The blue color mechanism on sapphires from different gem deposits before and after heating under oxidizing atmosphere

The blue color of sapphire is commonly related to the amount of Fe and Ti impurities replacing Al3+ in the Al2O3 structure. Generally, the color intensity on sapphires is related to the gem deposits including the basaltic-related and metamorphic-related ones. The color of sapphires has been changed after heating under oxidizing atmosphere. However, the explanation about the color mechanism from some previous research contradicted each other and it was still wondered. For this reason, this research is focused on the role of Fe and Ti oxidation states as well as the blue color mechanism on sapphires before and after heating under oxidizing atmosphere. In this study, the sapphire samples were collected from different gem deposits including basaltic-related sapphires from Kanchanaburi province, Thailand and metamorphic-related ones from Sri Lanka before and after heating at 1100 °C under oxidizing atmosphere. As a result, the blue color on sapphires before heating can be described as a hole color center assigned to Fe3+-Ti4+ mixed acceptor states inside an energy band gap that could receive an electron from the valence band for charge-balancing after excitation. After heating, the basaltic-related sapphires turned from dark blue to light blue and the metamorphic-related ones turned from light blue to colorless. The Fe3+-Ti4+ mixed acceptor states were decreased because a hole color center was filled by an electron from oxygen during the heating process instead of an electron from the valence band. Therefore, it can be concluded that the blue color mechanism on sapphires before and after heating under an oxidizing atmosphere can be explained by an energy band model involving the presence or absence of Fe3+-Ti4+ mixed acceptor states as well as a hole color center inside an energy band gap.


Introduction
Sapphire is an inorganic gemstone that belongs to the corundum species (Al 2 O 3 ).Generally, pure corundum is colorless.The colors of corundum including ruby and sapphire are caused by trace elements replacing Al 3+ in Al 2 O 3 structure [1][2][3].For example, the yellow color on yellow sapphire can be described by Fe 3+ impurity [4,5].While, the red color of ruby can be caused by Cr 3+ as dispersed metal ions [2,6].Especially, the blue color of blue sapphire can be caused by Fe-Ti pairs with many hypotheses either Fe 2+ -Ti 4+ intervalence charge transfer process [2,4] or Fe 3+ -Ti 4+ mix acceptor states [7].It was found that those theories contradict each other.Naturally, there are many color shades of blue sapphire from milky to dark blue based on color intensity relating to the content of trace elements and internal features as well as micro-and nano-inclusion [8].The blue intensity of blue sapphire can be changed after heat treatment [9][10][11][12][13][14][15][16] relating to the important parameters including oxidation-reduction atmospheric condition, heating temperature, soaking time and chemicals in gemstone [9].
Previously, heat treatment of blue sapphires by reducing condition could be an important factor in the changes in the color of blue sapphires from light to dark blue because when the temperature increase; Fe 3+ could be changed to Fe 2+ and increasingly produced blue colors by Fe 2+ -Ti 4+ intervalence charge transfer process [10].
Besides, heat treatment under oxidizing conditions can also be applied for changing the color of sapphires based on different heating temperatures [11,12].Low-temperature heat treatment (800 °C to 1100 °C) and high-temperature heat treatment (over 1100 °C to 1650 °C) could be produced different results.The dissolution of secondary-phase micro-crystal inclusion such as rutile silk could be involved only from high-temperature treatment affecting the increasing blue color intensity.However, the color of the sample will be faded after heat treatment at low temperatures.[11].Thus, the originated mineral inclusions on sapphires could play an important role in the blue color intensity after heating because some inclusions will be disappeared after heating at a high temperature [12].The iron and titanium dissolved from Fe-bearing and Ti-bearing mineral inclusions were the most important trace elements for the color-causing mechanism on blue sapphire after heating under oxidizing condition and the intensity of blue color after heating corresponded to the stronger and weaker intensity of the Fe 2+ -Ti 4+ IVCT absorption band [13].The slight changing from Fe 2+ to Fe 3+ on blue sapphire after heating under oxidizing atmosphere at 1400 °C were proposed by linear combination fitting calculation [14].However, the oxidation state of Fe and Ti on blue sapphire was reported as Fe 3+ and Ti 4+ before and after heating under oxidizing atmosphere no matter what temperature was used [15,16].The causes of color change and references can be summarized in table 1.
For this reason, the Fe 2+ -Ti 4+ IVCT could be questioned for clearly describing either the presence or the absence of blue coloration after heating under oxidizing condition.Therefore, this study aims to propose an explanation of the role of Fe and Ti oxidation states affecting the blue color mechanism on sapphires from different gem deposits before and after heating under oxidizing atmosphere.

