The occurrence phases and enrichment mechanism of rare earth elements in cobalt-rich crusts from Marcus-Wake Seamounts

1. Introduction

Cobalt-rich crusts are one of important mineral resources on the ocean floor, which are mainly distributed on the seamounts, ridges and plateaus with water depths of 1 500−4 000 m (Hein et al., 2000, 2013; Halbach et al., 2017; Lusty et al., 2018; Fukami et al., 2022). Cobalt-rich crusts are composed of manganese oxides and iron oxyhydroxides, which have great economic value for some strategic elements such as Co, Ni, Cu, Mo, Te and rare earth elements (REEs) (Hein et al., 2016; Marino et al., 2017; Azami et al., 2018; Josso et al., 2021; Zhou et al., 2022). Meanwhile, cobalt-rich crusts grow at very slow rates of 1−10 mm/Ma, and their growth process spans the entire Cenozoic period (Goto et al., 2017; Nishi et al., 2017; Josso et al., 2020; Konstantinova et al., 2022). Therefore, cobalt-rich crusts have been regarded as the concentrated sedimentary strata, which are the recorders of oceanic and climatic history over the past 60 Ma (Gueguen et al., 2016, 2021; Josso et al., 2019; Charles et al., 2020; Konstantinova et al., 2020).

The major minerals of cobalt-rich crusts are vernadite (δ-MnO₂) and amorphous ferric oxyhydroxide (FeOOH) (Hein and Koschinsky, 2014; Conrad et al., 2017; Zhong et al., 2017; Jiang et al., 2020), and the secondary minerals include quartz, feldspar and carbonate fluorapatite (CFA). Koschinsky et al. (Koschinsky and Halbach, 1995; Koschinsky and Hein, 2003; Koschinsky et al., 2020) divide the mineral phases of cobalt-rich crusts into the following four occurrence phases: L1, adsorbed cations (easily leachable) and carbonate phase; L2, Mn oxide phase (easily reducible); L3, Fe oxyhydroxide phase (moderately reducible); L4, residual phase (hardly soluble) with silicates, crystalline oxides, and CFA. Previous studies suggested that a small quantity of alkali metals and alkaline earth metals are mainly enriched in the adsorbed and carbonate phases, while most other elements are mainly enriched in Fe-Mn oxide phases by means of surface adsorption and coprecipitation (Khanchuk et al., 2015; Mikhailik et al., 2017; Surya et al., 2020). Nath et al. (1994) proposed that the REE enrichment in cobalt-rich crusts is influenced by ferric oxide/hydroxide (FeOOH) and Mn oxide (δ-MnO₂), REEs are first adsorbated by FeOOH and then oxidized by δ-MnO₂. The leaching experiment of cobalt-rich crusts on REEs showed that (Bai et al., 2004), amorphous ferric oxyhydroxide has a strong adsorption capacity for REEs, and is the main occurrence phase of REEs. Xu et al. (2008) reported that about 66.4% of REEs are mainly enriched in the Mn oxide phase, which perhaps is associated with the higher Mn content in its Mn oxide phase. Bau and Koschinsky (2009) found through experiments that in non-phosphatized crusts, about 61.73%−81.63% of REEs are enriched in the Fe oxide phase, only about 14.07%−34.05% of REEs are enriched in the Mn oxide phase. Ren et al. (2011a) proposed that in the new non-phosphatized crusts, REE and Y are mainly enriched in the Mn oxide phase, and yet in the old phosphatized crusts, REE and Y mainly exist as separate mineral phases (independent of carbonate fluorapatite), which may be rare earth phosphate minerals. Mohwinkel et al. (2014) found through selective leaching experiments that the REEs enrichment is mainly influenced by the adsorption of the Fe oxide phase, but the contribution of the Mn oxide phase is relatively small. Recently, Huang et al. (2022) presented some different mechanisms to explain this strong enrichment in the Fe-Mn oxide phases, and the experiment showed that the metal elements enter the crusts by surface complexation, redox reactions, and lattice substitution. Among them, Ce³⁺ in seawater is oxidized to Ce⁴⁺ and enters the Fe oxide phase as CeO₂ precipitation, but other REEs first form anionic complexes and then enter the Fe oxide phase through adsorption.

As above mentioned, these results reflect the complexity of the enrichment mechanism in cobalt-rich crusts, and the enrichment mechanism of REEs is still unclear, so the cobalt-rich crust samples from Marcus-Wake Seamounts in the Pacific are chosen as the research object. In this study, we analyze the mineral composition and REE contents of the samples, and discuss the source, occurrence phases and enrichment mechanism of REEs. This study provides a reference for the metallogenesis of cobalt-rich crusts in the Pacific.

2. Materials and methods

2.1 Sample descriptions

The cobalt-rich crusts samples (No. XD3) are collected from Marcus-Wake Seamounts in the Pacific using the dredge during the DY115-18 expedition of R/V Dayang Yihao in 2006. The location of the sample XD3 is 19.79455°N and 157.31595°E (Fig. 1), with a water depth of 2 450 m.

Figure  1.  Location of the cobalt-rich crust XD3.

