Analyzing the share of amorphous silica in mixtures with different soil minerals using fourier transform infrared spectroscopy and PLSR chemometrics

Oliver Hunfeld; Ruth H Ellerbrock; Mathias Stein; Carsten W Müller; Jörg Schaller

doi:10.1038/s41598-026-45511-3

. 2026 Mar 25;16:9969. doi: 10.1038/s41598-026-45511-3

Analyzing the share of amorphous silica in mixtures with different soil minerals using fourier transform infrared spectroscopy and PLSR chemometrics

Oliver Hunfeld ¹, Ruth H Ellerbrock ¹, Mathias Stein ^1,², Carsten W Müller ³, Jörg Schaller ^1,^4,^✉

PMCID: PMC13022021 PMID: 41882072

Abstract

Amorphous silica (ASi) improves key soil functions and crop productivity but is difficult to quantify due to complex mineral mixtures and time-consuming chemical analyses. This study explored the possibility of using Fourier-transform infrared spectroscopy (FTIR) in combination with partial least-squares regression (PLSR) to estimate the ASi content in samples of mineral mixtures. For this purpose, mixtures of different pedogenic minerals (kaolin and montmorillonite) with known ASi content were produced and analysed using FTIR spectroscopy. Based on these data, a PLSR model was used to predict the ASi concentration based on the FTIR spectra. The results show that the model is capable of estimating ASi content in simple mineral mixtures with high accuracy. This suggests that FTIR, combined with PLSR, could be a promising method for the rapid and cost-effective determination of ASi in environmental samples. Future studies should investigate how the method performs with more complex mixtures and natural soil samples, and how factors such as mineral weathering and sample origin influence the accuracy of the prediction.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-026-45511-3.

Keywords: Amorphous silica, Fourier-transform infrared spectroscopy, Pedogenic minerals, Quantification, Silicon

Subject terms: Chemistry, Environmental sciences, Solid Earth sciences

Introduction

In the last decade it was shown that amorphous silica (ASi) significantly improves ecosystem processes such as nutrient availability¹, water retention², soil aggregate stability³ and crop production^4,5. Most recently, ASi was suggested to be a potential key for sustainable and resilient crop production⁶. In soils ASi exists in mixture with minerals including crystalline minerals (e.g. quartz, feldspars), short-range ordered species (e.g., imogolite, allophanic minerals) and amorphous silica of biogenic or minerogenic origin. However, the quantification of ASi in soils as a mixture of different minerals with ASi relies on wet chemical extractions. Such extractions are time-consuming and often lack selectivity and effectiveness⁷.

More recently, Fourier-transform infrared spectroscopy (FTIR) analyses were successfully applied to distinguish between short-range ordered aluminosilicates (SROAS), pure amorphous silica (ASi) and minerals like quartz⁸. Other studies indicate difference between quartz sand and ASi like glasses^9,10, the latter by using Raman spectroscopy. Additionally¹¹, interpreted FTIR data by using energy-loss functions to determine crystal-symmetry based vibrational modes of ASi species and quartz. However, systematic studies on the relation between infrared (IR) spectral features and how they can be used to differentiate between solid Si species differing in the crystallinity of the Si tetrahedral network are limited. A recent study suggests the potential of FTIR for determining ASi content in mixtures with minerals such as quartz¹². This study showed that the absorption band around 800 cm^− 1 can be used to distinguish between crystalline minerals like quartz and ASi, because the absorption band at 800 cm^− 1 shifts towards lower wavenumber and becomes smaller, with increasing crystallinity. Every mineral may show specific absorption bands reflecting differences in the structure of the mineral¹². Another study has shown that FTIR can distinguish between 1 and 1 and 2 − 1 layer silicates¹³. Other studies showed FTIR spectra for different clay minerals¹⁴.

However, the accuracy of ASi determination in mineral mixtures is not clear yet as well as if such analysis is successful for other minerals dominant in soils, for example aluminosilicates such as kaolin and montmorillonite as important examples for clay minerals in soils¹⁵. Furthermore, it is not clear if different silicates as main constituent of soils (kaolinite, montmorillonite, olivine, illite, vermiculite and biotite)¹⁵ can be distinguished from each other by FTIR or if different forms of ASi (Aerosil 300, rice husks, natural amorphous silica, Sipernat 310 and Sipernat 50)¹⁶ show different FTIR patterns. In addition, it is not clear yet if minerals show different FTIR patterns if from different sources and sites.

This study analyzed following important soil minerals: the isolated silicate olivine, the 2-layer clay mineral kaolin from 3 different sites, the 3-layer clay minerals montmorillonite from 2 different sites, vermiculite, illite and the mica biotite¹⁵ enabling comprehensive soil analysis. Furthermore, the amorphous silicates such as Sipernat310, Sipernat50, Aerosil300 and different natural amorphous silicates were analyzed.

In this study, we aimed to identify specific absorption bands in FTIR spectra that can differentiate between Si minerals or ASi forms. Further, we test if the share of ASi in mixture with different minerals can be determined by partial least-squares regression (PLSR), as PLSR is a popular modelling technique used in chemometric analyses and is commonly used for quantitative spectral analyses.

Results and discussion

The spectra of all minerals show in general absorption bands in three regions: I: WN 3800 to 2700 cm^− 1, II: 1300 to 850 cm^− 1 and III: 650 to 400 cm^− 1. To allow for comparison in all spectra, the maximum of the most intense band was normalised to one.

General band assignment

The broad band at WN region I (3800 to 2700 cm^− 1) is caused by O-H groups that are part of minerals, and water (O-H band¹²). Its broadness results from the formation of hydrogen bridges between the different O-H groups. The bands at WN region II (1300 to 850 cm^− 1) are caused by stretching vibrations of Si-O groups within the Si-tetrahedra (Si-O bands¹²). These stretching vibrations of Si-O groups are caused by Si-O-Si, Si-OH/Si-O⁻ or Al-OH groups. The Si-O-Si stretching vibrations occur between WN 1300 and 1000 cm^− 1, while the Si-OH/Si-O⁻ and Al-OH stretching vibrations occur between WN 1050 and 850 cm^− 1^12,17^,18. The bands at WN region III (600 to 400 cm^− 1) are caused by Si-O-Si bending vibrations (δ-Si-O-Si band¹⁹) The small band at the WN range 850 to 750 cm^− 1 (black arrow in Fig. 2) is caused by symmetric Si-O-Si stretching vibrations¹². In Minerals containing aluminium, WN region III can also show a second band (650 to 450 cm^− 1) caused by Al-O-Si deformation vibrations (δ-Al-O-Si band¹⁸).

