Lithium molybdenum purple bronze is a chemical compound with formula Li
17, that is, a mixed oxide of molybdenum and lithium. It can be obtained as flat crystals with a purple-red color and metallic sheen (hence the "purple bronze" name).
This compound is one of several molybdenum bronzes with general formula A
z where A is an alkali metal or thallium Tl. It stands out among them (and also among the sub-class of "purple" molybdenum bronzes) for its peculiar electrical properties, including a marked anisotropy that makes it a "quasi-1D" conductor, and a metal-to-insulator transition as it is cooled below 30 K.
The compound was first obtained by Martha Greenblatt and others by a temperature gradient flux technique. In a typical preparation, a stoichometric melt of Li
2 and MoO
3 is maintained in a temperature gradient from 490 to 640 °C oven 15 cm in vacuum over several days. Excess reagents are dissolved with a hot potassium carbonate solution releasing metallic-purple plate-like crystals, a couple mm wide and less than a mm thick.
The crystal structure of Li
17 was determined by Onoda and others through single-crystal X-ray diffraction. The crystal system is monoclinic, with approximate unit cell dimensions a = 1.2762 nm, b = 0.5523 nm, and c = 0.9499 nm, with angle β = 90.61°, volume V = 0.6695 nm3 and Z = 2. In typical crystals, a is the shortest dimension (perpendicular to the plates) and b the longest. The density is 4.24 g/cm3. The structure is rather different from that of potassium molybdenum purple bronze K
17, except that both are organized in layers. The difference may be explained by the relative sizes of the K+
The unit cell contains six crystallographically independent molybdenum sites. One-third of the molybdenum atoms are surrounded by four oxygens, two thirds are surrounded by six oxygens. The crystal is a stack of slabs; each slab consists of three layers of distorted MoO
6 octahedra sharing corners. The lithium ions are inserted in the large vacant sites between the slabs. There are zigzag chains of alternating molybdenum and oxygen atoms extending along the b axis.
Lithium molybdenum purple bronze is quite different than the sodium, potassium and thallium analogs. It has a three-dimensional crystal structure, but a pseudo-one-dimensional (1D) metallic character, eventually becoming a superconductor at about 2 K Its properties are most spectacular below 5 meV. The Tomonaga-Luttinger liquid theory has been invoked to explain its anomalous behavior.
At room temperature, Greenblatt and others (in 1984) measured the resistivity of lithium purple bronze along the a, b and c axes as 2.47 Ω cm, 0.0095 Ω cm, and on the order of 0.25 Ω cm, respectively. The conductivities would be in the ratio 1:250:10, which would make this compound an almost one-dimensional conductor. However, Da Luz and others (2007) measured 0.079, 0.018, and 0.050 Ω cm, respectively, which corresponds to conductivity ratios 1:6:2.4 for a:b:c; whereas H. Chen and others (2010) measured 0.854, 0.016, and 0.0645 Ω cm, respectively, which correspond to conductivity ratios of 1:53:13.
This anisotropy has been attributed to the crystal structure, specifically to the zig-zag chains of molybdenum and oxygen atoms 
The resistivity along all three axes increases linearly with temperature from about 30 K to 300 K, as in a metal. This is anomalous since such a law is expected above the Debye temperature (= 400 K for this compound) The resistivity ratios along the three axes are preserved in that range.
As the lithium purple bronze is cooled from 30 K to 20, it changes abruptly to an insulator. After reaching a minimum at about 24 K, the resistivity increases 10-fold and becomes somewhat more isotropic, with conductivities 1:25:14. The anisotropy is partially restored if a magnetic field is applied perpendicular to the b axis. The transition may be related to the onset of a charge density wave. Santos and others have observed that the thermal expansion coefficient is largest along the a axis, so cooling will bring the conducting chains closer together, leading to a dimensional cross-over. The theory of Luttinger liquids then predicts such behavior. Anyway, as of 2010 there was no consensus explanation for this transition.
Lithium molybdenum purple bronze becomes superconductor between 1 and 2 K.
Li0.9Mo6O17, due to spin–charge separation, can have a much higher thermal conductivity than predicted by the Wiedemann-Franz law. 
The magnetoresistance of lithium purple bronze is negative when the magnetic field is applied along the b-axis, but large and positive when the field is applied along the a-axis and the c-axis.