DownLoad: Full-Size Img PowerPoint

The sample XD3 is a tabular crust with obvious three-layer structure, which can be sub-divided into dense layer, loose layer and denser layer from top to bottom, and the bedrock is pyroclastic rock (Fig. 2). Of them, the upper dense layer and the middle loose layer are the new crust, and the lower dense layer is the old crust (Fig. 2). Along the growth profile of the crust, five structural layers are divided from top to bottom, followed by upper dense layer, lower dense layer, loose layer, upper denser layer and lower denser layer, marked as XD3(Ⅰ), XD3(Ⅱ), XD3(Ⅲ), XD3(Ⅳ) and XD3(Ⅴ), respectively. Then, 30 micro layers are further obtained by stratification, with sampling spacing of about 4−6 mm, marked as XD3(1) to XD3(30). The sample descriptions of the structural layers and micro layers are shown in Table 1. After drying the sample, grind it to 200 mesh by an agate mortar and place it in a clean sample bag for later use.

Figure  2.  Photograph of the cobalt-rich crust XD3.

DownLoad: Full-Size Img PowerPoint

Table  1.  Description of the cobalt-rich crust XD3

Structural layers number	Micro layers number	Structural layers	Depth/mm	Sample description
XD3(Ⅰ)	XD3(1)	dense layer	0−4	surface oolitic, black brown, dendritic structure
	XD3(2)	dense layer	4−8	black brown, dendritic structure
	XD3(3)	dense layer	8−12	black brown, dendritic structure
	XD3(4)	dense layer	12−14	black brown, dendritic structure
XD3(Ⅱ)	XD3(5)	dense layer	14−19	black brown, columnar structure
	XD3(6)	dense layer	19−24	black brown, columnar structure
XD3(Ⅲ)	XD3(7)	loose layer	24−27	black brown, columnar structure, many clays
	XD3(8)	loose layer	27−30	black brown, columnar structure, many clays
	XD3(9)	loose layer	30−34	black brown, columnar structure, many clays
	XD3(10)	loose layer	34−38	black brown, columnar structure, many clays
	XD3(11)	loose layer	38−42	black brown, columnar structure, many clays
	XD3(12)	loose layer	42−46	black brown, columnar structure, many clays
	XD3(13)	loose layer	46−50	black brown, columnar structure, many clays
	XD3(14)	loose layer	50−54	black brown, columnar structure, many clays
	XD3(15)	loose layer	54−56	black brown, columnar structure, many clays
XD3(Ⅳ)	XD3(16)	denser layer	56−58	Black, columnar structure, some phosphate veins
	XD3(17)	denser layer	58−63	Black, columnar structure, many phosphate veins
	XD3(18)	denser layer	63−67	Black, columnar structure, many phosphate veins
	XD3(19)	denser layer	67−73	Black, columnar structure, some phosphate veins
	XD3(20)	denser layer	73−79	Black, columnar structure, some phosphate veins
	XD3(21)	denser layer	79−83	Black, columnar structure, some phosphate veins
XD3(Ⅴ)	XD3(22)	denser layer	83−89	Bright black, lamellar structure, some phosphate veins
	XD3(23)	denser layer	89−93	Bright black, lamellar structure, some phosphate veins
	XD3(24)	denser layer	93−97	Bright black, lamellar structure, some phosphate veins
	XD3(25)	denser layer	97−101	Bright black, lamellar structure, some phosphate veins
	XD3(26)	denser layer	101−105	Bright black, lamellar structure, some phosphate veins
	XD3(27)	denser layer	105−109	Bright black, lamellar structure, some phosphate veins
	XD3(28)	denser layer	109−113	Bright black, lamellar structure, many phosphate veins
	XD3(29)	denser layer	113−117	Bright black, lamellar structure, many phosphate veins
	XD3(30)	denser layer	117−120	Bright black, lamellar structure, many phosphate veins

 | Show Table

DownLoad: CSV

2.2 Major and trace element analytical methods

Weigh the sample (50.00 ± 0.50) mg for each sample into the dissolved sample liner, then add 1 mL nitric acid, 1 mL hydrochloric acid and 1 mL hydrofluoric acid into the samples respectively, and then put them in a sealed steel sleeve and heat them in an oven at 190℃ for 48 h. After cooling, put it on an electric heating plate at 150℃and steam it dry, then add 3 mL of 30% (V/V) hydrochloric acid solution and 0.5 mL of 1 µg/g rhodium internal standard solution, and cover it and put it in a sealed steel sleeve, and heat it in an oven at 150℃ for 8 h. After cooling, use 2% (V/V) nitric acid solution to fix the volume to 50 g, to be measured. Major elements are determined by ICP-OES (Thermo Fisher, USA, iCAP 7400), and trace elements and REEs are determined by ICP-MS (iCAP RQ, Thermo Fisher, USA) at the Key Laboratory of Marine Geology and Metallogeny, First Institute of Oceanography, Ministry of Natural Resources (MNR). The standard materials GBW07337, GBW07338 and GBW07339 are used to monitor data quality in the analysis process. The relative errors of elements are less than 5%, and the recoveries are between 90% and 105%.

2.3 Chemical leaching method

According to the chemical leaching method (Koschinsky and Halbach, 1995; Koschinsky and Hein, 2003), four phases of cobalt-rich crusts are extracted at the Key Laboratory of Marine Geology and Metallogeny, First Institute of Oceanography, MNR, and the experimental steps are as follows.

