Ferrofluid Synthesis

Precipitate magnetite nanoparticles from iron salts and suspend them in oil to make a magnetic liquid

Advanced🕐120 min■High hazardprecipitationphysical-properties

Summary

You’ll precipitate¹ magnetite nanoparticles from iron salts, wrap each in a surfactant, and suspend them in oil to make a true ferrofluid — a liquid that spikes and climbs toward a magnet. You’ll be able to explain why the Fe²⁺:Fe³⁺ ratio must be 1:2, why particle size makes or breaks the fluid, and what the surfactant coating does.

History

Ferrofluids were invented by NASA engineer Steve Papell² in 1963, who was looking for a way to move rocket fuel in zero gravity — a liquid that a magnet could pull into position where gravity couldn’t. Suspending magnetite nanoparticles in kerosene gave a fuel that responded to a magnetic field while still flowing like an ordinary liquid.

The idea never flew on a rocket, but it found homes everywhere else: damping the voice coils in loudspeakers, sealing the spinning shafts in hard drives, and adding contrast in some medical imaging. The dramatic field of spikes that makes ferrofluid a favourite demonstration — the Rosensweig instability — was described by Martin D. Cowley and Ronald Rosensweig³ in 1967.

Hazards & preparation

Warning

PPE: safety glasses and gloves; work in a fume hood or outdoors.

• Ammonia gives off strong, irritating fumes — the precipitation step must be done with real ventilation.
• Ferric chloride is corrosive and stains skin and clothing orange-brown.
• Dry magnetite nanoparticles are a respiratory hazard — do not inhale the powder.
• Keep finished ferrofluid away from electronics and magnetic media; it stains fabric permanently.

Disposal: don’t pour ferrofluid (oil + iron) or iron/ammonia solutions down the drain. Neutralize the ammoniacal liquid, collect the iron solids, and bin or hold for hazmat collection. See the Safety page.

Materials

What	How much	Notes
Ferrous sulfate (FeSO₄·7H₂O)	4 g	Fe²⁺ source
Ferric chloride (FeCl₃·6H₂O)	8 g	Fe³⁺ source; or use 4.8 g of anhydrous FeCl₃ (same moles, no crystal water)
Ammonia solution (~10%)	40–50 mL	Precipitant; NaOH also works. If using 30% ammonia, use 15–18 mL instead
Oleic acid	3–4 mL	Surfactant — available from art supply, soap-making, or chemistry suppliers
Mineral oil or light vegetable oil	30 mL	Carrier fluid
Distilled water	~200 mL	For rinsing
Strong neodymium magnet	1	N42 or stronger; larger is better for demonstrations
Glass beaker, 250 mL	1
Hotplate or stove		For heating to ~80°C
Safety glasses and gloves		Ferric chloride stains; ammonia is irritant

Note on oleic acid: See the oleic acid inventory page for sourcing. Without it, the nanoparticles will not be hydrophobic and will not transfer into oil — you will be left with a water-based magnetic sludge rather than a true ferrofluid.

Procedure

Part 1 — Precipitate Magnetite

1. Dissolve 4 g of ferrous sulfate in 50 mL of distilled water immediately before use — Fe²⁺ oxidizes to Fe³⁺ on standing, which shifts the ratio and produces a weaker product. Dissolve 8 g of ferric chloride hexahydrate (or 4.8 g of anhydrous FeCl₃) in 50 mL of distilled water. Combine the two solutions in the 250 mL beaker. The Fe²⁺:Fe³⁺ molar ratio is approximately 1:2, the stoichiometric ratio for Fe₃O₄.
2. Heat the combined solution to about 80°C on a hotplate. Keep stirring.
3. In a fume hood or outdoors, add the ammonia solution steadily while stirring continuously — roughly 1 mL every 10 seconds, over about 5 minutes. A dense black precipitate of magnetite forms as you add it. Do not dump all the ammonia in at once, and do not go so slowly that the precipitation drags on for more than 10 minutes. Addition rate controls particle size: a sudden excess of base causes chaotic nucleation; excessively slow addition gives particles time to coarsen. Steady addition with good stirring produces the small, uniform particles needed for a stable ferrofluid.
4. Continue stirring for 2–3 minutes after the last ammonia addition. The suspension should be completely black with no brown or orange tint.

