X-ray Photoelectron Spectroscopy (XPS) Study of Layered Double Hydroxides with Different Exchangeable Anions

1. Introduction

Hydrotalcites, also referred to as layered double hydroxides (LDHs), are categorized as anionic clays due to their distinctive layered architecture [1,2]. This structure is characterized by positively charged hydroxy layers, in stark contrast to cationic (silicate) clays such as smectites, which possess a negative layer charge. Within the interlayer space of these positively charged layers, anions are present to balance the overall charge. A more comprehensive understanding of hydrotalcite’s structure reveals positively charged hydroxide layers that resemble those found in brucite [Mg(OH)₂]. In this configuration, a portion of the Mg²⁺ ions can be substituted with trivalent metals, such as Al³⁺, while charge-compensating anions, typically in a hydrated form, occupy the interstitial regions between the hydroxide layers [2,3].

The composition of LDH is very versatile, commonly represented by the formula [M²⁺_1-xM³⁺_x(OH)₂][A^n–]_x/n·yH₂O, where M²⁺ and M³⁺ denote the divalent and trivalent cations residing in octahedral coordination within the hydroxide layers. The variable x usually ranges from 0.17 to 0.33, indicating the proportion of M³⁺ ions. Notably, it is imperative that the ionic radii of M²⁺ and M³⁺ do not exhibit significant disparity—specifically, a deviation of no more than 30%—to maintain structural integrity. For instance, the ionic radii for Mg²⁺ and Al³⁺ are approximately 0.65 Å and 0.45 Å, respectively. Furthermore, sound chemical principles dictate that the solubility products, S₁ for M²⁺(OH)₂ and S₂ for M²⁺CO₃, must adhere to the conditions where the difference pS₁ – pS₂ is maintained within the range of 0 and 10, ensuring the stability of the hydrotalcite phase [2,3,4].

The variety of anions that can exist within the LDH framework is relatively unconstrained, provided that these anions do not engage in complexation with the cations within the octahedral sheets during the synthesis process. Such complexation would inhibit the formation of a viable LDH structure [5]. The inherent flexibility in both cationic and anionic compositions of hydrotalcites enables the engineering of tailor-made materials suitable for a broad array of applications. These may include their utilization as basic catalysts or precursors for synthesizing mixed metal oxide catalysts. Additionally, hydrotalcites serve functional roles as adsorbents, fillers, UV stabilization agents, chloride scavengers, and thermal stabilizers, positioning them as versatile components in various industrial and technological contexts [6,7,8,9,10,11].

Infrared spectroscopy, along with the less frequently employed Raman spectroscopy, has proven instrumental in elucidating the structural and compositional characteristics of LDH containing varied cations and anions. These spectroscopic techniques have facilitated comprehensive examinations of anionic intercalation and thermal degradation processes in these materials, primarily serving to identify a range of exchangeable anions. Among the anions detected in such studies are carbonate (CO₃^2–), chloride (Cl^–), perchlorate (ClO₄^–), nitrate (NO₃^–), sulfate (SO₄^2–), arsenate (AsO₄^3-), vanadate (VO₄^3-), molybdate (MoO₄^2-) and chromate (CrO₄^2–) [12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27]. Moreover, the investigation has extended to more complex anionic entities, including anionic silica (SiO(OH)₃^–) [28] and larger polyoxometalate ions as well as hexacyanoferrate complexes (Fe(CN)₆^n–) [see e.g., [29,30,31,32,33,34,35]].

The unique vibrational fingerprints associated with each of these anions are discernible through their respective infrared and Raman spectra. These spectral signatures provide crucial insights into the interactions and substitutions occurring within the hydrotalcite matrix. Through rigorous analysis using these spectroscopic techniques, researchers have been able to correlate the presence of specific anions with structural modifications within the LDH framework. This knowledge not only enhances the understanding of the properties of LDH but also informs potential applications in fields such as catalysis, ion exchange, and environmental remediation, where the selective uptake and release of ions play a critical role. Such advancements underscore the relevance of infrared and Raman spectroscopy in characterizing the multifaceted nature of hydrotalcites and their layered structures.

X-ray photoelectron spectroscopy (XPS) has established itself as the predominant technique for investigating surface phenomena. However, there remains a notable paucity of research focusing on the bulk atomic structure and chemical states of minerals. This oversight is particularly striking given that the vast majority—often exceeding 90%—of the intensity of XPS lines is derived from the bulk material when employing conventional laboratory X-ray sources, such as the Al Kα source at 1486.6 eV. This discrepancy invites further inquiry into the collective properties of minerals that are not fully captured by surface-centric methodologies.

Historically, the XPS analysis of various mineralogical components has been documented in the literature, encompassing a diverse range of materials including feldspars, clays, aluminum oxy(hydroxide) phases, arsenates, and phosphates. Notable contributions have highlighted the unique electronic and structural characteristics present in these minerals, but comprehensive studies that simultaneously account for both surface and bulk properties remain limited. It is essential to broaden the scope of investigation to include these often-ignored bulk properties to achieve a better understanding of mineral behavior. So far, limited use has been made of XPS as an analytical tool to study LDH in detail and generally stops at just providing binding energy values [see e.g., [36,37,38,39,40,41,42]]. As part of an ongoing study on LDH this study aims to better understand the X-ray photoelectron spectroscopy (XPS) of the interlayer anions in LDH to complement the existing Infrared and Raman spectral data of LDH.

2. Materials and Methods

The synthesis of hydrotalcite, characterized by the theoretical composition Mg₆Al₂(OH)₁₆CO₃.nH₂O, was performed utilizing the procedure outlined by Kloprogge and Frost [43]. The procedure involves the gradual simultaneous introduction of a solution containing aluminum nitrate (0.25 M) and magnesium nitrate (0.75 M) alongside a mixed solution of NaOH (2.00 M) and Na2CO3 (0.125 M), while vigorous stirring. The pH is maintained at approximately 10 throughout the process. The product was washed to eliminate excess salt and dried at 60 °C.