Fig. 2 — (a) FTIR spectra of Kaolin(A), Kaolin(B) and Kaolin(C) from WN range 4000 to 2800 cm^-1 and 1500 to 400 cm^-1(b) detail enlargement from 1200 to 900 cm^-1.

Amorphous silica

The FTIR spectra of Sipernat50, Sipernat310, Aerosil300, natural amorphous silica and rice husk show strong similarities. The bands with band maxima at 3438, 969 and 470 cm^− 1 deviate slightly in intensity (Fig. 1). These differences in the WN of the band maxima are lower than the spectral resolution of r = 2 cm^− 1. The FTIR spectra of both Sipernat (310 and 50) and Aerosil300 (Fig. 1; Figure S1−3; n = 2) shows a broad O-H absorption band at WN region I (3700 to 2700 cm^− 1) with a maximum at 3439 cm^− 1 (Sipernat 310) and at WN 3438 cm^− 1(Sipernate 50). The Si-O band (WN region II) has a maximum located at WN 1096 cm^− 1 (Sipernat 310) and at 1097 cm^− 1(Sipernate 50) (Table 1). This band has a shoulder located around 1155 cm^− 1 for Sipernat 310. The δ-Si-O-Si band (WN region III) shows a maximum at 471 cm^− 1 (Sipernat 310) and at 470 cm^− 1(Sipernate 50 and Aerosil300). Additionally, the FTIR spectra of Sipernat310 shows two less pronounced bands with maxima at 967 (WN region II) and 801 cm^− 1 (black arrow 1 in Fig. 1) (Table 1). The maxima of the bands in WN regions II and III Aerosil300 are located at 1104 and 470 cm^− 1 deviate less than 8 cm^− 1 from the respective band in the FTIR spectra of Sipernat310. Additionally, the intensity of the absorption bands varies in intensity. While the FTIR spectra of Sipernat310 shows a band with a maximum at 967 cm^− 1, the FTIR spectra of Aerosil300 only shows a small shoulder, because pyrogenic manufactured Aerosil300 is poor of hydroxyl (OH) groups.

Fig. 1 — FTIR of amorphous Silica from wavenumber (WN) range 4000 to 2800 cm^-1 and 1500 to 400 cm^-1.

Table 1.

Maxima of sorption bands in FTIR spectra of minerals associated with their WN range (1 between layers, 2 in layers). The absorption peaks correspond to those highlighted in Figures 1, 2 and 3.

	valence vibration			deformation vibration
	region I	region II	arrow marked	region III
	ν- OH	ν- Si-O	ν- Si-O-Si	δ- Al-O-Si	δ - Si-O-Si
Reference	¹²	¹²	¹²	¹⁸	¹²
WN	3800-2700	1300-850	850-750	650-500	550-400
Sipernat310	I₁3439 hydrogen bridges	II ₁ 1096 ν-Si-O II ₂ 967 ν-Si-O	1 801 ν-Si-O-Si		III 471 δ-Si-O-Si
Sipernat50	I₁ 3438 hydrogen bridges	II ₁ 1097 ν-Si-O II ₂ 969 ν-Si-O	1 798 ν-Si-O-Si		III 470 δ-Si-O-Si
Aerosil	I₁ 3438 hydrogen bridges	II ₁ 1104 ν-Si-O II ₂ 987 ν-Si-O	1 809 ν-Si-O-Si		III 470 δ-Si-O-Si
Rice husk	I₁ 3448 hydrogen bridges	II ₁ 1104 ν-Si-O	1 803 ν-Si-O-Si		III 469 δ-Si-O-Si
Natural amorphous silica	I₁ 3438 hydrogen bridges	II ₁ 1092 ν-Si-O II ₂ 945 ν-Si-O	1 799 ν-Si-O-Si		III 466 δ-Si-O-Si
Kaolin(A)	I_a 3694 Si-OH¹ I _b3618 Si-OH² I_c 3439 ν- Al- OH hydrogen bridges	II _a 1114 ν-Si-O II _b 1032 ν-Si-O II _c 1009 ν-Si-O II _d 913 ν-Si-O	a 796 ν-Si-O-Si	III_a539 δ-Al-O-Si	III_b470 δ-Si-O-Si
Kaolin(B)	I _a3696 Si-OH¹ I _b3619 Si-OH² I_c 3448 ν- Al-OH hydrogen bridges	II _a 1113 ν-Si-O II _b 1031 ν-Si-O II _c 1007 ν-Si-O II _d 912 ν-Si-O	a 789 ν-Si-O-Si	III_a537 δ-Al-O-Si	III_b469 δ-Si-O-Si
Kaolin(C)	I_a 3694 Si-OH¹ I _b3619 Si-OH² c 3448 ν- Al- OH hydrogen bridges	II _a 1115 ν-Si-O II _b 1032 ν-Si-O II _c 1010 ν-Si-O II _d 913 ν-Si-O	a 797 ν-Si-O-Si	III_a539 δ-Al-O-Si	III_b470 δ-Si-O-Si
Montmorillonite (A)	I₁ 3630 Si-OH I₂ 3421 ν- Al- OH hydrogen bridges	II ₁ 1090 ν-Si-O II ₂ 1048 ν-Si-O	1 796 ν-Si-O-Si	III₁ 522 δ-Al-O-Si	III₂471 δ-Si-O-Si
Montmorillonite (B)	I₁ 3621 Si-OH I₂ 3438 ν- Al- OH hydrogen bridges	II 1048 ν-Si-O	1 798 ν-Si-O-Si	III₁525 δ-Al-O-Si	III₂468 δ-Si-O-Si
Illite	I 3438 hydrogen bridges	II 1025 ν-Si-O		III₁ 524 δ-Al-O-Si	III₂ 471 δ-Si-O-Si
Vermiculite	I 3414 hydrogen bridges	II 1000 ν-Si-O			III 450 δ-Si-O-Si
Biotite	I 3438 hydrogen bridges	II 1004 ν-Si-O			III 461 δ-Si-O-Si
Olivine	I₁ 3438 hydrogen bridges	II ₁ 983 ν-Si-O II ₂ 885 ν-Si-O II ₃ 839 ν-Si-O			III₁ 605 δ-Si-O-Si* III₂ 504 δ-Si-O-Si* III₃ 415 δ-Si-O-Si*