Materials
According to the geological origin, blue sapphire sources were related to the basaltic origin and metamorphic one.For example, the basaltic-related blue sapphire deposits were Thailand, China, Nigeria, etc whereas the metamorphic ones were Sri Lanka, Vietnam, Madagascar, etc [17].Generally, major differences and appearance of rough sapphire from different origins could be identified from the color intensity based on chemical concentration of trace elements as well as the type of mineral inclusion relating with the associated mineral from the originated host rock, i.e., either basaltic rock or metamorphic rock [1].Mostly, the basaltic-related blue sapphires show darker blue intensity than metamorphic ones due to the higher Fe content [5].
In this study, the sapphire samples were collected from different gem deposits including Kanchanaburi province, Thailand (KAN; figure 1(a)) and Sri Lanka (SLK; figure 1(b)).Twenty samples from both sources were polished for preparing a smooth surface with a suitable sample thickness around 1 mm to 2 mm.After that, the samples were cleaned with plain water and were put aside to dry before the experiment.

Methodology
The sapphire samples were examined using basic gemological instruments before and after heating under oxidizing condition consisting of hydrostatic balance to examine specific gravity, a refractometer to examine refractive indices, a polariscope used to study optical characteristics, a gems microscope to verify internal features also known as inclusions, GIA gem set and colorimeter use for identifying color as a visual color code and color in terms of universal number, respectively.The colorimeter was AvaSpec-2048 spectrometer equipped with a light source and an integrating sphere.The measurement range was 275 nm to 1100 nm with specular component include (SCI).The CIELAB color indices was calculated by a module integrated with the AvaSoft7.5software.The CIE 1931 standard observer matching function (2˚) and the spectral power distribution illuminant source of D65 were selected.It should be note that this type of colorimeter should be used for weak or no fluorescence sample.After that, advanced gemological instruments were applied to the samples consisting including EDXRF spectrophotometer to analyze the chemical composition [18], the internal features focused on types of mineral inclusions of the samples were observed beneath gems microscope by Laser-Raman spectroscopy as well as the micro inclusions were observed by SEM-EDS [8], the absorption spectra of the samples that related with the cause of color on blue sapphire before and after heating were determined by UVvis-NIR spectrophotometer [19], FTIR spectrometer was used for analyzing the functional group in molecule i.e., C-H C-O and O-H [15] and XAS technique was focused on Fe K-edge and Ti K-edge XANES spectra to determine Fe and Ti oxidation state [7,15,16].The data from XAS spectra were analyzed by Athena for determining the Fe and Ti oxidation states [20] and Larch software for studying the site symmetry of Fe by characteristics of pre-edge features [21,22].Moreover, the energy band gaps (E g ) of the samples were calculated using a Tauc plot diagram [23].
For the heating experiment, the sapphire samples from Kanchanaburi province, Thailand and Sri Lanka were heated under oxidizing atmosphere at 1100 °C with 5 °C/minute increasing rate and 1 h soaking time by an electric furnace.
Then, the data will be analyzed and applied for creating an energy band model.Finally, the model will be used for describing the blue color mechanism on sapphires from different gem deposits before and after heating under oxidizing atmosphere.

