Heat Treating Ruby and Sapphire Technical Aspects
Heat Treating Ruby and Sapphire Technical Aspects
Heat Treating Ruby and Sapphire Technical Aspects
Fall 1981
121
for each process will depend on the specific material being heated; considerable variation must
be expected with corundum from different localities.
Nine possible treatment modes are summarized in table 1 and examined in turn below. Identifying characteristics of such treatment methods
have been discussed in detail by Crowningshield
and Nassau (1981);therefore, they will be touched
on here only briefly.
It should be noted that no consensus has yet
been reached as to whether any of these types of
treatment need to be disclosed to the buyer. The
parallel is sometimes drawn that since heat treatment is not customarily disclosed with other
gemstones, such as heated blue aquamarine, why
should such disclosure be necessary with sapphires and ruby? There may be validity in viewing
some types of treatment in this way, especially
those that mimic natural processes (as is the case
with aquamarine). However, other treatment
methods-diffusion in particular-do
not have
any parallel in nature. The results of diffusion
treatment, for example, are readily recognized,
and remind one more of Lechleitner synthetic
emerald overgrowth on a natural beryl than of
heated aquamarine.
PROCESS 1:
DEVELOPMENT OF
POTENTIAL ASTERISM
Some natural sapphire and ruby, as well as some
synthetic material intended for star use, contains
a significant amount of titanium oxide. If such
corundum cools fairly rapidly from its growth
conditions, the material remains clear; the titanium oxide is in "solid solution" in the form of
T i f i (titanium sesquioxide) replacing some of
the A120y(aluminum oxide).
If such material, which typically contains only
a few tenths of a percent titanium oxide, is held
at between llOO and 1500 for some time
(say, 1300 for 24 hours), particularly under
mildly oxidizing conditions, the TizOg converts
to TiOz (titanium dioxide) as follows:
Specific
processa
1. Moderate temperature
(130OoC)
2 . High temperature
(160OoC), rapid cooling
3. Reductive heating
(16OO0C)
4. Oxidative heating
(16OO0C)
5. Extended heating
(1800C
Result
Develops potential asterism
Removes silk and asterism
Develops potential blue color
Diminishes blue color
Diminishes Verneuil banding
and strain
6. ?
Produces asterismb
7. Add TiO,
8. Add Ti02 and/or Fe203 Produces blue colorb
9. Add Cr203,NiO, etc.
Produces other colorsb
Introduces fingerprint
inclusionsb
122
Fall 1981
Fall 1981
123
124
Fall 1981
This change can be achieved by an extended heating of the stone in a hydrogen atmosphere for
equation (5) or by packing the material in charcoal, graphite, or another substance that produces
carbon (such as mineral oil, sugar, and the like),
so that combustion with only a small amount of
air produces carbon monoxide
where a and b refer to different sites in the crystal. In each instance, a single electron is transferred from one atom to another atom and back
again.
It is important to note that both processes require that some of the iron be in the divalent ferrous, Fe2+,state. Also, sufficient quantities of iron
and titanium must be present in the original stone
to produce a deep blue.
A sapphire that contains adequate amounts of
iron and titanium oxides but is too highly oxidized, so that not enough ferrous iron is present,
may be pale blue, green, yellow, or colorless in its
original state. Such material may be heated in a
strongly reducing environment to convert some
Fe3+to Fez+,as follows:
The result is the slow removal of one of the essential coloring ingredients, Fe2+,on the left side
of equations 3 and 4, thus lightening the blue
color. If the process is continued long enough, a
virtually colorless stone may result. This treatment has been described by Jobbins (1971))Eigen-
Fall 1981
125
J
-,
126
4
-
-.
Fall 1981
which can lead to reduction rather than the desired oxidation if not performed carefully.
PROCESS 5 :
DIMINISHING VERNEUIL
BANDING A N D STRAIN
Extended heatings (for many days) at sufficiently
high temperatures (1600and up), such as those
associated with many of the processes discussed
here, will result in a reduction of strain and will
permit some smoothing of color irregularities.
The curved striations typical of Verneuil-grown
synthetic corundum originate from irregularities
in the growth process; their main feature is a variation in the concentration of the impurities
(Nassau, 1980a). Heating for an extended period
permits a very slow diffusion to even out some of
this variation, with the degree of improvement
depending on the duration of the heating process.
This lengthy, high-temperature process is said
to be performed in Bangkok on synthetic Verneuil
blue sapphire. When used in conjunction with
process 6 (below),it results in improved color and
less prominent curved growth lines. It is also
more difficult to observe a positive Plato test in
stones treated in this manner (Crowningshield
and Nassau, 1981).A similar procedure should be
possible in ruby and other colored corundum. The
small gas bubbles associated with the Verneuil
technique probably cannot be removed in this
manner.
PROCESS 6:
INTRODUCING FINGERPRINT
INCLUSIONS
As reported by Crowningshield and Nassau (19811,
some samples of "heat-treated natural blue sapphire" obtained from Bangkok turned out to be
Vemeuil synthetic sapphire with induced fingerprint inclusions. Both Verneuil synthetic ruby
and pink sapphire showing fingerprints were also
reported by Crowningshield (1980). Judging from
the characteristics of these stones, it is clear that
an extended heat treatment similar to that of process 5 was involved.
