on the same side as the outlet — the same left or right half of the tank. It is worth understanding why, because the reasoning matters more than the rule.

If your outlet is in the back-left corner firing right, the jet travels right along the back wall, descends the right end wall, crosses the substrate from right to left, and rises up the left end wall. The gyre's natural return point is the left side of the tank — the same side the outlet is on. Placing the inlet here (front-left, or mid-left wall) means water has already completed the entire circuit before being drawn in. The gyre is fully intact.

The placement to be most cautious about is the back of the far end — in this example, the back-right corner. The jet fires from back-left and arrives directly at back-right. An inlet there can draw water straight back before the gyre has formed at all — the right end wall, the substrate, and the front of the tank see almost no circulation. This is a genuine short circuit.

A front-of-the-opposite-side position (front-right in this example) is less clear-cut: the jet has to travel the full back wall, descend, and cross the substrate before reaching it, so it is not a true short circuit. Many hobbyists use this layout with acceptable results. It tends to work better in shorter or higher-flow tanks; in long, lower-flow tanks the same-side placement is more reliable for ensuring the near half of the tank circulates properly.

The practical rule: whichever corner your outlet is in, keep the inlet on the same side of the tank (same left or right half), positioned at the front or mid-wall so it catches water as the gyre completes its loop. Avoid placing it directly at the back of the far end wall, directly in the jet's path.

Should the inlet be at the front or back of the glass?

Front is perfectly fine — and for most setups it is actually the better choice. The outlet is what generates the gyre, so it has to fire parallel to a long wall (and, for looks, we keep it at the back); the inlet has no such job — it is simply drawing water in, not generating flow. A front-placed inlet on the same side as the outlet catches water as the gyre completes its return loop and keeps the intake hidden behind foreground plants. Back-corner placement on the same side also works, provided it is not immediately adjacent to the outlet.

What matters is the left/right position — same side as the outlet, not the far end. Whether the inlet is front or back on that side is largely a practical and aesthetic decision.

How far apart, and at what height?

Distance: inlet and outlet should be on the same side of the tank (same left or right half), but as far apart as practical within that side — front or mid-wall rather than immediately adjacent to the outlet. The maximum separation within the same side is what you are aiming for.

Height of the inlet: mid-water to low — roughly one-third to one-half of the tank depth from the substrate. Two placements to avoid:

• At the substrate: pre-filter sponges placed directly at the bottom disturb fine substrate, pull up plant roots in a mature tank, and clog significantly faster. Even 3–5 cm clearance makes a noticeable difference in maintenance frequency.
• At the surface: an intake positioned near the waterline draws primarily from the surface film rather than the full water column. In a CO₂-injected tank this is counterproductive — you want the surface as calm as possible, and a surface inlet adds unnecessary agitation on the same side as the outlet, compounding CO₂ loss.

Spray bars: a different approach

Spray bars distribute flow along their length rather than concentrating it in a single jet. This makes them inherently better at covering width, but they don't drive a gyre in the same way — the distributed flow tends to produce a broad sheet rather than a concentrated circulation cell.

For spray bars, the common recommendation is: mount along the back wall or one side wall, angled downward toward the substrate rather than at the surface. A moderate downward angle creates a sheet of flow that pushes water along the bottom and generates a return current up the opposite wall — a weaker version of the gyre effect, but adequate for many tanks. The exact angle depends on tank depth and flow rate; the goal is to avoid surface disturbance while still reaching the lower half of the water column.

Spray bars aimed toward the surface are among the most aggressive CO₂ off-gassers in common use. If yours is pointed at or near the waterline, expect to lose a meaningful fraction of your CO₂ before it reaches the plants.

Scaling for tank size

• Under 80 litres: a single back-corner lily pipe covering the full length of the tank is sufficient. The gyre is compact enough to cover the whole volume in one pass.
• 80–200 litres: back-corner placement still works well. Aim for at least 10× turnover — even a perfectly aimed outlet with too little flow will leave slow spots in a tank this size.
• 200+ litres or very long tanks: consider two outlets in opposite back corners, each firing toward the centre of the tank. This produces a figure-eight circulation that covers both halves without the two jets opposing each other.

For the bigger picture on flow rates — how many times per hour the tank volume should turn over, and why that number matters for plant growth — see the guide to flow in a planted tank. If you have persistent algae in corners or on hardscape near the filter outlet, CO₂ starvation in dead spots is often the cause — the black beard algae guide covers how to diagnose and treat it.