To minimize the incorporation of carbonate during the synthesis of hydrotalcites with other anions, several measures were employed: deionized water was boiled prior to use, NaOH pellets were thoroughly rinsed, and the synthesis was conducted in a nitrogen atmosphere. A mixed solution containing aluminum and magnesium nitrates was prepared with concentrations of [Al³⁺] at 0.25 M and [Mg²⁺] at 0.75 M. Additionally, a separate mixed solution of sodium hydroxide was formulated with a hydroxide concentration of [OH⁻] at 2 M, incorporating the desired anion at the specified concentration. Both solutions were transferred into distinct vessels and subjected to nitrogen purging for 20 minutes, with all solutes dissolved in freshly decarbonated water. The cationic solution was delivered to the anionic phase using a peristaltic pump at a flow rate of 40 mL/min, while maintaining the pH above 9. The resulting mixture was subjected to aging at 75 °C for 18 hours in a nitrogen atmosphere. Following this, the precipitate was filtered using room temperature decarbonated water to eliminate nitrate ions and subsequently dried in a vacuum desiccator over several days. This method facilitated the synthesis of hydrotalcites containing various anions in the interlayer. Phase composition was evaluated via X-ray diffraction, and chemical composition was determined through energy-dispersive X-ray spectroscopy (EDXA) analyses [44].

X-ray photoelectron spectroscopy (XPS) data were acquired on a Kratos AXIS Ultra (Kratos Analytical, Manchester, U.K.) utilizing a monochromatic 225 W Al X-ray source and conducted under ultrahigh vacuum conditions. Initial survey scans were executed within the energy range of 0 to 1200 eV, employing a dwell time of 100 milliseconds, a pass energy of 160 eV, and a step size of 1 eV over a single sweep. Subsequent high-resolution analyses were performed by increasing the number of sweeps, reducing the pass energy to 20 eV with a step increment of 100 meV, and extending the dwell time to 250 milliseconds. For calibration purposes, the adventitious C 1s peak at 284.8 eV was utilized to correct all spectra.

3. Results

3.1. Layer Structure

Figure 1 shows the survey scan for a normal LDH with carbonate as the interlayer anion. It is characterized by a single Mg 2p peak at 50.1 eV, a single Al 2p peak at 74.5 eV (Figure 2), as well as a single O 1s peak at 531.6 eV. High resolution scans of the O 1s exhibits three peaks at 530.5 eV, 531.6 eV and 532.4 eV assigned to oxygen atoms in the interlayer carbonate anion-, hydroxyl groups and interlayer water, respectively. These values are in excellent agreement with the results reported by Kloprogge and Wood [45]. They also showed the Mg 2p for Mg(OH)₂ at 49.3 eV, which is at a slightly lower binding energy (BE) than for LDH as shown here, caused by the partial substitution of Mg²⁺ by Al³⁺. The LDH Al 2p peak at 74.5 eV is close to the BE observed for gibbsite, Al(OH)₃ (74.4 eV), bayerite, Al(OH)₃ (75.0 eV), boehmite (73.9 eV), AlOOH, and pseudoboehmite, AlOOH (74.3 eV) [46]. Peng, et al. [47] observed comparable BE values for LDH with Mg 2p at 50.29 eV, Al 2p at 74.31 eV.

3.2. Carbonate Anion

The most often studied interlayer anionic group in LDH, due to its strong affinity to the hydroxide layers, is CO₃^2–. Going from the free anion in solution changes can be anticipated when the CO₃^2– is present in the interlayer space of LDH due to interactions with interlayer H₂O and/or OH groups at the interlayer surfaces of the LDH layers.

The C 1s high resolution spectrum of the carbonate anion shows a peak at a BE value of 288.6 eV (Figure 3 left). This peak is due to the interlayer anion. The other peaks are due to surface adsorbed organic matter, the socalled “rubbish” carbon.The corresponding O 1 peak was found at 530.5 eV (Figure 3 right) and the ratio of O/C is about 3.10, close to the expected ratio of 3 for CO₃^2-. The BE value is slighlty lower compared to pure carbonate minerals such as magnesite which has a C 1s at 290.0 eV, dolomite 289.8 eV, and calcite 289.5 eV [45]. This shift in BE is probaly associated with interactions with the interlayer water molecules and the restrictions imposed by the LDH layers.

In earlier work Kloprogge, Wharton, Hickley and Frost [19] have shown by infrared and Raman spectroscopy that there is an interaction with interlayer water as shown by a water-carbonate bridging mode around 3000-3100 cm^-1. Rey, et al. [48] observed that heating Mg/Al-LDH to 100 °C resulted in the disappearance of this water-carbonate bridging mode together with the water OH-bending mode around 1616 cm^-1 due to dehydration. Relative to free CO₃^2- a shift toward lower wavenumbers was detected. Based on X-ray diffraction, ionic size, and charge density, it has been generally accepted that the CO₃^2- anion is present within the interlayer space with its C₃ axis perpendicular to the interlayer surfaces of the LDH hydroxide layers [49].Due to these restrictions in the interlayer space the symmetry of the carbonate anion decreased from D_3h to C_2s causing the activation of the IR inactive ν₁ mode around 1050–1060 cm^-1. In addition, the ν₃ showed a splitting of 30–60 cm^-1.Though IR and Raman clearly show that there are interactions between the interlayer anion and the LDH layers as well as water and that the movement of the interlayer anion is restricted, XPS shows that the binding energy of the C 1s is less than for a pure carbonate mineral indicative that the carbonate anion still has some degree of freedom inside the interlayer space.

3.3. Phosphate Anion

The interlayer phosphate anion is characterized by the P 2p3/1 peak with a BE of 133.0 eV. The corresponding P 2p1/2 is found at a BE value of 132.1 eV (Figure 4 left). The synthesis of the phosphate containing LDH was performed at 3 different pH values (9.3, 11.9 and 12.5). No difference in the BE values was observed but a significant higher amount of phosphate was present in the interlayer space of the LDH in th lowest pH sample. The BE value is similar to some common phosphate minerals, such as monazite, autunite, vivianite and apatite, which have P 2p3/2 Be values around 133.2 to 134 eV [50]. On the other hand the BE of H₃PO₄ is slightly higher than observed here at 135.2 eV. The same is true for NaH₂PO₄ with a BE of 134.0 to 134.2 eV [51]. Wang, Cai, Han, Fang, Chen and Tan [22] reported the P 2p at approximately 134 eV for LDH containing hexametaphosphate. At the same time they observed a shift in the Mg 2p and Al 2p towards higher BE. They ascribed this shift in BE to the fact that the Mg²⁺/Al³⁺ in the LDH hydroxide layers can extract electron density from the negatively charged oxygen atoms in hexametaphosphate [52]. Gupta, Saifuddin, Kim and Kim [52] found a BE shift in the Zn 2p of 0.04 eV which is considered to be within the experimental error of XPS, while the Fe 2p showed a larger shift in BE of about 0.2 eV. In this study no shift was observed for the Mg 2p, and a minor shift toward a higher BE value of 0.1 eV for the Al 2p.

