Discarded oyster shell, Glyphosate [N-(Phosphonomethyl) glycine, C₃H₈NO₅P, 95%] was from Shi-Feng Biotechnology Co., Ltd., Shanghai, PRC; Tetrabutyl orthotitanate (TBOT), Absolute ethanol (C₂H₅OH) as solvent, Cerium nitrate hexahydrate [Ce(NO₃)₃ • 6H₂O] and urea (CH₄N₂O) were used as the precursor of cerium and nitrogen, respectively, hydrochloric acid (HCl). All chemicals were analytical grade purity and can be directly applied without any further treatment.

The surface of the discarded oyster shells was removed with a steel wire ball, polished to a white colour with a cutting machine and cut into small pieces, dried in a blast oven at 65°C, crushed in a sealed sample miller for 3 min–5 min, sieved, soaked in 0.1% dilute hydrochloric acid, vacuum dried and calcined in a muffle furnace at 900°C to obtain modified oyster shell powder, which was sealed and stored.

CeNT@Oys was prepared by a simple hydrothermal method in a typical procedure as follows: Solutions A was prepared by mixing 36 mL of anhydrous ethanol and a quantitative amount of modified oyster shell powder by ultrasonic shaking and then adding 10 mL of tetrabutyl titanate to it and stirring evenly. Solution B was then slowly added to Solution A at a rate of 1 drop/s and stirred for two hours at room temperature. To 120°C in a vacuum oven for 16 h and then removed and cooled, the resulting precipitate was filtered several times with deionised water and dried overnight at 60°C. To allow complete crystallisation of the sample, the semi-manufactured goods was calcined in a muffle furnace for three hours at a fixed temperature.

The resulting composite nano-photocatalyst is abbreviated as x-CeNT@Oys-y, where x denotes the Ce doping molar ratio (%), and y denotes the calcination temperature (°C). In addition, Single N-doped (N-TiO₂@Oys), Single Ce-dope (Ce-TiO₂@Oys), catalysts unloaded with modified oyster shell powder and pure TiO₂, were all prepared uniformly using the above procedure.

The crystal phase composition and crystallinity of composites with different calcination temperatures were analyzed by analyzing the data of powder X-ray diffraction (Rigaku, Cu Kα) 2θ range between 20° to 80°. The samples morphology tests were carried out by scanning electron microscope and specific surface area and pore size distribution measurements were carried out by a surface area and porosity analyser (QUANTACHROME, United States). The existence state of elements is analyzed and studied by a Thermo Scientific Escalab 250Xi X-ray photoelectron spectroscopy (XPS), which the source gun type was Al Kα and the test spot area was 500 µm. The surface functional groups of the photocatalysts were determined and analysed by FTIR spectra and Raman spectra. To determine the optical properties of the samples, absorption spectra in the wavelength range 200 nm–800 nm were determined using UV-vis Diffuse reflectance spectra (Hitachi, Japan) for different cerium doping amount. The measurements of the amperometric i-t curves were performed on an electrochemical workstation (CHI660E, CH Instruments, Austin, United States) where the photocurrent light source was provided by a xenon lamp with the cut-off filter.

The liquid chromatography-mass spectrometer (LC-MS) was used for the qualitative analysis of glyphosate and Aminomethyl phosphonic acid (SN/T 1923–2007). In a typical procedure, glyphosate samples were derivatised using FMOC-Cl acetonitrile solution, pH adjusted with borate buffer and methanol was added to obtain HPLC injection conditions. The final degradation products NH₄ ⁺-N and PO₄ ^3- were determined spectrophotometrically using a nano reagent (HJ 535–2009) and a molybdenum antimony anti-spectrophotometric method (GB 11893–89) to analyse the presumed photodegradation steps of glyphosate.