Open in a new tab

*can contain ν-Mg-O vibrations¹⁹

The FTIR spectra of natural amorphous silica (nASi) (Fig. 1; n = 1) shows a broad O-H band at WN region I (3750 to 2800 cm^− 1) with the maximum at WN 3438 cm^− 1. The Si-O band (WN region II) has a maximum located at 1092 cm^− 1 (Table 1). The δ-Si-O-Si band within WN region III is located at 466 cm^− 1. The FTIR spectra of nASi are similar to that of Sipernat310. Both have a broad band between 3700 and 3000 cm^− 1. But the O-H band from nASi is less intense. The bands with maxima at 1092 (WN region II) and 466 cm^− 1(WN region III) deviate less than 8 cm^− 1 in WN and slightly in intensity. With the intensity of the FTIR spectra of nASi being higher. The band maxima at 945 cm^− 1 deviates by 22 cm^− 1 towards lower WN and has a stronger Intensity compared to the FTIR spectra of Sipernat310. The shoulder at WN 1151 cm^− 1 is also more pronounced. For Rice husk the FTIR spectra (Fig. 1; Figure S4; n = 1) shows a broad O-H band at WN region I (3750 to 3000 cm^− 1) with the maximum at WN 3448 cm^− 1 (Table 1). The Si-O band (WN region II) has a maximum located at 1104 cm^− 1. The δ-Si-O-Si band within WN region III is located at 469 cm^− 1. The FTIR spectra of Rice husk is rather similar to that of Sipernat310. The O-H band from Rice husk is less intense compared to the O-H band of Sipernat310. The bands with maxima at 1104 (WN region II) and 469 cm^− 1 (WN region III) deviate less than 8 cm^− 1 in WN and slightly in intensity. With the intensity of the FTIR spectra of Rice husk being higher. The FTIR spectra of Sipernat310 shows a band maxima at 967 cm^− 1, which the FTIR spectra of rice husk doesn’t have. The repeated measurements of all different amorphous silica are in strong agreement with each other but deviate slightly in intensity. FTIR spectra of Sipernat310, Sipernat50 and Aerosil300 are in agreement with spectra from the literature¹². The FTIR spectra of natural amorphous silica is in agreement with other studies²⁰. As well as the FTIR spectra of rice husk²¹.

FTIR spectra of clay minerals

The FTIR spectra of kaolin(A) (Fig. 2a; Figure S5; n = 2) shows pronounced O-H bands at WN region I (3750 to 3000 cm^− 1) with two pronounced band maxima located at 3694 (I_a) and 3619 cm^− 1 (I_b). Band (I) is caused by Si-OH stretching vibrations, where the Si-OH groups are located between the different layers of kaolin. The band (II) also caused by Si-OH stretching vibrations, where the Si-OH groups are in the layers of Kaolin¹⁸. Two smaller bands are located between band I and II. WN region I showed another broad band with a band maximum at 3439 cm^− 1 (I_c) (Table 1). WN region II has two pronounced bands with maxima located at 1032 (II_b) and 1009 cm^− 1 (II_c) (Table 1). Bands II_b and II_c are each accompanied by a second but less pronounced band maximum at 1114 (II_a) and 913 cm^− 1 (II_d). Band II_a has a second maximum and band II_d has a shoulder (Fig. 2b; red arrows). WN region III shows two sharp and intense bands with maxima located at 539 (III_a) and 470 cm^− 1 (III_b). The FTIR spectra of kaolin(B) (Fig. 2a; Figure S6; n = 2) shows pronounced O-H bands in the WN region I (3750 to 3000 cm^− 1) with band maxima at 3696 and 3619 cm^− 1 and a broad band with a band maximum at 3448 cm^− 1. The Si-O band in WN region II is a double band with the maxima at 1031 and 1007 cm^− 1 (Table 1). WN region III shows two pronounced bands with maxima located at 537 cm^− 1 and 469 cm^− 1. The FTIR spectra of kaolin(B) and kaolin(A) show strong similarities. The band maxima are almost located at the same WN (deviation between the WN is smaller than the resolution (r = 2 cm^− 1)). But the smaller band (band II_a) surrounding the double band with a maximum at 1113 cm^− 1 is not accompanied by a second maxima (Fig. 2b; red arrow). The FTIR spectra of kaolin(C) (Fig. 2; Figure S7; n = 2) shows pronounced O-H bands in the WN region I (3750 to 3000 cm^− 1) with maxima at 3694 and 3619 cm^− 1 and a broad band with a band maximum at 3448 cm^− 1. In WN region II the Si-O band is a double band with the maxima at 1032 and 1010 cm^− 1 (Table 1). WN region III shows two pronounced bands with maxima at 539 cm^− 1 and 470 cm^− 1. The FTIR spectra of kaolin(C) (Fig. 2; n = 2) show strong similarities with kaolin(A). The wavenumbers of the band maxima are in the same wavenumber range (deviation between the WN is smaller than the Resolution (r= 2 cm^− 1)) and the intensities of the band maxima are very similar, except for the band maxima at 1115 (II_a) and 1103 cm^− 1, where the intensity of kaolin(A) is higher than the intensity of Kaolin(C) (Fig. 2.b; red arrow).

In the FTIR spectrum of the sample kaolin C the lower intensity at WN 1114 cm^− 1 as compared to that in the spectra of kaolin A or kaolin B (Fig. 2b) may be caused by differences in cation composition of the investigated kaolin samples. The kaolin samples investigated here are collected at different sites and can be assumed to be different in their cation composition¹⁹. Differences in cation composition cause differences in the electron density distribution within the kaolin´s structure which in turn affect the dipole-moment of the kaolin functional groups. Since the dipole-moment is related to the extinction coefficient in Lambert Beer’s law²², that describes the relation between absorption band intensity and substance concentration²³, the cation induces difference in the dipole-moment may explain the differences in absorption band intensity observed here. A comparable difference was shown by Sánchez-Sánchez, et al²⁴. in FTIR spectra of Na₂CO₃ compared with that of CaCO₃: The intensity of one carbonate absorption band is smaller for the Na₂CO₃ as compared to CaCO₃.