At present, nothing definite is lznown about
the treatment used with synthetic stones to mimic
the fingerprint inclusions of their natural counterparts. The simultaneous occurrence of fingerprints, curved but weakened Verneuil banding,
and occasional gas bubbles is clear evidence of
such a treatment. According to some unsubstan-
tiated reports, a flux-type chemical such as sodium carbonate or borax may assist in this process. It should be noted that the formation of
inclusions seems to be limited to a region close
to the surface of the stone.
THE DIFFUSION MECHANISM
Diffusion in solids is a mechanism by which atoms may be moved from one region to another.
The amount of movement increases with both
the temperature and the length of the heating.
Atoms of oxygen or hydrogen can move very rapidly in corundum, which explains why the effects
of processes 3 and 4 will penetrate fully throughout a stone in as little as a few hours in some
cases (and no more than a day or so in others).
The formation or removal of the rutile needles of
silk and asterism in titanium-containing corundum is also diffusion controlled. Although titanium diffuses very slowly, the distances involved
are so small, only a few micrometers, that processes 1 and 2 as well require only a day or so to
be effective. The banding in Verneuil-grown corundum is much coarser; this explains why extremely long heating would be required for the
near-total removal of the banding by process 5,
which involves the movement of the slowly diffusing color-causing transition metals such as
chromium, iron, titanium, and nickel. It is not
lznown if a total removal of the banding is possible, since other factors (the dislocation structure, for example) may prevent this from occurring.
The movement of color- and star-forming atoms into and within corundum is a very slow process; as a result, the effects of processes 7, 8, and
9, discussed below, are limited to a relatively thin
skin on and just below the surface, typically to a
depth of a few tenths of a millimeter. Very high
temperatures must be used to obtain significant
penetration in a reasonable time, since the fuel
costs for these forms of heat treatment are considerable. As a consequence of the high temperatures required and the thin film that results,
these treatment methods cannot be performed on
rough material but must be applied to a preform
or a cut stone; even so, only the lightest of polishing (or repolishing, since the surface is roughened by the treatment) can be used or the affected
skin will be completely removed.
It is the localization of the effect of these treatment methods just below the surface of the stone,
Fall 1981
127
the restrictions on polishing, and the high temperatures required that provide the clues to identifying that these processes have been used. Viewing such stones while they are immersed in
methylene iodide reveals both the localized effect
and a blotchiness from the combination of uneven diffusion and light repolishing (figures 7 and
8). Other signs are stress fractures, pockmarlzed
faces, and abnormal girdles as discussed under
processes 2, 3, and 4, and shown in figures 2
Figure 9. "Bleeding" of color around cavities
and fractures in a diffusion-treated sapphire.
Magnified 15 X. Photomicrograph by
Robert E. Kane.
128
Fall 1981
eting; it can be recognized in stones by the unnaturally sharp stars caused by very fine rutile
needles, by uneven color, by "bleeding" of the
color at pits and fractures, and by other evidence
of high-temperature treatment such as stress fractures and pockmarlzed surfaces.
PROCESS 8:
ADDING BLUE BY DIFFUSION
If insufficient iron oxide or titanium oxide is present in a colorless, yellow, green, or pale blue sapphire, it is possible to add either or both of the
missing ingredients by diffusion. A reducing atmosphere as inprocess 3 is required, or a separate
reductive heating step must follow oxidative diffusion. This type of diffusion is extremely slow,
so that even with extended heating the penetration will be shallow. The result is a relatively thin
skin of dark blue. This process was described in
detail by Carr and Nisevich (Linde patent) in 1975
for the combination with process 7, and subsequently as a separate process (Carr and Nisevich,
1976, 1977). It appears to be in wide use (Crowningshield, 1980; Fryer, 1981), both on faceted
stones and on cabochons. Identification includes
immersion in methylene iodide as well as the
other clues described above for process 7.
When Linde stopped producing synthetic gems
about 1975, these U.S. patents were assigned to
the Astrid Corporation Ltd. of Hong Kong, a firm
set up to take over Linde's star corundum stock.
PROCESS 9:
ADDING COLORS OTHER THAN
BLUE BY DIFFUSION
Just as diffusion of iron and titanium oxide can
produce a blue slzin, so can other color-causing
impurities also be diffused, again as described by
Carr and Nisevich (1975, 1976, 1977). Thus, the
diffusion of chromium produces a red slzin, nickel
gives yellow, chromium plus nickel creates the
pinkish orange "padparadscha," and so on; the
colors that will be produced by the diffusion of
different substances are well known from sapphire synthesis (Nassau, 1980a). Orange-red diffused stones (figure 10) were examined in detail
by Crowningshield in 1979. There may be problems associated with oxidation/reduction conditions if variable valence ions such as iron are involved; the identifying characteristics are the same
as those given for processes 7 and 8.
The colors produced by these diffusion processes are just as stable as the natural and synthetic colors produced by the same impurities.