(1) Exchangeable ions and carbonate phase (referred to as carbonate phase): weigh 1 g of sample, put it in a reaction bottle, add 30 mL of 1 mol/L acetic acid solution (pH = 2.5), and shake at 25℃ for 5 h, after centrifugation of the extract, to be measured.

(2) Mn oxide phase: after washing the residue in step (1), add 175 mL of 0.1 mol/L hydroxylamine hydrochloride solution (pH = 3.7) and shake at 25℃ for 24 h, after centrifugation of the extract, to be measured.

(3) Fe oxide phase: after washing the residue in step (2), add 175 mL of 0.2 mol/L ammonium oxalate solution (pH = 3.5) and shake at 25℃ for 24 h, after centrifugation of the extract, to be measured.

(4) Residual phase: after washing the residue in step (3), transfer to the dissolved sample liner, major elements and trace elements are analyzed according to the Section 2.2. During the experiment, the elements recoveries are between 90% and 105%, and the relative errors are less than 5%. The acetic acid, hydroxylamine hydrochloride, oxalic acid and ammonium oxalate used in the experiment are all superior purity, and the nitric acid, hydrofluoric acid are all obtained by secondary azeotropic distillation, and the water is secondary deionized water.

2.4 Mineral identification method

The cobalt-rich crust samples are placed into a special sample carrier cup and pressed into thin slices. X-ray diffractometer (XRD, Nippon Science Corporation, Japan, D/MAX2500HB+/PC) is used for scanning analysis. Studies on mineral identification of samples were performed at the Key Laboratory of Marine Geology and Metallogeny, First Institute of Oceanography, MNR.

3. Results

3.1 Mineralogical characteristics

The XRD analytical results of cobalt-rich crusts from Marcus-Wake Seamounts are shown in Fig. 3. The major crystalline mineral is vernadite (δ-MnO₂), which is found in all cobalt-rich crust samples, and many amorphous ferric oxyhydroxides (FeOOH) are contained in cobalt-rich crusts, as Hein et al. (2000, 2013) suggested. The main detrital minerals of cobalt-rich crusts include quartz and plagioclase. The new crusts XD3(3), XD3(6), and XD3(13) contain more detrital minerals than the old crusts XD3(18) and XD3(26), indicating that cobalt-rich crusts receive more terrigenous detrital materials during forming new crusts (Hein et al., 2000; Conrad et al., 2017). The old crusts XD3(18) and XD3(26) contain more CFA, while the new crusts are almost absent, indicating that cobalt-rich crusts are affected by phosphatization during forming old crusts (Koschinsky et al., 1997; Hein et al., 2016). Therefore, the main mineral phases of cobalt-rich crusts can be sub-divided into Mn mineral phase, Fe mineral phase, detrital mineral phase, and phosphate phase.

Figure  3.  X-ray diffraction pattern of the cobalt-rich crusts.

V-vernadite, Q-quartz, Pl-plagioclase, CFA-carbonate fluorapatite.

DownLoad: Full-Size Img PowerPoint

3.2 Major oxides and trace element characteristics

The major oxides and trace element contents of cobalt-rich crusts from Marcus-Wake Seamounts are shown in Table 2. Of the major elements, Mn and Fe contents are the highest, with Mn ranging from 10.83% to 28.76% and Fe ranging from 6.14% to 18.86%. Next are CaO and P₂O₅, with CaO ranging from 2.87% to 27.74% and P₂O₅ ranging from 0.58% to 16.75%. Then there are Al₂O₃, Na₂O, MgO, TiO₂ and K₂O, with their contents ranging from 0.29% to 2.55%. Again, there are Co, Ni, Ba, Sr, Pb and Cu, with their contents ranging from 0.04% to 0.98%. Finally, there are V, Zn, Mo and Zr, with their contents ranging from 202 µg/g to 1 024 µg/g.

