*4.1. The Coupled Substitution B(F, OH)Si*−*1, O*−*<sup>1</sup>*

The incorporation of boron in olivine is mainly related to the coupled substitution of boron and H for silicon [6,21–24]. The BO3 group lies on the (O3−O1−O3) face of the tetrahedral site, inclined at 17◦ from the (x,y) plane, and the H atom is bonded to the O2 atom [24]. The O2−H group points either to the O1 oxygen or out of the tetrahedral site. The former configuration is more stable than the latter [24]. The introduction of B promotes the formation of defects in the crystal structure of forsterite and may result in a distortion of the tetrahedral site [22,34], which is interpreted as the formation of terminal B−(OH) or B−F bonds or the perturbation of the B−O−M linkage. Guo et al. [35] pointed out that the substitution of B for Si may result in the reduction in cell parameters and cell volume of forsterite.

The results of single crystal X-ray diffraction analysis indicate that the unit-cell parameters (a, b, and c) and unit-cell volume of forsterite in Jian forsterite jade are much smaller than those of known olivine. The small crystal parameters are rather due to an increased content of Mg. It's well known that lattice parameters linearly decrease with increasing forsterite [3]. However, we compared the lattice parameters of B-rich forsterite (Fo = 99.7) with pure forsterite (Fo = 100) (Figure 4) [1,13]. The lattice parameters of the former are smaller than those of the latter. Similar phenomena were described in B-rich diopside [34]. Halenius et al. [34] presented that a B−Si substitution in B-rich diopside decreases the T site volume and leads to disruption of one of the T−O bonds compared with end-member diopside. In our study, VT (Å3) in B-rich forsterite (2.196 Å3) is smaller than that of pure forsterite (2.21 Å3). Thus, we speculated that the incorporation of B in olivine slightly reduces the lattice parameters, especially decreases the T site volume and affects the neighboring M1 and M2 sites. Due to the low concentration of B, the X-ray diffraction techniques can't resolve the true coordination of B in forsterite. Nevertheless, the occurrence of the BO3 group is further demonstrated by our spectra data.

B incorporation into forsterite has no effect on the Raman spectrum. In our study, the difference between B-rich and B-free forsterite Raman spectrum is caused by the different magnesian contents. No evident bands of B−O vibration appear in the Raman spectrum. While significant differences are observed between the B-rich and B-free forsterite IR spectrum. B-rich forsterite displays extra five bands at 761, 1168, 1259, 1303, and 3593 cm−<sup>1</sup> compared with B-free forsterite. The vibrational bands at 1200–1400 cm−<sup>1</sup> are related to B–O vibrations [22,24,34]. Halenius et al. [34] reported the existence of the replacement

of SiO4 by BO3 in synthetic minerals (i.e., diopside and forsterite). The bands at 761, 1168, 1259, and 1303 cm−<sup>1</sup> are consistent with the bands of vibration modes of the BO3 group [22,24,34]. In addition, strong OH band at 3696 cm−<sup>1</sup> and weaker band at 3593 cm−<sup>1</sup> are displayed in B-rich forsterite. OH groups commonly exist in nominally anhydrous minerals through a variety of pathways (commonly known as "water") [23,36,37]. Due to the variety and complexity of the infrared spectra of olivine associated with O-H stretching modes, uncertainties remain in the determination of the OH bands associated with the B(OH)Si–1O–1 substitution in B-rich olivine. Ingrin et al. [24] presented that the OH bands at 3704 cm−<sup>1</sup> (//z), 3598 cm−<sup>1</sup> (//x,y), and 3525 cm−<sup>1</sup> (//x) are associated with the B(OH)Si–1O–1 substitution in synthetic forsterite and natural olivine. Matsyuk and Langer [38] noted that two bands at 3672 and 3535 cm−<sup>1</sup> are assigned to boron-related defects. Gose et al. [37] speculated that the OH defect at 3597 cm−<sup>1</sup> may be linked to the coupled B(OH)Si−1O−<sup>1</sup> substitution. Thus, the OH band at 3593 cm−<sup>1</sup> is most likely associated with the B(F, OH)Si−1O−<sup>1</sup> substitution in our B-rich forsterite. The OH band at 3696 cm−<sup>1</sup> is assigned to OH stretching vibrations of OH defects. Therefore, the IR spectra of B-rich forsterite provided evidence that the incorporation of B into forsterite is mainly caused by the coupled substitution of boron and H for silicon. Theoretically, the ratio of the contents of B to H is close to the 1:1 trend expected from the B(F, OH)Si–1O–1 substitution. However, the contents of B (1773.4–1795.91 ppm) in our sample are higher than that of water (1450(300)) wt ppm estimated by Wang et al. [25]. The deviation from the 1:1 trend can be explained by additional fluorine (usually 0.13–0.18 wt% in our sample) [6].

