Anatase vs rutile
Both are TiO2 and tetragonal. Anatase: hardness 5.5 to 6, density 3.79 to 3.97, uniaxial negative. Rutile: 6 to 6.5, 4.23, positive. 6 of 15 properties overlap.
Anatase and rutile are both TiO2 and both tetragonal, but they are different minerals. Anatase: hardness 5.5 to 6, density 3.79 to 3.97, uniaxial negative, birefringence 0.0732, white to pale yellow streak. Rutile: 6 to 6.5, 4.23, uniaxial positive, 0.286 to 0.296, pale brown streak. In water a 20.00 g piece reads 14.72 to 14.96 g against 15.27 g.
In short
- Same formula, different mineral. The September 2026 IMA-CNMNC Master List lists both as TiO2: anatase status A, 1962 s.p., France; rutile G, 1803, Spain. The Handbook of Mineralogy sheets call each trimorphous with the other and brookite. Both are tetragonal, point group 4/m 2/m 2/m, but in different space groups: anatase I41/amd, rutile P42/mnm.
- The overlap ledger. Set the two sheets side by side on 15 properties and ask whether any printed number or word is shared. 6 overlap: crystal system, point group, hardness, colour, lustre and habit. 9 do not: density, in-water weight, optic sign, ω, ε, birefringence, streak, cleavage forms and twin forms. What overlaps is what the eye sees first.
- The clean separators, in numbers. Density 3.79 to 3.97 against 4.23: a 20.00 g piece reads 14.72 to 14.96 g in water as anatase, 15.27 g as rutile, 0.31 g apart at the nearest. Optic sign negative against positive. Birefringence 0.0732 against 0.286 to 0.296, worked. Cleavage perfect on {001} and {011} against {110} good and {100} moderate. Twins rare on {112} against common on {011}, geniculated.
- Documented failure: the visible properties. Hardness 5.5 to 6 meets rutile’s 6 to 6.5 at 6. Both sheets list reddish brown, pale yellow and black. Both are adamantine. Rutile is rarely dipyramidal; anatase is less commonly prismatic. The ω indices are only 0.044 apart; ε, 0.411.
- The decision. Weigh a solid piece, look for the elbow twins and striated prisms the rutile sheet describes, and streak it: white to pale yellow says anatase, pale brown to yellowish brown says rutile. The species pages are anatase and rutile; brookite, the third TiO2, is on the rutile page’s ladder.
| Property | Anatase (sheet) | Rutile (sheet) | Overlap? |
|---|---|---|---|
| Crystal system | Tetragonal | Tetragonal | Yes |
| Point group | 4/m 2/m 2/m | 4/m 2/m 2/m | Yes |
| Hardness | 5.5–6 | 6–6.5 | Yes, at 6 |
| Colour | Brown, pale yellow or reddish brown, indigo, black; pale green, pale lilac, gray, rarely nearly colorless | Reddish brown, red, pale yellow, pale blue, violet, rarely grass-green; black if high in Nb–Ta | Yes: reddish brown, pale yellow, black |
| Lustre | Adamantine to splendent, metallic | Adamantine to submetallic | Yes: adamantine |
| Habit | Typically acute dipyramidal; obtuse pyramidal or tabular; less commonly prismatic; to 3.75 cm | Prismatic, elongated and striated; equant, rarely dipyramidal; massive; to 25 cm | Yes: both can be prismatic or dipyramidal |
| Density | 3.79–3.97 | 4.23, increasing with Nb–Ta | No: 0.26 apart |
| 20.00 g piece in water | 14.72–14.96 g | 15.27 g | No: 0.31 g apart |
| Optic sign | Uniaxial (−); anomalously biaxial in deeply colored crystals | Uniaxial (+) | No |
| ω | 2.5612 | 2.605–2.613 | No: 0.044 apart |
| ε | 2.4880 | 2.899–2.901 | No: 0.411 apart |
| Birefringence, worked | 0.0732 | 0.286–0.296 | No: 3.9 to 4.0 times |
| Streak | White to pale yellow | Pale brown, yellowish brown; gray, greenish black with high Nb–Ta | No shared word |
| Cleavage | Perfect on {001} and {011} | {110} good; {100} moderate; {111} in traces | No shared form |
| Twinning | Rare, on {112} | On {011}, common, or {031}; geniculated, reticulated, cyclic | No shared form |
Are anatase and rutile the same mineral?
