The C 1s high-resolution spectrum of the carbonate anion shows a peak at a BE value of 288.6 eV (Figure 4 left). This peak is due to the interlayer anion. The other peaks are due to surface-adsorbed organic matter, the so-called “rubbish” carbon. The corresponding O 1 peak was found at 530.5 eV (Figure 4 right), and the ratio of O/C is about 3.10, close to the expected ratio of 3 for CO₃²⁻. The BE value is slightly lower compared to pure carbonate minerals such as magnesite, which has a C 1s at 290.0 eV, dolomite at 289.8 eV, and calcite at 289.5 eV [45]. This shift in BE is probably 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 CO₃²⁻ a shift toward lower wavenumbers was detected. Based on X-ray diffraction, ionic size, and charge density, it has been generally accepted that the CO₃²⁻ 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.

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 5 left). The synthesis of the phosphate-containing LDH was performed at three different pH values (9.3, 11.9 and 12.5). No difference in the BE values was observed, but a significantly higher amount of phosphate was present in the interlayer space of the LDH in the 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 that the P 2p was 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 there was 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 PO₄²⁻ at 531.6 eV (Figure 5 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 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 presumably, this is the P-O in the phosphate, as there are no M-O modes in the LDH hydroxide layer. These assignments do not seem to be correct as, generally, O 1s for PO₄ are always found at much higher BE.

LDHs intercalated with PO₄³⁻ have been shown to form different phases with very different basal spacings: 0.84 nm [20], 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 PO₄³⁻ between the LDH hydroxide 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 PO₄³⁻ 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, claimed the presence of not only PO₄³⁻ in the interlayer space but also HPO₄²⁻. Though the spectrum provided is of rather low quality and they do not show the band fit, they claim that the broadband 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 HPO₄²⁻ 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 attributed to HPO₄²⁻ 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 HPO₄²⁻ in addition to PO₄³⁻ [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 synthesized at pH 12.5 exhibited a strong, sharp band at 960 cm⁻¹ with a broader band at 1026 cm⁻¹ assigned to the PO₄³⁻ symmetric and antisymmetric stretching vibrational modes identical to that of the PO₄³⁻ 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 detected 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 PO₄³⁻ ion has a strong band at 936 cm⁻¹, while HPO₄²⁻ anion exhibits a strong band at 990 cm⁻¹. Therefore, the spectrum can be interpreted as a combination of the PO₄³⁻ anion (964 and 1033 cm⁻¹) and the HPO₄²⁻ 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 PO₄³⁻ 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 of 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 HPO₄²⁻ have been claimed [52,53]. For the two samples synthesized at pH 12.5 and 11.9, Raman spectroscopy showed that this was not the case [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 us to distinguish between PO₄³⁻ and HPO₄²⁻ in the three samples studied.

In naturally occurring LDH, other than the commonly present CO₃²⁻ anion, SO₄²⁻ 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 6 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 slightly lower than observed for common sulfate minerals such as baryte (169.2 eV), anhydrite (168.9 eV), or gypsum (169.2 eV) [45], indicating that the sulfate in the LDH interlayer space is less strongly interacting with the LDH hydroxide layers compared to a sulfate 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 sulfate group but also with oxygen in interlayer water that has a similar BE. This value is comparable with H₂SO_4, 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 sulfates by Wahlqvist and Shchukarev [64].

Bish and Livingstone [61] described for honessite the SO₄²⁻ 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 SO₄²⁻ 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], p. 140). Earlier work in our group has shown that the infrared spectrum of Mg/Al-LDH containing SO₄²⁻ in 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, which was 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 SO₄²⁻ ν₃ antisymmetric stretching mode. A weak band at 986 cm⁻¹ was assigned to the corresponding ν₁ symmetric stretching mode. They also reported that carrboydite exhibited 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 broadband is visible around 1134 cm⁻¹. They indicated that when SO₄²⁻ 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 sulfate anion. This is not compatible with the infrared spectrum, where all four modes are present. For analogous reasons, the C₃ site symmetry, as proposed 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] concluded 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 sulfate anion from T_d to C_2v or even lower symmetry.