To investigate the photocatalytic activity of the CeNT@Oys photocatalyst, the photocatalytic degradation of PMG in aqueous solution was carried out under the irradiation of the CEL-S500 simulated sunlight xenon lamp (CEAuLight Ltd.). The aqueous solution of glyphosate (50 mg/L) was first sonicated with 100 mg photocatalyst for 15 min, the mixture was magnetically stirred in the darkroom to reach adsorption-desorption equilibrium and then the xenon light source was switched on. The phosphate content was determined based on the molybdenum antimony anti spectrophotometry by UV-vis (Shimadzu, UV-1800). The degradation efficiency of organic phosphorus in glyphosate was calculated according to the equation C_t/C₀, and the degradation efficiency of PMG was obtained indirectly from the inorganic phosphorus conversion rate. Where C_t is the phosphate content of the aqueous solution of PMG at moment t and C₀ is the total phosphorus content of PMG at the initial moment.

Figure 1 shows the TiO₂-500°C and CeNT@Oys-y photocatalysts at different calcination temperatures, all samples have more distinctive characteristic diffraction peaks at 2θ are 25.26, 37.74, 48.00, 53.84, 55.02, and 62.64, corresponding to anatase TiO₂ (101), (004), (200), (105), (211) and (204) crystalline peaks, respectively (JCPSD No. 21–1,272).

As the calcination temperature increased, the diffraction peaks of CeNT@Oys composite nano-photocatalyst are sharper and more crystalline, with the rutile phase appearing at 600°C. However, no CeO₂ crystalline phase was found in all samples, probably because the Ce doping concentration is too low to reach the low limit of XRD detection. In addition, calcined at 500°C, the rutile phase is almost absent from the diffraction peaks of CeNT@Oys, but is present in TiO₂. This suggest that the doping of cerium and nitrogen inhibits the transformation of the anatase phase to the rutile phase (Tang et al., 2018). As the Ce ionic radious (Ce³⁺ = 0.111 nm, Ce⁴⁺ = 0.101 nm) are much larger than those of Ti⁴⁺ (0.068 nm), therefore, theoretically, the Ce³⁺/Ce⁴⁺ is difficult to change the lattice structure of TiO₂, and can only be dispersed on the crystal surface or doped in the interstices (Matějová et al., 2014).

Supplementary Table S1 lists the lattice parameters, the full width at half maxima (FWHM) and crystal type for all samples, and calculates the average grain diameter based on the FWHM of main diffraction peak (101) the full width at half maxima (FWHM) based on Debye–Scherrer’s equation (Tu et al., 2021). Firstly, it is observed that the increase of calcination temperature, the grain size increases with decreasing FWHM, which is mainly caused by the densification of the material due to the high surface energy of the sample particles. Secondly, Comparing TiO₂-500°C and CeNT@Oys-500°C, it can be seen that the doping of Ce and N leads to the reduction of grain size. It has been reported that Ce and N doping can also effectively inhibit the phase transformation of TiO₂. The occurrence of grain boundary segregation by cerium ions doping can inhibit microcrystal growth (Singaram et al., 2017), while N doping directly leads to the reduction of Ti-O bonding and the disruption of rutile phase structural connections, thus inhibiting them growth, which is also responsible for the reduction of grain size. In addition, the lattice parameters a and b decrease slightly with increasing temperature, while c increases more significantly. Unfortunately, no significant crystalline phases are detected due to the low doping of the modified oyster shell powder. Supplementary Figure S1 plots the XRD peaks of the modified oyster shell powder, whose XRD spectrum is dominated by the diffraction peaks of Ca (OH)₂ and CaCO₃.

Figures 2A–C shows SEM micrographs of different samples. The modified oyster shell powder shows a layered plate-like microstructure. Figure 2B shows that the Ce-N-TiO₂ nanoparticles are uniformly distributed on the surface of the modified oyster shell powder, and there is some aggregation phenomenon. After magnification, the Ce-N-TiO₂ nanoparticles show a uniform particle size. The average particle size of 25.04 nm ± 9.02 nm was obtained from the statistical analysis of the particles in Figure 2D, which is slightly higher compared to the XRD calculation results because the Scherrer formula only imputes the particle size of scattered particles, while the presence of aggregation behaviour of particles in photographs taken by scanning electron microscopy was counted by computer.

