CeramicsWeb banner

Home      Education   Glazes   Materials   Video   Publications   Software   Links       Main Index

Titania in Glazes

(c) 1994 Karl Platt
(placed on the CeramicsWeb by permission of the author)

Titania (TiO2) per se is sparsely soluble in silicate rich melts
(such as ceramic glaze), and owing to this it has mainly been used
throughout history as an opacifier in glasses, glazes and enamels of
all type. Presently, it is also widely used in white paper as a
pigment (often along with clay as a filler). Titania (TiO2) has
effects marked on colors produced by the usual transition elements (V,
Fe, Cr, Co, etc), and it may also produce interesting blue tones on
its own. This latter effect has been more widely used in ceramic
glazes than anywhere else in the fire arts.

When added to an SiO2 rich parent glass, to the extent that Ti can be
dissolved by the given composition, TiO2 (as it were) assumes a spot
in the Si-O glassy network effectively replacing Si -- Ti-O, in
effect, replaces Si-O. In this role Ti would serve to weaken the
silicate network somewhat, and, indeed, Ti has found use as such in
the manufacture of glasses as diverse as fused quartz and vitreous
enamel -- it lowers viscosity in the former, and adds chemical
durability to the latter.

In the sort of raw parent glaze called for by "rutile-blue" glazes
seen in previous Clayart threads, we could surmise that from 3-7 wt%
of TiO2 might be accomodated before certain opacity was given.
Compared with ZrSiO4 (Opax or Ultrox), TiO2 is a less efficient
opacifier -- and one with collateral effects on coloration.

In a garden variety silicate melt, small amounts of Ti, when present
in the melt as Ti4+, give no color. Rather, the desirable purple-blue
color is given by Ti3+ (Ti2O3 in "oxide" terms) which must be
"reduced" from Ti4+. This is not so simple as employing a smoky fire,
but typically requires a glaze/glass composition having a notable
amount of B2O3 and a relatively low alkali content (or high acidity).
It would seem that the best results would be with compositions that
are trisilicates or better -- i.e. where the SiO2 in the empirical
formula is 3 or more. In any event, the higher the better, which
suggests that higher temperature glazes (Cone 7+) would furnish the
most reliable parents for the Ti3+ blue. The limiting factor for SiO2
content will be the amount of B2O3 necessary to furnish a
sufficiently weak glassy network so to allow the Ti4+ to snag an
electron (become "reduced") when a rich flame is in the kiln to become
the colorful Ti3+.

The Ti3+ color demands the use of Boron (B) in the glaze. B2O3 added
to a glaze/glass at the expense of SiO2 in amounts <10% will mainly
have B ions arranging themselves between 4 oxygens just as with Si4+
or Ti4+. However, the strength with which B holds those oxygens is
quite a bit weaker than that seen with Si-O or Ti-O. As such adding
B2O3 to a silicate glass at the expense of SiO2 will tend to raise
thermal expansion, increase fluidity and so on. Note, though, that
when substituting B2O3 for alkali, it would have somewhat the opposite
effect, as B2O3 is a glass former. When B2O3 replacement for SiO2
exceeds 10 wt%, the B ions rearrange themselves between 3 O's. This
structure is vastly weaker than the previous arrangement where B and
Si resided between 4 O's. The effect of this greatly disrupted glassy
network is to show well lowered melting temperature and chemical
durability coincident with vastly increasing thermal expansion -- The
glaze would tend toward being runny, shiny, soft and crazed.

In the interest of developing the Ti3+ purple-blue it is desirable to
have the SiO2 network subtantially perturbed by the presence of B2O3,
say, on the order of 3-7 wt% or so. High fluidity is usual for a glaze
supporting the Ti3+ color. Adding too much B2O3, however, for SiO2
reduces the acidity of the melt, which operates against the
development of the purple-blue.

