Gemmological Properties of Opals from Spencer, Idaho, USA

Abstract

Fifteen opal samples of gem quality from the Spencer Opal Mines in Idaho (USA) were examined. These opals display play-of-colour, with three of them having a pink body colour. The analysis was conducted using standard gemmological tools, along with non-destructive spectroscopic, chemical, and imaging methods. All samples were identified as opal-A, with their play-of-colour resulting from a regular stacking of hydrated silica spheres. However, at times, the spheres are large to allow diffraction. Opal-A are only occasionally found in nature as gems. Inclusion observed in the studied samples included alunite (KAl₃(SO₄)₂(OH)₆), anhydrite (CaSO₄), and iron-rich spherules (hematite). The pink body colour in the studied specimens is caused by the scattering of light from reddish dendrites. The samples show significant stratification, with fluorescence ranging from green to blue to orange, the latter of which represents a new phenomenon in opals not associated with quinones. Additionally, the orange fluorescence is accompanied by emissions from Sm²⁺ in anhydrite inclusions.

1. Introduction

Opal is a hydrated, poorly crystallized silica mineral, used in jewellery mostly when it displays play-of-colour (a.k.a. “precious opal”). This play-of-colour is the result of the diffraction of light on a network of spheres or lepispheres all of the same size, when the diameter is within a certain range (about 250 nm) [1,2,3]. The body (or mass) colour of opals may vary. It typically ranges from white to black, and from orange to red in the case of “fire” opals [4,5]. Opals might also occur in other colours, including pink, which is mainly linked with mineral inclusions [6,7]. At present, the main producing countries of gem-quality precious opals are Australia, Ethiopia, Brazil, and Mexico [8]. However, other countries produce gem opals to a lesser extent, including Honduras, Indonesia, Slovakia and the United States (US). In the US, several deposits can be found, with the best-known in Nevada, Oregon, Idaho, and Louisiana. These gems are less used in jewellery but appreciated by mineral collectors; however, scientific studies on them are scarce.

In Idaho, the main opal mine is located approximately 8 km (5 miles) east of the small town of Spencer, in a rhyolite (called the “Snake River Plain”, see Figure 1A). It is the result of an intense volcanic activity dating from the Pliocene to Quaternary periods, and it crosses the southern half of the state of Idaho [9]. This plain extends for several hundred kilometres from the border with Oregon to the western boundary of Yellowstone National Park in Wyoming. Opal was formed by the successive deposition of silica in cracks and cavities through the infiltration of meteoritic (i.e., rainfall), and not meteoritic, water. This water, heated by local volcanism at the time, dissolved silica in the rhyolite as well as the vitrophyre and, as it settled, deposited a gel which progressively transformed into opal [9,10].

Opal mining in the Spencer area began in the late 1890s with various mines; several outcrops of common and precious opal have been discovered in the state. However, none of these mines had the commercial success of deposits outside US. The Spencer Opal Mine is an open pit on a hillside (Figure 1B). This mining site was discovered by chance in 1948. The opals from this mine can be found on the market polished and, in some cases, even set in jewellery, notably at the Tucson show [11], although this is still rare. Many are sold as handcrafted jewellery made directly on the mining site, at the “Opal Country Cafe and Gift Shop and Spencer Opal Mines” (mine, shop and restaurant, https://www.spenceropalmines.com/ accessed on 14 February 2025). It has been also mentioned that most of opals with play-of-colour are thin, and triplets are made to better use them in jewellery [5].

The Spencer Opal Mine is the only operating in the region today. It produces opals with play-of-colour. Their body colour is most often white, but less frequently yellow to ochre and, more rarely, pink or even purple (Figure 2A,B). These colours are thought to be due to manganese and iron (Erin Mortensen, pers. comm. 2024). Extremely rare star opals have also occasionally been found in this mine [12]. The first SEM observations of “volcanic” opals from Idaho, which were made in 1964 [1], show indirectly that these are opal-A. At this time, though, the A or CT identification of opals and the resulting difference in structure were not yet known. These were established by X-ray diffraction in the 70s, with the first Raman studies of opals appearing in 1997 [13]. The purpose of this article is to better document the little-studied opals from the area near Spencer, Idaho, US.

