1. Introduction
Carbon is the fourth most abundant element in the solar system, after hydrogen, helium, and oxygen [
1,
2]. It is a crucial element for life on Earth, and its chemical compounds are the main energy sources for human beings. Earth’s carbon is mostly stored in mantle and sedimentary reservoirs, with a present-day mass of about 3 × 10
19–2.4 × 10
20 kg of carbon stored in the mantle and a mass of about 5.4 × 10
19–8.7 × 10
19 kg surficial carbon [
3,
4]. Research on Earth’s carbon cycle processes mainly focuses on cycles of two different timescales: ① the short-term carbon cycle, which involves years to tens of thousands of years of carbon transfer among surficial reservoirs including the atmosphere, oceans, soils, and biosphere; and ② the long-term carbon cycle, which involves many millions to billions of years of carbon transfer between rocks and surficial reservoirs [
4–
7]. Both the short-term and long-term carbon cycles play important roles in global climate change, which is a life-and-death factor for life on Earth. The short-term carbon cycle has been studied relatively extensively, whereas the long-term carbon cycle—and especially the deep carbon cycle—is still a very open question due to the rarity of research objects from the Earth’s deep mantle. Mantle peridotites are commonly found in various orogenic belts in the world and preserve important information about the mantle carbon reservoir
[8]. However, due to the extremely low solubility of carbon in mantle minerals such as olivine, diopside, and pyrope, the carbon budget of the mantle is dominated by carbonates and diamond [
9,
10], making them important candidates for investigating the carbon cycle in the Earth’s mantle [
11–
13].
Diamond is well known to have been found in kimberlites, lamproites, and related placer deposits, which are sources of gem-quality diamonds
[14]. Diamonds also occur in lamprophyres
[15], ultrahigh-pressure (UHP) metamorphic rocks [
16–
18], and meteorites [
19–
21]; however, due to their small size and poor abundance, these diamonds have little commercial value but are of great scientific interest. Since the first microdiamond was recovered from the Luobusa ophiolite in the Tibet Autonomous Region (Tibet)
[22], diamonds have been successively discovered in a series of ophiolites of different orogenic belts [
23–
26]. With increasing evidence for the existence of diamond in ophiolitic peridotites and chromitites [
27,
28], a new occurrence of diamond was proposed: namely, ophiolite-hosted diamond (also known as ophiolitic diamond)
[29].
Subducted slabs may reach the mantle transition zone (MTZ) at 410–670 km
[30], and even the core-mantle boundary (CMB) at a depth of 2900 km [
31–
33]. Most hydrous minerals in subducted slabs become unstable with increasing depth and are replaced by high-pressure minerals, while the water in their crystalline structure is released into surrounding fluids or melts. This process has been widely recognized based on evidence from surface geology, geophysical data, and high-temperature/high-pressure experiments
[34]. Current knowledge of the fate of these deeply subducted materials within and near the MTZ or the CMB is still in its infancy. Mineral inclusions in diamonds provide a glimpse of mantle processes and the fate of recycled crustal material at various mantle depths [
35,
36]. At present, most knowledge related to the recycling of subducted material comes from diamonds and their mineral inclusions in kimberlites and related rocks from old cratons [
1,
13,
37–
40]. However, diamonds and recycled crustal minerals newly discovered in peridotites and chromitites from different ophiolites (Fig. 1) [
22,
24,
26,
28,
41–
45] have provided a new window for probing carbon and other crustal material cycling in the deep mantle.
Although we provide a brief description of different occurrences of diamonds in this paper, our emphasis is on the occurrence of natural and in situ diamonds in ophiolitic peridotites and chromitites, the significance of these occurrences for the petrogenesis of the oceanic upper mantle, and the processes by which diamonds and high-pressure/high-temperature equivalents of subducted crustal material are integrated into chromitites–peridotites at MTZ depths. We summarize the geochemical characteristics and possible carbon sources of these unusual diamond occurrences in ophiolites and present a revised geodynamic model for their generation and transportation to shallow mantle depths in the oceanic lithosphere. Our results and interpretations have strong implications for the origin of ophiolite and related chromitites, and for revealing the carbon cycle on Earth.
2. Occurrences of diamond on Earth
Diamonds of different occurrences generally vary in terms of color, crystal size, crystal form, isotopic composition, and so forth.
2.1. Diamonds in kimberlite and related igneous rocks
Diamonds form at depths of over 150 km in the Earth’s mantle
[1], and are carried to the Earth’s surface by three main types of rocks: kimberlite, lamproite, and lamprophyre [
14,
37,
46]. Kimberlite and lamproite are the main sources of gem-quality diamonds in the world, and are distributed in ancient cratons in Australia, South Africa, Siberia, the North Atlantic, and North China [
47–
51]. The lamprophyre rocks in the Wawa and Abitibi subprovinces of Canada host the world’s oldest diamond deposits (∼2.67 Ga)
[52].
