Library › Part C · Article 23 in the series · Level 2 · 12 min read

Nutrient film technique: growing roots in a moving film

Standfirst: After reading this article you will be able to explain why a nutrient film technique (NFT) channel only works while the film of solution stays a few millimetres deep, choose a slope, a flow rate and a channel length from that mechanism, size the catchment tank and pump for a stated number of plants, and recognise the three ways an NFT system fails.

Before you read: Anatomy of a hydroponic system, Dissolved oxygen in the root zone, Managing a recirculating solution over time.

The film: water from below, air from above

In the nutrient film technique, plants sit in holes along the top of a channel, their roots lie on its floor, and a shallow stream of nutrient solution runs continuously down the channel from a pump at the high end to a tank at the low end^1,2. The method was developed by Allen Cooper at the Glasshouse Crops Research Institute in England from 1965¹. Resh reports that the word “film” was chosen on purpose, to stress that the solution flowing past the roots must be very shallow so that the roots receive enough oxygen¹.

The mechanism is a division of labour within the root mat. The lower surface of the mat lies in the film and takes up water and nutrients; the upper surface sits in the moist air of the channel and takes up oxygen, which diffuses far faster through air than through water, as Dissolved oxygen in the root zone explains. Mason states the rule that follows: the depth of solution should normally be no more than a few millimetres, and the major part of the root mat should not be submerged³. A deep film drowns the mat.

Every design parameter below exists to keep the film thin, even, oxygenated and cool from the first plant to the last.

Slope: enough to drain, not so much that the top runs dry

Gravity moves the film, so the channel must slope. If it is too flat, or has a low spot, the solution ponds and the film deepens over the roots. In Cooper’s first systems, slopes from 1% to 4% were tested, and the steeper slopes reduced the effect of localised depressions¹. Resh now gives 2% to 3% for lettuce channels on benches and, for his cascade systems, a minimum of 2% for adequate flow¹. Jones gives a gentler figure, usually about 1%².

Mason explains why the range is wide: it depends on the channel floor. With a capillary mat on a flat base a slope of 1 in 150 or more is required³; without a mat it can be as little as 1 in 200 to 1 in 250³; and many growers now use short channels with adjustable slopes as steep as 1 in 60 or 1 in 40, to prevent ponding as the root mat develops and slows the flow³.

Steeper is not free. For a given flow rate, faster water occupies a smaller cross-section, so a steep channel carries a thinner and narrower stream past the first plants; this is this article’s inference from continuity of flow, not a measured result.

Flow rate: feed the far end without deepening the film

Flow rate sets both the depth of the film and the supply to the last plant. Mason’s rule is to feed the channel as fast as possible without raising the depth above a few millimetres³. Jones recommends about 1 litre per minute down each trough², and the Saskatchewan guide uses the same 1 L/min, warning that too little flow depletes nutrients and oxygen along the channel while too much wastes pumping capacity and may interfere with the uptake of some nutrients⁵. Virginia Tech’s extension guide gives 3 to 5 US gallons per hour into each channel⁴, which is 11 to 19 L/h, or about 0.2 to 0.3 L/min, by our conversion.

A lettuce trial falls between those figures. Al-Tawaha and co-workers ran channels on a 2% to 3% slope at 10, 20 and 30 L/h with plants 15 cm apart, and the middle rate gave the highest head mass, 236.67 g, against 152.67 g at 10 L/h and 134.67 g at 30 L/h⁶. The flow is set per channel with an emitter or valve; the Saskatchewan guide uses pressure-compensated emitters of 0.5 to 1 L/min behind a 100-mesh filter, because emitters plug with debris and algae⁵.

Channel length and temperature: what changes along the film

As the film runs past one root mat after another, the upstream plants take oxygen and nutrients out of it, and the solution warms if the channel is in the sun. Jones notes that plants at the upper end can reduce the oxygen or nutrient content enough to affect growth at the lower end, and he limits troughs to 9 m². Resh sets the limit at 15 m, beyond which insufficient oxygenation, blockage by roots and nutrient gradients may occur, and calls about 4.6 m the maximum practical length¹. Mason also gives 15 m and notes that shorter lengths have advantages in warm climates³.