2. Materials and Methods

In this study, 15 opals were examined (presented in Figure 3 and

Table 1. Description of opals studied. All samples have a white body colour (unless otherwise stated) with spots exhibiting play-of-colour.

Sample	Dimensions (mm)	Weight	Comments
HU001	63.83 × 48.80 × 19.43	88.07 g	Evidence of matrix
HU002	52.32 × 37.58 × 15.27	48.33 g	Evidence of matrix
HU003	22.23 × 18.93 × 4.81	16.01 ct	Pink body colour
HU004	14.10 × 13.22 × 3.54	4.38 ct	Pink body colour
HU005	25.94 × 11.57 × 6.62	3.19 g	Evidence of matrix
HU006	32.19 × 24.97 × 12.01	13.17 g	Evidence of matrix
HU007	34.84 × 23.34 × 8.49	8.78 g	Evidence of matrix
HU008	26.55 × 25.27 × 5.87	29.86 ct	-
HU009	10.84 × 9.94 × 4.32	2.54 ct	-
HU010	11.19 × 9.20 × 4.13	1.94 ct	-
HU011	10.78 × 5.51 × 2.18	0.76 ct	-
HU012	63.53 × 43.36 × 13.02	46.12 g	Evidence of matrix
HU013	85.45 × 71.62 × 50.12	348 g	Evidence of matrix
HU014	19.04 × 15.17 × 11.26	4.55 g	Evidence of matrix
HU015	12.38 × 12.37 × 7.28	1.48 g	Evidence of matrix + Pink body colour

). The samples were purchased from 3 different sources (see Acknowledgments), with most of the samples acquired from Devan Harris (an American lapidary specializing in Idaho opals). The study samples were mined in the 70s and 80s and have only recently been polished (D. Harris, pers. comm. 2022).

Standard gemmological tools were used for the study of the specimens. Refractive index and specific gravity were measured using a Topcon refractometer and a Mettler Toledo JP1603C scale (Mettler-Toledo International Inc., Columbus, OH, USA); the specific gravity was taken on samples without a matrix. Luminescence was observed under a 6 W ultraviolet (UV) lamp (Vilber Lourmat -VL-6.LC; VILBER, Marne-la-Vallée, France) with long-wave (LWUV, 365 nm) and short-wave (SWUV, 254 nm) ultraviolet light, inside a CN-6 dark room (with 10 cm distance between the sample and the lamp). Luminescence was compared to 5 diamond luminescence standards (none, weak, moderate, strong and very strong luminescence).

The exact nature of the 15 samples and the identification of inclusions in samples HU001, HU002, HU005, HU006, HU008 and HU013 were established using a Renishaw inVia Raman (Renishaw plc, Wotton-under-Edge, Gloucestershire, UK) equipped with a 514 nm DPSS laser (diode-pumped solid-state; Cobolt AB -Hübner photonics-, Solna, Sweden) with a spectral resolution of 1.5 cm⁻¹ (grating of 1800 L/mm) and coupled to a Leica DMLM optical microscope (Leica Camera AG, Wetzlar, Germany) with a ×50 lens. The laser power on the samples was 4 mW for these spectra, and no polarizer was used during the Raman measurements. Raman spectra were obtained from the 100 to 4000 cm⁻¹ region with 1 to 10 accumulation and 10 to 15 s exposure time. A Bruker MultiRam Fourier-Transform (FT) Raman (Bruker Corporation, Billerica, MA, USA) with a liquid nitrogen-cooled detector equipped with a 1064 nm laser with a resolution of 2 cm⁻¹ and an acquisition time of 1000 scans was also used to acquire spectra from 100 to 4000 cm⁻¹ on samples HU003 and HU004 using the IMN’s characterization platform (PLASSMAT, Nantes, France).