No. They share a formula, TiO2, and are two species. The September 2026 IMA list carries anatase, status A, 1962 s.p., France, and rutile, status G, 1803, Spain, as separate rows; a script finds 5 rows written TiO2 in all, as rutile sets out. The Handbook anatase sheet calls anatase trimorphous with brookite and rutile, and the rutile sheet says the same of rutile. Same atoms, different arrangement: anatase space group I41/amd, rutile P42/mnm.
Why the question gets asked. On the sheets they agree on more than they differ at a glance: both tetragonal with point group 4/m 2/m 2/m, both adamantine, both found brown, reddish brown, pale yellow or black, and each on the other’s Association line. The anatase sheet says it is usually secondary, derived from other titanium-bearing minerals; the rutile sheet calls rutile a common high-temperature, high-pressure accessory mineral. Both are common detrital minerals.
What this page adds. The rutile page’s TiO2 ladder weighs all three polymorphs; the anatase page works the pyramid angle and the cell. This page asks one narrower question: of everything the two sheets print, what overlaps and what does not.
How do you tell anatase from rutile? The overlap ledger
Of 15 properties on the two sheets, 6 overlap and 9 do not. Method: each property was transcribed from both sheets, the indices and forms checked against the rendered pages, and a script asked whether the two entries share any number (ranges that touch count) or any word. Overlap: crystal system, point group, hardness (at 6), colour (reddish brown, pale yellow, black), lustre (adamantine) and habit (each sheet allows the other’s shape). No overlap: density, in-water weight, optic sign, ω, ε, birefringence, streak, cleavage forms and twin forms.
The separators a collector can use. Weight: a 20.00 g solid piece reads 14.72 to 14.96 g in water as anatase and 15.27 g as rutile, 0.31 g apart at the nearest, with the method on how to measure specific gravity at home. Streak: white to pale yellow against pale brown to yellowish brown, on how to use a streak plate. Twins: the rutile sheet describes common geniculated, reticulated and cyclic twins; the anatase sheet says twinning is rare, on {112}. Cleavage: anatase perfect on {001} and {011}; rutile good on {110}.
The separators that need an instrument. Anatase is uniaxial negative with ω 2.5612 and ε 2.4880; rutile uniaxial positive with ω 2.605 to 2.613 and ε 2.899 to 2.901. Worked, the birefringence is 0.0732 against 0.286 to 0.296, so rutile’s birefringence is 3.9 to 4.0 times anatase’s. All four indices are above 2.4.
Where the easy tests fail
Documented failure: hardness meets at 6. Anatase is 5.5 to 6, rutile 6 to 6.5. A point that scratches one at hardness 6 may scratch the other; the scale cannot separate a soft rutile from a hard anatase. The method and its limits are on how to test mineral hardness properly.
Documented failure: colour and shape. Both sheets list reddish brown, pale yellow and black. Habit crosses over in both directions: the rutile sheet says equant, rarely dipyramidal, and the anatase sheet obtuse pyramidal or tabular, less commonly prismatic. A steep double pyramid is anatase’s usual form, and the anatase page works its angle, but a single crystal’s shape is a hint, not a verdict.
Documented failure: the streak near yellow, and Nb–Ta. Anatase’s streak runs to pale yellow and rutile’s to yellowish brown: on a pale streak the words are close. The rutile sheet says rutile is black if high in Nb–Ta, with a gray to greenish black streak, and that its density increases with Nb–Ta, which widens the weight gap rather than closing it. The anatase sheet says deeply colored crystals can be anomalously biaxial. The magnet does not help either: USGS left its one anatase sample behind at 1.70 amp and pulled rutile only in a tail, set out on is anatase magnetic.
Which is it on my label?
Read the crystal first, then confirm with the balance and the streak. Striated prisms, elbow twins or a reticulated net point to rutile, as its sheet describes it; a sharp double pyramid, or a tabular plate with perfect cleavage on {001}, points to anatase. Then weigh: 14.72 to 14.96 g for a 20.00 g solid piece is anatase, 15.27 g rutile. Then streak: white to pale yellow, or pale brown.
Where they come from. The two sheets share Cavradi, Tavetsch, Graubünden, Switzerland, among their named localities. The anatase sheet adds the Virtuous Lady mine, Devon, Fron Oleu near Tremadog, Wales, and carbonatites in Minas Gerais; the rutile sheet Graves Mountain, Georgia, Magnet Cove, Arkansas, and Conquista and Ibitiara, Bahia, Brazil. A locality is a clue, not a test.
On the label. Write the species the readings support, and write TiO2 polymorph if they do not settle it. The identification questions are indexed at the collecting guides. Nothing on this site is offered for sale; the wanted list is the only commercial route.