Near-Infrared Spectroscopic Study of Secondary Minerals in the Oxidation Zones of Copper-Bearing Deposits

Abstract

This study measured the infrared spectra of secondary minerals in the oxidation zones of three types of copper ores: dioptase, malachite, and azurite, and assigned the peak positions of OH stretching vibrations and the origins of OH combination vibrations. Dioptase contains three types of water molecules with different orientations within its ring channels, which exhibit six kinds of OH stretching vibrations in the 3000–3600 cm⁻¹ range; the bond length range is 2.652 to 2.887 Å. Among them, the 3443 cm⁻¹ band shows strong near-infrared activity and combines with Si–O vibrations or OH bending vibrations in the structure, resulting in five combination vibration peaks in the 4000–5000 cm⁻¹ range. Malachite contains two inequivalent hydroxyls in its structure, leading to two OH stretching vibrations in the high-frequency region located at 3314 and 3402 cm⁻¹, respectively. Azurite contains only one type of hydroxyl, and thus only one characteristic OH stretching vibration is present at 3424 cm⁻¹. The OH stretching vibrations of malachite and azurite mainly combine with [CO₃]²⁻ vibrations or OH bending vibrations, leading to six and five combination peaks in the OH combination vibration region, respectively. By analyzing the combination of peak positions at 4341 cm⁻¹ in the near-infrared spectrum, the merged OH bending vibration at 921 cm⁻¹ in azurite was discovered. Spectroscopic research on secondary minerals can better provide a basis for ore exploration and geological remote sensing.

1. Introduction

Copper is one of the earliest metals utilized by humans and continues to play an irreplaceable role in modern human society. Copper ores come in various types, such as porphyry, skarn, and shale, among others [1]. The characteristics and scales of these different types of deposits are highly complex. Utilizing secondary copper minerals in the oxidation zone for exploration is an effective method of prospecting. Typically, in the upper oxidation zone of copper ore deposits, minerals such as cuprite, chalcopyrite, azurite, and malachite are found, which can coexist and transform into each other under certain conditions. They typically exhibit vivid colors, making them easily identifiable. They also inherit the geochemical characteristics of the original deposits [2]. These minerals are also commonly used in pigments and are frequently encountered in archaeology and material science. Those with good form and color can also be used as gemstones or mineral specimens.

Dioptase [Cu₆(Si₆O₁₈)·6H₂O] is a copper silicate hydroxide with six-membered (Si₆O₁₈) rings. It is almost exclusively restricted in its occurrence to the oxidation zones of copper-bearing deposits in arid climates. Its ring structure is composed of six silicate tetrahedra that fold into a triangular ring, with six water molecules forming a distorted version of this ring [3]. These water molecules are situated within the silicate ring structure, which is coordinated by copper atoms both laterally and longitudinally (Figure 1a) [4].

Spectroscopy is an invaluable tool for investigating the crystal structures of minerals. Changes observed in infrared and Raman spectroscopy indicate that the crystal structure of dioptase remains largely unchanged at 600 °C [4,5,6] or under pressures of 14.5 GPa [7,8]. This suggests that dioptase has a high degree of stability. The gradual loss of water molecules within channels over a broad temperature range as the temperature increases is also noteworthy. A study by Ruan [9] of infrared and Raman spectroscopy of dioptase from Namibia and the Democratic Republic of the Congo revealed that the vibrational frequencies of the ring anions increase with widening of the Si–O–Si bond angles. As a result, variations in composition can lead to minor changes in the crystal structure framework, which in turn cause shifts in the frequency and intensity of the spectral bands that reflect the crystal structure framework. This characteristic can be exploited to discern compositional differences through spectral variations, thus differentiating the environmental conditions under which the mineral formed.

Malachite [Cu₂(CO₃)(OH)₂] and azurite [Cu₃(CO₃)₂(OH)₂] are two fundamental copper carbonate minerals characterized by their chain-like structures. [CO₃] groups form these chain structures by coordinating with cations and (OH) polyhedral (as Figure 1b,c). Current spectroscopic research on malachite and azurite predominantly focuses on Raman spectroscopy [10,11,12], which allows for a detailed assignment of Raman bands to corresponding vibration groups. Xu [13] and colleagues also studied the Raman spectral changes of azurite under high pressure at room temperature, providing insights into the behavior of hydroxide under extreme conditions. However, it is difficult to distinguish secondary copper minerals from surrounding minerals, especially clay minerals, using only Raman spectroscopy. Their spectra often exhibit a high level of background noise, which can obscure Raman signals. And the peak position may shift due to the very close characteristic wavenumber [12].