Part 2 — Coat with Surfactant

5. While the suspension is still hot (keep at ~80°C), add 3–4 mL of oleic acid directly to the black suspension. Stir vigorously for 5–10 minutes. The oleic acid molecules adsorb onto the magnetite surface — the carboxylate head binds to Fe²⁺/Fe³⁺ sites on the surface, and the C₁₈ hydrocarbon tail faces outward.
6. The suspension will start to look slightly iridescent or oily at the surface. Hold a magnet to the outside of the beaker — if the particles respond (clumping toward the magnet), the basic precipitation worked. At this stage the particles are still in water.

Part 3 — Transfer to Oil

7. Add 30 mL of mineral oil to the beaker and stir. Then allow the mixture to cool to room temperature.
8. Hold a strong magnet to the bottom of the beaker. The coated particles are now hydrophobic and carry a thin oil coating; the magnet pulls this magnetic oil-and-particle phase down and pins it against the bottom, while the spent water/ammonia — now holding no particles — separates above it and clears. (Bulk mineral oil on its own would float, but the dense magnetite dominates, so the magnet keeps the magnetic phase at the bottom.)
9. While the magnet holds the particles, carefully pour off (decant) as much of the water/ammonia layer as possible.
10. Remove the magnet. Add 50 mL of distilled water, stir, then reapply the magnet and decant again. Repeat 2–3 times to remove residual salts and excess oleic acid.
11. The remaining oily black liquid in the beaker is your ferrofluid.

What you should see

The ammonia turns the amber iron solution instantly, densely black — that’s the magnetite. Once coated and transferred, you have a glossy black oil that, held over a strong magnet, doesn’t just pull toward it but erupts into a field of sharp, regularly-spaced spikes (the Rosensweig instability) and climbs the container walls.

Symptom	Likely cause	Fix
Particles won’t transfer to oil	Too little oleic acid coating	Reheat to 80°C, add another 1–2 mL oleic acid, stir 10 min, retry
Fluid settles quickly / sludgy	Particles too large (poor/uneven stirring)	No fix after the fact — remake with steady stirring and steady ammonia addition
Weak magnetic response	Low yield / wrong Fe²⁺:Fe³⁺ ratio	Check the ~1:2 ratio; precipitate hot
Fluid too thick	Too much oleic acid / too little oil	Add more carrier oil and stir

The reactions

Magnetite precipitates when base is added to the 1:2 mix of Fe²⁺ and Fe³⁺:

\[\ce{Fe^2+ + 2Fe^3+ + 8NH3 + 4H2O -> Fe3O4 v + 8NH4+}\]

iron(II) and iron(III) ions in a 1:2 ratio combine under ammonia to precipitate black magnetite (Fe₃O₄).

The recipe uses ammonia in deliberate excess — roughly 2–2.5× the stoichiometric amount — to drive the pH well above 9 and precipitate the magnetite completely; the leftover ammonia is removed in the later washes.

Alternative Method — Advanced Nanoparticle Route

A more involved route than the basic procedure above, but it produces sharper spikes and a more stable suspension. It uses PCB etchant as the iron source (or dry iron salts — see the Part A alternative below), precipitates the magnetite on a larger scale with an overhead stirrer, and separates the coated particles with acid rather than heat. Yield is roughly 8–9 g of dry magnetite powder → ~10 mL of ferrofluid. Note the time: 120 badge covers only the basic method — this route spans several hours plus overnight (or multi-day) drying.

Warning

Additional hazards — advanced route only. This route is meaningfully more dangerous than the basic method (hence the raised hazard badge). A fume hood or genuine outdoor ventilation is mandatory.

• Concentrated (~28%) ammonia gives off heavy, choking fumes and is corrosive to eyes and lungs.
• Concentrated hydrochloric acid is corrosive; adding it to the ammoniacal mixture releases a cloud of ammonium chloride smoke (and any acid-on-base addition is exothermic). Add slowly, downwind, with full eye protection.
• Kerosene and isopropanol are flammable — keep them and their vapours well away from the hotplate/stove and any flame; never heat the finished kerosene fluid.
• Concentrated ammonia and concentrated HCl must never be stored or handled near each other.
• Gloves, goggles, and long sleeves throughout; an adult runs every step.