Diamonds in kimberlite and related rocks mainly occur in three forms—namely, monocrystalline, polycrystalline, and fibrous/coated forms
[14]. Monocrystalline diamonds generally show an octahedral, cubic, or cubo-octahedral crystal shape, and are often cut and polished to make gems (Fig. 2(a))
[46]. Polycrystalline diamonds can be classified into different subtypes including framesite, bort, ballas, and carbonado
[53]. Due to the high densities of the grain boundaries, polycrystalline diamonds are preferred over monocrystalline diamonds as super-hard abrasives [
54,
55]. Fibrous/coated diamonds are mixed polycrystalline-monocrystalline diamonds with a monocrystalline core mantled by a polycrystalline layer enclosing abundant fluid inclusions
[14].
Only about 1% of these mantle-derived diamonds contain mineral inclusions such as olivine, pyroxene, chromite, garnet, sulfides, metal alloys, and fluid inclusions
[56]. These mineral inclusions preserve information about the pressure–temperature (
P–
T) conditions of diamond formation, the formation ages, and the host rock types of diamonds [
56–
58].
P–
T calculations of silicate mineral inclusions indicate that diamonds mainly form at a depth of 140–200 km in the lithospheric mantle of cratons [
58,
59]. However, an increasing number of observations show that diamonds may also form in the asthenosphere, MTZ, lower mantle, or even CMB [
60–
63]. These sublithospheric diamonds make up only about 1% of the world’s diamond population, and contain mineral inclusions such as majorite, perovskite, stishovite, and periclase
[57].
Based on mineralogy and chemical compositions, mineral inclusions in mantle-derived diamonds can be divided into three suites: ① the peridotitic/ultramafic suite (65%), ② the eclogitic suite (33%), and ③ the websteritic suite (2%)
[56]. The peridotitic suite can be further subdivided into the lherzolitic (13%), harzburgitic (86%), and wehrlitic (1%) suites. Different suites of mineral inclusions reflect different rock types of mantle source rock during diamond formation. The chemical compositions of silicate minerals, such as Cr
2O
3 in garnet and Cr# (= Cr/(Cr + Al) × 100) of clinopyroxene, can be used to distinguish the different mineral suites in diamonds
[56]. Diamonds with different mineral inclusion suites also vary in chemical composition.
According to the review by Cartigny
[37], worldwide diamonds have a carbon isotopic composition with δ
13C (= [(
13C/
12C
sample)/(
13C/
12C
RM − 1)] × 1000) values ranging between −38.5‰ and +5.0‰. Peridotitic and eclogitic diamonds both have a peak δ
13C value of about −5.0‰, which is consistent with that of the normal mantle carbon reservoir [
14,
37]. However, eclogitic diamonds have a larger δ
13C range (from −38.5‰ to +2.7‰) than that of peridotitic diamonds (from −26.4‰ to +0.2‰). About 34% of eclogitic diamonds have quite light carbon isotopic composition with δ
13C values < −10.0‰, while this percentage is only about 2% for peridotitic diamonds. The difference in the carbon isotopic composition of peridotitic and eclogitic diamonds indicates a varied carbon source.
2.2. Diamonds in ultrahigh-pressure metamorphic rocks
Microdiamonds also occur in crustal rocks that have experienced UHP metamorphism (> ∼4 GPa) due to deep subduction (> ∼140 km) [
17,
64]. It is amazing that these UHP rocks can been brought back to the surface again after subducting to such extreme depth. Metamorphic diamonds were first discovered by Rozen et al.
[18] in eclogites of the Precambrian Kokchetav Massif (Fig. 2(b)); however, their metamorphic origin was not accepted until the discovery of
in situ diamonds as inclusions in zircons
[65]. Since then, metamorphic diamonds were consecutively recovered in the Dabie Mountains and North Qaidam, China [
16,
66]; the Western Gneiss Region, Norway [
67,
68]; the Erzgebirge Massif, Germany [
69,
70]; the Rhodope Massif, Greece
[71]; the Bohemian Massif, Czech Republic [
72,
73]; and the Pohorje region, Austria
[74].
Metamorphic diamonds mainly occur as mineral inclusions in garnet, clinopyroxene, zircon, kyanite, zoisite, dolomite, and Mg-calcite from UHP metamorphic rocks [
16,
65,
66,
75]. These diamonds may have a cuboid, cube-octahedral, or octahedral shape. Unlike mantle-derived diamonds, metamorphic diamonds are usually quite small, at about 1–60 μm in diameter
[17]. In rare cases, these diamonds can reach several hundred micrometers in diameter [
65,
66].
Diamonds formed in metamorphic rocks have δ
13C values ranging from −30‰ to −3‰
[37]. Most metamorphic diamonds have a much lighter carbon isotopic composition than those of the mantle carbon reservoir (δ
13C = −5‰ ± 3‰) [
14,
65]. According to the statistical work by Cartigny
[37], metamorphic diamonds have exclusively positive δ
15N values. Carbon and nitrogen isotopic compositions suggest that the material source for metamorphic diamonds is subducted metasediments
[76]. The noble gas contents of metamorphic diamonds in the Kokchetav Massif have been analyzed by Verchovsky et al.
[77]. The results showed that the saturation concentration of
4He (5.6 × 10
−4 cm
3·g
−1, STP) is among the very highest observed in any terrestrial diamonds
[77]. Sumino et al.