Table  2.  Major and trace element contents in cobalt-rich crusts

Sample number	Mn/ %	Fe/ %	CaO/ %	P2O5/ %	Al2O3/ %	Na2O/ %	K2O/ %	MgO/ %	TiO2/ %	Co/ %	Cu/ %	Ni/ %	Ba/ %	Sr/ %	Pb/ %	V/ (µg·g−1)	Zn/ (µg·g−1)	Zr/ (µg·g−1)	Mo/ (µg·g−1)
XD3(1)	19.78	18.62	2.96	0.90	1.99	2.34	0.62	1.76	1.55	0.53	0.04	0.28	0.11	0.14	0.17	618	502	577	443
XD3(2)	19.12	18.86	2.87	0.85	1.71	2.31	0.62	1.70	1.61	0.52	0.04	0.28	0.12	0.14	0.16	617	500	620	487
XD3(3)	20.62	18.22	2.92	0.82	1.38	2.23	0.56	1.69	1.63	0.54	0.05	0.32	0.13	0.14	0.17	635	528	597	540
XD3(4)	21.26	17.73	3.07	0.82	1.01	2.29	0.51	1.71	1.62	0.56	0.05	0.35	0.13	0.15	0.18	676	558	530	607
XD3(5)	24.20	15.38	3.34	0.74	0.69	2.30	0.49	1.76	1.86	0.63	0.07	0.43	0.13	0.15	0.18	683	580	412	709
XD3(6)	25.87	13.96	3.50	0.67	0.62	2.27	0.53	1.83	2.09	0.64	0.09	0.53	0.14	0.15	0.18	672	648	412	702
XD3(7)	26.67	13.12	3.48	0.58	0.73	2.32	0.57	1.88	2.12	0.61	0.11	0.61	0.15	0.15	0.17	605	692	493	659
XD3(8)	25.06	13.59	3.42	0.62	0.76	2.24	0.59	1.82	2.12	0.62	0.12	0.55	0.15	0.15	0.16	586	667	534	597
XD3(9)	24.99	13.28	3.39	0.59	0.87	2.24	0.59	1.84	2.00	0.67	0.13	0.54	0.15	0.14	0.15	587	676	550	610
XD3(10)	25.45	12.31	3.69	0.79	1.38	2.32	0.69	2.01	2.05	0.77	0.15	0.62	0.16	0.14	0.13	576	747	548	627
XD3(11)	23.85	11.90	4.28	1.31	1.35	2.19	0.76	2.00	2.17	0.75	0.15	0.60	0.17	0.13	0.13	553	771	577	571
XD3(12)	26.70	11.03	4.43	1.26	1.22	2.28	0.79	2.24	1.92	0.73	0.18	0.80	0.18	0.13	0.12	587	920	559	721
XD3(13)	26.70	10.64	4.59	1.33	1.23	2.38	0.78	2.24	1.77	0.67	0.19	0.83	0.18	0.13	0.12	584	926	540	758
XD3(14)	28.76	11.29	4.52	1.11	0.95	2.55	0.76	2.37	1.63	0.63	0.21	0.98	0.19	0.13	0.13	591	1024	503	833
XD3(15)	27.29	10.68	5.62	1.84	0.75	2.38	0.70	2.24	1.51	0.58	0.20	0.94	0.18	0.13	0.13	548	974	464	805
XD3(16)	23.46	9.31	10.80	5.20	0.87	2.37	0.65	2.00	1.28	0.42	0.17	0.80	0.16	0.13	0.11	488	830	449	676
XD3(17)	13.97	6.14	26.12	15.49	1.27	1.86	0.56	1.35	0.91	0.20	0.10	0.40	0.10	0.13	0.06	282	456	376	320
XD3(18)	10.83	7.08	27.74	16.75	1.23	1.79	0.56	1.18	0.95	0.15	0.08	0.23	0.10	0.13	0.07	283	364	386	202
XD3(19)	20.92	11.96	12.38	6.08	1.33	2.06	0.62	1.62	1.88	0.39	0.15	0.36	0.21	0.15	0.13	590	615	664	482
XD3(20)	20.96	14.42	10.42	4.93	0.97	2.12	0.48	1.48	2.05	0.38	0.15	0.31	0.25	0.17	0.16	708	640	807	505
XD3(21)	21.26	14.51	15.14	8.20	0.69	1.95	0.56	1.48	1.11	0.35	0.16	0.40	0.19	0.16	0.14	666	598	508	770
XD3(22)	22.73	14.59	9.01	3.95	0.58	1.91	0.46	1.42	1.59	0.34	0.13	0.29	0.25	0.18	0.22	874	671	566	792
XD3(23)	20.79	13.27	12.30	5.85	0.55	1.90	0.45	1.33	1.44	0.30	0.11	0.26	0.24	0.18	0.20	780	609	502	701
XD3(24)	21.62	11.74	13.23	6.47	0.42	1.89	0.46	1.32	1.37	0.29	0.10	0.28	0.22	0.18	0.18	752	569	417	683
XD3(25)	20.42	11.89	13.66	7.24	0.57	1.93	0.43	1.28	1.47	0.26	0.09	0.26	0.23	0.18	0.18	732	560	408	666
XD3(26)	19.84	11.83	15.63	8.12	0.40	1.80	0.41	1.22	1.50	0.22	0.07	0.23	0.23	0.18	0.18	719	556	383	600
XD3(27)	21.22	12.61	11.24	5.40	0.34	1.87	0.43	1.29	1.68	0.26	0.06	0.25	0.26	0.18	0.20	798	618	376	693
XD3(28)	20.92	12.74	11.09	5.24	0.31	1.80	0.43	1.28	1.78	0.26	0.05	0.24	0.28	0.19	0.20	818	639	366	736
XD3(29)	21.50	13.40	11.62	5.55	0.29	1.79	0.42	1.27	1.65	0.27	0.05	0.22	0.30	0.19	0.20	818	642	336	778
XD3(30)	20.62	13.98	12.90	6.65	0.51	1.67	0.38	1.23	1.55	0.26	0.05	0.22	0.31	0.19	0.22	728	606	392	652

 | Show Table

DownLoad: CSV

The REE contents of cobalt-rich crusts from Marcus-Wake Seamounts are shown in Table 3. Cobalt-rich crusts are enriched in REE, with REE ranging from 1 563 µg/g to 3 238 µg/g. Among all REEs, Ce content is higher than other elements, with Ce ranging from 790 µg/g to 1 722 µg/g, accounting for about 50% of the total REEs. LREE contents range from 1 271 µg/g to 2 639 µg/g, and HREE contents range from 236 µg/g to 599 µg/g, thus the LREE/HREE ratio is 2.95−5.80, indicating that LREE are higher than HREE. The δCe ranges from 1.24 to 2.22, and δCe >1, indicating that the cobalt-rich crusts have positive Ce anomaly. The δEu ranges from 0.89 to 1.04, and δEu ≈ 1, indicating that the cobalt-rich crusts have no obvious Eu anomaly. In addition, the average content of REEs in the upper crust is 2 032 µg/g, and that in the middle crust is 1 722 µg/g, and that in the lower crust is 2 405 µg/g, so REE contents of the old crusts are higher than those of the new crusts.