Mason reports how severe the oxygen gradient can be: dissolved oxygen has been recorded falling from 7 to 8 ppm to 0 ppm along channels as short as 10 m carrying crops with high oxygen demand such as cucumbers, melons and tomatoes³. Warm solution makes this worse, because it holds less oxygen while the plant needs more³. Mason’s target is 18 °C to 22 °C for most plants³; he warns that solution overheats inside channels in direct sun before the crop shades them, and that black polythene gullies in summer sun can “cook” the roots of a young crop³.

The channel, its inlet, its outlet and its cover

A commercial lettuce channel is about 10 cm wide and 4 cm deep, with a ribbed floor that keeps the solution flowing down the centre¹. Mason gives 10 to 15 cm of width for a single row of plants and 25 to 30 cm for a double row, and where the floor is flat a capillary mat is sometimes laid in the channel so that the solution spreads across the full width instead of running down one side³. Lettuce is spaced at about 20 cm along the channel with channels 15 to 18 cm apart from centre to centre¹.

The solution is pumped to headers at the high ends of the channels, enters each channel through a small flexible line, flows down by gravity, and drains at the low end into a catchment pipe that returns it to the tank¹. Channels need covers, because light inside the channel grows algae^1,2.

Intermittent flow

Flow is normally continuous, but Mason describes an alternative: the solution runs for a few minutes, then stops for 15 to 60 minutes before flowing again, to raise aeration in the root zone for crops that need it, especially in established plants whose roots have clogged the channel and slowed the flow³. The pause drains the film from the mat so that air reaches roots that a deepening film had covered.

The catchment tank: the only buffer you have

Everything the plants drink, and everything they change in the solution, shows up in the tank, because the film itself holds almost nothing. Virginia Tech’s rule is a quarter to one US gallon of tank capacity per plant, with an example of a 25-gallon tank for 50 plants, and the advice to oversize the tank to buffer against pH and nutrient changes⁴. By our conversion of⁴ that is 0.95 to 3.8 L per plant, and 95 L for 50 plants. A larger tank drifts more slowly in EC, pH and temperature; the drift itself is the subject of Managing a recirculating solution over time.

The pump

The pump must deliver the sum of the channel flows, lifted from the tank’s water surface to the inlet headers, plus the friction of the pipework; flow and lift together fix the point on the pump’s curve⁴. For a hobby system of several channels, the Saskatchewan guide lists a submersible pump of 10 L/min as sufficient⁵. Virginia Tech advises keeping a backup pump in case of pump failure⁴.

Start-up and the root mat

A seedling in a cube above a channel floor cannot drink until its roots reach the film, so the crop is started elsewhere. In Resh’s lettuce sequence, seed is sown in rockwool or foam cubes, grown 18 days on a propagation bench, moved to nursery channels with holes at 5 cm for 10 to 12 days, then set in finishing channels for 20 to 25 days, about 55 days from sowing to harvest¹.

The mat then grows until it becomes the problem. Jones describes the flow diminishing as the mat thickens, and the solution moving over the top and down the outer edge of the mat instead of through it². Resh warns that long-term crops with large root systems plug the gutters, cutting flow and oxygen and leading to blossom-end rot and death of the plant¹. NFT is today used principally for lettuce and herbs while vine crops moved to rockwool¹, the route that Drip irrigation and substrate culture covers.

How NFT systems fail

The pump stops. The film is a few millimetres deep³, so when the flow stops the channel drains and the roots keep only the moisture held in the mat. Mason notes that a channel filled with coarse aggregate keeps roots moist “for a period” after the flow stops, an advantage a bare channel does not have³. A backup pump⁴ is the defence.