Additional Materials

What	How much	Notes
PCB etchant (40% ferric chloride solution)	~50 mL	Fe³⁺ source; sold at electronics stores
Steel wool (fine grade)	a small handful	Reduces Fe³⁺ to Fe²⁺
— or, starting from powder (see Part A alternative) —
Ferric chloride (FeCl₃·6H₂O)	30 g	Fe³⁺ source; or 18 g of anhydrous FeCl₃
Ferrous sulfate (FeSO₄·7H₂O)	18 g	Fe²⁺ source; dissolve fresh
Ammonia solution (~28% concentrated)	~41 g total	33 g (Part B) + 5 g (Part B) + ~3 g (Part C soap)
Oleic acid	5 g	Surfactant
Hydrochloric acid (muriatic acid)	~17 mL concentrated	Dilute 1:2 with water before use
Isopropanol (rubbing alcohol)	~50 mL	For washing
Kerosene	~8 mL	Carrier fluid; evaporates more slowly than mineral oil
Calcium chloride	a cup	Desiccant for drying
Overhead stirrer or powerful drill mixer		A magnetic stir bar is not strong or fast enough
Vacuum chamber + pump		For drying; can substitute air-drying over several days
Coffee filter and funnel
Strong neodymium magnet	1+	For magnetic separation
Addition funnel or separatory funnel	1	For controlled acid/base addition

Part A — Prepare the Fe²⁺ Solution

The etchant is ~40% FeCl₃ (Fe³⁺). To get Fe²⁺, reduce it with steel wool:

\[\ce{2Fe^3+ + Fe^0 -> 3Fe^2+}\]

This matters because the reaction converts 2 mol Fe³⁺ into 3 mol Fe²⁺, so the Fe²⁺ concentration in the final solution is 1.5× what you might naively expect. Procedures that ignore this end up with the wrong Fe²⁺:Fe³⁺ ratio later, which degrades magnetite quality.

1. Pour about 65 mL of the PCB etchant into a beaker and add a small handful of steel wool. The solution immediately begins reacting (exothermic) and shifts from dark red-orange to green as Fe³⁺ is consumed. Keep adding small amounts of steel wool until the reaction slows.
2. Stir occasionally and let it sit for at least 30 minutes, or overnight for certainty. The result should be a green solution with no red-orange colour remaining — and, as a more reliable check than colour alone, a little undissolved steel wool should remain, confirming the iron was in excess and essentially all the Fe³⁺ was reduced. If iron(II) chloride precipitates (white-green solid), add more distilled water.
3. Filter through a coffee filter into a storage bottle. Wash the steel wool and filter with distilled water. Top up to a known volume (e.g. 100 mL) with distilled water — this lets you calculate the relative concentration.
4. Drop a few iron nails into the storage bottle. The Fe²⁺ solution oxidises back to Fe³⁺ on standing; the iron nails reduce it back down, keeping it stable.

Part A (alternative) — Starting from Dry Iron Salts

If PCB etchant is unavailable, skip the steel wool reduction and dissolve the iron salts directly. The amounts below are scaled to match the etchant method and hit the 1.7:1 Fe³⁺:Fe²⁺ target ratio.

1. Dissolve 30 g of FeCl₃·6H₂O (or 18 g of anhydrous FeCl₃) in 100 mL of distilled water — this is your Fe³⁺ solution. Separately, dissolve 18 g of FeSO₄·7H₂O in 100 mL of distilled water immediately before use; Fe²⁺ oxidises to Fe³⁺ on standing and should not be prepared in advance. Do not add iron nails.

Proceed to Part B. In step 6, add the entire Fe²⁺ solution, then the entire Fe³⁺ solution, to the 3 L of water — instead of the measured etchant volumes described there.

Part B — Precipitate Magnetite

The stoichiometric Fe³⁺:Fe²⁺ ratio for magnetite is 2:1, but fast stirring introduces air and oxidises some Fe²⁺ back to Fe³⁺. Using a ratio of 1.7:1 Fe³⁺:Fe²⁺ gives a buffer that absorbs this oxidation without producing excess Fe²⁺ at the end.