Table  3.  REE contents in cobalt-rich crusts

Sample number	La/ (µg·g−1)	Ce/ (µg·g−1)	Pr/ (µg·g−1)	Nd/ (µg·g−1)	Sm/ (µg·g−1)	Eu/ (µg·g−1)	Gd/ (µg·g−1)	Tb/ (µg·g−1)	Dy/ (µg·g−1)	Ho/ (µg·g−1)	Er/ (µg·g−1)	Tm/ (µg·g−1)	Yb/ (µg·g−1)	Lu/ (µg·g−1)	Y/ (µg·g−1)	REE/ (µg·g−1)	LREE/ (µg·g−1)	HREE/ (µg·g−1)	LREE/ HREE	${\text{δ}} $Ce	${\text{δ}} $Eu
XD3(1)	288	904	67.23	270	58.14	14.00	64.01	9.69	57.37	11.02	30.39	4.22	27.73	3.95	205	2015	1601	414	3.87	1.42	1.01
XD3(2)	303	866	65.55	265	55.74	13.70	61.73	9.48	57.96	11.16	31.08	4.33	28.08	4.09	210	1987	1569	418	3.76	1.34	1.03
XD3(3)	323	847	66.61	271	55.74	13.61	62.85	9.66	60.96	11.88	33.19	4.68	30.13	4.40	215	2009	1577	432	3.65	1.25	1.01
XD3(4)	328	857	68.00	273	56.10	13.73	63.42	9.72	60.54	11.83	32.89	4.66	30.33	4.34	218	2033	1597	436	3.67	1.24	1.01
XD3(5)	328	951	73.97	289	60.84	14.51	65.10	9.91	60.31	11.58	32.38	4.59	29.17	4.26	187	2121	1717	404	4.25	1.33	1.01
XD3(6)	290	958	70.88	271	58.22	13.62	60.50	9.10	53.31	9.94	27.44	3.92	25.49	3.63	170	2025	1662	363	4.58	1.46	1.01
XD3(7)	239	985	58.75	221	48.26	11.14	50.27	7.39	42.49	7.81	21.57	3.18	20.65	2.95	131	1850	1562	288	5.43	1.81	0.99
XD3(8)	227	964	54.58	203	44.18	10.42	48.06	6.76	38.80	7.18	19.94	2.89	19.15	2.70	131	1781	1504	277	5.44	1.88	0.99
XD3(9)	198	909	46.48	172	36.95	8.75	41.50	5.69	33.14	6.07	17.01	2.48	16.50	2.34	112	1608	1371	236	5.80	2.07	0.98
XD3(10)	200	927	46.29	177	38.18	9.17	44.10	6.15	35.70	6.83	19.15	2.73	17.45	2.53	129	1660	1397	263	5.31	2.10	0.98
XD3(11)	222	903	54.76	217	46.61	11.29	52.64	7.57	45.70	8.92	24.63	3.37	20.97	2.97	193	1815	1455	360	4.04	1.78	1.00
XD3(12)	229	936	54.13	211	44.89	11.11	51.90	7.49	45.85	9.09	25.04	3.42	21.54	3.11	192	1845	1486	360	4.13	1.83	1.01
XD3(13)	223	919	50.36	199	42.13	10.32	50.21	7.14	43.01	8.60	24.28	3.36	21.13	3.02	195	1799	1444	355	4.06	1.89	0.99
XD3(14)	219	797	48.50	187	38.40	9.75	44.27	5.86	38.36	6.73	18.17	2.61	16.49	2.43	145	1579	1299	280	4.64	1.68	1.04
XD3(15)	218	790	45.12	173	35.46	9.16	42.55	5.59	37.28	6.78	18.89	2.76	17.71	2.65	157	1563	1271	292	4.36	1.73	1.04
XD3(16)	228	900	37.90	147	28.32	7.43	38.05	5.08	34.41	8.20	25.55	3.94	26.48	4.16	308	1803	1349	453	2.98	2.07	0.99
XD3(17)	228	894	39.07	155	30.50	7.66	38.86	5.45	36.03	8.08	24.62	3.66	24.32	3.90	314	1812	1354	459	2.95	2.03	0.98
XD3(18)	249	1052	49.77	198	38.88	9.61	47.21	6.54	40.03	8.65	24.77	3.44	22.85	3.51	319	2074	1598	476	3.35	2.05	0.98
XD3(19)	266	1170	61.11	241	49.14	11.98	57.34	7.67	43.92	8.51	23.12	3.08	19.09	2.78	212	2176	1799	377	4.77	2.00	0.99
XD3(20)	303	1358	68.21	269	53.74	13.09	65.82	8.53	49.25	9.38	25.38	3.33	21.08	3.00	237	2488	2065	423	4.88	2.06	0.97
XD3(21)	296	1202	54.20	207	39.35	9.66	48.70	6.80	42.21	8.38	24.08	3.51	23.26	3.41	268	2237	1 809	428	4.22	2.04	0.97
XD3(22)	288	1234	49.36	183	32.86	8.25	46.44	5.81	33.99	6.63	18.74	2.71	17.70	2.58	256	2186	1796	391	4.60	2.22	0.93
XD3(23)	312	1340	58.77	221	38.74	9.54	54.87	7.14	43.89	9.05	25.70	3.79	24.62	3.62	248	2400	1 980	420	4.71	2.13	0.91
XD3(24)	266	1160	52.33	198	34.39	8.76	48.79	6.56	41.48	8.70	25.25	3.70	24.03	3.54	284	2165	1720	446	3.86	2.13	0.94
XD3(25)	275	1197	51.61	192	33.26	8.45	48.34	6.24	38.97	8.09	23.18	3.37	22.06	3.25	245	2156	1757	398	4.41	2.17	0.93
XD3(26)	332	1406	60.49	223	38.00	9.25	53.16	6.66	40.57	8.32	23.80	3.40	22.43	3.30	240	2471	2069	402	5.15	2.14	0.90
XD3(27)	353	1532	65.13	239	40.28	9.70	57.09	6.89	40.76	8.42	24.42	3.45	23.00	3.42	278	2685	2239	445	5.03	2.18	0.89
XD3(28)	373	1662	78.07	290	50.79	12.27	69.67	9.01	54.64	11.60	33.59	4.79	31.09	4.61	276	2962	2467	495	4.99	2.11	0.91
XD3(29)	419	1722	89.34	330	58.02	14.21	77.13	10.34	64.50	12.80	40.10	5.62	36.44	5.32	341	3225	2632	593	4.44	1.93	0.93
XD3(30)	424	1711	89.64	341	59.49	14.60	78.26	10.32	64.41	12.69	39.69	5.60	35.90	5.28	347	3238	2639	599	4.41	1.91	0.94
Note: LREE = La + Ce + Pr + Nd + Sm + Eu, HREE = Gd + Tb + Dy + Ho + Er + Tm + Yb + Lu + Y, ${\text{δ}} $Ce = 2CeN/(LaN+ PrN), ${\text{δ}} $Eu = 2EuN/(SmN + GdN), LaN, CeN, PrN, SmN, EuN and GdN are normalized by North American shale (NASC), and NASC data are from Wang et al. (1989).