#### *4.2. The Rarity of Mg- and B-Rich Forsterite*

The composition of olivine in the mantle and magmatic rocks is typically Fo85~96 [5,39]. Olivine with nearly forsterite end composition is rare and can only be formed in special geological environments. For example, rare forsterite (Fo = 97–99) inclusion in magnesiochromite was reported by Xiong et al. [40], and Majumdar et al. [41] found a pseudomorph rim of Mg-rich olivine (Fo = 98) in serpentinized dunite. Blondes et al. [42] found several olivine grains with Fo at up to 99.8 in multiple primitive basaltic lava flows from Big Pine Volcanic Field (Inyo, CA, USA). Nekrylov et al. [43] reported high-Mg olivine (Fo value up to 99.8) from magnesian skarns and silicate marbles from different locations. In addition, boron-rich olivines are less well-known. Most natural olivine samples have very low concentrations of boron [19,20]. Sykes et al. [22] noted that olivine from the Tayozhnoye iron deposit, Siberia, Russia, contains substantial B2O3 (1.11–1.35 wt%). Majumdar et al. [41] reported a boron content up to 10.4 ppm in high-magnesium olivine. Nekrylov et al. [43] presented that B concentrations of olivine from magnesian skarns and silicate marbles vary from 23 to 856 ppm. However, the natural forsterite tested in this study surpasses all olivines known in geological objects in magnesium and boron content (Figure 2). The formation of B-rich forsterite remains disputed.

Olivine with high boron content is generally considered to be associated with the metasomatism of boron-rich fluids [22,43]. The B-rich forsterite may be derived from mantle peridotite metasomatized by B-rich fluids. However, the composition of olivine in mantle peridotite is typical Fo86–92 [5]. Moreover, mantle olivine has high Ni, Mn, and Co concentrations (Ni, 2040–3310 ppm; Mn, 460–850 ppm; Co, 87–137 ppm) [26]. However, the trace element compositions of B-rich forsterite are evidently inconsistent with the mantle olivine. The above speculation about the origin seems to be incorrect. In addition, Nekrylov et al. [43] presented that olivine from magnesian skarns and silicate marbles are enriched in B and depleted in Ni, Co, and Cr concentrations, which are consistent with our trace element characteristics. According to the description of the geology setting in Wang et al. [25], the formation in which B-rich forsterite forms contains various altered felsic and granitic rocks, dolomitic marble, and serpentinized olivine marble. Moreover, previous study confirms that the formation is a boron-bearing sequence [25]. Thus, we suggest that B-rich forsterite is derived from the metamorphism of magnesian marbles. The origin and formation process of B-rich forsterite needs further study.

#### **5. Conclusions**


**Supplementary Materials:** The following supporting information can be downloaded at: https://www. mdpi.com/article/10.3390/cryst12070975/s1, crystallographic information files (CIF).

**Author Contributions:** Conceptualization, M.H.; methodology, M.H. and B.P.; software, B.P. and S.W.; formal analysis, B.P., S.W. and J.F.; investigation, B.P. and S.W.; data curation, B.P.; writing—original draft preparation, B.P., S.W. and J.F.; writing—review and editing, M.H. and M.Y.; supervision, M.H. and M.Y.; All authors have read and agreed to the published version of the manuscript.

**Funding:** National Mineral Rock and Fossil Specimens Resource Center.

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** Not applicable.

**Acknowledgments:** We thank Yingliang Jiang and Haiyang Yu from Ji'an Xinli Mining Co, LTD (Jilin, China) for providing samples for this study. We thank Xi Liu and Qiangwei Su for fruitful discussions and suggestions.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**