Figure 3 shows the N₂ adsorption-desorption isothermal curves and the inserted image shows the pore size distribution plot. According to the IUPAC classification analysis, both Ce-N-TiO₂ and CeNT@Oys have type IV isothermal adsorption curves and there is an adsorption hysteresis in the relative pressure range of approximately 0.6–0.9 for P/P₀ and the presence of an H₂-type hysteresis loop, indicating the presence of a mesoporous structure (Shaari et al., 2012), which often appears in the interstitial pore structure of dense spherical particles. The adsorption saturation platform is observed in the curve when the relative pressure is greater than 0.9, indicating that the pore size distribution of the sample is uniform. The specific surface area of Ce-N-TiO₂ before and after doping with modified oyster powder is 61.649 m²/g and 75.301 m²/g, respectively, indicating that to a certain extent the doping of modified oyster powder increased the specific surface area of Ce-N-TiO₂. The XRD characterization results show that the doping of Ce and N leads to the reduction of TiO₂ grain size, thereby increasing the active sites of the catalyst. Based on the Barret-Joyner-Halenda method, the pore size distribution was estimated from the adsorption branch of the nitrogen isotherm, which favourably demonstrated that most of the sample pores were mesoporous structures with average pore centres of 15.4 nm and 12.2 nm, respectively.

Figure 4A shows the infrared spectrogram of TiO₂ and CeNT@Oys. One of the most pronounced absorption peaks is located in the 460 cm⁻¹–520 cm⁻¹ range, which is mainly attributed to the stretching vibration of the Ti-O bond. The wider and weaker peaks and valleys of the co-doped catalysts may be caused by the doping of the metal Ce affecting the Ti-O bond formation. The presence of a Ce-Ti-O bond at around 600 cm⁻¹ has been reported (Maarisetty and Baral 2019), but no distinctive peak appears due to the low Ce doping in this experiment, although the broadening of the trough at this point also demonstrates the presence of a Ce-Ti-O bond. The peak at 1,620 cm⁻¹ and the broad peak within 3,200 cm⁻¹–3,400 cm⁻¹ represent the bending vibration and stretching vibration of water molecules physically or chemically adsorbed on the surface and hydroxyl groups at different active sites, respectively. Further, The intensity of the co-doped absorption peak is significantly higher than that of the pure sample, indicating that it absorbs more hydroxyl groups on the surface, which is conducive to the photocatalytic reaction. The peak at 1,420 cm⁻¹ and the broad peak at 3,180 cm⁻¹ are attributed to the bending vibration and stretching vibration of the N-H bond. The weak peak at 2,350 cm⁻¹ in the spectrum of the pure sample is a characteristic CO₂ absorption peak due to the incomplete evaporation of the ethanol solvent during preparation (He et al., 2021).

Raman spectroscopy was used to further study the changes of molecular vibration after modification. The Raman activity mode of anatase TiO₂ is (A₁+2B₁+3E_g) (Nasir et al., 2014), and Figure 4B shows that both samples have significant vibrational peaks at about 145 cm⁻¹ (E_g), 196 cm⁻¹ (E_g), 396 cm⁻¹ (B₁), 514 cm⁻¹ (A₁+B₁) and 638 cm⁻¹ (E_g). The peaks at 395.56 cm⁻¹ and 514.92 cm⁻¹ represent the stretching vibrations of the O-Ti-O bond, and the peaks of 145 cm⁻¹ and 638 cm⁻¹ can be attributed to the bending vibrations of this bond.