 | Show Table

DownLoad: CSV

4. Discussion

4.1 Origin of cobalt-rich crusts

Marine ferromanganese deposits can be generally sub-divided into three origins: hydrogenetic, diagenetic and hydrothermal (Halbach, 1986; Liu and Wang, 2021). The metal ions of hydrogenetic ferromanganese deposits come from seawater, and the main components are the Fe-Mn oxides precipitated by colloidal particles in seawater. They are usually formed under strong oxidation condition, and the growth rate is very slow, generally ranging from 1 mm/Ma to 10 mm/Ma. REE contents are above 1 500 µg/g, and REE distribution patterns normalized by North American shale (NASC) show positive Ce and negative Y anomalies (Hein et al., 2013, Bau et al., 2014). The metal ions of diagenetic ferromanganese deposits come from the pore water in the sediment or sediment-water interface under suboxidation condition (Marino et al., 2019). They are usually formed under weak oxidation condition, and REE contents are generally lower than those of diagenetic Fe-Mn deposits, about 1 000 µg/g, and REE distribution patterns normalized by NASC show negative Ce and Y anomalies (Marino et al., 2019). The metal ions of hydrothermal ferromanganese deposits come from the hydrothermal fluids that erupt from the seafloor at medium-low temperature, with the fastest growth rate, and REE contents generally are lower than 100 µg/g. REE distribution patterns normalized by NASC show negative Ce and positive Y anomalies (Pelleter et al., 2017). In this study, REE contents of cobalt-rich crusts are 1 563−3 238 µg/g, δCe > 1 and δY < 1 in the non-phosphatized crusts, indicating positive Ce anomaly and negative Y anomaly, all of which are consistent with hydrogenetic Fe-Mn crusts.

According to the Fe-(Co + Cu + Ni) × 10-Mn ternary diagram (Bonatti et al., 1972) and the (Cu + Ni) × 15- (Zr + Y + Ce) × 100-(Fe +Mn)/4 ternary diagram (Josso et al., 2017), the origin of ferromanganese deposits can be distinguished. The results show that (Fig. 4), the sample points are all plotted within the hydrogenetic area, which indicate that the cobalt-rich crusts are hydrogenetic and are not affected by submarine hydrothermal activity and diagenesis. Meanwhile, according to δCe-Nd and δCe-Y_N/Ho_N correlation diagram (Bau et al., 2014), the origin of ferromanganese deposits can also be distinguished. The results show that (Fig. 5), Nd > 100 µg/g, δCe > 1 and Y_N/Ho_N < 1.5, the sample points are all plotted within the hydrogenetic area, which indicate that the cobalt-rich crusts are hydrogenetic and are not affected by submarine hydrothermal activity and diagenesis.