The O 1s is characterized by three peaks associated with the layer hydroxyl groups at 530.9 eV, interlayer water at 532.4 eV and the interlayer PO₄^2- at 531.6 eV (Figure 4 right). Again the BE value for the oxygen in the phosphate anion is comparable to that found for various phosphate minerals, e.g., monazite (531.2 eV), vivianite (531.2 eV), amblygonite (531.6 ev) and apatite (531.0 eV) [50]. Similar BE values are also listed in the NIST database, with e.g., AlPO₄ at 532.8 eV [51]. Wang, Cai, Han, Fang, Chen and Tan [22] only reported a general position for the O 1s at about 532 eV, but did not attempt to provide details fitting of the different oxygen species present. Gupta, Saifuddin, Kim and Kim [52] fitted the O 1s with three bands with the O 1s OH at 530.1 eV, slightly lower than observed here. In contrast though, the other two peaks are at very different BE values with an extremely broad peak at 530.9 eV assigned to H₂O, -O-Nand O-C, while a third peak at very low BE of 528.4 eV assigned to M-O. No explanation is given for what M is but presumebaly this is the P-O in the phosphate as there are no M-O modes in the LDH hydroxyde layer. These assignments don’t seem to be correct as generally O 1s for PO₄ is always found at much higher BE.

LDHs intercalated with PO₄^3- have been shown to form different phases with very different basal spacings: 0.84 nm [20], and 0.78 nm, 0.80 nm, 1.19 nm [17] and 1.11 nm [53], depending on the synthesis conditions such as Mg/Al ratio, pH, etc.. This spread in basal spacings can be explained by different positions adopted by the PO₄^3- between the LDH hydroxyde layers (perpendicular, inclined and planar). Wang, Cai, Han, Fang, Chen and Tan [22] even indicate that the hexametaphosphate complex stays intact within the interlayer space of the LDH after exchange. Mid-infrared and Raman spectra have been reported in a limited number of papers. Benício, et al. [54] reported the IR-active bands in the 1050 cm^-1, 870 cm^-1 and 550 cm^-1 regions, which correspond to ν₃, ν_1, and ν₄ vibrations [55,56]. The PO₄^3- anion is tetrahedral with T_d symmetry, thus there are four normal modes of vibration, all of which are Raman active. However only the triply degenerate ν₃(F₂) and ν₄(F₂) modes are infrared active. The observation of the ν₁ mode must be due to either the presence of some nitrate or carbonate instead [19]. In contrast, Benício, Constantino, Pinto, Vergütz, Tronto and Da Costa [53] in a paper a year earlier claim the presence of not only PO₄^3- in the interlayer space but also HPO₄^2-. Though the spectrum provided is of rather low quality and they don’t show the band fit, they claim that the broad band in the 1020 cm^-1 region, consists of the ν₃ vibrational mode of PO₄³⁻ anion and two additional bands assigned to the ν₃ and ν₂ modes of HPO₄^2- anion [57]. In contrast, Gupta, Saifuddin, Kim and Kim [52] assigned the bands at 1045 cm^-1 and 565 cm^-1 corresponding to the v₃ and v₄ band vibrations of HPO₄²⁻ or H₂PO₄⁻ without further explanation. Wang, Cai, Han, Fang, Chen and Tan [22] reported only a single band at 1088 cm^-1 attributed to the vibration of P-O. Shabanian, et al. [58] reported that three bands attributedto HPO₄^2- coalesce into a single broad band around 1056 cm^-1. In addition, a shoulder band was detected at 870 cm^-1 near the theoretical antisymmetric stretching mode of P–OH. Others also described the characteristic band related to phosphate in the interlayer of Zn/Al LDH at 1040 cm^-1 [59] and 1048 cm^-1 [60].

In earlier work we reported on the Raman spectra of phosphate-LDH and addressed the possibility of determining the presence of H₂PO₄^- and HPO₄^2- in addition to PO₄^3- [17]. The Raman spectra will change if hydrogen is linked to the phosphate unit as the symmetry will change from T_d to C_3v to C_2v. This will cause a loss of degeneracy and other bands may become visible. The sample synthesised at pH 12.5 exhibited an strong sharp band at 960 cm^-1 with a broader band at 1026 cm^-1 assigned to the PO₄^3- symmetric and antisymmetric stretching vibrational modes identical to that of the PO₄^3- ion. Similar bands were observed for the pH 11.9 sample at 957 cm^-1 and 1032 cm^-1. Benício, Eulálio, Guimarães, Pinto, Costa and Tronto [54] reported the formation of bands in the region between 940 cm-1 and 960 cm-1 after exchange with phosphate. Slight changes were also deted in the band positions in the 1050 cm^-1 region. A different Raman spectrum was obtained for the pH 9.3 sample with four bands at 964, 989, 1033 and 1138 cm^-1. In aqueous systems the PO₄^3- ion has a strong band at 936 cm^-1, while HPO₄^2- anion exhibits a strong band at 990 cm^-1. Therefore the spectrum can be interpreted as being a combination of the PO₄^3- anion (964 and 1033 cm^-1) and the HPO₄^2- ion (989 and 1138 cm^-1). However it is questionable whether these are actually present in the interlayer space of LDH as XRD showed the material mainly as amorphous. For the oxyanion PO₄^3- the symmetric bending mode (ν₂) and the ν₄ mode are observed at 420 cm^-1 and 567 cm^-1, respectively. For the phosphate LDH at pH 12.5 a band at 474 cm^-1 was observed. A similar band was found at, 472 cm^-1 for the pH 11.9 sample, while a broad band around 441 cm^-1 was present in the Raman spectrum of the pH 9.3 sample.