[78] reanalyzed the noble gas composition of Kokchetav diamonds by crushing the diamonds
in vacuo. The results showed that the inclusion-hosted
3He/
4He ratio is about (3.3–6.5) × 10
−5, which is higher than that of the mid-ocean ridge basalt (MORB)-source mantle, but consistent with the maximum value of ocean island basalts (OIBs)
[78]. Thus, it has been suggested that a plume-like component was involved in the formation of the Kokchetav diamonds
[78].
2.3. Diamonds in meteorites and related impact craters
Diamonds related to meteorites can be classified into two groups: ① meteoritic diamonds, which occur as presolar grains in meteorites (Fig. 2(c)) [
19,
20,
79–
82]; and ② impact diamonds, which form under the high-pressure and high-temperature conditions produced when meteorites impact the Earth’s surface [
83–
85].
Meteoritic diamonds are usually only a few nanometers in diameter and are thus also known as “nanodiamonds”
[20]. The anomalous noble gas isotopic compositions in meteorites led to the discovery of nanodiamonds [
79,
86]. Diamonds in meteorites were first recognized by Lewis et al.
[19] when checking the residue carbonaceous material from the Allende, Murchison, and Indarch meteorites after acid treatment
[19]. Meteoritic diamonds mainly have a cubic crystal structure, although rare cases show a hexagonal symmetry. Russell et al.
[87] analyzed the isotopic composition of diamonds from primitive chondrites. These diamonds have δ
13C values ranging from −38‰ to −32‰, a δ
15N of −348‰ ± 7‰, and nitrogen contents of 7–13000 ppm, which are obviously distinct from those of terrestrial diamonds [
37,
87]. By comparison, transmission electron microscope (TEM) studies of meteoritic diamonds with synthesized nanodiamonds produced by high-pressure shock waves and vapor-deposition processes suggest that these presolar diamonds in meteorites have a circumstellar condensation origin and form through low-pressure deposition processes [
20,
21].
Impact diamonds have been discovered in impact melt rocks and breccias in several terrestrial impact craters in Russia [
84,
88], Ukraine
[89], and Germany [
83,
90]. Impact diamonds have a relatively large size, ranging from tens of nanometers to several hundred micrometers, or even millimeters, in diameter [
83,
84]. Diamonds from the Popigai impact crater in Russia have been analyzed for carbon and nitrogen composition. These diamonds have δ
13C values of −20‰ to −8‰, δ
15N values (= [(
15N/
14N
sample)/(
15N/
14N
RM) − 1] × 1000) of −6‰ to −2‰, and nitrogen contents of 5–50 ppm, all of which differ from those of presolar diamonds [
84,
87]. Similarities in mineralogical, isotopic, and crystallographic characteristics between impact diamonds and graphite in the Popigai impact crater indicate that these diamonds have transformed from associated graphite in the solid state
[84]. A study on diamond and graphite from the Ries impact crater demonstrated a similar formation mechanism for impact diamonds
[91].
2.4. Diamonds in ophiolitic peridotites and chromitites
Diamonds were first recovered from podiform chromitites of the Luobusa and Dongqiao ophiolites in Tibet, China, in 1981. Since then, diamonds have been successively discovered in ophiolitic peridotites and chromitite of different orogenic belts (Fig. 1). These diamond-bearing ophiolites mainly include: ① the Luobusa, Xigaze, Dongbo, Burang, Dangqiong, and Dongqiao massifs in Tibet, China [
22,
28,
41,
92,
93]; ② the Sartohay ophiolite in the Xinjiang Uygur Autonomous Region, China
[94]; ③ the Hegenshan ophiolite in the Inner Mongolia Autonomous Region, China
[95]; ④ the Ray-Iz ophiolite in Polar Urals, Russia
[44]; ⑤ the Pozanti–Karsanti (also called Aladag) ophiolite in Turkey
[24]; ⑥ the Mirdita ophiolite in Albania [
26,
96]; and ⑦ the Myitkyina ophiolite in Myanmar
[25]. The abundance of diamonds varies considerably between different ophiolites. Xu et al.
[92] discovered more than 1000 grains of diamonds in 1100 kg of chromitites from the Luobusa ophiolite (Fig. 2(d)), which is about 1 grain per kilogram. In the Hegenshan ophiolites, about 130 diamond grains were recovered from 2000 kg of chromitites, which is about 0.06 grain per kilogram
[95]. Lian et al.
[24] recovered more than 100 grains of diamond from 500 kg of chromitite from the Pozanti–Karsanti ophiolite, which is about 0.2 grain per kilogram. However, ophiolites such as the Mirdita
[97], Sartohay
[94], and Myitkyina
[25] ophiolites have quite low abundances of diamonds. Chen et al.
[25] only recovered 6 grains of diamond from 1540 kg of peridotites in the Myitkyina ophiolite.