Results and discussion
The gemological properties of sapphire samples from both gem deposits using basic gemological instruments were summarized in that the specific gravity ranged from 3.90 to 3.99, refractive indices were 1.762 and 1.770, and optical characteristics were double refraction (uniaxial).
The color code of sapphire samples before heating experiments using the GIA gem set was observed that the sapphire samples from Kanchanaburi province, Thailand were mostly found as B3/1 as well, however, two samples were shown as VB7/3 (dark, very slightly grayish, violetish blue) and another one sample was shown as B5/1 (medium, grayish, blue).Besides, the samples from Sri Lanka were only shown as B3/1 (light, grayish, blue).
After heating at 1100 °C under oxidizing atmosphere, the samples from both gem deposits were either decreased blue intensity or unchanged based on the trace element content that related to blue coloration.Although the blue intensity seemed to be decreasing, however, the color code of sapphire samples from Kanchanaburi province, Thailand was still mainly found as B3/1 except only one sample was shown as VSlgr 2/2 (very light, slightly grayish, very slightly greenish blue).The sapphire samples from Sri Lanka were still only B3/1 (light, grayish, blue) as well.However, it could be noted that two Sri Lankan samples including SLK2 and SLK5 were destroyed after the heating process because there were a lot of fractures spreading all over inside them.
Since the color observation by the naked eyes with GIA gem set could not be properly acceptable.Because of no red color appearance, we found that there were weak fluorescence due to low Cr 3+ in the samples [24].The CIELAB color indices of sapphire samples using colorimeter as shown in table 2 presented L * , a * and b * .Before the heating experiment as following figures 2(a) and 3(a), the samples from Kanchanaburi province, Thailand could be approached to -b more than the samples from Sri Lanka.These values corresponded to the color of the samples that were visually observed by the naked eyes as the samples from Kanchanaburi province, Thailand showed higher blue intensity than Sri Lanka.However, the sapphire samples from Sri Lanka showed more brightness than Kanchanaburi province, Thailand related to the milky to light blue color.
After heating at 1100 °C under oxidizing atmosphere, the colorimeter showed L * , a * and b * of CIELAB color indices of sapphire samples as figures 2(b) and 3(b).The samples from Kanchanaburi province, Thailand and Sri Lanka could be approached to zero position referring to decrease blue intensity as well.However, the brightness of the samples from both deposits could be involved.The chemical composition of the sapphire samples was indicated in table 3 showing the percentage by weight of major composition (Al 2 O 3 ) and impurities (Fe 2 O 3 , TiO 2 and Cr 2 O 3 ).The trace elements including Fe and Ti could be an important role in causing the blue color on sapphire samples.According to figure 4(a), the sapphire samples from Kanchanaburi province, Thailand (basaltic-related origin) had more Fe content than those from Sri Lanka (metamorphic-related origin) based on their geological deposits.Thus, the sapphire samples from Kanchanaburi province, Thailand were mainly darker blue color than those from Sri Lanka.On the other hand, the Kanchanaburi sapphire samples had lower Ti content than Sri Lankan ones evidencing on the Sri Lankan sapphires showed more turbidity related to Ti-bearing mineral inclusions.After heating at 1100 °C under oxidizing atmosphere, there was barely an unchanged chemical composition in terms of semiquantitative analysis as shown in figure 4(b).
In this study, the internal features inside the sapphire samples were focused on a type of mineral inclusions by Laser-Raman spectroscope as shown in table 4. According to the morphology and Raman spectrum, it could be found that there was a group of rutile mineral inclusion and negative crystal referred to as corundum inside the sapphire samples from Kanchanaburi province, Thailand.In addition, there was also a group of mineral inclusion, rutile crystal and hematite mineral inclusion among those from Sri Lanka.
The representative micro inclusions were observed by SEM-EDS.To identify various types of mineral inclusions before heating as shown in table 5, the EDS points had to be clicked for this micro inclusion profile estimate.It was found that micro inclusions such as hematite inclusion was observed inside the sapphire samples from Kanchanaburi province, Thailand.Besides, hematite and rutile inclusions were found in sapphire samples from Sri Lanka.After heating at 1100 °C, the micro inclusions were partially melted due to the heating temperature was not high enough for disseminating completely the micro inclusions in the samples.For example, the morphology of rutile inclusion generally remained because it will be absolutely dissolved at 1200 °C-1350 °C [8].