The FTIR spectra of montmorillonite(A) (Fig. 3a; Figure S8; n = 2) showed a broad O-H band at WN range I (3750 to 3000 cm⁻1) with two band maxima located at 3630 cm^− 1 (I₁) and 3421 cm^− 1 (I₂). In WN range II the Si-O band has two maxima located at 1090 and 1048 cm^− 1. Band II is accompanied by 2 smaller band maxima located at 1200 and 918 cm^− 1 (Table 1). WN range III shows a sharp intense δ-Si-O-Si band with a maximum located at 471 cm^− 1 (III₂). Band III₂ is accompanied by a sharp and less intense Al-O-Si band with a maximum located at 522 cm^− 1 (III₁). The FTIR spectra of montmorillonite(B) (Fig. 3a; Figure S9; n = 2) shows a broad O-H band at WN range I (3750 to 2700 cm⁻1) with two band maxima located at 3621 cm^− 1 (I₁) and 3438 cm^− 1 (I₂) (Table 1). WN range II sows the Si-O band with a band maximum located at 1048 cm^− 1. WN range III shows a sharp intense δ-Si-O-Si band with a maximum located at 468 cm^− 1 (III₂). Band III₂ is accompanied by a sharp and less intense Al-O-Si band with a maximum located at 525 cm^− 1 (III₁) (Table 1). The FTIR spectra of montmorillonite(B) shows similarities to the FTIR spectra of montmorillonite(A). Montmorillonite(A) is a bentonite, which is why there are deviations between montmorillonite(A) and montmorillonite(B). The biggest deviation is, that the Si-O band of montmorillonite(B) shows only a single band maximum at WN 1048 cm^− 1 (II). While the Si-O band in montmorillonite(A) has 2 band maxima (II₁ and II₂) (see Fig. 3b). This is probably because of the other minerals containing in montmorillonite(A). Other deviations are the band maxima at 3621 (I₁) deviates by 9 cm^− 1 (4.5 times resolution) from montmorillonite(A). Also band (I₁) has a small shoulder at 3691 cm^− 1. The band located at 3438 cm^− 1 (I₂) deviates by 17 cm^− 1 (8.5 times resolution) from montmorillonite(A). Band I from montmorillonite(B) is also higher in intensity than band I from montmorillonite(A). This could be because of different external OH groups from water in the samples. In the WN range III the band maxima of montmorillonite(B) are within the resolutions of the band maxima of montmorillonite(A). The intensity of the bands in WN range III is lower compared to montmorillonite(A).

The FTIR spectra of illite and vermiculite showed similarities. Both have a band maximum located in WN region I at a similar WN, but the intensity of the vermiculite band is much higher (see Fig. S10-S12). The illite band also shows a second band maximum in region I of the WN at a similar wavelength to montmorillonite (A). Illite and vermiculite show a similar band in WN region II, with a small WN shift towards higher WN for illite (Table 1). In WN region III, illite shows similarities with the montmorillonite (A) FTIR spectra, but the intensity of the bands is much lower. Vermiculite shows only one band, with a fine structure visible in the maximum of the band.

The replicate measurements of kaolin, montmorillonite, vermiculite and illite are in strong agreement with each other, only changing in intensity for some bands. The FTIR spectra of all analysed kaolin¹⁸, montmorillonite²⁵, vermiculite and illite^17,²⁶^,27, are in agreement with spectra from the literature.

Primary silicates

Both biotite and olivine showed a small band in WN region I with low intensities (Figure S13-15) (Table 1). In WN region II, the FITR spectra of biotite and olivine are dissimilar: biotite shows a single band, while olivine shows multiple bands. However, the smaller olivine band is at a similar wavelength to the biotite band. In WN region III, olivine and biotite show no similarities, with olivine having three band maxima and biotite having only one at a different wavelength (Table 1). The replicate measurements of the individual minerals biotite and olivine are in strong agreement with each other and comparable with spectra from the literature²⁸^,29.

Comparison between silicate minerals

When comparing the silicate minerals and the amorphous silicates large differences in the FTIR spectra were found (Figs. 1 and 4). Multiple structural differences of the amorphous silicates are visible in the FTIR spectra (Fig. 1).

Structural differences being e.g. the number of layers, as can be seen in Fig. 4, which compares the two-layer clay mineral kaolin with the three-layer clay mineral montmorillonite. The difference in the FTIR spectra of kaolin and montmorillonite suggests that the structure has a significant impact on the FTIR spectra. Another structural difference is the different stages in the weathering process (see Figure S13, in which primary silicates such as olivine and biotite are compared with the clay mineral montmorillonite). Biotite is a three-layer mineral like montmorillonite, whereas olivine is an nesosilicate or orthosilicate. The FTIR spectra of montmorillonite showed a more intense O-H band, and the ν- Si-O-Si bands of all three minerals are at different WN, with montmorillonite having the highest WN and olivine the lowest. Another structural difference is that different cation concentrations in various clay minerals (Figure S10) show different FTIR spectra.

Illite, vermiculite and montmorillonite are all three-layer clay minerals with different cation concentrations¹⁵. The FTIR spectra of all three clay minerals showed similarities (Figure S10). The O-H band of vermiculite was the most intense. The Si-O-Si bands of the three-layer clay minerals were slightly shifted, with the maximum of the montmorillonite band having the highest WN and the maximum of the vermiculite band having the lowest WN. The difference between the band maxima of vermiculite and montmorillonite are 48 cm^− 1 WN. The structural differences are assumed to be responsible for the differences in the FTIR spectra. However, a study in which all parameters except the one under investigation are consistent is needed to confirm this. Such a study would allow the impact of weathering, cation concentration and the number of layers to be tested independently.

Comparison of mixed samples

Note that all spectra have been normalised to the band maximum at WN 1032 cm^− 1. During the formation of the mixed samples a rotating mixer was used to prepare the mixed samples. However, the different densities of the samples used can lead to stronger sedimentation of the densest. In addition, Sipernat is highly electrostatic. This means that, even when using a rotating mixer for 24 h, heterogeneous mixtures may occur. This may result in discrepancies between the actual and expected concentrations of the samples.

The spectra of kaolin and Sipernat310 mixtures showed differences compared to that of kaolin (Fig. 5). The most distinct differences with increasing Sipernat310 content were observed for the bands at WN 3448 to 3439 cm^− 1 and WN 1115 to 1099 cm^− 1. The intensity of these bands increases with Sipernat310 content, because they are characteristic for Sipernat310. Additionally, the WN of these band maxima shifted from 3448 to 3439 cm^− 1 and from 1115 to 1099 cm^− 1. The bands typical for kaolin located at WN 796, 469 and 432 cm^− 1 decrease with increasing Sipernat310 content, indicating the relative decrease of kaolin in the mixtures. The decrease in the intensity of bands typical for kaolin is relatively lower as compared to the increase in the intensity of bands typical for Sipernat310, as the relative percentage increase for Sipernat310 is significantly higher than the relative percentage decrease for kaolin. In relation to the band at WN 1032 cm^− 1 no statement can be made, since the spectra where normalized to this band.