The ultraviolet–visible (UV–Vis) spectra were obtained with a Jasco 670 UV-Vis (JASCO Corporation, Hachioji, Tokyo, Japan), with a spectral range of 250 to 850 nm, an interval of 2 nm, and a scan rate of 50 nm/min in reflectance on samples HU003 and HU004. In order to avoid the influence of light diffraction, the spectra were taken in areas without play-of-colour. Various measurements were taken per sample depending on their size and varying the positions.

Luminescence spectra were acquired using an EXA from MAGILABS (Magilabs Oy (Ltd.), Helsinki, Finland). The excitation wavelengths were 254 nm for short-wave UV (SWUV) and 365 nm for long-wave UV (LWUV). A Horiba Jobin-Yvon Fluorolog 3 (Horiba Jobin-Yvon GmbH, Oberursel, Germany) from Nantes Université was also used for excitation spectra.

Microstructure was imaged with a JEOL JSM 7600F (JEOL Ltd., Tokyo, Japan) scanning electron microscope (SEM) equipped with a field-emission electron gun using the IMN’s characterization platform (PLASSMAT, Nantes, France). Samples HU009, HU014, and HU015 only were imaged. They were first treated with hydrofluoric acid diluted to 10% volume and then coated with platinum to eliminate charging.

3. Results and Discussion

All studied samples are opaque and with a white (N9 to N10 [5]) body colour, with slightly yellow as well as pink areas (HU003, HU004 and HU015). Nine out of the fifteen samples still have a matrix in various proportions (see again Table 1). All of them present play-of-colour (although difficult to photograph), most often forming pinpoints and rarely larger patches.

With the unaided eye, it was also observed that the studied opals are made of thin coloured strata (Figure 4A) which frequently show a play-of-colour. Five samples have thin cracks on the surface (HU001, HU002, HU003, HU004, HU006, HU007, HU012 and HU013), a sign of possible destabilisation by cracking. Numerous red spherules were observed in the studied opals (e.g., HU003, HU004, HU014 and HU015), sometimes being agglomerated (Figure 4B,C). These inclusions were identified as hematite-rich through Raman spectroscopy (see below). They form thin layers (Figure 5A) and scatter light depending on the illumination’s orientation, giving the opal a pinkish colour. Epigenetic orangey-red dendrites were observed in large quantities in the cracks in the two pink opals and partially or completely contribute to their body colour (Figure 5B1,B2). It is important to note that the body colour of many opals, often without play-of-colour, is attributed to inclusions [6,7]. A grey-gold coloured crystal, identified as pyrite with marcasite by Raman spectroscopy, was also observed (Figure 6). Numerous transparent and colourless rhombohedral to pseudo-cubic crystals were also observed (Figure 7A). The same crystals were also observed in a high concentration in the contact zone between the opal and the rhyolite. This distribution suggests that they are probably protogenetic or syngenetic, but not epigenetic. Indeed, if they had been epigenetic, they should have a more homogeneous distribution in the opal and not be concentrated at the edges. Raman spectroscopy demonstrated that they are alunite KAl₃(SO₄)₂(OH)₆ (see below).

Anhydrite (CaSO₄) was also identified in the samples HU002 and HU013 (Figure 7B; see again the Raman spectra below) in areas with a slightly chalky appearance. The crystals are transparent and near colourless. A “fingerprint” appearance (Figure 7C) is also on samples HU002, HU006 and HU007, as the plane formed by silica spherules is very close to the polishing plane of the opal (i.e., the vicinal plane). The intersection of the two planes forms clearly visible “stair steps”.