Ophiolites are relics of the ancient oceanic lithosphere that were tectonically emplaced into continental margins
[45]. An intact ophiolitic sequence includes the following units, from bottom to top: mantle peridotites, ultramafic and mafic cumulate rocks, sheeted dykes, pillow lavas, and sedimentary cover. Chromitites may occur in both mantle and crustal horizon in the ophiolitic sequence [
24,
98]. It is well accepted that peridotites, chromitites, and the oceanic crustal derivatives in ophiolitic sequence form at shallow depths (60–80 km) beneath a spreading center in a mid-ocean ridge setting or above a suprasubduction zone [
99–
101]. However, diamonds form at depths greater than 150 km or at pressures greater than 5 GPa
[37]; therefore, the formation conditions for diamond and ophiolite are incompatible with each other. The discovery of diamonds in ophiolitic chromitites and peridotites have received much skepticism. In the early stages of this research, when diamonds were being recovered from mineral concentrates rather than as
in situ minerals, scholars questioned whether these diamonds to be contaminants during mineral separation processes. Therefore, conducting a comparison study between ophiolitic diamonds and synthetic diamonds and looking for
in situ diamond grains in ophiolites are of great importance to prove the natural origin of these diamonds.
Howell et al.
[28] reported an independent discovery of diamonds from several hand-specimens sampled from the Luobusa ophiolite. The diversity in diamond sizes ruled out the possibility of contamination from drill bits or saws
[28]. Systematic comparison studies of isotopic and trace element compositions between the Luobusa diamonds and synthetic diamonds revealed a natural rather than anthropogenic contamination origin for these diamonds. Yang et al.
[41] reported the first discovery of an
in situ diamond as an inclusion enclosed by an OsIr alloy. However, this diamond grain is only 1–2 μm in size, which is much smaller than the diamonds from mineral concentrates
[29]. Due to the rarity of diamonds in peridotite and chromitite, it is extremely difficult to discover
in situ diamonds. After extensive work checking polished chromitite pieces for more than a year, six
in situ diamonds were finally discovered in the Luobusa and Ray-Iz chromitites
[44]. These
in situ diamonds are exclusively enclosed by chromite and associated with amorphous carbon
[29]. The diamonds hosted by chromite are similar to the diamonds recovered from mineral concentrates.
3. Characteristics of diamonds in ophiolitic peridotites and chromitites
Despite their occurrences in ophiolites of different age and tectonic setting, these diamonds share some common features in terms of morphology, internal structure, and chemical composition.
3.1. Morphology and diamond type
Ophiolite-hosted diamonds mainly range from colorless to light yellow. Most of the recovered diamonds are 50–500 μm in size, with a few as large as 700 μm (Fig. 2(d)). These diamonds include both monocrystalline and polycrystalline crystals, and no coated crystals have ever been recovered. Most diamonds are euhedral minerals with sharp crystal edges, while some grains are rounded with eroded or pitted surfaces (Fig. 3) [
24,
92]. Ophiolite-hosted diamonds predominantly show a cubo-octahedral morphology, with a small portion having a perfect octahedral shape (Fig. 3(a, b)).
Nitrogen is the most common and important impurity in diamonds. Based on the nitrogen content, diamonds can be classified into Type I (nitrogen concentration > 20 ppm) and Type II (nitrogen concentration < 20 ppm) diamonds
[102]. According to the aggregation form of the nitrogen atoms, Type I diamonds can be further subdivided into Ia and Ib subgroups
[102]. Secondary ion mass spectrometry (SIMS) and Fourier transform infrared spectroscopy (FTIR) can both be used to measure the nitrogen concentration in diamonds, but only FTIR can obtain information about the nitrogen aggregation form. FTIR analyses of diamonds from the Ray-Iz and Luobusa ophiolites indicate that these diamonds are Type Ib with a single substitutional center, implying a relatively short mantle residence time [
28,
103].
3.2. Internal structure and mineral inclusions
Diamonds with both cuboid and octahedral growth habits are known as “mix-habit diamonds” [
104,
105]. Ophiolite-hosted diamonds show growth sectors with different brightness in cathodoluminescence (CL) images (Fig. 4(a, b)). A bright sector represents an octahedral growth sector with a layered-growth characteristic, and a dull sector represents a cuboid growth sector
[28]. These ophiolite-hosted diamonds are mainly mix-habit diamonds.
Both nanometer-sized and micrometer-sized mineral inclusions have been observed in diamonds from the Luobusa, Ray-Iz, Hegenshan, and Pozanti–Karsanti ophiolites [
44,
92,
95,
106,
107]. These inclusions have been studied using scanning electron microscope (SEM), FTIR, Raman spectroscopy, focused ion beam (FIB), and TEM techniques. Recognized mineral inclusions include: ① metal alloys, including Ni–Mn–Si–Co, Ni–Mn–Co, Cr–Fe, and Fe–Si; ② silicate minerals, including albite, Mn-garnet, Mn-olivine, CaMn-wollastonite, and CaMn-perovskite; ③ oxides, including chromite, hematite, magnetite, and coesite; and ④ fluid inclusions (Fig. 4) [
28,
44,
106]. Lian et al.
[106] reported a mineral inclusion assemblage of (Ca
0.81Mn
0.19)SiO
3, Ni–Mn–Co alloy, and a nanometer-sized quenched fluid phase in diamonds from the Pozanti–Karsanti chromitite (Fig. 4(d)). As such a mineral combination has never been reported in a synthetic diamond, the discovery of this mineral inclusion assemblage provides further evidence for the natural origin of ophiolite-hosted diamonds
[106].