The spectra of montmorillonite and Sipernat310 mixtures showed differences compared to that of montmorillonite (Fig. 6). The most distinct differences with increasing Sipernat310 content were observed for the bands at WN 3431, 1042, 523 and 471 cm^− 1. The intensity of the band at 3431 cm^− 1 increased with Sipernat310 content, since this band is more intense in pure Sipernat310 compared to the spectra of montmorillonite(A). The bands typical for montmorillonite(A) located at WN 1042, 918, 624, 523 and 471 cm^− 1 decreased with increasing Sipernat310 content, indicating the relative decrease of montmorillonite in the mixtures. The decrease in the intensity of bands typical for montmorillonite(A) is similar compared to the increase in the intensity of bands typical for Sipernat310. In relation to the band at WN 1092 cm^− 1 no statement can be made, since the spectra were normalized to this band. The decrease in the intensity of bands typical for montmorillonite(A) is similar compared to the increase in the intensity of bands typical for Sipernat310. In relation to the band at WN 1092 cm-1 no statement can be made, since the spectra where normalized to this band.

Differences between kaolin(A) plus sipernat310 and montmorillonite(A) plus sipernat310

Comparing mixtures of Sipernat310 plus kaolin and Sipernat310 plus montmorillonite shows that band maxima vary depending on the concentration of Sipernat310 (Fig. 7). The variation in intensity is significantly greater in the kaolin mixtures than in the montmorillonite mixtures.

Fig. 7 — FTIR spectra of mixed samples with Sipernat310 plus kaolin(A) plus montmorillonite(A) from WN range 4000 to 2800 cm^-1 and 1500 to 400 cm^-1.

The FTIR spectra of the mixture of Sipernat310 plus montmorillonite plus kaolin shows features from the kaolin spectra at 3694 to 3619, 1034, 913, 537 and 470 cm^− 1. It also shows features of the Sipernat310 and montmorillonite FTIR spectra at WN 3438, 1096, 796, 470. The differentiation between these two is not as distinct as the distinction from kaolin.

Analysing the mixed samples with partial least squares regression (using TQ analyst)

Three replicates of each mixture were used to calculate the ASi concentration of the samples (Fig. 8). These included mixtures of Sipernat310 plus kaolin, Sipernat310 plus montmorillonite, mixture of Sipernat310 plus montmorillonite plus kaolin and the raw spectra of Sipernat310, kaolin(A) and montmorillonite(A). Replicates A and B of the mixtures were used for calibration and replicate C for validation. The model had a Root Mean Square Error of Calibration (RMSEC) of 2.08 and a Root Mean Square Error of Prediction (RMSEP) of 3.08 for calculating the ASi contents (Fig. 8; Figure S19-S20; Table 2), with an correlation coefficient of R = 0.99. Hence, our model could predict the share of ASi (Sipernat310) and different minerals in mixtures very well.

Table 2.

Modelled share (TQ Analyst) of mixture (33:33:33; 50:25:25; 1:2:2) of Sipernat310 (ASi), montmorillonite and kaolin compared to the share determined by FTIR in three replicates (A, B, C).

1-1-1	Sipernat310	Kaolin(A)	Montmorillonit(A)
expected	33.3	33.3	33.3
A	32.97	32.31	33.66
B	32.1	32.73	34.79
C	34.33	30.18	36.04
1-0.5.5-0.5	Sipernat310	Kaolin(A)	Montmorillonit(A)
expected	50	25	25
A	44.2	29.04	26.89
B	49.62	24.5	25.63
C	50.89	26.55	23.49
1–2-2	Sipernat310	Kaolin(A)	Montmorillonit(A)
expected	20	40	40
A	17.18	41.1	41.14
B	17.23	40.77	42.85
C	20.57	38.41	41.53

Open in a new tab

Additionally, we built a model using all the three replicates of the two and three component mixtures (except for the three component mixture of 40% kaolin, 40% montmorillonite and 20% Sipernat310) for a calibration curve that showed a Root Mean Square Error of Calibration (Rmsec) of 3.45 and a correlation coefficient of 0.99. When using the spectral data of the three component mixture of 40% kaolin, 40% montmorillonite and 20% Sipernat310 the spectral data were found to allow an estimation of the Sipernat310 weight in the respective mixtures with a Root Mean Squared Error of Prediction of 1.65 with values for the 20% Sipernat310 of 21.37 (sample A), 17.51 (sample B) and 19.77 (sample C).

Conclusion

The study demonstrated that the ASi content of simple mineral mixtures can be calculated with a high degree of accuracy using FTIR and partial least square regression. The next step is to test how minerals from different sites differ and how this affects the analysis of ASi concentration. It would also be useful to investigate how mixtures behave when they become more complex due to the addition of more minerals, as well as changes in environmental parameters such as pH value, temperature and water content. Another study investigated the use of FTIR and PLS to analyse more complex sediment samples. Silicate minerals proved to be a challenge, as they have similar absorption bands in similar WN ranges³⁰. Further research on these topics would be useful.

Materials and methods

The Sipernat310, Sipernat50 and Aerosil300 samples (all Evonik Operations AG, Germany) were used as offered by the company (white powders with particle sizes between 2 and 10 μm). As minerals kaolin from Erbslöh (Germany) (K(A)), kaolin from Quarzwerke (Kemmlitz, Germany) (K(B)) and kaolin from Gluhivzi (Ukraine) (K(C)), Bentonit from Sigma-Aldrich (Türkiye) with an montmorillonite concentration around 85% (M(A)) and montmorillonite from Sigma-Aldrich (Germany) (M(B)), vermiculite from Pasiora, olivine from Dieter Welsch GmbH (Germany), biotite from Kremer (Canada), and natural amorphous silicon extracted from plants using³¹, illite and rice husks (Green Sugar Technologies, Dresden, Germany) were used.

All mineral samples were gently ground by hand in an agate mortar for about 5 min to obtain a powder (particles sizes between 2 and about 10 μm) and subsequently analyzed using an IS20 FTIR spectrometer (Thermo Fisher Scientific, Dreieich, Germany). The different samples were carefully mixed with KBr in different ratios depending on how they interact with the FTIR (shown in Table 3).

Table 3.

Ratios of minerals to KBr powder used for FTIR analysis.

Mixing ratio: sample: KBr	1:50 (2 mg sample per 98 mg KBr)	1:100 (1 mg sample per 99 mg KBr)	1:200 (mg sample per 199 mg KBr)
Sample	Vermiculite	Montmorillonite	Kaolin
	Biotite	Sipernat50	Illite
	Olivine		rice husk
			Sipernat310
			Aerosil300
			natural ASi

Open in a new tab

The mixtures were dried for 12 h over silica gel in a desiccator to ensure consistent water content, after which they were pressed into pellets. This can result in inhomogeneities in the pellets, which can falsify the measurements¹⁴. The resulting KBr pellets containing the minerals were analyzed using transmission mode²³. All spectra were recorded in triplicate at a resolution of 2 cm^–1, with 120 scans (= 120 repetitions of a single spectra) to acquire the absorption spectra over a wavenumber range of 4000 to 400 cm^–1. All spectra were corrected against atmosphere³², and auto-baseline-corrected using Thermo Fisher scientific software (OMNIC). The auto-baseline-corrected spectra was analyzed for bending and stretching vibrations caused by the SiO₄ tetrahedron and the AlO₆ octahedra.