Combining Raman with an infrared spectrum can solve the recognition problem more effectively. Schmidt [14] studied the infrared and Raman spectra of basic copper salts, and detailed the changes following deuterization and ¹³C substitution. Schuiskii [15] conducted an infrared spectroscopic study of natural and synthetic malachite in the range of 400–4000 cm⁻¹, meticulously assigning peaks to their respective vibration groups. The near-infrared region (NIR, generally 4000–12,500 cm⁻¹) can reflect the vibration characteristics of combination bands of hydroxyl and metal ions, as well as the overtone bands of water and some functional groups in minerals. However, there is still a relative scarcity of research on infrared, especially near-infrared, spectroscopy for these minerals.

Figure 1. (a) The silicate ring structure and internal water molecules of dioptase. Note that water molecules and the silica ring are not in the same plane (modified according to Ribbe [3]); (b) the crystal structure of malachite, in which there are two kinds of hydroxyl groups (modified according to Süsse [16]); (c) the crystal structure of azurite (modified according to Belokoneva [17]). O–H bonds in three mineral structures are marked in green.

Figure 1. (a) The silicate ring structure and internal water molecules of dioptase. Note that water molecules and the silica ring are not in the same plane (modified according to Ribbe [3]); (b) the crystal structure of malachite, in which there are two kinds of hydroxyl groups (modified according to Süsse [16]); (c) the crystal structure of azurite (modified according to Belokoneva [17]). O–H bonds in three mineral structures are marked in green.

This study investigates the mid-infrared (MIR) and near-infrared spectra of dioptase, malachite, and azurite. The findings are applicable to the identification of and research on minerals and archaeology. Although X-ray fluorescence spectrometry (XRF), X-ray powder diffraction (XRD), and other technologies can also distinguish the minerals, infrared technology may be more convenient and safer. Additionally, the correlation between mid-infrared and near-infrared spectra is an intriguing area of study. The research also provides a detailed discussion on the assignment of OH vibrational spectral peaks in the near-infrared region, attempting to rationalize the origins of these composition bands. Careful measurement and interpretation of the infrared spectra of secondary copper minerals may be crucial for mineral exploration, research, and the interpretation of geological remote sensing data.

2. Materials and Methods

Samples were sourced from the National Mineral Rock and Fossil Resource Center (NIMRF) of China (Figure 2). Dioptase (CT-dio) is a single crystal originating from the Tsumeb region in Namibia, exhibiting an idiomorphic rhombohedral form. Malachite (KO-mal) and azurite (LT-azu) are aggregates collected from the Daye area in Hubei, China. Azurite forms a sphere with columnar single crystals visible under magnification; malachite crystals are fibrous and have an overall crustaceous structure.

Isomorphism commonly occurs in minerals, and impurity elements can affect the results of spectroscopic experiments to varying degrees. Before this study, the samples were analyzed using micro-X-ray fluorescence, and the results indicated that the samples generally contained elements such as Al, Fe, Ca, Cl, etc., at generally low concentrations (

Table 1. Mineral composition of the samples.

Sample	Color	Minerals	SG	Isomorphism (w%)
CT-dio	Green	Dioptase	3.32	Mg-0.21%, Al-0.51%, Ca-0.14%, Fe-0.04%, S-0.17%, Cl-0.53%
KQ-mal	Green	Malachite	4.22	Mg-0.02%, Mn-0.01%, Ca-0.02%, Fe-0.01%, Si-0.20%, Cl-0.09%
LT-azu	Blue	Azurite	3.84	K-0.49%, Al-1.35%, Ca-0.11%, Fe-0.14%, Si-3.16%, Ti-0.06%

). The azurite sample contained higher amounts of silicon and aluminum, which may be due to the azurite sample being an aggregate with a small amount of country rock or inclusions mixed in. Before testing, the samples were pulverized and purified under a microscope to ensure the accuracy of the spectra.