Of the mineral inclusions found in ophiolite-hosted diamonds, CaMn-perovskite is of great importance. Perovskite (Ca-perovskite, CaSiO
3) is an important mineral in both the MTZ and the lower mantle
[108]. Its modal mineral proportions are inferred to be about 5 vol% and 25 vol% in the pyrolitic mantle and subducted MORB materials, respectively, under lower mantle conditions [
108,
109]. Ca-silicate perovskite was discovered as mineral inclusions in diamonds from podiform chromitite of the Hegenshan and Sartohay ophiolites in China. The mineral inclusions were identified by FIB and TEM analyses. The calculated mineral formula of the mineral is Ca
0.8Mn
0.2SiO
3—that is, Mn-bearing Ca-silicate perovskite. TEM diffraction data reveal that the Hegenshan inclusion has
d-spacings and angles between adjacent lattice planes consistent with those of Ca-silicate perovskite with an orthorhombic structure
[110]. In comparison with other research results
[111], the
P–
T formation conditions of these inclusions in Hegenshan diamonds are estimated to be
P > 20 GPa and
T > 1600 °C. Different from the Ca-silicate minerals in Hegenshan and Sartohay ophiolite-hosted diamonds, the (Ca
0.81Mn
0.19)SiO
3 mineral inclusions in the Pozanti–Karsanti diamonds have a crystal structure of wollastonite rather than UHP perovskite, indicating that this inclusion may be a retrograde phase of UHP perovskite
[106].
3.3. Isotopic and trace element compositions
Cartigny
[37] and Shirey et al.
[14] have reviewed the carbon and nitrogen isotopic composition and nitrogen concentration of diamonds of different occurrences (Fig. 5) [
14,
37,
106]. However, in comparison with mantle-derived and metamorphic diamonds, the chemical compositions of ophiolite-hosted diamonds have not been well studied. Thus far, only diamonds from the Luobusa [
28,
29,
103], Ray-Iz
[44], and Pozanti–Karsanti
[106] ophiolites have been analyzed by SIMS for carbon and nitrogen compositions. We have compiled the available chemical composition data for the ophiolite-hosted diamonds, and compared these data with those of other types of diamonds (Fig. 5).
Ophiolite-hosted diamonds are characterized by quite light carbon isotopic composition, with δ
13C values ranging from −29‰ to −18‰ and a primary δ
13C mode at −25‰. These values are much lower than those of diamonds from either kimberlites/lamproite (mainly from −2‰ to −8‰) or UHP metamorphic belts (mainly from −5‰ to −18‰). The nitrogen isotopic composition and nitrogen concentration of ophiolite-hosted diamonds are quite different from those of UHP metamorphic diamonds, but are consistent with those of peridotitic and eclogitic diamonds (Fig. 5). Different models such as ① the primordial inhomogeneity model [
1,
112], ② the isotopic fractionation model [
113,
114], and ③ the subduction model [
35,
115] have been proposed to explain the origin of low δ
13C values in diamonds. Considering the characteristics of ophiolite-hosted diamond, Lian et al.
[106] proposed that the
13C-depleted carbon signature of ophiolite diamonds was inherited from previously subducted crustal matter.
Trace element concentration has been analyzed by Howell et al.
[28] for the Luobusa diamonds, which have light rare earth element (LREE)-enriched trace element patterns similar to those of some kimberlitic fibrous diamonds, but are characterized by strong negative anomalies in strontium (Sr), samarium (Sm), europium (Eu), and ytterbium (Yb). The trace element patterns are clearly distinct from those of synthetic diamonds
[28].
3.4. In situ diamonds
In situ diamond grains have been found in polished sections of chromitite ore from both the Luobusa and Ray-Iz ophiolites
[44]. Carbon element mapping and laser Raman studies on these
in situ microdiamonds show that they are enclosed in spherical or irregular patches of amorphous carbon hosted in chromite grains (Fig. 6).
Diamonds and amorphous carbon collectively form round inclusions. Such an inclusion usually contains a diamond in the center of about 0.2 mm in size and a few small pieces of chromite, along with numerous tiny silicate minerals (Fig. 6). The small angular pieces of chromite (Fig. 6(e)) in the inclusion are compositionally similar to the chromite grains surrounding the inclusions. The diamond in this inclusion is euhedral in shape, with a sharp edge within the surrounding amorphous carbon (Fig. 6). The spatial relationship clearly indicates that the diamond formed first, and the amorphous carbon subsequently cooled rapidly under decreasing temperatures. These textures likely formed when relatively hotter carbon fluids with diamond crystals entered cooler chromite grains. Highly pressurized fluids moving at high velocities appear to have shattered the host into smaller pieces.
3.5. Associated UHP, super-reduced, and crustal minerals
In addition to diamonds, other UHP minerals (e.g., pseudomorph stishovite, coesite, and qingsongite) and crustal minerals (e.g., zircons and quartz) have been discovered in different ophiolites, as summarized in Table 1 [
22,
24–
26,
29,
41,
43,
44,
92–
95,
116]. Yang et al.