Mixed samples consisting of Sipernat310 plus kaolin (A), Sipernat310 plus montmorillonite (A) and Sipernat310 plus kaolin (A) plus montmorillonite (A) were created.

The minerals kaolin (A) and montmorillonite (A) were used because they are both common minerals in soil, and the FTIR spectra of kaolin are distinctively different from that of Sipernat310 and can easily be identified. Whereas, the FTIR spectra of montmorillonite are more similar to the FTIR spectra of Sipernat310. Such it was expected that kaolin would be easier to distinguish from the Sipernat310 spectra and that the limitations of this method would be better tested using a mineral with a similar FTIR spectrum to Sipernat310, such as montmorillonite. For interpreting the spectra of the mixtures, we focused on absorption bands typical for Sipernat310 (at WN 796, 469 and 432 cm^− 1; ¹⁹) and kaolin since overlapping absorption bands may hamper an assignment towards defined organic or inorganic components²⁴.

The mixtures were created using the clay mineral concentrations shown in Table 4 and Table 5. These mixtures were mixed using a rotating mixer for 24 h. The mixtures were analysed using the same procedure described above.

Table 4.

Mixtures of different concentrations of ASi (Sipernat 310) added to kaolin and montmorillonite used to create prediction.

Two component mixtures
Mixture name	Kaolin	Sipernat310	Mixture name	Montmorillonite	Sipernat310
Kaolin 50%	50 g	50 g	Montmorillonite 50%	50 g	50 g
Kaolin 62.5	62.5	37.5	Montmorillonite 62.5%	62.5 g	37.5 g
Kaolin 75%	75	25	Montmorillonite 75%	75 g	25 g
Kaolin 87.5%	87.5	12.5	Montmorillonite 87.5%	87.5 g	12.5 g
Kaolin 90%	90	10	Montmorillonite 90%	90 g	10 g
Kaolin 95%	95	5	Montmorillonite 95%	95 g	5 g
Kaolin 97%	97	3	Montmorillonite 97%	97 g	3 g
Kaolin 98%	98	2	Montmorillonite 98%	98 g	2 g
Kaolin 99%	99	1	Montmorillonite 99%	99 g	1 g

Open in a new tab

Table 5.

Mixtures of different concentrations of ASi (Sipernat 310) added to kaolin and montmorillonite used to create prediction models”.

Three component mixtures
Mixture name	Sipernat310 (S)	Kaolin (K)	Montmorillonite (M)
S-K-M (20–40-40)	20 g	40 g	40 g
S-K-M (33-33-33)	33.3 g	33.3 g	33.3 g
S-K-M (50-25-25)	50 g	25 g	25 g

Open in a new tab

Partial least-squares regression was applied using TQ Analyst (Thermo Fisher) to calibrate the spectral data with the reference values (i.e., percentual addition of ASi). These models were used to predict ASi contents of the mineral mixtures. All mixtures with ASi (Sipernat 310) were prepared in three replicates. The spectral data of the 1 st and 2nd replicate of the two component mixtures (kaolin plus ASi; montmorillonite plus ASi) were used to determine calibration curves that allows to estimate the %-weights of ASi in the respective mixtures. The calibration curve was determined by the help of statistics offered by the TQ analyst software package (https://www.thermofisher.com/order/catalog/product/de/de/IQLAADGABZFAIXMAVF). Afterwards, these calibration curves were used to estimate the ASi weight from the spectral data of (i) the 3rd replicate of the two component mixtures and from those (ii) of the three components mixture (kaolin plus montmorillonite plus ASi) (Validation). Additionally, we tested a model using all the three replicates of the two and three component mixtures (except for the three component mixture of 40% kaolin, 40% montmorillonite and 20% Sipernat310) for a calibration curve Sipernat310-kaolin-montmorillonite (20%−40%−40%) for validation Table 5.

Supplementary Information

Below is the link to the electronic supplementary material.

Supplementary Material 1^{(905.6KB, pdf)}

Author contributions

JS and MS conceived the study. OH did the measurements, calculation and wrote the first draft of the manuscript with help of RE. JS, RE, OH, MS and CWM revised the manuscript and gave their approval to the final version of the manuscript.

Funding

Open Access funding enabled and organized by Projekt DEAL.

Data availability

The datasets used and/or analyzed during the current study is available from the corresponding author on reasonable request.

Declarations

Competing interests

The authors declare no competing interests.