The Ti3+ color is typically produced in glazes in which Ti is added in
as Rutile. Rutile is, generally speaking, pretty trashy stuff.
Depending on where it came from, and whether or not any benefication
was given to it, Fe2O3 may occupy up to 15 wt% of the material. In
shooting for the Blue-purple color, the presence Fe probably has less
to do with any absence of anticipated color than the diminished
amount of Ti in a given weight of the raw material. Although I'd
venture there are probably a few not fully appreciated effects Fe has
on redox within the molten glaze. To a point, when a Ti deficient lot
of rutile is on- hand, it is sometimes useful to add a bit of TiO2 as
such as a supplement to ensure enough Ti makes it into solution with
the glaze to give a color of decent density. As with any solution
color in glazes, the purple-blue given by Ti3+ is more pure when
produced in a glass/glaze with K2O ( K+1) predominating as the

In the event that Ghastly Borate is used to provide B2O3, there are a
host of factors which can operate to modify the amount of B2O3
actually available in the fusion -- not the least of which is soluble
borate or variable composition. Ghastly Borate is, like Pewtile,
rarely consistent lot-to-lot. There is no compelling reason for anyone
to use Ghastly Borate in Ti3+ glazes -- or for that matter in any
other. It is the pariah material. Its ilk mandated the use of frits
even in ancient times and to use it is to invite absurd results.

If it must be used test it and when testing a new lot of this stuff,
the goal is to determine whether the powder is rich in either CaO or
B2O3. This is more or less simply learned by making a substitution
trial and seeing whether the test glaze is harder, softer or the same
as with the old lot. A harder melt indicates more CaO and vice-versa.
The result will suggest the appropriate adjustment needed for the new
lot of material. Excess B2O3, again, reduces the acidity of the melt,
which is not in the interest of producing Ti3+, and insufficient B2O3
leaves the glassy network too firm to permit the Ti to fetch an
electron on reduction firing.

Simply put, using Pewtile and Ghastly Borate in the same glaze is
asking for it. It would be of interest, I believe, for someone to run
a series of trials incorporating the Ti into a borosilicate frit as a
precursor to the Ti3+ blue.

With that out of the way... the presence of Al2O3 in the glaze serves
to add durability and to control viscosity and thermal expansion. It
has relatively little effect on color development. Al2O3 is, however
important to consistent results. Given half a chance Ti3+ would
happily revert to become the either colorless Ti4+ or to precipitate
as TiO2 -- depending on concentration and the cooling curve. The
glaze, therefore, should be compounded to have a steep as practical
temperature/viscosity relation. This will retard the ability of the
electron snatched by the Ti during reduction from going elsewhere on
cooling. This suggests the use of as much Al2O3 as is consistent with
the needed results and the use of CaO in favor of, but not necessarily
to the exclusion of any other RO component -- such as MgO. Practical
factors (and intuition) will typically dictate the rest.

Ti has substantial effects on the colors produced by the standard
group of transition elements added to glasses/glazes as colorants.
These are sometimes revealed when off- spec. clay with high amounts of
tramp Ti is used in body or glaze. Ti, in sufficient amounts, has
the habit of shoving the coloring ions into the silicate network where
they attempt to replace Si, and in so doing will alter the more usual,
if you will, colors given by these elements. The teal blue of Cu2+ is
shifted to green, Fe3+ furnishes brown (widely used for pill bottles),
U6+ turns bright yellow and its fluoresence is diminished.

It has been my observation that Ti3+ glazes were most interesting when
containing Ti in amounts slightly exceeding its maximum solubility for
the given parent composition. Here lovely play with opalesence and
color is had. The amount of Ti available to be forced in to a position
to yield color is at a maximum, and just enough excess Ti exists to
lend shading, texture and depth. The use of phosphates in glazes
compounded to give Ti3+ blue enhances these interesting effects.

P2O5, Like SiO2 or B2O3, will form glass all by itself if melted and
cooled by usual means. None of these glasses mix very well at all.
That is, even though one might mix and melt a composition containing
all 3 glassformers, the likely result will be opalesence or
turbidity. It doesn't take much P2O5 to enhance these effects, usually
less than 1 wt%. However, P2O5 is volatile in the fire and an excess
of this added to the batch will ensure enough remains to give the
intended result.

All of this, of course, ignores the nephelauxetic effect, or the
photosensitivity of glazes from which crystalline rutile is
precipitated, which we'll save for later.

Karl P. Platt 11/94 All Rights Reserved



back to the top

CeramicsWeb bottom border

Home     Education   Glazes   Materials   Video   Publications   Software   Links       Main Index