The film gets hot. Dissolved oxygen falls as temperature rises while root demand climbs, and the 10 m channel that went from 7 to 8 ppm to 0 ppm is the warning³. Shade in the early weeks and short channels in warm climates follow from that³.

Disease travels in the water. Every channel drains into the same tank and is fed from it again. Stouvenakers and co-workers list NFT among the systems that favour Pythium and Fusarium because of easy spread by water recirculation, possible recontamination after a disinfection step, and low oxygen in the solution; Pythium zoospores move freely and actively in liquid water⁸. Resh’s growers rinse harvested channels and soak them in a 10% bleach solution for at least an hour before reuse¹.

Worked example: sixty lettuce plants

Suppose you want sixty heads of lettuce in finishing channels. Each input is sourced; each result is our arithmetic.

• Channels. Resh’s benches use channels about 4 m long with lettuce at 20 cm¹, so one channel holds 4 ÷ 0.2 = 20 plants and three channels hold 60.
• Flow. At 1 L/min per channel^2,5, the pump supplies 3 L/min, or 180 L/h.
• Pump. With the channel inlets 1.0 m above the tank water level, a height chosen for this example, the pump must deliver the 3 L/min from^2,5 at more than 1 m of head plus pipe friction; the 10 L/min pump of the Saskatchewan guide⁵ has a margin of three times.
• Tank. At 0.95 to 3.8 L per plant, our conversion of⁴, the tank holds 57 to 227 L. Following the advice to oversize⁴, take 227 L; the pump then passes the whole tank through the channels every 227 ÷ 180 = 1.3 hours, our arithmetic from^2,4,5.
• Daily water loss. Kratky reports that a lettuce crop in his non-circulating tanks, harvested about 6 to 7 weeks from seeding, consumes 3 to 6 L of solution per plant⁷. If all of that were drunk during the 20 to 25 day finishing stage¹, an overestimate that errs on the safe side, the rate would be 0.12 to 0.3 L per plant per day, or 7 to 18 L per day for sixty plants, our estimate from^1,7.
• Top-up. That loss is 3% to 8% of the 227 L tank per day but 13% to 32% of a 57 L tank, our arithmetic from^4,7. The small tank needs water every day and its EC moves quickly; the large one can wait several days.

Vertical towers: a variant with a medium-filled core

Vertical columns stack plants above one another on a wall or frame. Touliatos, Dodd and McAinsh compared upright columns, each container holding 130 g of perlite and fed by a sprinkler at the top so that solution dripped down by gravity, with horizontal pipes of the same medium⁹; the columns produced 13.8 times more lettuce per square metre of floor, but the heads were smaller, 95 g against 138 g, and the bottom layer produced 43% less than the top as light fell from 491 to 134 µmol m⁻² s⁻¹ down the column⁹. Commercial towers of the ZipGrow type hold a matrix medium of recycled PET instead¹¹; the floor-area gain comes with a light gradient, and the medium adds a surface that solids can clog.

PonicLabs has run NFT twice. In 2017 a single air pump aerated a 25 L jug, moved biofilm carriers and lifted the solution to the top of an NFT channel, from which it returned by gravity, and the system grew lettuces, pak choi, coriander and parsley on urine and vermicompost leachate¹⁰. In 2019 four slanted ZipGrow towers fed from a biofilm reactor grew lettuce, basil, chard and snap pea to harvest size on a biogas slurry, but solids in the slurry repeatedly clogged the feed pipes and the matrix medium¹¹. Both were nutrient trials (see the sustainable-hydroponics pages); here they stand only as observations.