It has been reported (Choudhury et al., 2013) that the main reason for the variation in peak height, width and position of the Raman spectra is the different size and defects of the nanoparticles. The Raman vibrational peak intensity was significantly reduced and the peak width became larger due to the doping of Ce and N. The peak of 196 cm⁻¹ position of the co-doped sample was shifted to a higher wave number as can be seen from Figure 4B, and the occurrence of this phenomenon can be attributed to the following two points. Dhanalakshmi et al. (2017) analysed that the shift of the peak of 196 cm⁻¹ to higher wave numbers is due to the presence of oxygen vacancies. Therefore, the ionic radii of Ce and Ti are different and when Ce is doped in the lattice gap it may causes oxygen vacancy defects leading to lattice distortion. Secondly, The formation of O-Ti-N bond by N doping leads to the decrease of Ti-O bond strength and the change of vibration peak intensity and width. In addition, XRD analysis shows that co-doping leads to a reduction in the particle size of the sample, which also causes a change in the vibrational peak.

X-ray photoelectron spectroscopy was used to determine the surface chemical composition and binding state of the sample. The full spectrum of the CeNT@Oys catalyst is shown in Figure 5A, indicating that the sample contains the elements Ti, O, C, Ca, N and Ce, with Ti and O being the dominant elements in the sample, while C may be residual carbon due to incomplete evaporation of the preparation reagent anhydrous ethanol and carbon-based contamination of the XPS detection equipment (Matějová et al., 2014; He et al., 2021).

Figures 5B–F show the high-resolution spectra based on the XPS Peak41 software split-peak fit with charge correction for the C 1s main peak binding energy (284.80 eV). Figure 5B shows that the O 1s are deconvoluted into a double peak, with the characteristic peaks (529.79eV, 530.07eV) representing lattice oxygen (O-Ti-O) and an additional peak (531.79eV, 531.89eV) attributed to the presence of chemisorbed oxygen or hydroxyl groups. Notably, the peak at 531.89eV for the composite nano-photocatalyst is higher than that of pure TiO₂, indicating more hydroxyl groups and adsorbed oxygen on the surface of the sample after co-doping, which undoubtedly facilitates the photoinduced electron-hole pairs transfer and optimises the photocatalytic performance. Figure 5C shows the Ti 2p high-resolution spectrum, with a difference of 5.6eV between the double peaks consistent with Ti⁴⁺of anatase TiO₂, indicating that Ti is only present in the form of Ti⁴⁺, and there is no Ti³⁺. It is not difficult to find that the peak positions of Ti 2p and O 1s are both shifted towards the lower binding energy by 0.2 eV, which is mainly due to the electronegativity of N being less than that of O (Wang Y. et al., 2011b), and the N doping leads to an increase in the electron cloud density around Ti and O, which reduces the binding energy, and the peak positions will then shift negatively.

Figure 5D shows the Ca 2p high-resolution spectrum. After fitting the split peaks, the presence of a Ca-O bond at 347.16 eV in Ca 2p_3/2, a Ca-O bond at 350.63 eV and a Ca-OH bond at 351.22 eV in Ca 2p_1/2 indicates that the presence of the modified oyster shell powder profits the adsorption of hydroxyl groups.

Previous studies have shown that there are slight differences regarding the analysis of N 1s energy spectra in N-TiO₂. Most researchers believe that the N 1s binding energy is mainly related to the N in TiO₂ doping form or the N-TiO₂ synthesis method (Senthilnathan and Philip 2010). The TiN characteristic peak is near 396 eV, and the characteristic peak at 396 eV–398 eV is attributed to the O-Ti-N bond or N-Ti-N bond formed by the replacement of lattice oxygen by nitrogen species, and the characteristic peak at 398 eV–401 eV is attributed to the interstitial N species. In this studies, no N 1s peak for Ti-N was detected in the N1s fitted curve at 396 eV–398 eV, and no TiN (111) and (220) diffraction peaks were detected in XRD (Lee et al., 2013). It is clear that the peak at 399.76 eV is formed by te structural features of N in the interstitial form forming Ti-N-O or N-O-Ti. Similarly, 401.29 eV may be associated with the presence of N-H bonds (Chen and Liu 2017; Bian et al., 2021).