Fourier-transform infrared spectroscopy (FTIR) measurements were performed at room temperature, using a Bruker Tensor II spectrometer, Germany at the laboratory of NIMRF. The samples were ground into 200 mesh powders in an agate mortar and dried at a low temperature (60–80 °C, which is not enough to affect the water molecules in silicate rings or structural hydroxyl) to remove surface-adsorbed water. The spectra were collected in transmission mode in a tablet of KBr. The mid-infrared scanning range was 400~4000 cm⁻¹, with a resolution of 4 cm⁻¹. The near-infrared scanning range was 4000~8000 cm⁻¹, with a resolution of 8 cm⁻¹. Each spectrum was averaged from 64 scans to improve the signal-to-noise ratio. If a mineral particle size is larger than or close to the infrared wavelength, the infrared light will be affected by multiple scattering within the particles, resulting in an interference effect that reduces the signal strength [18,19]. In this study, the ratio of sample to KBr was increased to enhance the signal and achieve satisfactory results.

3. Results

As shown in Figure 3, vibration of the silicon–oxygen framework is mainly in 400–1200 cm⁻¹ of dioptase, and peaks of weak hydroxyl bending vibration appear in 1400–1700 cm⁻¹. Malachite and azurite show various peaks of Cu–O stretching and [CO₃] stretching vibrations in 400–1600 cm⁻¹. The obvious difference is that malachite exhibits two types of OH stretching vibration in the high-frequency region, whereas azurite has only one.

The 3000–3600 cm⁻¹ region corresponds to the O–H stretching vibration area. Due to the superposition of multiple spectral peaks, the shape of the spectrum is relatively complex. Therefore, peak fitting was performed to determine the accurate number and position of spectral peaks, as shown in Figure 4. The peak fitting process utilized Peakfit v4.12 software, employing a combination of Gauss + Lorentz Area functions, with a correlation coefficient r² > 0.98 between the fitting function and the original function. The fitting results were verified and calibrated using the second derivative [20,21,22]. The specific method is the same as Wu [23].

Dioptase in this region exhibits six characteristic peaks, with the strongest absorption located at 3373 cm⁻¹. Frost [4] attributes all peaks to stretching vibrations of water in silicate rings. The different vibrational frequencies are due to the varying distances between the O–O and H–O bonds [24,25,26,27,28]. He tested the OH stretching vibration position and calculated the corresponding hydrogen bond distance of dioptase from different origins. In this paper, the OH stretching vibration peak position and hydrogen bond distance were fitted by data from Frost [4] (shown in Table S1 and Figure S1):

y = 6.4240 × 10⁻⁷ x² − 0.0037x + 7.9978

(1)

where x is the OH stretching vibration peak position, and y is the corresponding hydrogen bond distance (Å). The formula coefficient takes four digits after the decimal point to meet the accuracy requirements; r² = 0.941. The hydrogen bond distance corresponding to the OH stretching vibration of dioptase in this paper was calculated according to the regression equation. The bond length range is 2.652–2.887 Å. See

Table 2. The MIR bands of copper mineral samples and their assignments.

Band Assignment of CT-dio [4,5,6,9,29,30]	Band Position (cm−1)	Band Assignment of KQ-mal [10,15,31,32,33]	Band Position (cm−1)	Band Assignment of LT-azu [10,33]	Band Position (cm−1)
M–O vibration, Si–O ring bending vibration, and both coupled vibration	411 456	Lattice modes Cu–X Stretch	428
O–Si–O bending vibration	516 574 608 677	Lattice modes Cu–OH Stretch	485 524	Lattice modes Cu–OH Stretch	458 532
Si–O stretching vibration	778 886 934 955 997 1024	Lattice modes Cu–O Stretch	505 572 582	Lattice modes Cu–O Stretch	492
		Outer [CO3] stretching vibration ν4	713 750	[CO3] stretching vibration ν4	744 768
		[CO3] bending vibration	776
		[CO3] symmetric stretching vibration ν2	820	[CO3] symmetric stretching vibration ν2	817
		OH bending vibration	874 1047	OH bending vibration	921 * 953 1026
		[CO3] stretching vibration ν1	1096	[CO3] stretching vibration ν1	1089
OH (H2O) bending vibration	1425 1492 1620 1641 1681	[CO3] antisymmetric stretching vibration ν3	1391 1418 1494 1519	[CO3] antisymmetric stretching vibration ν3	837 1415 1464 1492
OH (H2O) stretching vibration (2.6522)	3018	OH (H2O) stretching vibration	3229	OH (H2O) stretching vibration	3355
OH (H2O) stretching vibration (2.6813)	3137	OH2 stretching vibration	3314	OH (H2O) stretching vibration	3387
OH (H2O) stretching vibration (2.7171)	3231	OH1 stretching vibration	3402	OH (H2O) stretching vibration	3401
OH (H2O) stretching vibration (2.7927)	3373	M–OH(no Cu) stretching vibration	3494	OH stretching vibration	3424
OH (H2O) stretching vibration (2.8395)	3443			M–OH(no Cu) stretching vibration	3481
OH (H2O) stretching vibration (2.8870)	3506