Key numbers

Common mistakes

1. Running the film deep so that the far end is well fed. A deep film submerges the mat and cuts its oxygen supply, which is the failure Cooper’s design set out to avoid^1,3.
2. Building one long channel instead of several short ones. Oxygen and nutrients fall along the film, and dissolved oxygen has reached zero within 10 m under a demanding crop³.
3. Sizing the tank to the pump rather than to the plants. The tank is the only buffer against drift and against a stopped pump, and the per-plant rule exists for that reason⁴.
4. Growing tomatoes or cucumbers in lettuce channels. Their root mats plug the gutter and raise the film, leading to blossom-end rot and plant death¹.
5. Ignoring the shared return. Zoospores move actively in water, and recirculation carries them from one channel to every other⁸.

Summary

NFT works because a film a few millimetres deep lets the root mat drink from below while breathing from above. Slope keeps the film moving, flow rate feeds the last plant without deepening the film, and channel length is limited because oxygen, nutrients and coolness are used up along the way. The tank is the system’s only buffer, so it is sized per plant and oversized when in doubt, and the pump is chosen for the summed channel flow at the lift required. Seedlings are raised until their roots reach the film, and the mat they grow will eventually block it, which is why NFT suits lettuce and herbs. A stopped pump, a hot film and a waterborne pathogen are the three failures; a backup pump, short cool channels and clean channels are the answers.

Next in the series: Ebb and flow.

References

1. Resh, H. M. (2013). Hydroponic Food Production: A Definitive Guidebook for the Advanced Home Gardener and the Commercial Hydroponic Grower, 7th ed. CRC Press, Boca Raton. Chapter 6, Nutrient Film Technique.
2. Jones, J. B. Jr. (2005). Hydroponics: A Practical Guide for the Soilless Grower, 2nd ed. CRC Press, Boca Raton. Chapter 9, Nutrient Film Technique.
3. Mason, J. Commercial Hydroponics, 3rd ed. ACS Distance Education. ISBN 978-0-9871022-2-5.
4. Mullins, C., Vallotton, A., Latimer, J., Sperry, T. and Scoggins, H. (2023). Hydroponic Production of Edible Crops: Nutrient Film Technique (NFT) Systems. Virginia Cooperative Extension, publication SPES-463. https://www.pubs.ext.vt.edu/content/pubs_ext_vt_edu/en/SPES/spes-463/spes-463.html
5. University of Saskatchewan, Grow with USask (2025). Hydroponic Leafy Greens. https://gardening.usask.ca/documents/grow-with-us-pdfs/hydroponicleafygreens_website2025.pdf
6. Al-Tawaha, A. R., Al-Karaki, G., Al-Tawaha, A. R., Sirajuddin, S. N., Makhadmeh, I., Megat Wahab, P. E., Youssef, R. A., Al Sultan, W. and Massadeh, A. (2018). Effect of water flow rate on quantity and quality of lettuce (Lactuca sativa L.) in nutrient film technique (NFT) under hydroponics conditions. Bulgarian Journal of Agricultural Science, 24(5), 793–800. https://agrojournal.org/24/05-09.pdf
7. Kratky, B. A. (2009). Three non-circulating hydroponic methods for growing lettuce. Acta Horticulturae, 843, 65–72.
8. Stouvenakers, G., Dapprich, P., Massart, S. and Jijakli, M. H. (2019). Plant pathogens and control strategies in aquaponics. In Goddek, S., Joyce, A., Kotzen, B. and Burnell, G. M. (eds), Aquaponics Food Production Systems, Springer, Cham. https://doi.org/10.1007/978-3-030-15943-6_14
9. Touliatos, D., Dodd, I. C. and McAinsh, M. (2016). Vertical farming increases lettuce yield per unit area compared to conventional horizontal hydroponics. Food and Energy Security, 5(3), 184–191. https://doi.org/10.1002/fes3.83
10. Sánchez, H. (2017). NFT Anthroponics system with airlift recirculation. PonicLabs technical report, Malmö.
11. Sánchez, H. (2022). Using green waste biogas slurry as nutrient source for a NFT hydroponics system combined with an MBBR. PonicLabs technical report, Lund.

This article was drafted with the help of AI tools and reviewed and edited by PonicLabs. Sources are listed above; tell us if you find an error.