Refractive indexes were from 1.42 to 1.45, while specific gravity ranged from 1.96 to 2.13 for the six samples with no matrix (see

Table 2. Refractive index and specific gravity of opals studied in this work.

Sample	Refractive Index	Specific Gravity
HU003	1.42	2.09
HU004	1.43	1.96
HU008	1.45	2.10
HU009	1.45	2.11
HU010	1.43	2.10
HU011	1.44	2.13

). This is in the expected range for opals. The UV fluorescence of the most studied samples ranged from inert to moderate blue under UV lamp excitation (Figure 8). Samples HU001, HU002, HU006, HU008, and HU013 show a moderate to weak pink to orange reaction, being slightly stronger in LWUV than in SWUV, situated in spots with a slightly different appearance and texture under daylight (Figure 9). These areas correspond to the presence of inclusions (identified as anhydrite, see Raman spectroscopy). In addition, green fluorescence is also visible in zones (especially under SWUV, see again Figure 8) on samples HU001, HU002, HU008, and HU014. Finally, nine samples (HU001, HU002, HU005, HU006, HU007, HU008, HU011, HU012, and HU013) had a faint persistent blue-green luminescence lasting more than 5 s (sometimes zoned), after excitation ceased. The three studied pink opals (HU003, HU004, and HU015) showed a green reaction. All the luminescence colours mentioned above have previously been described in other opals [3].

In Figure 10, a typical FT Raman spectrum of Idaho opals is presented from 100 to 2100 cm⁻¹, (sample HU004). All samples analysed with Raman spectroscopy presented comparable spectra. All Raman spectra show a broad main band with an apparent maximum at around 440 cm⁻¹. This consists of two large bands, the more intense at around 460 cm⁻¹ and the less intense at around 410 cm⁻¹. These bands are attributed to Si-O bending vibration, which is characteristic of opal-A. A band centred from 380 to 420 cm⁻¹ with a shoulder at around 495 cm⁻¹ is characteristic for opal-A, whereas it is centred from 325 to 340 cm⁻¹ for opal-CT with a shoulder at around 415 cm⁻¹ [3,13,14,15,16,17]. The aspect of the main band as a doublet has never been observed before. Bands were also observed at around 790 cm⁻¹ due to Si-O symmetric stretching vibration; around 970 cm⁻¹ (i.e., the stretching of silanols linked to the silica structure [18,19,20,21]); around 1064 and 1210 cm⁻¹, which correspond to asymmetric Si-O stretching [13,14,15]; and at around 1640 cm⁻¹ (symmetric H-O-H bending [13,15,16]). No band was observed above 1700 cm⁻¹. All the samples examined are opal-A, which is a rare case of opal-A being used as a gem [22,23,24,25,26,27].

The Raman spectra of inclusions observed under the microscope are shown in Figure 11 (inclusion from Figure 4C), Figure 12 (inclusion from Figure 6), Figure 13 (inclusion from Figure 7A), and Figure 14 (inclusion from Figure 7B). Two types of inclusions stand out: those rich in iron (hematite, pyrite/marcasite) and those derived from evaporites (alunite and anhydrite). The presence of alunite and anhydrite confirms the hypothesis that these opals were formed in saline environments and evaporites [10].

In Figure 15, the UV–Vis diffuse reflectance spectra of pink-coloured opals from Idaho are presented. They show a broad complex absorption (i.e., a decrease in reflectance) centred at around 500 nm, extending from 350 to about 600 nm. The sample HU003 shows weak features at 435, 493, 533 and 558 nm. The UV–Vis spectra obtained here can be compared with spectra of pink opals from Mexico, Peru, and France. While the pink opals from Peru and Mexico are coloured by fossil quinones absorbed by palygorskite [28,29], the pink opals from Quincy (France) owe their colouring to sepiolite itself coloured by quinones [30,31,32,33]. The general shape of the absorption is consistent with quinones, and not much else. Nevertheless, in most pink opals, the pink areas luminesce orange, which is not observed here. The cause of the pink colour in the Idaho opal remains unexplained. The presence of quinones in a phyllosilicate, e.g., sepiolite, remains an option.