Figure 5F shows an XPS diagram of the Ce3d split-peak fit. Due to the different hybridisation of the Ce 4f and O 2p orbitals and the different electronic valence states of Ce³⁺ and Ce⁴⁺, the Ce 3d spectrum splits the 3d_5/2 spin-orbit state labeled v and the 3d_3/2 spin-orbit state labeled u. The inverse fold product was performed according to the rules constructed by Burroughs et al. (1976). The red solid fitted peaks represent Ce⁴⁺ and the blue dashed fitted peaks represent Ce³⁺. Taking Ce 3d_5/2 as an example, v, v'' and v''' are the characteristic peaks representing Ce⁴⁺ in the 3d_5/2 electronic levels, where v and v''' are attributed to the allocation of two or one electron from the O 2p orbital to the Ce 4f orbital, forming the (5d6s)⁰4f² O 2p⁴ and (5d6s)⁰4f¹ O 2p⁵ hybrid configurations. v''' is caused by the primary photoelectric emission of Ce⁴⁺ (Xu et al., 2006), forming the (5d6s)⁰4f⁰ O 2p⁶ electronic configuration. v₀ and v' are the 3d_5/2 electronic levels that construct the characteristic Ce³⁺ peak. Where v' represents the (5d6s)⁰4f¹ O 2p⁶ unhybridised configuration, and v₀ represents (5d6s)⁰4f² O 2p⁵. It is inferred that Ce³⁺ and Ce⁴⁺ are present in the catalyst samples, combined with XRD analysis results it is clear that Ce can only be present in the grain boundary gaps or on the surface, forming a Ce-O-Ti bond structure.

The optical properties of the x%-CeNT@Oys catalysts were characterized by UV-vis diffuse reflectance spectrum. As shown in Figure 6, TiO₂ exhibit high absorbance in the UV region due to the photoexcited carrier leap in the anatase band gap and no absorbance in the visible region. In contrast, the co-doped samples show significantly enhanced light absorption in the UV and visible regions, and the significant red-shift of CeNT@Oys compared to TiO₂ is mainly attributed to the reduction in its forbidden band width.

Firstly, Ce doping forms the Ce 4f impurity level below the Ti 3d minimum in the TiO₂ conduction band, thus reducing the forbidden band width (Xu et al., 2002), and the different electronic structure of Ce⁴⁺(4f⁰5d⁰)/Ce³⁺(4f¹5d⁰) results in a better electron-hole separation. The redshift effect at the absorption edge becomes more and more pronounced with increasing Ce doping, unfortunately, it was shown (Ikeda et al., 2001) that when Ce doping exceeds a certain amount, it may not improve the photocatalyst degradation activity even if the redshift is more pronounced leading to a smaller band gap. Secondly, N doping forms an N 2p impurity energy level above the O 2p in the TiO₂ valence band (Barolo et al., 2012), and ultimately the co-doped sample significantly reduces the band gap energy and improves the photocatalytic performance with the synergistic effect of both. Ce and N doping not only allows the sample to make full use of visible light energy, but also facilitates the improvement of photogenerated electron-hole pair separation efficiency.

Based on the Kubelka-Muna equation F(R)=(1-R)²/2R (Spadavecchia et al., 2010; López and Gómez 2011), the absorption function was calculated from the reflectance data and the curve was plotted with hv as the horizontal coordinate and [F(R) • E]^1/2 as the vertical coordinate as shown in the Supplementary Figure S2. The band gap energies of TiO₂, N-TiO₂@Oys, 0.5%-CeNT@Oys and 1.0%-CeNT@Oys samples are 3.20 eV, 3.00 eV, 2.91 eV, and 2.69 eV, corresponding to absorption edges of 387 nm, 413 nm, 426 nm, and 461 nm, respectively.