Figure  4.  Ternary diagram in cobalt-rich crusts: a. Fe-(Co + Cu + Ni) × 10-Mn; b. (Cu + Ni) × 5-(Zr + Y + Ce) × 100- (Fe + Mn)/4 (Asavin et al., 2010; Hein et al., 2012, 2013; Usui et al., 2017; Sousa et al., 2021; Zhou et al., 2022)

DownLoad: Full-Size Img PowerPoint

Figure  5.  Element correlation diagram in cobalt-rich crusts: a. δCe vs. Nd; b. δCe vs. Y_N/Ho_N. δCe = 2Ce_N/(La_N + Pr_N). La_N, Ce_N, Pr_N, Y_N and Ho_N are normalized by NASC, and NASC data are from Wang et al. (1989).

DownLoad: Full-Size Img PowerPoint

NASC-normalized REE distribution patterns show that (Fig. 6), the non-phosphatized crusts have positive Ce and negative Y anomalies, while the phosphatized crusts have positive Ce and Y anomalies. The increase of Y content in the phosphatized crusts is caused by the phosphate mixing (Koschinsky et al., 1997; Koschinsky and Hein, 2003; Ren et al., 2016). Seawater shows negative Ce and positive Y anomalies, and REE pattern of the non-phosphatized crusts mirrors that of seawater, indicating that REEs in seawater are consumed during forming cobalt-rich crusts. Therefore, REEs of cobalt-rich crusts mainly come from seawater. In addition, Fe-Mn oxides and marine phosphates are two typical marine authigenic components (Pan et al., 2002, 2005), and their REEs come from seawater. The phosphate rock on seamount has a REE pattern similar to seawater, showing negative Ce and positive Y anomalies, indicating that marine phosphates inherit the REE pattern of seawater. So the unique REE pattern of seawater is the result of the balance between supply and consumption (Ren et al., 2017).

Figure  6.  NASC-normalized REE distribution patterns for cobalt-rich crusts: NASC and seawater data are from Wang et al (1989), non-phosphatized crusts, phosphatized crusts and phosphate rock data are from this study.

DownLoad: Full-Size Img PowerPoint

4.2 Enrichment mechanism of REEs

In order to reveal the occurrence phases and enrichment mechanism of REEs in cobalt-rich crusts from Marcus-Wake Seamounts, scientists have separated the following four phases (Koschinsky and Halbach, 1995; Koschinsky and Hein, 2003; Gao et al., 2015, 2023): carbonate phase, Mn oxide phase, Fe oxide phase and residual phase. The phase distribution of its elements in structural layers is shown in Fig. 7. In the non-phosphatized crusts XD3(Ⅰ), XD3(Ⅱ), and XD3(Ⅲ), Na, K, Ca, Mg, and Sr are mainly enriched in the carbonate phase, and Mn, Ba, Sr, Co and Ni are mainly enriched in the Mn oxide phase, and Fe, P, Ti, Cu, Pb, V, Zn, Zr, Mo and REEs are mainly enriched in the Fe oxide phase, and Al and K are mainly enriched in the residual phase. Thus, about 67% of REEs are enriched in the Fe oxide phase, and about 17% in the Mn oxide phase, and only about 13% in the residual phase, and yet the contents of these elements in carbonate phase is relatively small. In contrast, in the phosphatized crusts XD3(Ⅳ) and XD3(Ⅴ), the distribution ratios of Ca, P and REEs in the residual phase increase, and the proportion ratios of REEs increase from 13% to 64%, Ca from 0.54% to 55%, and P from 11% to 40%. Therefore, in the phosphatized crusts, about 64% of REEs are enriched in the residual phase containing carbonate fluorapatite, but the influence of Fe and Mn oxide phases is greatly reduced, which is consistent with the results of Koschinsky et al. (Koschinsky and Hein, 2003; Koschinsky et al., 2020).

Figure  7.  The occurrence phase of elements in structural layers of cobalt-rich crusts.

DownLoad: Full-Size Img PowerPoint

The cobalt-rich crusts are hydrogenetic, and the elements mainly come from seawater (Hein et al., 2016; Halbach et al., 2017). According to the results of mineral identification, the carbonate phase is mainly composed of calcite and other minerals, and Na, K, Ca, Mg and Sr in seawater mainly exist in the form of free cations and then enter the carbonate phase (together with carbonate minerals) through ion exchange or adsorption (Koschinsky and Hein, 2003; Bai et al., 2004). The Mn oxide phase is mainly composed of vernadite (δ-MnO₂), and Ba, Sr, Co and Ni in seawater are enriched in the Mn oxide phase by adsorption during the formation of Mn colloidal particles (He et al., 2005, 2011). The Fe oxide phase is mainly composed of amorphous ferric oxyhydroxide (FeOOH), and P, Ti, Cu, Pb, V, Zn, Zr and REEs in seawater are enriched in the Fe oxide phase by adsorption during the formation of Fe colloidal particles (Koschinsky and Hein, 2003; Koschinsky et al., 2020). The residual phase is mainly composed of quartz, feldspar and other minerals, containing silicon aluminate such as Si, Al and K, which are mainly derived from terrigenous detrital materials and enriched in cobalt-rich crusts by seawater exchange (Hein et al., 2016).