Infrared and Raman spectroscopy of phosphate-LDH has shown that the nature of the anion present in the interlayer space is affected by the synthesis conditions. In this study three different pH values were studied. Despite the fact that in the literature the presence of H₂PO₄^- and HPO₄^2- have been claimed [52,53]. For the two samples synthesised at pH 12.5 and 11.9 this was shown not to be the case by Raman spectroscopy [17]. This is further confirmed by the XPS results that show only a single P 2p 3/2 peak as well as by the O/P ratio which in both cases is close to 4. XPS did not allow to distinguish between PO₄^3- and HPO₄^2- in the three samples studied.

3.4. Sulphate Anion

In naturally occurring LDH, other than the commonly present CO₃^2- anion, SO₄^2- forms an additional important interlayer anion found in minerals such as honessite (Ni₆Fe₂SO₄(OH)₁₆·4H₂O), hydrohonessite (Ni₆Fe₂SO₄(OH)₁₆·7H₂O), and carrboydite ((Ni,Cu)₁₄Al₉(SO₄,CO₃)₆(OH)₄₃·7H₂O) [61,62,63]. Figure 5 shows the S 2p and O 1s of sulphate-LDH. The S 2p is characterized by a single S 2p3/1 peak with a BE of 168.3 eV. This BE is slighly lower than observed for common sulphate minerals such as baryte (169.2 eV), anhydrite (168.9 eV), or gypsum (169.2 eV) [45], indicating that the sulphate in the LDH interlayer space is less strongly interacting with the LDH hydroxide layers compared to a sulphate group in a rigid crystal structure. In the O 1s a single peak at 532.4 eV is associated not only with the oxygen in the sulphate group but also with oxygen in interlayer water that has a similar BE. This value is comparable with H₂SO₄ which has an O 1s BE of 532.5 eV and Al₂(SO₄)₃ of 532.4 eV [51]. Similar results were also reported for group 1A sulphates by Wahlqvist and Shchukarev [64]

Bish and Livingstone [61] described for honessite the SO₄^2- ν₁, ν₂, ν₃, and ν₄ infrared active modes at 980, 500, 1140, and 650 cm^-1, respectively. The ν₃ mode is obviously split though no separate band positions were provided. Since all four modes are infrared active the symmetry of the SO₄^2- anion was lowered from T_d (free anion) to either C₃ or C_3v, causing the activation of the two infrared inactive modes together with the splitting of the ν₃ mode [65]. Earlier work in our group has shown that the infrared spectrum of Mg/Al-LDH containing SO₄^2- in the the interlayer space exhibits a strong but broad ν₃ mode at 1126 cm^-1, a ν₄ mode at 614 cm^-1, and a very weak ν₁ mode at 981 cm^-1 [19]. The ν₂ mode was not observed as a distinct band, in contrast to honessite and hydrohonnesite, due to its overlap with the O-M-O bending vibration of the LDH hydroxide layer around 450 cm⁻¹ [66]. Fahami and Beall [67] reported only a band at a significantly lower wavenumber position around 1066 cm^-1 were attributed to the bending mode of interlayer sulfate ions. Liu and Yang [68] described a single band for Zn/Al-LDH at 1113 cm-1 which is much closer to the value observed by Kloprogge, Wharton, Hickley and Frost [19]. Frost, et al. [69] described for glaucocerinite (Zn,Cu)₁₀Al₆(SO₄)₃(OH)₃₂∙18H₂O, a naturally occurring LDH mineral, three strong bands at 1053, 1078 and 1109 cm^-1 which was attributed to the SO₄^2- ν₃ antisymmetric stretching mode. A weak band at 986 cm^-1 was assigned to the correcsponding ν₁ symmetric stretching mode. They also reported that carrboydite exhibted three similar bands at 1088, 1021 and 978 cm^-1 with the first two bands due to ν₃ antisymmetric stretching modes and the last band to the ν₁ symmetric stretching mode.

Kloprogge, Wharton, Hickley and Frost [19] showed that the Raman spectrum of sulphate-LDH consisted of a very intense ν₁ mode at 982 cm^-1 together with medium intensity ν₂ and ν₄ modes around 453 and 611 cm^-1, respectively. The ν₃ mode was not observed as a separate band, even though a broad bandwas visible around 1134 cm^-1. They indicaed that when SO₄^2- is located in the interlayer space of LDH the infrared ν₂ mode may become inactive, while at the same time the ν₁ mode becomes active. In contrast to the other modes the ν₃ is considerably broader indicating the possible existence of two overlapping bands due to splitting. Dutta and Puri [70] proposed a D₂ site symmetry for the sulphate anion. This is not compatible with the infrared spectrum where all four modes are present. For analogous reasons the C₃ site symmetry as propsed by Bish [71] is not compatible with the Raman spectrum. Hence, using the combined observations in both the infrared and Raman spectra Kloprogge, Wharton, Hickley and Frost [19] conclued that the site symmetry is most probably C_2v with ν₁ infrared and Raman active, ν₂ infrared and Raman active, and ν₃ and ν₄ infrared and Raman active. A similar conclusion was reached by Lin, et al. [72] for Ni_xZn_6−xAl₂(OH)₁₆(SO₄)·4H₂O, while Frost, Theiss, López and Scholz [69] concluded for glaucocerinite a reduction in symmetry of the sulphate anion from T_d to C_2v or even lower symmetry.

It is clear from the infrared and Raman data that the sulphate anion when present in the interlayer space is restricted in its movement and has a lower site symmetry due to the interactions with the LDH hydroxide layers. This is also reflected in the XPS data where the S 2p clearly shows a slightly lower BE compared to structural sulphate in mineral structures. This confirms that the sulphate anion is no longer a completely free anion, but it is to a certain extent interacting with the hydroxide layers.

