Gem News International Gems & Gemology, Fall 2021, Vol. 57, No. 3

Heat Treatment Effects on the Behavior of the 3161 cm–1 Feature in Low-Iron Metamorphic Yellow Sapphire

FTIR peak height plots of sapphire before and after heating.
Figure 1. Plot of FTIR peak height at 3161 cm–1 of representative yellow sapphires from Sri Lanka (above) and Madagascar (bottom), before and after heating at different temperatures between 500° and 1050°C for six hours.

Heating is the most common treatment to improve the color and/or clarity of corundum. Untreated natural yellow sapphires from metamorphic deposits often have low color saturation; therefore, heat treatment is typically performed to intensify the yellow coloration. However, the chemical composition of the stones, the heating temperatures employed, the duration of the treatment process, as well as the composition of the furnace atmosphere—reducing or oxidizing—are also important factors that influence the color alteration (K. Nassau, “Heat treating ruby and sapphire: Technical aspects,” Fall 1981 G&G, pp. 121–131; J.L. Emmett et al., “Treatments,” in R.W. Hughes, Ed., Ruby & Sapphire: A Gemologist’s Guide, 2017, pp. 197–247).

The identification of heat treatment in a yellow sapphire is challenging for gemologists. Apart from visible inclusion changes, Fourier-transform infrared (FTIR) absorption spectra are one of the few remaining clues to identify the treatment of low iron content metamorphic yellow sapphires. These sapphires as formed are acceptor dominated when [Mg2+] > [Ti4+] + [Si4+] with the charge compensation being accomplished totally or partially by hydrogen (H+) (E.V. Dubinsky et al., “A quantitative description of the causes of color in corundum,” Spring 2020 G&G, pp. 2–28). In corundum, H+ forms a bond with oxygen creating OH. It is the OH that is responsible for the broad peak at 3161 cm–1 (and the series of peaks at 3161, 3242, and 3355 cm–1) when associated with Mg2+ (C.P. Smith et al., “Infrared spectra of gem corundum,” Fall 2006 G&G, pp. 92–93). However, the origin of the 3161 cm–1 peak is not well understood. The article by N. Fukatsu et al. (“Incorporation of hydrogen into magnesium-doped α-alumina,” Solid State Ionics, Vol. 162, 2003, pp. 147–159) showed that when OH is incorporated into Mg2+-doped synthetic sapphire, the 3161 cm–1 peak was not present but a broad band appeared in the 3000 cm–1 region. If the hydrogen is eliminated either in nature or in the laboratory, the trapped hole (h) forms to provide charge compensation. The trapped hole pairs with Fe3+, creating a trapped-hole-Fe3+ (h-Fe3+) pair, which is a very strong yellow chromophore in corundum. In this report, the authors aim to study the effects of heat treatment in changing the 3161 cm–1 feature and creating the yellow coloration.

Twelve yellow sapphires reportedly from Sri Lanka and Madagascar that initially displayed a weak to strong 3161 cm–1 peak or a 3161 cm–1 series in FTIR were selected for heat treatment in an oxidizing atmosphere (air) at 500°, 700°, 900°, and 1050°C for a fixed duration of six hours at each temperature. For a final step, some stones were heated at 1550°C in pure oxygen to assure the complete elimination of hydrogen. The samples were fabricated as optical wafers with at least two polished surfaces perpendicular and/or parallel to the c-axis. The samples possess a sufficiently large and clean area to get high-quality FTIR spectra. After each heat treatment step, FTIR spectra were collected at the same position for each sample by fixing it in an identical position in the sample holder.

Heat treatment can cause the H+ to move to different sites by diffusion or at sufficiently high temperatures to diffuse completely out of the sample. The diffusion coefficient of hydrogen in corundum is exponentially dependent on temperature. After heat treatments at 500°C and 700°C, the intensities of 3161 cm–1 peaks and the color appearance of the stones were essentially unaltered. Thus, there is little outward diffusion for a sample of this size. However, heating at 900°C and above for six hours in air will begin to diffuse the hydrogen out of the lattice. As the hydrogen begins to diffuse out, trapped holes form and pair with Fe3+ to maintain charge compensation, increasing the yellow coloration.

At the same time, the amplitude of the 3161 cm–1 feature will decrease proportionally. The test at 1550°C for six hours was conducted in pure oxygen to eliminate any possibility of the water vapor in air contributing hydrogen. With this last step, the outward diffusion of all the hydrogen in the stone was complete, eliminating OH peaks in FTIR spectra and maximizing the yellow trapped-hole coloration.

Figure 1 shows these heating processes as applied to stones from Sri Lanka with a strong 3161 cm–1 feature and to some of the stones from Madagascar with a weak 3161 cm–1 feature. As expected, both types exhibited a similar diffusion reduction in the initial hydrogen content with temperature and time. The coloration of some of the Madagascar stones was zoned, with areas colored by the h-Fe3+ pair, by h-Fe3+ plus Fe3+, and some by Fe3+ only, depending on the distributions of Fe3+ and h-Fe3+ in the stone. Usually, the zones in which h-Fe3+ is dominant will increase yellow coloration with high-temperature heat treatment by removing the hydrogen, whereas the areas with Fe3+ alone are rarely strongly zoned.

FTIR spectra of a Sri Lankan yellow sapphire before and after heating.
Figure 2. Sri Lankan sample no. 2801 showing a 3161 cm−1 series (red) in FTIR with initially light yellow color, then altered to 3000 cm−1 broad band series (blue) with a bit stronger yellow zone after heating at 900°C for six hours. Finally, the broad bands disappeared (green) and a much stronger yellow coloration was generated at 1550°C, thickness 3.748 mm, 190 ± 41 ppma Fe.

Interestingly, a characteristic peak at 3161 cm–1 in certain Sri Lankan samples could be occasionally transformed to the broadband series in 3000 cm–1 region with a broad band at 2625 cm–1 when heated at 900°C and above in air, as presented in figure 2. Previously, the 3000 cm–1 broad band series has been reported as indicative of heat treatment observed in Punsiri-type heated blue sapphires (G. DuToit et al., “Beryllium treated blue sapphires: Continuing market observations and update including the emergence of larger size stones,” GIA Research News, 2009, and could be observed in unheated high-Fe yellow sapphires from basalt-hosted deposits, such as Thailand and Australia (Fall 2016 GNI, pp. 325–327). However, the 3000 cm–1 broad band series could also be found in heat-treated Sri Lanka yellow sapphires, as seen from this study.

In summary, heat treatment was carried out in an oxidizing atmosphere to increase the trapped-hole color centers and deepen the yellow coloration. Sri Lankan samples colored mainly by the h-Fe3+ pair can produce stronger yellow coloration by heat treatment, but only at high temperatures (900°C and above). The heating temperature at 900°C starts to change the 3161 cm–1 peak/series in FTIR and the color appearance in metamorphic yellow sapphires. It may well be that heating for a much longer time at 900°C would more completely diffuse out the hydrogen. The amplitude of the 3161 cm–1 peak is insignificantly altered when heated at low temperature (below 700°C) but can be greatly reduced when heated at 900°C and above, and sometimes converted to the 3000 cm–1 broadband series, and then it disappears at much higher heating temperature. Although the 3000 cm–1 broadband series might not be indicative of heat treatment in basalt-related yellow sapphires, it is able to indicate heat treatment in Sri Lankan yellow sapphires. Careful consideration should be exercised when using FTIR to identify heat treatment in corundum.

Ungkhana Atikarnsakul is a staff gemologist at GIA in Bangkok. John L. Emmett is director of Crystal Chemistry in Brush Prairie, Washington and a consultant to GIA.