The luminescence of opals is relatively well understood [29]. Fluorescence emission spectra show a main band (apparent maximum) at around 432 nm for the LWUV blue-emitting opals (Figure 16) accompanied by shoulders at around 460, 500, and 568 nm. Excitation spectra reveal three major absorbers responsible for these emissions at around 317, 340, and 367 nm (Figure 17), all related to intrinsic silica defects. The blue fluorescence is intrinsic to opal [34] and is attributed to defects related to oxygen [35]. The emission band around 460 nm is attributed to an oxygen deficiency center (ODC) [36] and more precisely an ODC(II) corresponding to a non-relaxed configuration of a neutral oxygen vacancy. For the HU010 sample, the emission shoulder at 460 nm is preferentially excited at 367 nm compared to the main emission (Figure 18). An emission band at 414 nm is also attributed to defects intrinsic to silica [36,37,38] corresponding to a silicon bound to two oxygen atoms instead of four. This would correspond to the peak about 340 nm in the excitation spectrum [3]. These positions do not therefore correspond to the results obtained by previous studies [34,35,39,40,41,42]. It is interesting to note that some studies obtained emission spectra with an apparent maximum around 420 nm, while others recorded emission spectra with a maximum ranging from approximately 430 to 494 nm in opals from different localities [3,39]. This could be due to the presence of multiple bands with varying relative intensities, causing the apparent maxima to occur at different positions.

The green luminescence (mainly seen under SWUV) is characterised by a range of bands at around 504, 525, and 547 nm accompanied by two shoulders at around 572 and 602 nm (Figure 19). This result corresponds to the classical emission of uranyl ions (UO₂)²⁺, which are often seen in opal [35,41]. The orange-red fluorescence under LWUV and SWUV presents an emission spectrum consisting of a broad band centred around 630 nm, i.e., in the orange (Figure 20), accompanied by numerous, there are relatively sharp peaks between 680 and 820 nm (i.e., in the red and NIR part). Orange luminescence is rare in opals, and it is generally believed to be generated by quinones in pink opals [3,42,43]. Quinones are characterised by two major bands around 580 and 620 nm. This is very different from the present spectrum. In addition, in Idaho, the orange luminescence does not correspond to the pink colour, as is the case for quinone-coloured opals. Low-quality excitation spectra were obtained with difficulty for the weak emission around 630 nm, and they principally show a broad, asymmetric band with an apparent maximum around 400 nm and a medium-broad band centred near 550 nm. Such features are new to opal luminescence and could not be further interpreted. The relatively sharp peaks between 688 and 767 nm correspond to the emission spectrum of Sm²⁺ measured in anhydrite [44] (see again Figure 20). However, they are in the deep red and the near infrared, to which the eye is very minimally sensitive. Hence, they do not participate to the opal’s orange appearance [45]. Thus, the “pink-red” appearance is perhaps due more to diffusion of the orange fluorescence into the opal mass than to the secondary red emission component.

The microstructure of play-of-colour in opal is well known but not always documented [29]. Hence, three samples (Figure 21, Figure 22, Figure 23, Figure 24, Figure 25, Figure 26 and Figure 27) were examined using SEM on fresh breaks. The internal structures of all three opals analysed were composed of nanospheres. The regularity of these structures can be measured by the interstices (of varying widths) between the nanospheres that make them up. The silica sphere regularity and diameter vary greatly from one sample to another, and even within the same sample. Opal HU009, for example, is composed of irregular, disordered spheres (Figure 21; the interstices do not reveal any regular arrangement) but also, in other places, of regular arrangements of nanospheres (Figure 22). We can even see a change in the orientation of the planes of the stack of spheres in Figure 23. This cannot be linked with sample cutting plane orientation as the sample was treated with hydrofluoric acid before acquiring SEM images. Close examination of the shape of the interstices shows that the stack of spheres is more of a compact face-centred type in the parts analysed. The diameter of the spheres varies from approximately from 285 to 580 nm.