Multiple phosphatization events occur during the early growth of cobalt-rich crusts (Koschinsky et al., 1997; Hein et al., 2000). Due to the effect of phosphatization, Ca and P contents in the old crusts are generally high, and there are CFA minerals (Ren et al., 2017). During the growth of carbonate fluorapatite, some elements have undergone secondary enrichment, and REEs are transferred and enriched in the phosphate phase. Due to the similar radius of REE³⁺ and Ca²⁺, REE³⁺ can effectively combine with PO₄³⁻ instead of Ca²⁺ and form rare earth phosphate minerals (Koschinsky et al., 2020). In contrast, REE contents of phosphate selected from the crusts are much lower than those of the crusts, indicating that REE contents of phosphate in the crusts are not uniform, the rich rare earth phosphate is difficult to separate from the crusts. Consequently, the simulation results of the mixture of crusts and phosphates showed that (Ren et al., 2017), there should be rich rare earth phosphate in the phosphatized crusts, which may be disseminated in the crusts, representing a relatively slow phosphate precipitation process. The results of the cobalt-rich crusts from Magellan Seamounts showed that (Ren et al., 2011a), up to 42%−88% of REEs in the phosphatized crusts are contributed from phosphate, and it is inferred that there may be independent rare earth phosphate minerals. Therefore, further research is needed on the combination of rare earth elements with phosphate to form rare earth phosphate minerals.

4.3 Influencing factors of REE enrichment

The variation curves of element contents with depth in the growth profile of cobalt-rich crust sample XD3 are shown in Fig. 8. The results show that, REE contents increase from the new crusts to the old crusts, and there is a highest value in the old crusts. When CaO/P₂O₅ ratio is less than 2, it indicates that the cobalt-rich crusts have a phosphatization event (Koschinsky et al., 1997; Ren et al., 2016). In the study, CaO/P₂O₅ ratio of the old crusts ranges from 1.66 to 2.28, indicating that the phosphatization occurs in the old crusts. The P₂O₅/CaO ratio does not change much in the new crusts, but increases in the old crusts, and REE contents also increase, which indicates that phosphatization can promote the REE enrichment. When Mn/Fe ratio is less than 2.5, it indicates that the cobalt-rich crusts are hydrogenetic (Halbach et al., 1983; Hou et al., 2020). In the study, Mn/Fe ratio ranges from 1.01 to 2.56, indicating that the cobalt-rich crusts are hydrogenetic and REEs mainly come from seawater. Moreover, the δCe can indicate the redox environment of seawater during forming cobalt-rich crusts (Hein et al., 2000). During the growth period of the old crusts, δCe is higher, seawater is an oxidizing environment, and REE contents are higher relative to the new crusts, which indicates that the oxidizing environment of seawater can promote the REE enrichment. The variations of Sr contents can be used to indicate the degree of marine productivity (Li et al., 2011). During the growth period of the old crusts, Sr content is higher, marine productivity is higher, and REE contents are higher relative to the new crusts, which indicates that high marine productivity can promote the REE enrichment. Meanwhile, the variations of Co contents can indicate the growth rate of cobalt-rich crusts (Ren et al., 2011b). During the growth period of the old crusts, Co content is lower, the growth rate is slower, and REE contents are higher relative to the new crusts, which indicates that the slow growth rate can promote the REE enrichment. The variations of Al₂O₃ can indicate the sedimentation of terrigenous detrital materials (Cui et al., 2012). During the growth period of the new crusts, Al content is higher, terrigenous detrital materials are more, and REE contents are lower relative to the old crusts, which indicates that terrigenous detrital materials can dilute REE when they enter cobalt-rich crusts. As above mentioned, the variation of REEs is similar to that of P₂O₅/CaO, Mn/Fe, δCe and Sr, indicating that the oxidizing environment of seawater, high marine productivity, phosphatization, and slow growth rate can promote the REE enrichment.

Figure  8.  Variation curves of element contents with depth in the growth profile of cobalt-rich crusts.

DownLoad: Full-Size Img PowerPoint

5. Conclusions

This study analyzes the mineral composition and REE contents of cobalt-rich crusts from Marcus-Wake Seamounts in the Pacific, and discuss the source, occurrence phases and enrichment mechanism of REEs.

(1) The major crystalline mineral of cobalt-rich crusts from Marcus-Wake Seamounts is vernadite, and the secondary minerals include quartz, plagioclase and carbonate fluorapatite. Also many amorphous ferric oxyhydroxides are contained in cobalt-rich crusts.

(2) Mn and Fe contents are higher relative to other elements, and REE are enriched in the cobalt-rich crusts. REE contents of the old crusts are higher than those of the new crusts. The non-phosphatized crusts have positive Ce and negative Y anomalies, and yet the phosphatized crusts have positive Ce and positive Y anomalies, indicating that cobalt-rich crusts are hydrogenetic, and REE mainly come from seawater.

(3) The occurrence phases of elements are closely related to their mineral phases. In the non-phosphatized crusts, REEs are adsorbed by colloidal particles into the crusts, about 67% of REEs are enriched in the Fe oxide phase. In contrast, the phosphatized crusts are affected by the phosphatization, REEs may combine with phosphate to form rare earth phosphate minerals, and about 64% of REEs are enriched in the residual phase containing carbonate fluorapatite.

(4) The variation of REEs is similar to that of P₂O₅/CaO, Mn/Fe, δCe and Sr, indicating that the oxidizing environment of seawater, high marine productivity, phosphatization, and slow growth rate can promote the REE enrichment.