5. Add about 1 L of distilled water to a large container and start the mixer on a low setting.
6. Etchant route: add 20 mL of the Fe²⁺ solution (Part A), then 33 mL of the original Fe³⁺ etchant. (The reduced Fe²⁺ solution ends up with roughly the same iron concentration per mL as the etchant — the 1.5× boost from the 2Fe³⁺→3Fe²⁺ reduction is offset by the dilution to 100 mL — so ~33 mL etchant to 20 mL Fe²⁺ solution gives the target 1.7:1 Fe³⁺:Fe²⁺. Using only 25 mL would leave the ratio near 1.3:1.) Powder route: add the full Fe²⁺ solution then the full Fe³⁺ solution prepared in Part A (alternative), which are already scaled to 1.7:1. Either way, stir for 10 minutes.
7. Fill an addition funnel with 33 g of concentrated ammonia. Increase the mixer to about 1200 RPM. Open the funnel and add all the ammonia over about 20 minutes. The solution will turn brown (Fe(OH)₃ forming at pH > 4) and then black (magnetite forming at pH > 9). Slowly increase stirring to 1500 RPM as the mixture thickens.
8. Stir for another 15 minutes. While waiting, prepare the ammonium oleate soap (Part C below).
9. Add a final 5 g of concentrated ammonia quickly (within ~1 minute). Check that pH is just below 10.

Part C — Prepare and Apply the Surfactant

Unlike the basic method above (which adds oleic acid directly), this route first converts the oleic acid to ammonium oleate soap. The soap’s charged head binds more reliably to the magnetite surface.

10. In a small beaker, combine 5 g of oleic acid + 7 g distilled water + 3 g concentrated ammonia (~3.3 mL). The ammonia reacts immediately to form a thick soap. Stir until smooth and uniform.
11. Add all of the soap to the stirring magnetite suspension, rinsing the beaker with distilled water to get it all out. Stir for 20 minutes at increasing speed, finishing around 1900 RPM. The oleate molecules adsorb onto the magnetite surface with their hydrocarbon tails pointing outward, stabilising the particles.

Part D — Acid Precipitation

Most procedures remove the ammonium oleate with heat (~85°C, which drives off ammonia and regenerates oleic acid). Acid does the same thing faster, cooler, and — based on published comparisons — produces a better oleic acid coating.

12. Fill the addition funnel with dilute HCl (1 part concentrated HCl : 2 parts distilled water, ~10%). Add it slowly, checking pH every few minutes. The acid first neutralises excess ammonia (producing ammonium chloride), then attacks the ammonium oleate, regenerating insoluble oleic acid. The coated magnetite begins separating from the water.
13. Continue adding acid until pH reaches 6–7. The suspension will shift colour: foam clears, the magnetite visibly sinks, and the liquid above becomes much lighter. Turn off the mixer.

Part E — Wash and Dry

14. Decant as much water as possible. Hold a strong magnet to the outside of the container to pull the particles to one side, then pour off the water. Repeat 2–3 times.
15. Water washes (×3): Add distilled water, break up any clumps, hold the magnet for ~5 seconds (strongly magnetic particles pull out quickly; weakly magnetic ones are discarded with the rinse water). Decant. Repeat twice more.
16. Isopropanol washes (×3): Same process with rubbing alcohol. The first two washes may take 10–15 seconds for separation; by the third, separation is fast and the alcohol runs nearly colourless.
17. Transfer the wet magnetite to flat dishes (use a small amount of alcohol to rinse it out of the beakers).
18. Dry in a vacuum chamber over calcium chloride overnight. Without a vacuum chamber, spread the magnetite thinly and allow several days to air-dry; residual solvent will impair suspension into kerosene. The finished product is a dry, crumbly black powder. Do not inhale — nanoparticles are a respiratory hazard.

Part F — Suspend in Kerosene

19. Weigh out a batch of dry magnetite. Add kerosene at about 85% of the magnetite mass — e.g., 4 g magnetite → ~3.4 g of kerosene, which at kerosene’s density (~0.8 g/mL) is about 4.3 mL. The moment kerosene touches the powder, it begins absorbing and the liquid turns black.
20. Stir vigorously for 5–10 minutes. Use a magnet to pull remaining solids to the side and crush them against the beaker — this speeds up the last stubborn clumps. Aim to keep the concentration as high as possible; adding more kerosene is always possible later but dilutes the final fluid.
21. Test: hold a magnet under the dish. The fluid should flow freely toward the magnet and form sharp, evenly-spaced spikes with no increase in viscosity. If it thickens noticeably or partially solidifies, the particles are too large.

The Science

Magnetite (Fe₃O₄) is a mixed-valence iron oxide containing both Fe²⁺ and Fe³⁺ in a spinel crystal structure. It is a ferrimagnetic material — the magnetic moments of the two iron sites are aligned antiparallel but unequal, giving a net magnetic moment without an external field. This is what makes the nanoparticles individually magnetic rather than merely paramagnetic.