The corresponding hydrogen bond length (Å) is in brackets in column 1. * Not observed in the spectrum directly. See the discussion below for specific reasons.

for the specific distribution.

Malachite shows four peaks in this region, with the main absorptions located at 3314 and 3402 cm⁻¹. In the structure of malachite, there are two inequivalent hydroxyl groups with bond lengths of 0.92 ± 0.2 Å and 1.05 ± 0.14 Å [16]. Due to the increased bond length, the frequency of the stretching vibrations decreases, and the wavenumber correspondingly decreases. Thus, the peak at 3314 cm⁻¹ in malachite is assigned to the OH2 stretching vibration, and the peak at 3402 cm⁻¹ is assigned to the OH1 stretching vibration. The remaining two peaks located at 3229 and 3494 cm⁻¹ are attributed to the stretching vibrations of adsorbed water and the M–OH stretching vibrations caused by isomorphous substitution, as these two positions do not always exhibit characteristic peaks in other studies, or the shift and intensity of the absorption peaks near these positions vary significantly with different samples. Azurite contains only one type of hydroxyl group [17], and, therefore, the peak at 3424 cm⁻¹ is assigned to the OH stretching vibration, with the remaining peaks also considered to be stretching vibrations of adsorbed water or M–OH. The intensity of the 3481 cm⁻¹ peak is extremely weak and does not always exist in studies of azurite. It was assigned to metal (not Cu)–OH stretching vibration due to trace impurity elements. The specific assignments of the peak positions are shown in Table 2.

As depicted in Figure 5a, the near-infrared spectra of the three minerals primarily exhibit absorption bands in the 4000–5200 cm⁻¹ region, which are attributed to the combination of hydroxyl stretching vibrations and other vibrational absorptions. The spectral characteristics in the near-infrared region are not pronounced, with only azurite showing two discernible characteristic infrared peaks at 4241 and 4372 cm⁻¹. To further identify peak positions, the spectra were subjected to baseline correction and normalization, as illustrated in Figure 5b. Both malachite and azurite display two main absorption bands, within the ranges of 4000–4300 and 4300–4500 cm⁻¹, respectively. The characteristic bands of dioptase are considerably broader compared to the former two, with two main characteristic bands located in the ranges of 4000–4500 and 4500–5200 cm⁻¹. The fitting results present more detailed outcomes (as shown in Figure 6), where the three minerals consist of 5–7 bands in the respective regions, with specific peak positions detailed in

Table 3. The NIR bands and their corresponding MIR peaks.

Sample	Measured Peak Position (cm−1)	Fundamental Peaks Position (cm−1)	Assignments	Theoretical Peak Position (cm−1)	Error (cm−1)
CT-dio	4125	3443 + 677	OH stretching vibration + O–Si–O bending vibration	4120	5
	4211	3443 + 778	OH stretching vibration + Si–O stretching vibration	4221	−10
	4325	3443 + 886	OH stretching vibration + Si–O stretching vibration	4329	−4
	4700	3018 + 1681	OH stretching vibration + OH bending vibration	4699	1
	4810	3137 + 1681	OH stretching vibration + OH bending vibration	4818	−8
	4940	3443 + 1492	OH stretching vibration + OH bending vibration	4935	5
	5083	3443 + 1641	OH stretching vibration + OH bending vibration	5084	−1
KQ-mal	4090	3314 + 776	OH2 stretching vibration + [CO3] bending vibration	4090	0
	4130	3314 + 820	OH2 stretching vibration + [CO3] symmetric stretching vibration ν2	4134	−4
	4168	3402 + 776	OH1 stretching vibration + [CO3] bending vibration	4178	−10
	4232	3402 + 820	OH1 stretching vibration + [CO3] symmetric stretching vibration ν2	4222	10
	4400	3314 + 1096	OH2 stretching vibration + [CO3] stretching vibration ν1	4410	−10
	4508	3402 + 1096	OH1 stretching vibration + [CO3] stretching vibration ν1	4498	10
LT-azu	4192	3424 + 768	OH stretching vibration + [CO3] stretching vibration	4192	0
	4241	3424 + 817	OH stretching vibration + [CO3] symmetric stretching vibration	4241	0
	4261	3424 + 837	OH stretching vibration + [CO3] antisymmetric stretching vibration	4261	0
	4341	3424 + 921 *	OH stretching vibration + OH bending vibration	4245	−4
	4373	3424 + 953	OH stretching vibration + OH bending vibration	4377	−4