This contrasts with the yellow to orange color
produced by irradiating sapphire, which is unstable and will fade on exposure to light,
COMBINATION TREATMENTS
As suggested above, several of these processes can
be combined. Removal of silk and intensification
of color can be achieved in a pale sapphire by
heating the stone in a reducing atmosphere and
then cooling it rapidly, in a combination of processes 2 and 3. The removal of silk will produce
an improvement in color even in the absence of
any other changes, since scattered white light no
longer dilutes the color. Among the chemicals
employed in this process as used on milky white
to pale blue "geuda" Sri Lanlza sapphires are a red
liquid reported by Crowningshield (1980)and since
shown to be mineral oil (its role is described under process 3), the soda (sodium [bi-]carbonate?)
reported by Harder (1980)which is said to prevent
cracking of the stone and possibly to remove some
iron preferentially (?), and the "thick coating of
local paste" used in oil, gas, or electric furnaces
as mentioned by Gunaratne (1981).
Fall 1981
129
130
REFERENCES
Bratton R.J. (1971) Precipitation and hardening behavior of
Czochralski star sapphire. Journal of Applied Physics, Vol.
42, pp. 211-216.
Burdick J.N., Glenn J.W. (1949) Developing asterism in star
corundum. United States Patent 2,488,507, filed Deceinher 11, 1948, issued November 15, 1949.
Carr R.R., Nisevich S.D. (1975)Altering the appearance of sapphire crystals. United States Patent 3,897,529, filed December 15, 1969, issued July 29, 1975.
Carr R.R., Nisevich S,D. (1976)Altering the appearance of corundum crystals. United States Patent 3,950,596, filed
December 20, 1971, issued April 13, 1976.
Carr R.R., Nisevich S.D. (19771Altering the appearance of corundum crystals. United States Patent 4,039,726, filed
December 20, 1970, issued August 2, 1977.
Crowningshield R. (1979) Developments and highlights at
GIA's lab in New York. Gems a ) Gemology, Vol. 16, pp.
194-202.
Crowningshield R. (1980) Developments and highlights at
GIA's lab in New York. Gems &> Gemology, Vol. 16, pp.
315-323.
Crowningshield R., Nassau Kt (19811.Journal of Gemology, in
press.
Eigenmann K.,Gunthard H.H. (19721 Valence states, redox
reactions, and biparticle formation of Fe and Ti doped sapphire. Chemical Physics Letters, Vol. 139, pp. 58-61.
Eversole W.G,, Burdick J.N. (1954) Producing asteriated corundum crystals. United States Patent 2,690,630, filed
December 28, 1951, issued October 5, 1954.
Falkenberg R. [I9781 The Verneuil process. In C.H.L. Goodman, Ed., Crystal Growth, Vol. 2, Plenum Press, New
York.
Ferguson J., Fielding P.E. (1972)Origins of the colors of natural
yellow, blue, and green sapphires. Australian Journal o f
Chemistry, Vol. 25, pp. 1371-1385.
Fryer C. Ed. (1981)Gem Trade Lab notes. Gems &> Gemology,
Vol. 17, pp. 40-46,
Gunaratne H.S. (1981)'Geuda sapphiresf-their colouring elements and their reaction to heat. Journal of Gemmology,
Vol. 17, pp. 292-300.
Harder H. (1980) Edelsteine durch Brennen von Korunden.
Fortschritte der Mineralogie, Vol. 58, pp. 45-46.
Jobbins E.A. (1971)Heat-treatment of pale blue sapphire from
Malawi. Journal of Gemmology, Vol. 12, pp. 342-343.
Lang S.M., Fillmore C.L.,Maxwell L.H. (1952)Journal o f Research of the U.S. National Bureau of Standards, Vol. 48,
pp. 298-312.
Lehmann G., Harder H. (1970) Optical spectra of di- and trivalent iron in corundum. American Mineralogist, Vol. 55,
pp. 98- 105.
Fall 1981
Lejus A.M., Goldberg D., Revcolevschi A. (1966) New compounds formed between rutile, TiOa, and oxides of trivalent and quadrivalent metals. Cornptes rendus, Academic
des Sciences, Paris, Series C, Vol. 263, pp. 1223-1226.
Nassau K. (1968) O n the cause of asterism in star corundum.
American Mineralogist, Vol. 53, pp. 300-305.
Nassau K. (1980s) Gems Made By Man. Chilton Book Co.,
Radnor PA.
Nassau K. (1980b) The causes of color. Scientific American,
Vol. 243, pp. 124- 154,
Nilzolskaya L.V., Terekhova M., Samoilovich M.I. (1978) O n
the origin of natural sapphire color. Physics a n d Chemistry
of Minerals, Vol. 3, pp. 213-224.
Oughton J.H. (1971) Fakes and frauds-caveat emptor. Australian Gemmologist, Vol. 11, pp. 17-22.
Phillips D.S., Heuer A.H., Mitchell T.E. (1980) Precipitation
in star sapphire. 1. Identification of the precipitate. Philo-
Fall 1981
131