According to Laser-Raman spectroscope and SEM-EDS, it could be suggested that the presence of hematite and rutile mineral inclusions whether or not macro or micro inclusion could be the evidence for the Fe and Ti trace elements affecting the blue color of the sapphire samples.Furthermore, those mineral inclusions will be involved in the turbidity as well as transparency of the sapphire samples.The UV-vis-NIR absorption spectrum of the sapphire samples from both sources before and after heating at 1100 °C under oxidizing atmosphere were shown in figure 5.The typical absorption peaks at 377 nm and 450 nm were assigned to Fe 3+ /Fe 3+ while the peak position at 388 nm was assigned to single Fe 3+ [25][26][27].These sharp absorption peaks in blue regions responsible for the yellow color.These absorption peaks were generally found on the sapphire samples from both deposits before and after heating.According to the samples from Kanchanaburi province, Thailand, there were three absorption peaks including 580 nm, 710 nm, and 890 nm assigned to Fe-Ti pairs before the heating experiment (figure 5(a)) relating to blue coloration [7].These board absorption peaks in red regions responsible for the blue color intensity.It was noted that the peak at 890 nm was a significant position for basaltic sapphire deposit [11].After heating at 1100 °C under oxidizing atmosphere, the Fe-Ti pairs were still found at 710 nm and 890 nm whereas the peak at 580 nm was absent corresponding to the decreasing blue color (figure 5(b)).
In addition, the sapphire samples from Sri Lanka before heating showed the absorption peaks of Fe-Ti pairs at 580 nm and 710 nm without at 890 nm (figure 5(c)).It should be noted that the samples from Sri Lanka deposit (metamorphic-related sapphire) had lower Fe concentration as well as lighter blue than Kanchanaburi one (basaltic-related sapphire) and the peak at 890 nm was not found for metamorphic sapphire deposit [11].After heating at 1100 °C under oxidizing atmosphere, both Fe-Ti pairs absorption peaks at 580 nm and 710 nm disappeared and the color of the samples turned colorless (figure 5(d)).
Moreover, the energy band gaps (E g ) of the samples could be calculated by a Tauc plot method (an inserted diagram inside each figures 5(a)-(d)).Despite the color intensity of the samples had been decreased, the sapphire samples after heating at 1100 °C under oxidizing atmosphere showed nearer E g compared with the samples before heating not only from Kanchanaburi province, Thailand but also Sri Lanka deposit.It could be observed that the turbidity of the samples caused by the partial dissolution of mineral inclusions should be played an important role in the E g of the samples.
The FTIR spectra of sapphire samples from both deposits as shown in figure 6 were recorded in the region between 400 cm −1 to 4000 cm −1 .Before the heating experiment, the spectra showed absorption peaks of C-H stretching from organic matter at 2921 cm −1 and 2850 cm −1 , CO 2 at 2350 cm −1 and -Ti-OH stretching at 3309 cm −1 which are commonly found on sapphire structure especially the samples from Sri Lanka with high Ti content (> 0.02 wt%).Nevertheless, the presence of -Ti-OH stretching on the samples from Kanchanaburi province, Thailand was hardly due to low Ti content (<0.02 wt%).Thus, the amount of Ti content should be affected by the presence or absence of -Ti-OH stretching corresponding to the previous researches [15,16].
After heating at 1100 °C under oxidizing atmosphere, the band of molecular water at 3500-3100 cm −1 turned from a broad band to slightly sharp peaks for both sapphire deposits.The C-H stretching and CO 2 peaks were not changed.The -Ti-OH peak at 3309 cm −1 on sapphire structure as well as Ti-bearing mineral inclusions such as rutile was slightly decreased because the bonding between Ti and OH was weakened after the samples had undergone the heating process that affect not only the Ti from sapphire host structure but also the partial dissolution of rutile inclusion.Previously, the decrease of -Ti-OH stretching was reported that it should be related to the fading of blue color on blue sapphire after heating under oxidizing atmosphere [16].However, the coloring mechanism has been still unclearly described how Fe and Ti trace elements are manipulated on the sapphire structure after heating.
X-ray absorption spectroscopy (XAS) was applied to investigate the iron and titanium oxidation states of sapphire samples before and after heating at 1100 °C under oxidizing atmosphere for a better understanding of the role of trace elements affecting the color mechanism.In this study, the XAS technique was obtained from   Representative XAS spectra of sapphire samples from Kanchanaburi province, Thailand and Sri Lanka deposits were carried on using a fluorescent mode.To verify the oxidation states, the XAS spectra of Fe and Ti standards with actual oxidation states were measured in a transmission mode.Then, the XAS spectra of the samples and standards were analyzed by Athena software [20].
According to table 6 and figure 7, the Fe oxidation state of the samples before and after heating at 1100 °C under oxidizing atmosphere was still Fe 3+ compared with those of Fe referencing standards.The Fe standards were Fe foil, FeO and Fe 2 O 3 representing Fe°, Fe 2+ and Fe 3+ , respectively.The energy position (E 0 ) from Fe Kedge XANES spectra of sapphires from Kanchanaburi province, Thailand and Sri Lanka implied that the Fe 3+ oxidation state at around 7124 eV was unchanged before and after heating.Therefore, the pre-edge fit was deeply analyzed the oxidation state from the centroid pre-edge position as well as the site symmetry of Fe on sapphire samples using Larch software [28].