* Not observed in the spectrum directly. See the analysis for specific reasons.

. Malachite and dioptase exhibit a multitude of weak absorption peaks in the 7000–8000 cm⁻¹ range, which are attributed to the overtone peaks of hydroxyl or water molecule stretching vibrations associated with vibration groups [34,35].

4. Discussion

Assignment of OH Vibration in NIR Spectra

Both this study and previous works show that not all peaks participate in the combination of near-infrared regions (or the overtone vibration of higher wavenumbers). This may be because the combination is the superposition or cancellation of electric dipoles. By trial calculations of this experiment, we identified the most suitable near-infrared peak assignments, with specific combinations detailed in Table 3, and the error is controlled within 10 cm⁻¹ [23,36]. The signals of dioptase in the 4000–5200 cm⁻¹ region correspond to a combination of OH stretching vibrations with vibrations of the silicon–oxygen framework or OH bending vibrations. The most significant combination with the vibrations of the silicon–oxygen framework and bending vibrations of channel water molecules is the band at 3443 cm⁻¹, rather than the strongest peak at 3373 cm⁻¹. It is speculated that the orientation of the water molecule stretching vibration corresponding to 3443 cm⁻¹ is significantly different from other water molecule stretching vibrations [3], leading to a propensity to form a complex with other bands.

Malachite exhibits six bands in the 4000–4600 cm⁻¹ region, which belong to a combination of two types of OH stretching vibrations and three [CO₃]²⁻ vibrations, with water molecules and the stretching vibrations of M–OH from non-copper ions not participating in the combination vibrations. Among the four bands with higher wavenumbers, the errors between the theoretical and measured values can offset each other, which is considered to be due to the relative shift of the bands caused by mutual influence between the vibrations. In the crystal structure of azurite, only one type of hydroxyl group exists, and all the combination vibrations in the near-infrared region are related to the OH stretching vibration at the position of 3424 cm⁻¹. In addition to the [CO₃]²⁻ vibrations, combinations of hydroxyl stretching and bending vibrations are also observed.

It is worth noting that, based on the calculation of the OH stretching vibration at 3424 cm⁻¹ and the combination vibration at 4341 cm⁻¹, there should be an absorption vibration near 917 cm⁻¹. Upon detailed analysis of the nearby spectrum, a corresponding spectrum at 921 cm⁻¹ was observed, which the author attributes to the OH bending vibration, as shown in Figure 7, and which has both infrared and Raman activity [6]. This spectral band has hardly been mentioned in previous studies.

5. Conclusions

Dioptase exhibits six vibrational bands in the OH stretching vibration region, which are attributed to the vibrations of hydroxyl groups of water molecules with different orientations within the ring silicate structures of the channels. The combination of OH stretching vibrations and Si–O stretching vibrations with the bending vibrations of water molecules results in seven characteristic combination vibration peaks in the near-infrared region of 4000–5200 cm⁻¹. Both malachite and azurite are carbonate minerals. The infrared OH stretching vibration region, ranging between 3000–3500 cm⁻¹, features three types of characteristic peaks: water molecule OH stretching vibrations, structural OH stretching vibrations, and other structural OH stretching vibrations resulting from isomorphous substitution. The near-infrared characteristics appear within the ranges of 4000–4300 and 4300–4500 cm⁻¹, arising from the respective combinations of structural OH stretching vibrations and [CO₃]²⁻ stretching vibrations with OH bending vibrations.

Through the tracing of near-infrared peak positions, a new characteristic peak was identified at 921 cm⁻¹ in azurite. Due to its weak intensity and proximity to nearby spectral bands, it has rarely been mentioned before. The combined study of near-infrared and mid-infrared spectroscopy can be used to identify peaks that are difficult to observe in MIR spectroscopy.