The MoO₄-LDH is characterized by a single peak for the Mo 3d5/2 at a BE of 229.6 eV (Figure 7 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 the 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 7 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, and 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 MoO₄²⁻. The hardly visible shoulder at 231.7 eV was attributed to Mo(V) species, possibly caused by a charge transfer between (Mo⁶⁺ + O²⁻) 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 O²⁻ [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 slightly higher BE values than observed in this study and in comparison with other molybdate structures such as those 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 polymerized 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 of the MoO₄²⁻ 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 characterized 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 attributed 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 MoO₄²⁻ 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 MoO₄²⁻ 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 antisymmetric and symmetric stretching vibrations of Mo-O bonds, as suggested 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, which was attributed to the vibrations of Mo = O in polymolybdate Mo₇O₂₄⁶⁻. A band at 670 cm⁻¹ with a shoulder at 856 cm⁻¹ was assigned to the Mo–O–Mo stretching mode of Mo O₄²⁻ 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 MoO₄²⁻ symmetric stretching modes in the LDH were observed at 908 and 892 cm⁻¹. They were interpreted as being associated with two different MoO₄²⁻ 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 MoO₄²⁻ 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 the nature of the molybdate anion can change depending on the synthesis conditions, in particular, the pH. Under alkaline conditions, the MoO₄²⁻ 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⁺ + O²⁻) and (Mo⁵⁺ + O⁻), as suggested in some papers [74,75].

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 8 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 of Cr⁶⁺ in the CrO₄²⁻ 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 adsorption 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)₆³⁻ but this anion shows a strong affinity for carbonate [84], so the exchange of OH⁻ for CO₃²⁻ 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 Cr^III compounds occurs at slightly high BE and can not be distinguished from the O 1s OH peak of the LDH layer structure (Figure 8 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 polymerized to Cr₂O₇²⁻ 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 interaction of the CrO₄²⁻ 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 represents the Cr-OH symmetric and antisymmetric stretching mode and the shift is due to the restricted space within the LDH interlayer space. The IR interlayer CrO₄²⁻ 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 °C, a split of this band was observed, and two new bands appeared 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 CrO₄²⁻ 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 CrO₄²⁻ 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) ([65], p. 414).

The XPS Cr 2p high-resolution scan of the chromate-LDH provided evidence that upon aging, the nature of the interlayer anion changed, resulting in the reduction of Cr⁶⁺ to Cr³⁺. 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 CrO₄²⁻ 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⁺.

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)₆⁴⁻ while for Fe(CN)₆³⁻ it was observed at 397.4 eV (Figure 9 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)₆³⁻ 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. [89] 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 [90] 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)₆⁴⁻ LDH, the C 1s of the C-N bond was observed at a BE value of 283.5 eV while for Fe(CN)₆³⁻ LDH at was observed at 283.9 eV (Figure 9 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 [91] reported a band around 2000 cm⁻¹ was assigned to the CN stretching mode. In contrast, Idemura, Suzuki and Ono [32] 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 produce IR active bands at 2120 and 2040 cm⁻¹, respectively. After the intercalation of Fe(CN)₆⁴⁻ 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)₆⁴⁻ intercalated in LDH by Mao et al. [92]. Amini, Rahimpour and Jouyban [29] reported two bands at 2038 and 2113 cm⁻¹ for Fe(CN)₆³⁻ intercalated Ni/Al–LDH assigned to CN bound to Fe(II) and to CN bound to Fe(III), respectively. Panda et al. [93] described an intense sharp band at 2094 cm⁻¹ attributed to the CN stretching mode of hexacyanoferrate(III) anions, but in addition, two weak bands were observed around 2002 and 2163 cm⁻¹. The band at 2002 cm⁻¹ was interpreted as being due 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. [94] observed a band at 2111 cm⁻¹ was assigned to the CN stretching mode of hexacyanoferrate(III), and a second weak band at 2034 cm⁻¹ was 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 [89] described two sharp bands at 2035 ±1 cm⁻¹ assigned to the 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 CN-stretching 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. [95] attributed a band at 2030 cm⁻¹ when Fe(CN)₆³⁻ 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, Barriga, Rives and Ulibarri [35] 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 counter-cation. 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. [96] 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 [97,98], based upon which the origin of the two weaker bands at 2089 and 2060 cm⁻¹ can be assigned. Based on Jones [99] 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, they 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 have been much more dramatic, and the infrared spectrum much more complicated. Braterman et al. [100] and Boclair et al. [101] studied oriented samples of hexacyanoferrate intercalated LDH and compared those to randomly 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 lies 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 intercalated LDH were published by Kloprogge et al. [102] 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)₆⁴⁻ 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 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 expected that a number of these modes will overlap, causing the detection of fewer bands.