The petrological characteristics of the leucogranite samples were investigated using both transmitted light microscopy (Fig. 2) and X-ray fluorescence (XRF) spectroscopic imaging (Fig. 3). It should be noted that the XRF spectroscopic imaging and microscopic observations were performed on sections of different thicknesses from the same samples. Consequently, certain low-abundance minerals that could not be fully resolved or mapped in the XRF imagery were clearly observable under the microscope, resulting in slight discrepancies in the accessory mineral assemblages between the two methods.

Results show that sample XZ-BG is mainly composed of albite (35.61 wt %), quartz (31.75 wt %), and K-feldspar (26.86 wt %), with minor garnet (3.29 wt %), muscovite (3.28 wt %), and apatite (0.02 wt %). Under transmitted light, K-feldspar commonly exhibits a turbid appearance, and quartz inclusions can occasionally be observed within the grains (Fig. 2a). Twinning is commonly developed in albite (Fig. 2b). Quartz mainly occurs as anhedral granular crystals, characterized by abundant internal fractures, undulose extinction, and local dissolution textures. Muscovite typically occurs as flakes distributed along grain boundaries or filling fractures and displays golden-yellow, rose-red, or blue-violet interference colors under crossed polars. In addition, a small amount of amphibole is present (Fig. 2c) as anhedral grains showing golden-yellow pleochroism under plane-polarized light.

Sample XZ-WJ is dominated by quartz (33.64 wt %) and albite (32.18 wt %), followed by K-feldspar (17.99 wt %) and muscovite (13.42 wt %), with minor chlorite (2.69 wt %) and apatite (0.08 wt %). Quartz grains are generally coarser and better preserved, mostly ranging from 50 to 600 µm, with some grains reaching up to 7 mm, and locally showing dissolution by feldspars. Albite and K-feldspar are mostly subhedral and generally coarser than quartz (100–1000 µm) and may locally enclose quartz. Muscovite is relatively abundant and commonly occurs as elongate flakes. Garnet is scarce in this sample (0.08 wt %), and fractures are commonly filled with biotite (Fig. 2e).

Overall, the two samples show distinct petrographic differences. Compared with XZ-WJ, quartz in XZ-BG is generally finer-grained and more strongly affected by internal fracturing and undulose extinction, whereas XZ-WJ is characterized by coarser and better-preserved quartz, together with a markedly higher abundance of muscovite.

The major element compositions of the two samples are summarized in Table 1. XZ-BG contains 74.68 % SiO₂ and is enriched in total alkalis (ALK = 8.65 %), with K2O/Na2O = 1.57 and Al₂O₃ = 14.78 %. MgO is very low (0.099 %), whereas CaO and Fe₂O₃ are 0.25 % and 0.55 %, respectively. XZ-WJ contains 73.73 % SiO₂ and also has high total alkalis (ALK = 8.85 %) but with Na₂O exceeding K₂O (K2O/Na2O = 0.72) and slightly higher Al₂O₃ (15.16 %). Its MgO, Fe₂O₃, and CaO contents are 0.055 %, 0.11 %, and 0.28 %, respectively. Both samples have high A / CNK values (1.262 and 1.167), indicating peraluminous compositions, and plot in the high-K calc-alkaline S-type granite field (Fig. S2 in the Supplement).

In situ LA-ICP-MS spot analyses were performed on quartz grains from samples XZ-WJ and XZ-BG. The analytical results are reported in Table 2, and element–element plots based on individual spot data are provided in the Supplement (Fig. S3).

Quartz samples from XZ-BG contain 4.56–12.70 ppm Li, 3.06–36.91 ppm Na, and 40.92–72.59 ppm Al, with average concentrations of 7.82, 15.30, and 62.69 ppm, respectively. Ti concentrations range from 4.64 to 9.16 ppm, averaging 6.45 ppm, while Ge contents vary between 0.25 and 1.58 ppm, with a mean of 1.06 ppm. Additionally, K and Ca concentrations range from 0.65 to 20.28 ppm and from 10.26 to 128.10 ppm, respectively, with average values of 8.65 and 89.50 ppm. Quartz samples from XZ-WJ contain 5.12–13.58 ppm Li, 2.48–19.10 ppm Na, and 64.65–117.70 ppm Al, with average concentrations of 9.60, 7.78, and 89.95 ppm, respectively. Ti concentrations range from 2.52 to 5.16 ppm, averaging 3.87 ppm, while Ge contents vary between 0.48 and 1.91 ppm, with a mean of 1.19 ppm. K and Ca concentrations range from 5.31 to 19.63 and 10.90 to 152.60 ppm, respectively, with average values of 12.01 and 65.16 ppm.