Footnotes

Publisher’s note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

References

1.Schaller, J. et al. The importance of reactive silica for maintaining soil health. Plant Soil513, 1651–1662 (2025). [Google Scholar]
2.Schaller, J., Cramer, A., Carminati, A. & Zarebanadkouki, M. Biogenic amorphous silica as main driver for plant available water in soils. Sci. Rep.10, 2424 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
3.Barbosa, L. A. P., Stein, M., Gerke, H. H. & Schaller, J. Synergistic effects of organic carbon and silica in preserving structural stability of drying soils. Sci. Rep.10.1038/s41598-024-58916-9 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
4.Schaller, J. et al. Increased wheat yield and soil C stocks after silica fertilization at the field scale. Sci. Total Environ.887, 163986 (2023). [DOI] [PubMed] [Google Scholar]
5.Schaller, J. et al. Silica fertilization improved wheat performance and increased phosphorus concentrations during drought at the field scale. Sci. Rep.11, 20852 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
6.Schaller, J., Webber, H., Ewert, F., Stein, M. & Puppe, D. The transformation of agriculture towards a silicon improved sustainable and resilient crop production. npj Sustain. Agric.2, 27 (2024). [Google Scholar]
7.Stein, M., Puppe, D., Kaczorek, D., Buhtz, C. & Schaller, J. Silicon extraction from x-ray amorphous soil constituents: A method comparison of alkaline extracting agents. Front. Environ. Sci.10.3389/fenvs.2024.1389022 (2024). [Google Scholar]
8.Ellerbrock, R., Stein, M. & Schaller, J. Comparing amorphous silica, short-range-ordered silicates and silicic acid species by FTIR. Sci. Rep.12, 11708 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
9.McMillan, P. F. & Piriou, B. Raman spectroscopic studies of silicate and related glass structure: A review. Bull. Minéral.106, 57–75 (1983). [Google Scholar]
10.Liu, H., Kaya, H., Lin, Y. T., Ogrinc, A. & Kim, S. H. Vibrational spectroscopy analysis of silica and silicate glass networks. J. Am. Ceram. Soc.105, 2355–2384 (2022). [Google Scholar]
11.Gunde, M. K. Vibrational modes in amorphous silicon dioxide. Physica B292, 286–295 (2000). [Google Scholar]
12.Ellerbrock, R. H., Stein, M. & Schaller, J. Comparing silicon mineral species of different crystallinity using Fourier transform infrared spectroscopy. Frontiers in Environmental Chemistry5, 1462678 (2024). [Google Scholar]
13.Madejová, J. FTIR techniques in clay mineral studies. Vib. Spectrosc.31, 1–10 (2003). [Google Scholar]
14.Kaufhold, S., Hein, M., Dohrmann, R. & Ufer, K. Quantification of the mineralogical composition of clays using FTIR spectroscopy. Vib. Spectrosc.59, 29–39 (2012). [Google Scholar]
15.Amelung, W. Scheffer/Schachtschabel Lehrbuch der Bodenkunde (Springer-Verlag, 2018). [Google Scholar]
16.Schaller, J., Puppe, D., Kaczorek, D., Ellerbrock, R. & Sommer, M. Silicon cycling in soils revisited. Plants10, 295 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
17.Caccamo, M. T., Mavilia, G., Mavilia, L., Lombardo, D. & Magazù, S. Self-assembly processes in hydrated montmorillonite by FTIR investigations. Materials13, 1100 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
18.Vaculikova, L., Plevova, E., Vallova, S. & Koutnik, I. Characterization and differentiation of kaolinites from selected Czech deposits using infrared spectroscopy and differential thermal analysis. (2011).
19.Farmer, V. C. The infrared spectra of minerals. Mineralogical Soc. Monogr.4, 331–363 (1974). [Google Scholar]
20.Vogel, H., Meyer-Jacob, C., Thöle, L., Lippold, J. A. & Jaccard, S. L. Quantification of biogenic silica by means of Fourier transform infrared spectroscopy (FTIRS) in marine sediments. Limnol. oceanography: methods. 14, 828–838 (2016). [Google Scholar]
21.Atta, A., Jibril, B., Aderemi, B. & Adefila, S. Preparation of analcime from local kaolin and rice husk ash. Appl. Clay Sci.61, 8–13 (2012). [Google Scholar]
22.Nitzan, M. et al. Calibration-free pulse oximetry based on two wavelengths in the infrared—A preliminary study. Sensors14, 7420–7434 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]
23.Hesse, M., Meier, H. & Zeeh, B. Spektroskopische Methoden in der organischen Chemie (Georg Thieme Verlag, 2005). [Google Scholar]
24.Sánchez-Sánchez, A., Cerdán, M., Jordá, J., Amat, B. & Cortina, J. Characterization of soil mineralogy by FTIR: Application to the analysis of mineralogical changes in soils affected by vegetation patches. Plant. Soil.439, 447–458 (2019). [Google Scholar]
25.Banik, N. et al. Synthesis and characterization of organoclay modified with cetylpyridinium chloride. Bangladesh J Sci Ind Res50, 65–70 (2015). [Google Scholar]
26.Tran, L., Wu, P., Zhu, Y., Liu, S. & Zhu, N. Comparative study of Hg (II) adsorption by thiol-and hydroxyl-containing bifunctional Montmorillonite and vermiculite. Appl. Surf. Sci.356, 91–101 (2015). [Google Scholar]
27.Zviagina, B. B., Drits, V. A. & Dorzhieva, O. V. Distinguishing features and identification criteria for K-dioctahedral 1 M micas (illite-aluminoceladonite and illite-glauconite-celadonite series) from middle-infrared spectroscopy data. Minerals10, 153 (2020). [Google Scholar]
28.Wei, S. et al. The influence of metal cations on the dissolution and transformation of biotite. Mineralogical Magazine89(1), 12–24 (2025). [Google Scholar]
29.Burns, R. G. & Huggins, F. E. Cation determinative curves for Mg-Fe-Mn olivines from vibrational spectra. American Mineralogist: Journal of Earth and Planetary Materials57, 967–985 (1972). [Google Scholar]
30.Hahn, A. et al. Using Fourier transform infrared spectroscopy to determine mineral phases in sediments. Sediment. Geol.375, 27–35 (2018). [Google Scholar]
31.Puppe, D. et al. How big is the influence of biogenic silicon pools on short-term changes in water-soluble silicon in soils? Implications from a study of a 10-year-old soil-plant system. Biogeosciences10.5194/bg-14-5239-2017 (2017). [Google Scholar]
32.Haberhauer, G. & Gerzabek, M. Drift and transmission FT-IR spectroscopy of forest soils: An approach to determine decomposition processes of forest litter. Vib. Spectrosc.19, 413–417 (1999). [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary Material 1^{(905.6KB, pdf)}

Data Availability Statement

The datasets used and/or analyzed during the current study is available from the corresponding author on reasonable request.

[CR1] 1.Schaller, J. et al. The importance of reactive silica for maintaining soil health. Plant Soil513, 1651–1662 (2025). [Google Scholar]

[CR2] 2.Schaller, J., Cramer, A., Carminati, A. & Zarebanadkouki, M. Biogenic amorphous silica as main driver for plant available water in soils. Sci. Rep.10, 2424 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR3] 3.Barbosa, L. A. P., Stein, M., Gerke, H. H. & Schaller, J. Synergistic effects of organic carbon and silica in preserving structural stability of drying soils. Sci. Rep.10.1038/s41598-024-58916-9 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR4] 4.Schaller, J. et al. Increased wheat yield and soil C stocks after silica fertilization at the field scale. Sci. Total Environ.887, 163986 (2023). [DOI] [PubMed] [Google Scholar]

[CR5] 5.Schaller, J. et al. Silica fertilization improved wheat performance and increased phosphorus concentrations during drought at the field scale. Sci. Rep.11, 20852 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR6] 6.Schaller, J., Webber, H., Ewert, F., Stein, M. & Puppe, D. The transformation of agriculture towards a silicon improved sustainable and resilient crop production. npj Sustain. Agric.2, 27 (2024). [Google Scholar]

[CR7] 7.Stein, M., Puppe, D., Kaczorek, D., Buhtz, C. & Schaller, J. Silicon extraction from x-ray amorphous soil constituents: A method comparison of alkaline extracting agents. Front. Environ. Sci.10.3389/fenvs.2024.1389022 (2024). [Google Scholar]