[41] have observed lath-like grains of coesite and kyanite rimming a Fe–Ti alloy grain. The morphologies of individual coesite “crystal,” which is actually a mineral aggregate, are unlike the morphologies of known coesite grains formed in UHP metamorphic belts in the continental crust, but are similar to those of stishovite [
41,
117]. The polycrystalline nature and prismatic habit of these coesite crystals (which are several tens of micrometers long) strongly suggest that the observed coesites are replacements of stishovite (Fig. 7) [
24,
41–
43]. The presence of pseudomorph stishovites in our samples may imply pressures of greater than 9 GPa (∼300 km) and temperatures of 1000 °C for their formation conditions
[41].
Qingsongite, a natural analogue of cubic boron nitride (c-BN), has been found enclosed in lath-like kyanite and coesite from the Luobusa ophiolite (Fig. 7(b)). Other phases spatially associated with qingsongite include native iron, TiO
2 II (a high-pressure polymorph of rutile with a PbO
2 structure), amorphous carbon, and boron carbide of unknown stoichiometry
[41]. The accompanying mineral assemblages suggest qingsongite formation pressures of 10–15 GPa, assuming temperatures of about 1300 °C
[42]. Qingsongite is thus a typical UHP mineral, and its boron content is likely to have originated from the Earth’s surface via subduction.
Yamamoto et al.
[43] discovered numerous exsolution lamellae of diopside and coesite in chromites of the Luobusa podiform chromitites (Fig. 7(c)). This occurrence requires high solubility of SiO
2 and CaO in the host chromite. Chromite with diopside and coesite inclusions was interpreted to retrograde from a UHP CaFe
2O
4 (CF)-type structured polymorph, which is stable at pressures over 12.5 GPa (> 380 km)
[43]. These observations suggest that the source peridotites exposed along the Yarlung Zangbo Suture Zone (YZSZ) in southern Tibet may have originally been transported from the deep mantle (probably more than 380 km deep) to shallow depths beneath the Tethyan seafloor spreading system that produced these ophiolites. Recent experimental work has shown that chromite can form as deep as 410 km in the mantle, close to the MTZ
[118]. Thus, it appears that some ophiolitic chromitites have preserved the mineralogical evidence for their UHP origins and records
[29].
Super-reduced SiC minerals, which accompany diamonds, have been recovered from all the diamond-bearing ophiolites (Fig. 7(d)) [
24,
44,
92,
116]. The maximum oxygen fugacity (
fo
2) of SiC stability throughout the upper mantle is around five to six orders of magnitude below iron-wüstite oxygen buffer at equivalent temperature and pressure [
119,
120]. SiC minerals from ophiolite are generally light blue to blue in color, and are 50–300 μm in diameter (Fig. 7(d)). The existence of SiC in ophiolitic peridotite and chromitite suggests that these rocks have experienced a super-reduced condition. However, it is still under debate whether this super-reduced condition represents a grain-scale micro-environment [
121,
122] or is a common feature of Earth’s deep mantle [
44,
116].
Many typical crustal minerals, such as zircon, quartz, amphibole, and rutile, have also been recovered from podiform chromitites of the Luobusa and Dongqiao ophiolites in Tibet, the Oman ophiolite in the Southeast Arabian Peninsula, the Pozanti–Karsanti ophiolite in Turkey, and the Ray-Iz ophiolite in the Polar Urals, Russia (Fig. 7(e, f)) [
24,
44,
123,
124]. Most of the recovered zircons are sub-rounded grains with complex internal structures indicating polyphase growth. Their trace element contents and low-pressure inclusion assemblages (i.e., quartz, muscovite, K-feldspar, apatite, ilmenite, and rutile) indicate a continental crustal origin. Zircon grains have SIMS U–Pb ages that are generally much older than the host body (total range: 90–2500 Ma)
[123]. The presence of numerous crustal minerals, particularly zircons, suggests recycling of metasedimentary rocks on land into the deep mantle via subduction.
4. Origin of ophiolite-hosted diamonds
The discovery of diamonds and other unusual minerals in ophiolitic peridotites and chromitites has led to wide interest and debates on the origin of these minerals. Different models have been successively proposed by various scientific groups to explain the formation process of the diamonds and the hosting peridotites and chromitites (Fig. 8) [
44,
125,
126].
Overall, these models can be classified into three categories: ① the deep subduction-rapid exhumation model [
125,
127,
128]; ② the plume-related model [
44,
129]; and ③ the lightning-strike model [
130,
131]. In the deep subduction model, low-pressure rocks are subducted into the deep mantle (e.g., the MTZ or even the lower mantle), resulting in the formation of UHP minerals in ophiolitic peridotites and chromitites; these UHP mineral-bearing rocks are later transported into the shallow mantle via mantle convection (Fig. 8(a))
[125], or rise up in response to slab rollback in a collisional zone
[127]. Zhou et al.
[126] have suggested that ophiolite-hosted diamonds are of UHP metamorphic origin, and form in the subducting slab (Fig. 8(b)). However, ophiolite-hosted diamonds are clearly distinct from diamonds in UHP metamorphic rocks in terms of grain size and chemical composition
[106]. The plume-related model suggests that diamonds and other UHP minerals form in the transition zone and are carried up to the shallow mantle by plume upwelling (Fig. 8(c)) [
44,
129]. Yang et al.