Site symmetry of Fe such as coordination environments and Fe oxidation state could be obtained from the pre-edge fit considering from characteristics of pre-edge features [22,29].In this study, before heating found that trend of the integrated intensity of sapphire from Sri Lanka is higher than in Kanchanaburi province, Thailand, which tends to be in the center as shown in figure 8(a).Meanwhile, after heating at 1100 °C under oxidizing atmosphere both sapphire samples tended to decrease in centroid position as shown in figure 8(b) due to the effect of heat relating to the different Fe local atomic environment as well as the structural arrangement on sapphire.However, it could be suggested that Fe 3+ in all samples both before and after heating at 1100 °C under oxidizing atmosphere were predominantly 6-fold coordinated.
From the energy position and pre-edge centroid position, it can be confirmed that the presence or absence of blue color on sapphire has not been involved with the IVCT process [10,13,14] because there is no evidence of the transition among Fe 2+ and Fe 3+ oxidation state on the sapphire structure during the heating process.
Table 7 showed the Ti K-edge photon energy position (E 0 ) from Ti K-edge XANES spectra.Ti foil and TiO 2 were used as the Ti reference standards, which stood for Ti°and Ti 4+ , respectively.The Ti K-edge XANES Table 5. Representative micro inclusions observed from SEM-EDS of sapphire samples before and after heating at 1100 °C under oxidizing atmosphere.
spectra of the sapphire samples from Kanchanaburi province, Thailand and Sri Lanka were compared to the spectra of the Ti standards (figure 9).As a result, it could be reported that the Ti oxidation state of the samples from both deposits remained as Ti 4+ with the photon energy position at around 4980 eV similar to TiO 2 referencing standard.Consequently, it can be approved that the Ti 4+ oxidation state is stable on the sapphire structure although the fading blue color has appeared after the heating process.
According to the results, the presence or absence of blue color on sapphires before and after heating at 1100 °C under oxidizing atmosphere can be proposed by the energy band model.There are two different cases based on the geological gem deposits:-basaltic-related sapphires and metamorphic-related ones.Before heating, the blue color mechanism on sapphire from both sources can be explained by the Fe 3+ and Ti 4+ mixed acceptor  states.These energy states inside an energy band gap create a hole color center in the sapphire structure considering the ground state.Then, the energy states require to receive an electron from the valence band for charge-balancing after an excitation.After this process, the blue color is produced on sapphire samples.Generally, the basaltic-related sapphires are darker blue than metamorphic-related ones due to there is higher Fe content which is probably involved with the possibility of Fe 3+ -Ti 4+ pairs.Therefore, the color intensity is also significantly related to the amount of Fe 3+ -Ti 4+ mixed acceptor states.For sapphires from Kanchanaburi province Thailand (a representative basaltic-related origin) with considerably higher iron content, the blue color is caused by the Fe 3+ and Ti 4+ mixed acceptor states at 1.39 eV, 1.75 eV and 2.14 eV corresponding to absorptions at 890 nm, 710 nm and 580 nm, respectively with an energy band gap of 3.99 eV (figure 10(a)).For sapphire samples from Sri Lanka (a representative metamorphic-related origin) with lower iron content, the blue color in sapphires is caused by Fe 3+ and Ti 4+ mixed acceptor states located at 1.75 eV and 2.14 eV corresponding to 710 nm and 580 nm in UV-vis-NIR spectrum, respectively; which an energy band gap of 3.72 eV (figure 11(a)).The blue color mechanism causes by Fe 3+ and Ti 4+ mixed acceptor states.Meanwhile, the sapphire samples from Kanchanaburi province, Thailand show darker blue than those from Sri Lanka because there are higher Fe 3+ -Ti 4+ energy states for producing the color.
After heating at 1100 °C under oxidizing atmosphere, the intensity of the blue color on sapphires is weakened or absented since the amount of Fe 3+ -Ti 4+ mixed acceptor states have been reduced or damaged by heat, respectively and a hole color center in the sapphire structure has been filled with by an electron from oxygen during the heating process.For this reason, an electron from the valence band is not allowed into these energy states in the sapphire structure due to the unavailable of mixed acceptor states.
In this study, the basaltic-related sapphires are light blue whereas the metamorphic-related sapphires are colorless after heating at 1100 °C under oxidizing atmosphere because the Fe 3+ -Ti 4+ pairs have partially remained on basaltic-related sapphires while these pairs have been lost on metamorphic-related ones.Although, the energy band gaps of the sapphires do not change due to the partial dissolution of mineral inclusions affecting the increasing turbidity of the sapphires after heating.The sapphires from Kanchanaburi province, Thailand show a pale blue color caused by the Fe 3+ and Ti 4+ mixed accepter states located at 1.39 eV and 1.75 eV corresponding to absorptions at 890 nm and 710 nm, respectively with an energy band gap of 4.00 eV (figure 10(b)).On the other hand, the sapphires from Sri Lanka show colorless with an energy band gap of 3.73 eV (figure 11(b)).Therefore, it can be suggested that the presence or absence of blue coloration on sapphires before and after heating at 1100 °C under oxidizing atmosphere is caused by the presence or absence of Fe 3+ -Ti 4+ mixed accepter states as well as a hole color center explained by the energy band model.