Overall, compared with quartz from XZ-WJ, quartz from XZ-BG is characterized by lower Al, Li, and K concentrations but higher Ti and Na concentrations. Element–element plots (Fig. S3) illustrate the variability of trace element concentrations in quartz. In particular, Li, Na, and Ge display substantial scatter at a given Al concentration, and no well-defined linear relationships are apparent for these element pairs.

In the quartz lattice, Al is commonly incorporated through isomorphic substitution and generally exhibits the highest concentration among trace elements, providing significant environmental and petrogenetic information (Müller et al., 2012). Therefore, this study focuses on a graphical analysis of Al and its correlations with other trace elements in quartz (Fig. S3 in the Supplement).

During quartz crystallization, substitution of Si4+ by Al3+ requires charge compensation by coordinating ions (e.g., Li⁺, Na⁺, K⁺, and H⁺), which can lead to broadly coupled variations in these elements (Zhang et al., 2024b). In many quartz populations, Li⁺ may be an important compensating ion for Al3+ (Jourdan et al., 2009); however, in the present samples, the Al–Li/Na/K relationships are dispersed, suggesting that charge compensation is variable and/or that minor non-lattice contributions may affect some analyses.

Trace element variations in quartz are influenced by lattice incorporation and inclusions. Although laser in situ measurements avoided visible inclusions, measured impurities may not exclusively represent lattice elements. Typically, substitutional ions (Al3+, Fe3+, B3+) and interstitial ions (Li⁺, Na⁺, K⁺, P⁵⁺) occur in an approximately 1:1 ratio (Yang et al., 2022). In XZ-BG and XZ-WJ, the ratio (Al3+ + Fe3+ + B3+) / (Li⁺ + Na⁺ + K⁺ + P⁵⁺) ranges from 0.72 to 1.14 and from 1.52 to 2.34, with average values of 0.95 and 1.88, respectively. The deviation from unity, especially in XZ-WJ, indicates that the measured trivalent-cation budget is not fully balanced by the measured monovalent budget in many spots, which may reflect variable charge compensation behavior and/or contributions from submicroscopic inclusions (Xie et al., 2022).

Al, Ge, and Ti in quartz have been linked to magmatic differentiation in previous studies, where Al and Ge tend to increase whereas Ti decreases with progressive differentiation (Schrön et al., 1988; Zhang et al., 2024a). In the Dinggye leucogranites, Al-Ti and Ge-Ti show only weak negative tendencies, and Al-Ge shows a weak positive tendency (Fig. S3), consistent with published datasets (Rusk et al., 2008). In this study, quartz from XZ-WJ has higher mean Al than that from XZ-BG; given the scatter in spot data, this difference is interpreted as reflecting variations in crystallization domains.

Müller et al. (2009) and Wark and Watson (2006) showed that Ti content in quartz serves as a critical geochemical indicator for recording magmatic crystallization temperatures, exhibiting a systematic correlation with the degree of magmatic evolution (fractionation extent). Specifically, as magmatic differentiation progresses, the Ti content incorporated in late-crystallized quartz displays a characteristic declining trend (Larsen et al., 2004; Götze et al., 2004). This phenomenon provides a robust theoretical foundation for reconstructing the crystallization temperatures of granitic magmatic systems using Ti-in-quartz content. In this study, we applied the Ti-in-quartz thermometry formula (Sect. S1.7 in the Supplement) proposed by Hayden and Watson (2007), which is specifically designed for rutile-absent systems, to accurately calculate the crystallization temperatures of two samples of leucogranite from the Dinggye area. Table 2 results show that the crystallization temperature of leucogranite XZ-BG ranges from 526 to 569 °C, with an average temperature of approximately 542 °C; the crystallization temperature of leucogranite XZ-WJ ranges from 451 to 525 °C, with an average temperature of approximately 504 °C. The crystallization temperature of XZ-BG is higher than that of XZ-WJ. Based on previous theories, this temperature difference may suggest that XZ-WJ represents a relatively more evolved melt fraction formed during late-stage differentiation within the same magmatic system. Magma with a higher degree of differentiation generally has a longer evolution time and crystallizes later, and its final crystallization temperature is also lower. The crystallization temperatures of these two sets of leucogranite (451–569 °C) calculated in this study are significantly lower than the commonly accepted solidus temperature range for granite systems, which is generally considered to be between 650 and 700 °C (Steiner et al., 1975). It strongly proves that the leucogranite magma in the Himalayan crystallization region has the potential to evolve to temperatures below the solidus temperature of traditional granite magma systems. This means that magmas in this region can sustain its residual melt (or fluid) and ultimately crystallize into leucogranite at very low temperatures (below 650 °C), even below the solidus temperature for typical granite magmas. This low-temperature crystallization phenomenon may be closely related to the geological conditions of igneous crystallization in the Himalayan region. The source area is rich in water and volatiles (such as F and B), and these components are effective fluxes that can significantly lower the solidus temperature of magma (Tuttle and Bowen, 1958; Holtz et al., 1995; London, 1987; Gao et al., 2017). A shallow emplacement under low-pressure conditions can favor the stabilization of hydrous, low-melting-point minerals and thus promote crystallization at relatively low temperatures (Luth et al., 1964; Scaillet et al., 1995). In addition, slow cooling of the shallowly emplaced leucogranitic magma allows a volatile-rich residual melt to be maintained over a prolonged temperature interval. This extended residence time at low temperatures facilitates continued differentiation and promotes more complete crystallization of late-stage melts (Huang and Audétat, 2012). Therefore, the observed unusually low crystallization temperatures provide important constraints on melt evolution and emplacement processes in collisional orogenic belts, helping to clarify how leucogranitic magmas differentiate and solidify during crustal thickening and exhumation.