[23] reviewed the various observations and interpretations of these different models and developed a revised three-stage model, which includes: ① the generation of carbon-bearing fluids in the deep mantle; ② the formation of diamond-bearing chromites in the MTZ; and ③ the transport of chromites and other UHP minerals to shallow depths through plume upwelling (Fig. 9).
An interesting and innovative perspective on the origin of ophiolite-hosted diamonds and associated minerals was recently proposed by Ballhaus et al.
[132]. These scholars argued that the unusual mineral association discovered in ophiolitic peridotites and chromitites may form through lightning strikes
[132]. However, several lines of evidence do not support the lightning-strike model, as has been clarified by Griffin et al.
[131] and Yang et al.
[133]. These lines of evidence mainly include ① the common occurrence of UHP and super-reduced minerals in ophiolites of different elevations with little possibility of lightning strike [
106,
133]; ② the common existence of UHP minerals that are not likely to form through lightning strike [
131,
133]; ③ the abundant mineral inclusions in ophiolite-hosted diamonds, which were not produced in lightning-strike experiments
[133]; and ④ the large size of ophiolite-hosted diamonds, which cannot be produced through the chemical vapor-deposition process in several seconds’ time during a lightning strike
[131].
In summary, the origin of ophiolite-hosted diamonds and associated minerals is still a conundrum. To solve this conundrum, many fundamental questions—such as how and where these minerals first crystallized, how they were incorporated into ophiolitic peridotites and chromitites, and how they were preserved during transportation to the shallow mantle—need to be answered. To answer these questions, continued research work must be conducted on these unusual minerals and hosting ophiolites through mineralogical and geochemical analyses, high-temperature and high-pressure experiments, and numerical modeling.
5. A new window for probing deep carbon cycling
5.1. Crustal material influx and efflux in the deep mantle
Seismic tomography has shown that the oceanic lithosphere and continental crust may subduct into the MTZ depths (410–660 km). Some slabs may even penetrate through the 660 km seismic discontinuity and reach the CMB (2900 km), providing evidence for wholesale mantle convection [
31–
33,
134,
135]. On the basis of plate reconstructions and tomographic images, it has been shown that the edges of the largest heterogeneities in the deepest mantle have controlled the eruption of most Phanerozoic kimberlites as well as large igneous provinces [
136,
137]. Researchers have proposed that the plume-generation zones of large low shear velocity provinces are on the CMB, and have inferred that diamonds may be sampled by plumes from the CMB
[137].
Oceanic crust becomes integrated into the mantle in significant quantities as oceanic slabs plunge into the mantle in subduction zones. An experimental study demonstrated that the subducted upper continental crust could not exhume back to the surface once it reached a depth of 250 km, because at those pressures, the crustal material would be transformed into a jadeite-stishovitite lithology, which facilitates further sinking into the MTZ depths
[138]. Thermally equilibrated subducted MORB crust is denser than the average mantle at all depths [
138–
141]. Thus, subducted MORB slabs can readily penetrate the D″ region (the lowermost ∼150 km of the mantle) at the CMB.
Large amounts of water can be transported into the Earth’s deep mantle by hydrous minerals in subducted crust. At the CMB, the subducted water may react with iron, forming an extremely oxygen-rich layer composed of iron, iron dioxide, and iron hydride, which can be reflected by the seismic signature of the D″ layer [
142–
144]. Liu et al.
[143] produced hydrogen-bearing iron peroxide (FeO
2H
x) through a reaction between water and iron at 86 GPa and 2200 K. These FeO
2H
x phase minerals have sound velocities consistent with the ultralow-velocity zones (ULVZs) at the Earth’s CMB. Liu et al.
[143] pointed out that one-tenth the mass of the Earth’s ocean water is needed to produce the ULVZs through the water–iron reaction, making Earth’s CMB an important hydrogen and oxygen reservoir for life on Earth. Hu et al.
[142] identified a highly stable pyrite-structured iron oxide (FeO
2) at 76 GPa and 1800 K using first-principles calculations and experiments. Decomposition of the mineral goethite, FeOOH, under deep lower mantle conditions will produce FeO
2 and H
2 [142]. When FeO
2 moves upward into the middle lower mantle, it decomposes to form Fe
2O
3 and releases O
2, making FeO
2 a sporadic source of extra O
2 [142]. These critical experiments and theoretical calculations indicate the possibility that fluids formed in the deep Earth contain recycled hydrogen, oxygen, and—most likely—carbon as well, although no experimental work has been carried out for carbon yet.