3.5. Molybdate Anion

The MoO4-LDH is characterized by a single peak for the Mo 3d5/2 at a BE of 229.6 eV (Figure 6 left). This BE value is lower than e.g molybdenite, MoS₂, with a BE of 230.1 eV or wulfenite, PbMoO₄, with a BE of 231.9 eV [3]. Several molybdates, such as Al₂(MoO₄)₃ (232.5 eV), CaMoO₄ (232.8 eV), NiMoO₄ (233.0), etc. all have higher BE than observed for intercalated molybdate ions in LDH. This is similar to what has been observed before, an indication that the interactions between the molybdate anion and th LDH hydroxide layers are not as strong as when the molybdate is part of a rigid crystal structure. The corresponding O 1s peak is observed at 530.3 eV (Figure 6 right), which compares well with the value observed for wulfenite at 530.0 eV. The NIST database reports for H₂MoO₄ a BE of 530.7 eV, CaMoO₄ at 530,6 eV, NiMoO₄ at 530.9 eV, while for Al₂(MoO₄) the BE ranges from 530.7 to 531.0 eV [51]. Thao, et al. [73] assigned the Mo 3d5/2 peak at 232.6 eV to Mo(VI) in MoO₄^2-. The hardly visible shoulder at 231.7 eV was attributed to Mo(V) species possibly caused by a charge transfer between (Mo⁶⁺ + O^2-) and (Mo⁵⁺ + O^-) [74,75]. This shoulder was not observed in this study. Thao, Trung and Van Long [73] also reported the BE value of O 1s at 532.4 eV for the fresh molybdate-LDH assigned to the O^2- [35,38,45]. They concluded that it is probably composed of two overlapping photoelectron peaks at 532.4 and 531.5 eV. The peak at 532.4 eV was assigned to the metal hydroxides in the LDH layers, while the other peak at 531.6 eV was attributed to oxygen O^- in the oxomolybdenum. Both the Mo 3d5/2 and O 1s peaks for the MO42- anion are at slighly higher BE values than observed in this study and in comparison with other molybdate structures such as listed in Kloprogge and Wood [45] and the NIST database [51].

Previously, infrared and Raman spectroscopy on the same sample studied here have shown that molybdate anions are not polymerised in the interlayer space of the LDH because of the synthesis under alkaline conditions. The splitting of the vibrational modes of the molybdate anion in the vibrational spectra suggests a symmetry lowering due to interactions of the molybdate anion with the LDH hydroxide surfaces and interlayer water [16]. Palmer, Soisonard and Frost [21] reported a sharp Raman band at around 900 cm^-1 which was assigned to the MoO₄²⁻ symmetric stretching mode. A single broad IR band at approximately 830 cm^-1 was reported as characteristic for the MoO₄^2- anion, by Klemkaitė-Ramanauskė, et al. [76]. Mitchell and Wass [77] confirmed the finding of Adebajo, Musumeci, Kloprogge, Frost and Martens [16] that at pH above 7–8, the prevalent molybdate species was the MoO₄²⁻ ion characterised by an IR band at ca. 830 cm^-1, while LDH prepared at pH 4.5 and with high molybdate loadings had an IR band at ca. 920–930 cm^-1 and bands or shoulders near 890 cm^-1 suggesting the presence of polymolybdate. Nejati, et al. [78] reported a single band in the IR spectrum at 806 cm^-1 assigned to the antisymmetric stretching mode of Mo–O. Yu, et al. [79] observed a similar IR band attritubed to the antisymmetric mode of Mo–O–Mo in MoO₄²⁻ at 834 cm^-1. Their Raman spectrum showed the Mo–O symmetrical stretching mode in MoO₄²⁻ at 910 cm^-1, while two bands at 320 and 220 cm^-1 were attributed to the Mo=O bending and Mo–O–Mo deformation mode, respectively. A tetrahedral ion like MoO₄^2- has four vibrational modes if it has full Td symmetry. These consist of ν₁(A₁), ν₂(E), ν₃, and ν₄(F₂). The A₁ symmetric stretching mode and the E bending mode are Raman active only, while the F₂ stretching and bending modes are both IR and Raman active. For MoO₄^2- the fundamental modes are found at 894 cm^-1 (ν₁), 381 cm^-1 (ν₂), 833 cm^-1 (ν₃), and 318 cm^-1 (ν₄) [80]. Based on these fundamental modes it seems incorrect to attribute the IR bands at 823 cm^-1 and 635 cm^-1 to anti-symmetric and symmetric stretching vibrations of Mo-O bonds as suggested by Colombo, et al. [81]. The most complete description has probably been provided by Thao, Trung and Van Long [73]. In the IR spectrum they observed a broad band at 920 cm-1 attributed to the vibrations of Mo=O in polymolybdate Mo₇O₂₄^6-. A band at 670 cm^-1 with a shoulder at 856 cm^-1 was assigned to the Mo–O–Mo stretching mode of MoO42- in the interlayer space of LDH. It is more likely that the 670 cm^-1 band is associated with NO₃^{- -} in the system which is supported by the band at 1370 cm^-1. In the Raman spectra, the MoO₄^2- symmetric stretching modes in the LDH were observed at 908 and 892 cm^-1. They were interpreted as being associated with two different MoO₄^2- anionic species, the first one is hydrated and the other one is bonded to the brucite-like hydroxide surface of the LDH within the interlayer space [18]. A broad shoulder at 823 cm^-1 was ascribed to the MoO₄^2- antisymmetric stretching mode, while a band at 325 cm^-1 was attributed to the Mo-O bending mode [79].

Infrared and Raman spectroscopy have shown that nature of the molybdate anion can change depending on the synthesis conditions, in particular the pH. Under alkaline condition the MoO₄^2- anion is dominant, either in a hydrated state or forming a stronger interaction with the hydroxide layer of the LDH within the interlayer space as shown by the slight shift in band positions. Such small changes can not be observed in the XPS spectrum, which only shows that the Mo 3d5/2 peak is found at a slightly lower BE than for molybdate in a rigid crystal structure, indicating that, although there is some interaction with the LDH layer structure, the interaction is not as strong as in molybdate minerals such as wulfenite or other crystalline molybdate compounds [45,51]. In addition, no proof was found for the existence of charge transfer between (Mo⁶⁺ + O^2-) and (Mo⁵⁺ + O^-) as suggested in some papers [74,75].