Opal HU014 shows a very irregular structure, both in terms of sphere diameter and sphericity, at least in the areas analysed (Figure 24). A few spheres can be seen, but they are in the minority and scattered. The “spheres” are too irregular to have been measured. In opal HU015 (pink opal), an extremely regular structure was observed, made up of relatively perfect spheres. Their diameter ranges from around 660 to 880 nm, considering the “cutting plane” effect on the spheres (the measurements were taken “at the widest point”). The stacking is compact face-centred (Figure 25) and hexagonal compact (Figure 26, Figure 27, Figure 28, Figure 29 and Figure 30) depending on the zones analysed. We can also see the presence of cement (lighter in the image) binding the spheres together (Figure 26); however, this cement does not fill the interstices. Figure 27 shows a patchwork-like structure in which the orientation of the stack of spheres changes from one area to the next, with patches being less than 100 micrometres. In Figure 28, both regular and irregular structures, at bottom left of the image, were identified. Finally, the internal structure of some of the spheres is concentric, with layers of varying thickness (from 20 to about 100 nm) (Figure 29 and Figure 30).

SEM images of the three studied samples highlight several phenomena. For example, a single opal may have regions consisting of spheres with diameters that vary across different areas. Likewise, the orientation of these spherical stacks also changes from one area to another. These two factors contribute to the appearance of a patchwork of multiple diffraction colours within the same opal, even at the same location. The average diameters of the spheres measured in several areas were too large to allow diffraction, even though they were perfectly spherical and arranged in a highly regular manner [46]. Similarly, samples HU009 and HU014 even showed spheres that were irregular in diameter within the same volume, making diffraction impossible [47]. However, it is important to remember that the scales and field-of-view widths of the SEM can make it very difficult to locate diffracting zones precisely. The concentric structure observed in one of the samples has already been described in opals from Australia [1,3,47,48] and Slovakia [3]. Such a structure might be linked to the process of sphere formation by the successive deposition of layers of silica nanospheres on a central nucleus [49]. This process is probably itself the result of “slow” sphere growth [3].

4. Conclusions

Opals from Idaho with play-of-colour have been thoroughly studied, despite their relatively low production, with most opals coming from a single mine using artisanal methods. These opals typically have a white colour which can sometimes transition into pink. Inclusions of two sulphates, alunite and anhydrite, were detected alongside more typical iron-rich inclusions, particularly pyrite/marcasite and hematite. The presence of these sulphates suggests that the opals formed in saline environments and evaporites. Raman spectroscopy confirmed that all the samples analysed are opal-A, which is a rare instance of opal-A being used as a gem. SEM revealed a microstructure consisting of amorphous silica spheres of varying sizes and arrangements. The classical regular stacking of silica spheres was observed in nearly all samples, although areas exhibited disordered volumes, even in play-of-colour opals. Additionally, one sample displayed a remarkable arrangement of spheres that were too large to allow diffraction.

The pink colouration of Spencer opals was previously thought to be caused to manganese and iron, but this could not be confirmed in this study. The presence of quinones in phyllosilicates remains a possibility based on UV–visible spectroscopy, while the very weak Raman signal could be associated with sepiolite. Additionally, the lack of orange luminescence associated with the pink colour has been noted before. In the studied samples, the presence of various iron inclusions in large quantities and arranged in strata appears to cause the pink colour through light scattering. Idaho opals exhibit typical blue luminescence linked to intrinsic defects in silica, rare green emissions associated with uranyl, and a completely new type of orange luminescence linked to small areas with a distinct appearance. Furthermore, luminescence in the near-infrared was observed due to Sm²⁺ impurities in anhydrite inclusions.