5.2. Ophiolite-hosted diamond as a new window
Large amounts of organic and inorganic carbon have been carried by subducting slabs into the Earth’s mantle, which plays an important role in balancing CO
2 in the atmosphere. The average oceanic crust production rate from the Mesozoic to the present time is estimated to be about 25 km
3·a
−1 [145]. Assuming that this production rate has been constant during the last 4 billion years, the total amount of oceanic crust that has been subducted corresponds to at least 11% of the Earth’s mantle in volume, suggesting that large quantities of MORB resides in the mantle
[145]. On the basis of cosmochemical and geochemical arguments, the upper mantle is estimated to contain about 100 ppm carbon, while the undegassed lower mantle is estimated to have about 1000–3700 ppm carbon, making it a substantial carbon reservoir [
1,
146]. Carbon in the mantle has three main forms: ① oxidized species such as CO
2 and carbonate; ② neutral species such as graphite and diamond; and ③ reduced species such as CH
4, CO, and SiC
[8]. Of these different carbon species, diamond has the most extensive distribution in the Earth’s interior from the CMB to the upper mantle [
38,
57], and is thus of great significance in tracing the carbon cycle in the deep mantle.
It has long been known that current volcanoes in the world’s island arcs can release tremendous amounts of CO
2 from the mantle into the air and lavas [
147,
148]. Intrusions and lavas in this setting also contain large amounts of CO
2 and have light carbon isotopic composition, with δ
13C values between −25‰ and −30‰, which are indicative of the recycling of subducted organic materials [
147,
148]. More recently, ultrapotassic mantle-derived lavas (lamproites) from Gaussberg, Antarctica, have been interpreted as a product of the melting of deeply recycled (subducted) Archean-age metasediments in the MTZ
[149]. However, it is not clear how deep these materials have recycled in the mantle. Subduction-related volcano eruption can be seen as representing Type I carbon recycling (Fig. 9).
As a messenger from the Earth’s interior, diamonds preserve information on the fate of deep subducted material and the composition of the deep mantle. Most peridotitic, eclogitic, and fibrous/coated diamonds from around the world have carbon isotopic compositions within the normal mantle range, indicating that the mantle carbon reservoir is the main carbon source for diamond formation in the mantle. However, as stated before, a large proportion (∼34%) of the world’s eclogitic diamonds have δ
13C values outside of the normal mantle range. This finding cannot simply be explained by the isotopic fractionation process, and may be indicative of a recycled crustal carbon source [
14,
37]. Mineral inclusions in diamonds can provide information about the formation depths [
13,
56]. About 1% of mantle-derived diamonds worldwide are suggested to have formed beneath the lithospheric mantle [
56,
150]. Walter et al.
[13] reported kimberlitic diamonds with lower mantle mineral inclusion assemblage and a light carbon isotopic composition (mainly δ
13C < −15‰), confirming the deep mantle cycling of oceanic crust. This kimberlite eruption in the continent can be seen as representing Type II carbon recycling (Fig. 9).
More importantly, ophiolite-hosted diamonds are all characterized by a quite unique carbon isotopic composition with low δ
13C values (< −18‰). These diamonds have a light carbon isotopic composition that is consistent with carbon sources of subducted organic matter [
151,
152]. A small portion of eclogitic diamonds also have positive δ
13C values consistent with carbon sources of subducted inorganic matter, such as carbonate (Fig. 6)
[37]. Collectively, the isotopic compositions of ophiolite-hosted diamonds and other mantle-derived diamonds have proved that surface-derived carbon can be subducted into the Earth’s mantle at depths greater than 150 km. CaMn-perovskite/wollastonite inclusion exists in ophiolite-hosted diamonds, indicating a super-deep origin for this newly discovered occurrence of diamonds. However, the lower limit of the subduction depth needs to be further constrained. These oceanic-spreading-center-related ophiolite-hosted diamonds can be seen as representing Type III carbon recycling (Fig. 9).
In summary, systematic studies on traditional mantle-derived diamonds (e.g., peridotitic, eclogitic, and fibrous/coated types) have greatly contributed to our understanding of carbon in the Earth’s interior. Present studies have shown that diamonds have common occurrences in ophiolitic peridotites and chromitites, making the ophiolite-hosted diamond a new window for probing carbon cycling in the deep mantle.
6. Conclusions
Diamonds have been recovered from ophiolitic peridotites and chromitites in different orogenic belts. Diamonds of this newly discovered occurrence have been named ophiolite-hosted diamonds. Ophiolite-hosted diamonds from different ophiolites show many similarities in their morphology, mineral inclusion associations, and chemical compositions. The discovery of ophiolite-hosted diamonds provides evidence of a completely new geological environment for diamond formation in the Earth’s mantle. We posit that diamonds and associated UHP minerals have a common existence in the so-called suprasubduction zone (SSZ)-type or mid-ocean ridge (MOR)-type ophiolitic peridotites and chromitites. Whether such diamonds may also occur in ophiolites formed in different geodynamic settings (i.e., plume-proximal MOR or ocean–continent transition (OCT) zones) is a significant question that needs to be tested with further systematic field-based, structural, and geochemical studies.
The widespread occurrence of ophiolite-hosted diamonds and associated UHP minerals suggests that they may be a common feature of
in situ oceanic mantle, since oceanic spreading centers are widely distributed on the seafloor and major igneous accretion of the oceanic lithosphere occurs along such spreading centers
[153]. Ophiolites remain an important geological window that gives us a close look at mantle dynamics, and the newly recognized ophiolite-hosted diamonds have provided us with a new window for probing the Earth’s deep carbon recycling processes and revealing the compositions and processes of the deep mantle.
PDF
(4689KB)