3.6. Chromate Anion

The Cr 2p high resolution spectrum of the chromate intercalated LDH differs from the other anionic groups shown earlier in this paper in the sense that not one but two Cr 2p3/2 peaks are present (Figure 7 left). The first is a rather broad peak with a BE of 579.0 eV, while the second is much sharper and has a BE of 580.2 eV. The second peak is characteristic for Cr⁶⁺ in the CrO₄^2- anion, while the first peak at lower BE is associated with the presence of Cr³⁺, indicative of substantial reduction of the CrO₄ with time. Approximately 60 % of all Cr has been reduced based on the peak surface area ratio which is directly proportional to the amounts present. The mineral crocoite PbCr^VIO₄ showed a Cr 2p3/2 at 579.3 eV [45], while a compound such as BaCr^VIO₄ has a BE of 579.1 eV. In contrast Cr^III(OH)₃ has a lower BE of 577.1-577.4 eV [51]. Alidokht, et al. [82] observed that upon adsorbtion of chromate using LDH complete reduction of the Cr^VI to Cr^III took place with Cr 2p3/2 peak at a BE of 277.9 eV. A similar shift was also observed for the reduction of sodium chromate during ion etching in the XPS instrument by Treverton and Davies [83]. The Cr3+ is probably initially present in the form of Cr(OH)₆^3- but this anion shows a strong affinity for carbonate [84], so exchange of OH^- for CO₃^2- in the anion over time is well possible within the LDH interlayer space given that LDH itself also has a strong affinity for carbonate adsorption. The O 1s peak for the chromate is found around 530.2 eV, which is the same as that of crocoite, PbCrO₄ [45], and comparable to compounds such as Na₂CrO₄ at 530.0-530.3 eV, CaCrO₄ at 529.5 eV and Li₂CrO₄ at 530.3 eV [51]. The O 1s peak for CrIII compounds occur at slightly high BE and can not be distinguished from the O 1s OH peak of the LDH layer structure (Figure 7 right). The NIST database reports BE values of 530.8 to 531.6 for Cr(OH)₃ [51].

Vibrational spectroscopy (Raman and FTIR) results on the fresh sample showed that the initial chromate anions were not polymerised to Cr₂O₇^2- in the LDH interlayer space [16]. The Raman spectrum of the chromate anion in solution reveals distinct vibrational modes: the symmetric stretching mode (ν1) is observed at 848 cm⁻¹, the bending mode (ν2) at 348 cm⁻¹, the asymmetric stretching mode (ν3) at 884 cm⁻¹, and the out-of-plane bending mode (ν4) at 363 cm⁻¹.[80]. Frost, Musumeci, Martens, Adebajo and Bouzaid [18] reported the Raman spectrum of the same freshly prepared sample as later used in this study for the XPS analyses. They observed a strong band at 848 cm^-1 attributed to the ν₁ symmetric stretching vibrational mode, while two bands at 884 and 928 cm^-1 were assigned to the ν₃ antisymmetric stretching vibrational mode. The band at 474 cm^-1 was assigned to the ν₄ bending mode. Finally, two bands at 363 and 237 cm^-1 were interpreted as being the ν₂ bending modes. The splitting of the ν₃ and ν₂ modes is suggestive of an initial lowering of the symmetry of the chromate anions in the LDH interlayer space prior to partial reduction. The symmetry lowering must be taken into account through the interation of the CrO₄^2- anions with both interlayer water and the LDH hydroxide layer surfaces on both sides of the interlayer space. They did not give an explanation for a broad band at 821 cm^-1. Although this is lower than reported for fresh Cr(OH)₃ around 850 cm^-1 [85], it is possible that this band is associated with the start of the reduction of the Cr⁶⁺ to Cr³⁺ and represent the Cr-OH symmetric and antisymmetri stretching mode and the shift is due to the restricted space within the LDH interlayer space. The IR interlayer CrO₄^2- υ₃ (Cr-O) mode was reported at 870 cm^-1 by Del Arco, et al. [86]. while Prasanna and Vishnu Kamath [87] found the same band at 866 cm^-1. After heating between 100 and 200 ºC, a split of this band was observed and two new bands appared around 874 and 930 cm^-1 supporting a change in the anion symmetry from T_d to C_3v as a result of the interaction between the CrO₄^2- and the LDH hydroxide layers similar to what was observed in the Raman spectra by Frost, Musumeci, Martens, Adebajo and Bouzaid [18]. Prasanna, et al. [88] also observed in the mid-IR spectrum of intercalated CrO₄^2- two bands at 866 and 917 cm^-1. The tetrahedral chromate ion is characterized by two distinct IR-active vibrational modes: the antisymmetric stretch (ν₃) and the symmetric deformation mode (ν₄). The ν₄ is expected to be observed around 330 cm^-1 i.e., outside the mid-IR range and therefore not observed. The two bands represent the triply degenerate (F₂) mode of the chromate. With C_3v symmetry the antisymmetric stretching vibration splits into two modes (A₁ + E) [89].

The XPS Cr 2p high resolution scan of the chromate-LDH provided evidence that upon ageing the nature of the interlayer anion changed resulting in the reduction of Cr⁶⁺ to Cr ³⁺. Since the Raman spectrum of the same sample was taken shortly after synthesis, the spectrum did not show any significant evidence of this reduction [18]. Instead the Raman spectrum is dominated by the CrO₄^2- anion with a reduced symmetry, as has also been observed in the infrared spectra of fresh chromate-LDH samples. The XPS data were obtained about 6 months after the Raman spectra and by then approximately 60 % of the Cr⁶⁺ was reduced to Cr³⁺.

3.7. Hexacyanoferrate Anions

In earlier work by Kloprogge, Ponce and Ortillo [50] for the hexacyanoferrate(II) anion intercalated in LDH the N 1s peak was found at 397.1 eV for Fe(CN)₆^4- while for Fe(CN)₆^3- it was observed at 397.4 eV (Figure 8 left). The NIST database shows that Fe(CN)₆⁴ compounds have N 1s BE values of around 397.4 to 398.0 eV, while K₃Fe(CN)₆ had a BE of 398.1 eV [51], indicating that there is a minor difference in BE for N 1s in these two anions.