Conclusion
The blue color mechanism on sapphires before and after heating under oxidizing atmosphere at 1100 °C can be explained by the energy band model.The presence of Fe 3+ -Ti 4+ mixed accepter states inside the energy band gap  plays an important role in causing the blue color on sapphires.To produce a blue color, these states require to get an electron from the valence band to fill at a hole color center for charge-balancing after an excitation.Naturally, the blue intensity of sapphires is significantly related to the amount of Fe 3+ -Ti 4+ mixed accepter states as well as Fe concentration varying gemstone deposits such as Kanchanaburi province, Thailand and Sri Lanka referred to the basaltic-related origin and metamorphic-related one.After heating under oxidizing atmosphere at 1100 °C, the decreasing blue color on sapphires can be described by the absence of Fe 3+ -Ti 4+ mixed accepter states involving an unavailable space of a hole color center because it has completely fulfilled by an electron from oxygen during the heating process.

Figure 1 .
Figure 1.Sapphire samples in this study from Kanchanaburi province, Thailand (a) and Sri Lanka (b).

Figure 2 .
Figure 2. a * , b * and L * of CIELAB color indices of the sapphire samples from Kanchanaburi province, Thailand before heating (a) and after heating at 1100 °C under oxidizing atmosphere (b).

Figure 3 .
Figure 3. a * , b * and L * of CIELAB color indices of the sapphire samples from Sri Lanka before heating (a) and after heating at 1100 °C under oxidizing atmosphere (b).

Figure 4 .
Figure 4.Chemical concentration (%Fe 2 O 3 versus %TiO 2 ) of the sapphire samples from Kanchanaburi province, Thailand and Sri Lanka before heating (a) and after heating at 1100 °C under oxidizing atmosphere (b).

Figure 5 .
Figure 5. Representative UV-vis-NIR spectra with Tauc plots of sapphire samples from Kanchanaburi province, Thailand before heating (a) and after heating at 1100 °C under oxidizing atmosphere (b) compared with sapphire samples from Sri Lanka before heating (c) and after heating at 1100 °C under oxidizing atmosphere (d).

Figure 6 .
Figure 6.Representative FTIR spectra of sapphire samples before and after heating at 1100 °C under oxidizing atmosphere from Kanchanaburi province, Thailand (a) and Sri Lanka (b).

Figure 7 .
Figure 7. Representative Fe K-edge XANES spectra of sapphire samples before and after heating at 1100 °C under oxidizing atmosphere from Kanchanaburi province, Thailand and Sri Lanka compared with Fe referencing standards.

Figure 8 .
Figure 8. Relationship between pre-edge centroid energy position and integrated pre-edge intensity of sapphire samples from Kanchanaburi province, Thailand and Sri Lanka before heating experiment (a) and after heating at 1100 °C under oxidizing atmosphere (b).The uncertainties in the centroid position are 0.40 eV and the integrated intensity is 0.04.

Figure 9 .
Figure 9. Representative Ti K-edge XANES spectra of sapphire samples before and after heating at 1100 °C under oxidizing atmosphere from Kanchanaburi province, Thailand and Sri Lanka compared with Ti referencing standards.

Figure 11 .
Figure 11.An energy band model of Fe 3+ -Ti 4+ states in metamorphic-related sapphires from Sri Lanka before heating experiment (a) and after heating at 1100 °C under oxidizing atmosphere (b).

Figure 10 .
Figure 10.An energy band model of Fe 3+ -Ti 4+ states in basaltic-related sapphires from Kanchanaburi province, Thailand before heating experiment (a) and after heating at 1100 °C under oxidizing atmosphere (b).

Table 1 .
Summary causes of color change and references.

Table 2 .
CIELAB color indices of the sapphire samples.

Table 2 .
(Continued.) Remarks: n.d.refers to non-detected because the samples have been destroyed during the heating process.

Table 3 .
Semi-quantitative of major and trace element content in the sapphire samples measured by EDXRF.

Table 6 .
The oxidation state and centroid pre-edge position of sapphire samples compared to Fe referencing standards.