[CR8] 8.Ellerbrock, R., Stein, M. & Schaller, J. Comparing amorphous silica, short-range-ordered silicates and silicic acid species by FTIR. Sci. Rep.12, 11708 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR9] 9.McMillan, P. F. & Piriou, B. Raman spectroscopic studies of silicate and related glass structure: A review. Bull. Minéral.106, 57–75 (1983). [Google Scholar]

[CR10] 10.Liu, H., Kaya, H., Lin, Y. T., Ogrinc, A. & Kim, S. H. Vibrational spectroscopy analysis of silica and silicate glass networks. J. Am. Ceram. Soc.105, 2355–2384 (2022). [Google Scholar]

[CR11] 11.Gunde, M. K. Vibrational modes in amorphous silicon dioxide. Physica B292, 286–295 (2000). [Google Scholar]

[CR12] 12.Ellerbrock, R. H., Stein, M. & Schaller, J. Comparing silicon mineral species of different crystallinity using Fourier transform infrared spectroscopy. Frontiers in Environmental Chemistry5, 1462678 (2024). [Google Scholar]

[CR13] 13.Madejová, J. FTIR techniques in clay mineral studies. Vib. Spectrosc.31, 1–10 (2003). [Google Scholar]

[CR14] 14.Kaufhold, S., Hein, M., Dohrmann, R. & Ufer, K. Quantification of the mineralogical composition of clays using FTIR spectroscopy. Vib. Spectrosc.59, 29–39 (2012). [Google Scholar]

[CR15] 15.Amelung, W. Scheffer/Schachtschabel Lehrbuch der Bodenkunde (Springer-Verlag, 2018). [Google Scholar]

[CR16] 16.Schaller, J., Puppe, D., Kaczorek, D., Ellerbrock, R. & Sommer, M. Silicon cycling in soils revisited. Plants10, 295 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR17] 17.Caccamo, M. T., Mavilia, G., Mavilia, L., Lombardo, D. & Magazù, S. Self-assembly processes in hydrated montmorillonite by FTIR investigations. Materials13, 1100 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR18] 18.Vaculikova, L., Plevova, E., Vallova, S. & Koutnik, I. Characterization and differentiation of kaolinites from selected Czech deposits using infrared spectroscopy and differential thermal analysis. (2011).

[CR19] 19.Farmer, V. C. The infrared spectra of minerals. Mineralogical Soc. Monogr.4, 331–363 (1974). [Google Scholar]

[CR20] 20.Vogel, H., Meyer-Jacob, C., Thöle, L., Lippold, J. A. & Jaccard, S. L. Quantification of biogenic silica by means of Fourier transform infrared spectroscopy (FTIRS) in marine sediments. Limnol. oceanography: methods. 14, 828–838 (2016). [Google Scholar]

[CR21] 21.Atta, A., Jibril, B., Aderemi, B. & Adefila, S. Preparation of analcime from local kaolin and rice husk ash. Appl. Clay Sci.61, 8–13 (2012). [Google Scholar]

[CR22] 22.Nitzan, M. et al. Calibration-free pulse oximetry based on two wavelengths in the infrared—A preliminary study. Sensors14, 7420–7434 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR23] 23.Hesse, M., Meier, H. & Zeeh, B. Spektroskopische Methoden in der organischen Chemie (Georg Thieme Verlag, 2005). [Google Scholar]

[CR24] 24.Sánchez-Sánchez, A., Cerdán, M., Jordá, J., Amat, B. & Cortina, J. Characterization of soil mineralogy by FTIR: Application to the analysis of mineralogical changes in soils affected by vegetation patches. Plant. Soil.439, 447–458 (2019). [Google Scholar]

[CR25] 25.Banik, N. et al. Synthesis and characterization of organoclay modified with cetylpyridinium chloride. Bangladesh J Sci Ind Res50, 65–70 (2015). [Google Scholar]

[CR26] 26.Tran, L., Wu, P., Zhu, Y., Liu, S. & Zhu, N. Comparative study of Hg (II) adsorption by thiol-and hydroxyl-containing bifunctional Montmorillonite and vermiculite. Appl. Surf. Sci.356, 91–101 (2015). [Google Scholar]

[CR27] 27.Zviagina, B. B., Drits, V. A. & Dorzhieva, O. V. Distinguishing features and identification criteria for K-dioctahedral 1 M micas (illite-aluminoceladonite and illite-glauconite-celadonite series) from middle-infrared spectroscopy data. Minerals10, 153 (2020). [Google Scholar]

[CR28] 28.Wei, S. et al. The influence of metal cations on the dissolution and transformation of biotite. Mineralogical Magazine89(1), 12–24 (2025). [Google Scholar]

[CR29] 29.Burns, R. G. & Huggins, F. E. Cation determinative curves for Mg-Fe-Mn olivines from vibrational spectra. American Mineralogist: Journal of Earth and Planetary Materials57, 967–985 (1972). [Google Scholar]

[CR30] 30.Hahn, A. et al. Using Fourier transform infrared spectroscopy to determine mineral phases in sediments. Sediment. Geol.375, 27–35 (2018). [Google Scholar]

[CR31] 31.Puppe, D. et al. How big is the influence of biogenic silicon pools on short-term changes in water-soluble silicon in soils? Implications from a study of a 10-year-old soil-plant system. Biogeosciences10.5194/bg-14-5239-2017 (2017). [Google Scholar]

[CR32] 32.Haberhauer, G. & Gerzabek, M. Drift and transmission FT-IR spectroscopy of forest soils: An approach to determine decomposition processes of forest litter. Vib. Spectrosc.19, 413–417 (1999). [Google Scholar]

PERMALINK

Analyzing the share of amorphous silica in mixtures with different soil minerals using fourier transform infrared spectroscopy and PLSR chemometrics

Oliver Hunfeld

Ruth H Ellerbrock

Mathias Stein

Carsten W Müller

Jörg Schaller

Abstract

Supplementary Information

Introduction

Results and discussion

General band assignment

Fig. 2.

Amorphous silica

Fig. 1.

Table 1.

FTIR spectra of clay minerals

Fig. 3.

Primary silicates

Comparison between silicate minerals

Fig. 4.

Comparison of mixed samples

Fig. 5.

Fig. 6.

Differences between kaolin(A) plus sipernat310 and montmorillonite(A) plus sipernat310

Fig. 7.

Analysing the mixed samples with partial least squares regression (using TQ analyst)

Fig. 8.

Table 2.

Conclusion

Materials and methods

Table 3.

Table 4.

Table 5.

Supplementary Information

Author contributions

Funding

Data availability

Declarations

Competing interests

Footnotes

References

Associated Data

Supplementary Materials

Data Availability Statement

ACTIONS

PERMALINK

RESOURCES

Similar articles

Cited by other articles

Links to NCBI Databases