There were significant differences in the Fe 2p3/2 high resolution spectra with the Fe(CN)₆^3- exhibiting the main peak at a BE of 707.7 eV and a second peak at a BE of 709.5 eV. In contrast, the Fe(CN)₆⁴ spectrum exhibited only one peak at 707.8 eV (see Figure 14 in [50]). For K₄Fe(CN)₆ the NIST database shows a BE value of 707.1 and 708. 5 eV while for K₃Fe(CN)₆ it was reported at 709.6 eV [51]. Holgado, et al. [90] used Fe-XANES analysis to show that partial reduction of the Fe(III) in the Fe(CN)₆⁴ intercalated LDH had taken place. Moreover, Idemura, Suzuki and Ono [32] showed the reduction of Fe(III) in Fe(CN)₆⁴ complexes by Mössbauer spectroscopy. Yamashita and Hayes [91] found that the Fe 2p3/2 for Fe(II) has a BE of about 1 eV lower than Fe(III) in their oxides. The same is true for the potassium hexacyanoferrates and in this study of intercalated LDH with hexacyanoferrates. Kloprogge, Ponce and Ortillo [50] concluded that the Fe 2p3/2 peak at 707.7 eV is due to Fe(II) caused by partial reduction in the initial Fe(III) observed at 709.46 eV. In their study they did not report on the C 1s results for these two samples. For the Fe(CN)₆^4—LDH the C 1s of the C-N bond was observed at a BE value of 283.5 eV while for Fe(CN)₆^3—LDH at was observed at 283.9 eV (Figure 8 right). These BE values and the small difference between the two compare well with the values and differences reported for K₄Fe(CN)₆ at 283.5 eV and K₃Fe(CN)₆ at 283.9 eV [51].

The infrared and Raman spectra of hexacyanoferrate intercalated LDH have been reported for the last 30 years. Kikkawa and Koizumi [92] reported a band around 2000 cm^-1 assigned to the CN stretching mode. In contrast, Idemura, et al. [93] observed for the anion-exchanged LDH in an aqueous solution of K₃Fe(CN)₆ resulted in two bands at 2120 and 2040 cm^-1 in the CN stretching region, indicating that part of the cyanoferrate(III) complex was reduced to cyanoferrate(II) complex during the intercalation. They based this on the fact that pure K₃Fe(III)(CN)₆ and K₄Fe(II)(CN)₆.3H₂O produces IR active bands at 2120 and 2040 cm^-1, respectively. After the intercalation of Fe(CN)₆^4- in the LDH, no change in the oxidation state of Fe was detected. A band at 2040 ± 4 cm^-1 was the stretching vibration of ν(CN) of Fe(CN)₆^4- intercalated in LDH by Mao, et al. [94]. Amini, Rahimpour and Jouyban [29] reported two bands at 2038 and 2113 cm^-1 for Fe(CN)₆^3- intercalated Ni/Al–LDH assigned to CN bound to Fe(II) and to CN bound to Fe(III), respectively. Panda, et al. [95] described an intense sharp band at 2094 cm^-1 attributed to the CN stretching mode of hexacyanoferrate(III) anions, but in addition observed two weak bands around 2002 and 2163 cm^-1. The band at 2002 cm^-1 was interpreted as being do to the partial reduction of Fe³⁺ to Fe²⁺. The weak band at 2163 cm^-1 was thought to be the result of the formation of nickel ferricyanide, in which the hexacyanoferrate(III) anion is free from the LDH layers. Likewise, Meng, et al. [96] observed a band at 2111 cm^-1 assigned to the CN stretching mode of hexacyanoferrate(III) and a second weak band at 2034 cm^-1 attributed to the CN stretching mode of hexacyanoferrate(II), suggesting that a small amount of Fe³⁺ was reduced to Fe²⁺ in the LDH interlayer space. Holgado, Rives, Sanromán and Malet [90] described two sharp bands at 2035 ±1 cm^-1 assigned to CN stretching mode of hexacyanoferrate(II) and at 2120-2086 cm^-1 attributed to the CN stretching mode of hexacyanoferrate(III) intercalated in LDH.

Hansen and Koch [31] detected the CNstretching mode of hexacyanoferrate(II) and (III) at 2036 and 2112 cm^-1, respectively. In addition they observed a band at 2080 cm^-l, which they attributed to the presence of free cyanide anions in the interlayer space of the LDH, indicating that the hexacyanoferrate(II,III) cyanide ligands can be replaced with either water or hydroxyls. In addition, oxidation and ligand substitution of the hexacyanoferrate(II) were observed by Mössbauer spectroscopy. In contrast, Yao, et al. [97] attributed a band at 2030 cm^-1 when Fe(CN)₆^3- was incorporated in LDH to a CN⁻ group in which the bond strength of CN⁻ was weaker due to the interaction with the LDH hydroxide layer. The intensity of the band at 2030 cm^-1 was shown to increase upon aging. Crespo, et al. [98] remarked that regardless of the nature of the starting hexacyanoferrate, two separate or overlapping bands were observed in all cases. Furthermore, the two bands are found closer to each other (maximum Δν = 46 cm^-1) than for the pure potassium salts (Δν = 73 cm^-1)indicating that the samples underwent redox processes resulting in a mixture of Fe²⁺ and Fe³⁺ species. They indicated that it is common knowledge from coordination chemistry that the exact band position of the ν(CN) stretching mode for hexacyanoferrate(II) varies with the nature of the countercation. In the case of intercalated LDH hydrogen bonding between interlayer water molecules or the LDH hydroxide layer hydroxyl groups and the cyano groups may likewise have an effect on the exact band position. Fernández, et al. [99] observed the most intense band at 2118 cm^-1. In addition, a weak band was observed at 2042 cm^-1, while even weaker shoulders were detected at 2089 and 2060 cm^-1. The ν_CN band at 2042 cm^-1, indicates a partial Fe³⁺ to Fe²⁺ reduction within the LDH interlayer space similar to other studies. It is known that hexacyanoferrate anions are outer-sphere electron-transfer reductants or oxidants [100,101], based upon which the origin of the two weaker bands at 2089 and 2060 cm^-1 can be assigned. Based on Jones [102] the A_1g, E_g and T_1u ν_CN modes required by the O_h point group for Fe(CN)₆⁴⁻ are found at 2094, 2062 and 2044 cm^-1 in aqueous solution. The first two modes are infrared-forbidden, but in the restricted interlayer space of the LDH can become partially activated due to a decrease in symmetry, resulting in the two weak bands at 2089 and 2060 cm^-1, respectively. If grafting had happened (if the calculated interlayer space height was smaller than the size of the Fe(CN)₆⁴⁻ anion along the C₃ axis), the decrease in symmetry would be much more dramatic and the infrared spectrum much more complicated. Braterman, et al. [103] and Boclair, et al. [104] studied oriented samples of hexaxyanoferrate intercalated LDH and compared those to random oriented samples. The random oriented intercalated LDH exhibited two overlapping bands of similar intensity, with the higher wavenumber band being the broader of the two. This higher wavenumber band at 2041 cm^-l was still visible in the oriented sample, with unaltered band shape and width, but the sharper, lower wavenumber band at 2035 cm^-1 was no longer visible. Therefore, both these intense bands were interpreted to correlate with T_1u in O_h, and their separation is proof of a local reduction of symmetry. They attributed this effect to a change to D_3d symmetry by the ferrocyanide anion, resulting in a splitting of this mode into E_u and A_2u components. The symmetry reduction is a result of the orientation of the ferrocyanide anions in the interlayer space of the LDH, which lie with two opposed triangular faces of the coordination octahedron parallel to the internal surfaces of the LDH hydroxide layers. Under these circumstances, the E_u component is likely to be broader than the A_2u component. Furthermore, the E_u component is x,y-polarized in the molecular axis system, while the A_2u component is z-polarized. This explains the detected selectivity. The molecular three fold z-axis, the crystallographic c axis, and the direction of light propagation are all parallel in the oriented LDH sample.