At nanoparticle scale (below ~15 nm for magnetite), each particle is small enough to hold just one magnetic domain, its whole moment pointing one way. Shrink further and thermal energy can spontaneously flip that moment: the particle becomes superparamagnetic. (Being single-domain is necessary but not sufficient for superparamagnetism — the particle also has to be small enough for thermal energy to overcome the moment’s preferred direction; this is a simplification of a size-dependent effect.) In a field the moments align and the fluid magnetizes strongly; with the field off, the moments re-randomize by Néel relaxation — the internal moment flipping inside each fixed particle — so the ferrofluid keeps no residual magnetism and doesn’t lock together. (Whole-particle tumbling, Brownian relaxation, also contributes, but for particles this small the internal Néel flipping dominates.)

The surfactant coating creates a steric barrier: the hydrocarbon tails from neighboring particles repel each other, preventing agglomeration. This, combined with Brownian motion, keeps the particles permanently suspended.

Demonstrations

Rosensweig instability: Pour a small pool of ferrofluid into a shallow dish. Hold a strong magnet underneath and slowly raise it. At a critical field strength, the surface erupts into regular hexagonally-arranged spikes — the energy gained by elongating along the field lines outweighs the surface-tension cost of the deformation.

Magnet climbing: Tilt a glass jar of ferrofluid and touch a magnet to the outside — the fluid climbs the walls toward the magnet, apparently ignoring gravity.

Field-line mapping: Place the jar near (not touching) two magnets arranged north-to-south; the ferrofluid reveals the field’s shape.

Spike counting: Moving the magnet closer increases the field and changes the number of spikes. Count them at different heights.

Questions to Explore

1. Why must the Fe²⁺:Fe³⁺ ratio be exactly 1:2? Fe₃O₄ has one Fe²⁺ per two Fe³⁺. What forms if the ratio is off, and is it magnetic?

   Hint / answer

   Magnetite’s structure needs exactly that 1:2 mix. With excess Fe³⁺ you tend to get brown, weakly-magnetic iron(III) oxides/hydroxides (like Fe(OH)₃/maghemite); with excess Fe²⁺, other iron(II) products. Both give a poorer, weaker fluid than proper magnetite.

2. Why does particle size determine stability? Above ~15 nm particles clump; below ~5 nm they’re weakly magnetic. Why is the 5–15 nm window special?

   Hint / answer

   In that window each particle is a single magnetic domain — strongly magnetic in a field — yet small enough that thermal energy re-scrambles its magnetic moment (Néel relaxation) once the field is off, so they don’t stick together or settle. Too big: permanent clumping; too small: too little magnetism.

3. What is the Rosensweig instability? Why are the spikes regularly spaced, and why does a stronger field change their number rather than just their height?

   Hint / answer

   The fluid wants to stretch along the field lines (magnetic energy) but surface tension and gravity resist. The regular spike spacing is the pattern that best balances those forces. A stronger field shifts that balance, so the optimal number of spikes changes — it’s a pattern-forming instability, not just taller peaks.

4. What makes oleic acid the right surfactant? What if you used a shorter-chain fatty acid?

   Hint / answer

   Oleic acid’s long C₁₈ tail creates a thick cushion between particles that keeps them from touching and clumping. A shorter tail gives a thinner cushion, so particles approach closely enough to stick together — a less stable fluid that settles or gels.

5. Why oil rather than water? Could you make a water-based ferrofluid?

   Hint / answer

   The oleic-acid coating makes the particles greasy (hydrophobic), so they belong in oil. A water-based ferrofluid is possible but needs a different surfactant — one with a water-friendly outer surface (e.g. a charged or double-layer coating) — so the particles disperse in water instead of oil.

Going further

• Map a field. Use the ferrofluid to trace the field around bar magnets, horseshoe magnets, and speaker magnets.
• Compare carriers. If you have both, contrast a mineral-oil and a kerosene fluid for spike sharpness and how fast the carrier evaporates.
• It’s a colloid. Confirm the nanoparticle nature by relating it to the light scattering in the Tyndall Effect.

Footnotes

1. Precipitate — An insoluble solid that forms and separates out when two solutions are mixed.

2. Steve Papell — NASA engineer who invented ferrofluid in 1963 as a way to move rocket fuel in zero gravity.

3. Ronald Rosensweig — American engineer who described the spiking (Rosensweig) instability of ferrofluids.