The occurrence and concentration of impurities in high-purity-quartz (HPQ) raw ore are key factors determining the quality of the final product. Impurities are generally classified into three types: symbiotic gangue minerals, inclusions, and lattice impurities (Yang et al., 2022). Symbiotic gangue minerals coexist with quartz during mineralization, varying according to geological conditions. Common associated minerals include feldspar, mica, rutile, hematite, magnetite, tourmaline, fluorite, epidote, and calcite (Courba et al., 2025). Inclusions, captured during quartz crystallization, can be solid, liquid, or gaseous. They are classified as primary (formed during crystal growth on faces, most difficult to remove), pseudosecondary (formed during stress-induced cracks and subsequently trapped), and secondary (formed post-crystallization in intergranular cracks, easier to remove) (Shi et al., 2023). Inclusions can also be classified as mineral, melt, or fluid inclusions, with fluid inclusions – comprising H₂O, CO₂, CH₄, N₂, and hydrocarbons – being most abundant. Inclusions affect purification by introducing impurities and producing etch pits that generate bubbles in quartz glass. Quartz lattice impurities are formed when other metal or non-metal ions substitute silicon atoms in the quartz lattice, and they are closely related to the point defects in the quartz crystal (Stalder, 2021). The point defects in quartz crystals include vacancies, substitutional atoms, and interstitial atoms, with the latter two being the primary modes of lattice impurity element introduction. Foreign impurity ions such as Ti4+, Ge⁴⁺, Al3+, B3+, Fe3+, and P⁵⁺ substitute for Si4+, forming substitutional impurity elements. At the same time, when certain non-tetravalent ions replace Si4+, compensation ions like K⁺, Na⁺, and Li⁺ are introduced into the atomic gaps to maintain electrical charge balance, forming interstitial impurity elements (Larsen et al., 2004). These lattice impurities are difficult to remove from quartz using existing purification processes.

Considering the fact that leucogranite forms in high-temperature environments and the incorporation mechanisms of impurity elements in quartz, elements such as Ti and Al can easily enter the quartz lattice structure in large quantities. This makes leucogranite difficult to use as a raw material for purifying high-purity quartz. However, Himalayan leucogranite has a relatively low crystallization temperature, typically below 800 °C (Cisneros et al., 2020). Studies show that the zircon saturation temperature range for most Himalayan leucogranite is between 650 and 750 °C (Xie, 2023; Gao et al., 2024; Aikman et al., 2012). In this study, using the quartz-Ti thermometer, the formation temperature of leucogranite samples was calculated to be between 451 and 569 °C, indicating that the formation conditions are similar to those of pegmatitic crystallization temperatures and pressures, providing a source of high-purity-quartz raw material.

Table 3 shows that quartz grains from the XZ-BG and XZ-WJ leucogranite samples after the final purification process contain 99.9969 wt % and 99.9959 wt % SiO₂, respectively (Table 3). Al and Ti contents are typically used as indicators of quartz quality because these elements are among the most common and hardest-to-remove trace elements. Quartz grain samples were plotted in the Al–Ti high-purity-quartz classification diagram (Müller et al., 2007), and both samples fall within the high-purity-quartz field (Fig. 6). However, compared with the benchmark high-purity-quartz sand produced by Unimin (IOTA^® standard), regarded as a benchmark material in HPQ research and industry because of its exceptionally high purity and extremely low trace element contents, the residual Li and Ti contents of purified quartz grains are still relatively high (Fig. 5). This indicates that, although the quartz grains from Dinggye leucogranites could meet the HPQ field in terms of Al–Ti contents, a further reduction in Li and other trace elements such as Ti would be required to approach the IOTA^® benchmark level.