Two Raman spectroscopic studies on hexacyanoferrate intecalated LDH were published by Kloprogge, et al. [105] and Frost, Musumeci, Bouzaid, Adebajo, Martens and Theo Kloprogge [30]. Details about the symmetry of the hexacyanoferrate anion and its reduction and oxidation behavior after intercalation in LDH can be obtained by comparing the Raman spectra with the infrared spectra. Free Fe(CN)₆^4- will exhibit three vibrational modes with A_1g at 2098 cm^-1 and E_g at 2062 cm^-1 in the Raman spectrum and T_1u at 2044 cm^-1 in the infrared spectrum. For the intercalated LDH, they detected bands in the Raman spectrum at 2136, 2094 and 2065 cm^-1 for hexacyanoferrate(II) intercalated LDH, and at 2164, 2136, 2094 and 2059 cm^-1 for hexacyanferrate(III) intercalated LDH, clearly supporting the fact that a change in the site symmetry had taken place as well as partial reduction. Lowering of the site symmetry to C₃ after intercalation in LDH would predict four bands in the Raman spectra for both the hexacyanoferrate (II) and hexacyanoferrate(III) anions resulting in theoretically eight modes. In reality, it is to be expected that a number of these modes will overlap causing the detection of fewer bands.

It is clear in the case of hexacyanoferrate(III) intercalation in LDH generally results in a partial reduction of Fe3+ to Fe2+. In contrast to the other anionic groups studied here, the reduction of hexacyanoferrate(III) results in clear changes in the infrared and Raman spectra. These changes are also clearly visible in the XPS Fe 2p spectra. Integration of the peak areas for both the Fe 2p3/2 of Fe(II) and Fe(III) allows for a direct measure of the amount of reduction. In this sample 45 % of the Fe(III) has been reduced to Fe(II).

3.8. Influence of Interlayer Anions on the LDH Layer Structure

From the previous sections it is clear that the nature of the interlayer anion can be studied not only by vibrational spectroscopy (infrared and Raman) but also by XPS and additional information can be obtained. In most instances the BE of the non-oxygen atom is slighly lower compared to minerals in which the anion forms part of the rigid crystal crystal structure. In several instances evidence was found for partial reduction of the metal upon intercalation, which was not detected by vibrational spectroscopy. Since the interlayer anions are interaction, most likely through hydrogen bonds with the hydroxide layers of the LDH within the interlayer space, it is of interest to see if the atoms that make up the LDH layer structure are in any way affected by the nature of the interlayer anion. Table 1 provides an overview of the BE values obtained for Al 2p, Mg 2p and O 1s for the hydroxyl group. In all instances the BE values are very close though it seems that the MoO₄^2- and CrO₄^2- ions with their much heavier metals result in slightly lower BE values for Al 2p, Mg 2p and O 1s. The differences, however, are so small that they may be considered to be within the experimental error of XPS.

5. Conclusions

This comprehensive analysis of Layered Double Hydroxides (LDHs) intercalated with various anionic species elucidates the intricate interplay between interlayer anions and LDH structural dynamics. X-ray photoelectron spectroscopy (XPS) characterized the distinct binding energies for key elements, confirming the presence of carbonate, phosphate, sulfate, molybdate, chromate, and hexacyanoferrate anions within the interlayer space. The carbonate anion exhibited a notable reduction in binding energy compared to common carbonate minerals, attributed to interactions with interlayer water and hydroxyl groups, suggesting semi-constrained behavior within the LDH framework. In contrast, phosphate intercalation demonstrated stability without shifts in binding energies, while sulfate ions indicated limited movement, suggesting moderate interaction with hydroxide layers. Molybdate displayed weaker interactions than crystalline counterparts, reflecting the influence of LDH synthesis conditions. A significant observation was the reduction of chromate ions from Cr⁶⁺ to Cr³⁺, indicating dynamic redox behavior within the interlayer space. Furthermore, hexacyanoferrate anions revealed partial reduction, with distinctive spectral differences indicating site symmetry alterations. Collectively, these findings articulate the necessity for considering anion identity and synthesis parameters in the application and functional design of LDHs, highlighting their potential roles in catalysis, ion exchange, and environmental remediation. Further investigations into the relationships between interlayer anions and LDH stability are warranted to broaden the understanding of their reactivity and utility in advanced materials science.