The photogenerated carrier separation properties of the samples were analysed by comparing the photocurrent tests before and after TiO₂ doping with ions, with the usual photocurrent intensity representing the separation efficiency of the electron-hole pairs. Figure 7 records the instantaneous photocurrent response when switching the xenon lamp on and off five times. The photocurrent response signal of N-TiO₂@Oys is enhanced by a factor of 16.5 compared to TiO₂, and again by a factor of about 4 after continued doping with Ce. The introduction of N 2p hybridisation orbitals by N doping. Ulteriorly, Ce³⁺/Ce⁴⁺ is known as the “oxygen tank” due to its special redox properties (Chaker et al., 2020), Ce⁴⁺ has a strong electron capture ability to improve the separation efficiency of electron-hole pairs, while Ce³⁺ can absorb oxygen to form superoxide radicals to participate in photocatalytic reactions. It is well established that the appropriate amount of Ce and N doping can effectively improve the separation efficiency of photogenerated carriers.

In order to determine the active species acting in the photocatalytic reaction of CeNT@Oys composite nano-photocatalyst, quenching experiments were carried out under optimal reaction conditions, in which the addition of a certain amount of isopropyl alcohol (IPA), disodium ethylenediaminetetraacetate (EDTA-2Na) and p-benzoquinone (BQ) were used as chemical reagents to capture • OH, vacancies (h⁺) and • O₂ ⁻, respectively. The experimental results are shown in Figure 9A. The addition of EDTA-2Na severely affected the degradation efficiency of glyphosate, indicating that Holes or other active groups formed by reacting with holes are the main reactive species for glyphosate degradation. The addition of BQ and IPA also reduced the degradation efficiency of glyphosate by 27.9% and 14.48%, respectively. This phenomenon caused by the decrease in hole consumption and increase in electron-hole complex rate after the quenching of superoxide and hydroxyl radicals, and also indicates that both superoxide and hydroxyl radicals play an oxidative role in the degradation of glyphosate.

The electron paramagnetic resonance test using dimethyl pyridine N-oxide as a trapping agent was used to further demonstrate the presence of • OH and • O₂ ⁻ in the reaction system, as shown in Figures 9B, C. We found that there was no clear characteristic signal in the dark environment, however, However, with the irradiation of visible light, the characteristic signal of DMPO-• OH with a ratio of 1:2:2:1 and the characteristic signal of DMPO-• O₂ ⁻ with six peaks appeared obviously, and there was a good linear relationship between the increase in light reaction time and the consequent increase in radical signal intensity, indicating the presence of both radicals in the CeNT@Oys and glyphosate systems. This performance is in full agreement with the quenching experiments, further confirming that the CeNT@Oys composite nano-photocatalyst can effectively promote the formation of • OH and • O₂ ⁻.

Through the above series of material property characterization and quenching experiments and the EPR test, a reasonable analysis of the adsorption-photocatalytic degradation mechanism of CeNT@Oys composite nano-photocatalyst for glyphosate degradation in simulated sunlight is proposed as shown in Figure 10.

From XRD, XPS and other characterization, it can be seen that N is doped into TiO₂ in interstitial form, while Ce mostly stays on the lattice surface, with a few Ce entering the TiO₂ lattice in interstitial form. Combined with UV-vis DRS testing, it can be seen that N 2p orbitals are doped directly above the TiO₂ valence band and Ce 4f orbitals directly below the conduction band, as well as a small amount of Ce⁴⁺ oxide loaded on the surface of TiO₂.

Firstly, glyphosate molecules are adsorbed onto the photocatalysts surface and the adsorption phase was in accordance with a quasi-secondary kinetic model (Zhang et al., 2022), which can be divided into three phases: external liquid film diffusion dominated, intraparticle or pore diffusion and an adsorption equilibrium phase. Photocatalysis then proceeds. Since the doping of Ce and N excites the catalyst to act at visible wavelengths, the electron leap in the degradation of glyphosate by CeNT@Oys is mainly through 1) O 2p CB → Ti 2p VB; 2) N 2p hybrid orbital → Ti 2p VB; 3) O 2p CB → N 2p hybrid orbital → Ti 2p VB; 4) O 2p CB → Ce 4f hybrid orbital → Ti 2p VB is achieved in four forms. The doping of Ce can effectively reduce the electron-hole complexation rate, because the photogenerated carriers are partly excited to Ce 4f hybrid orbital. In summary, under the irradiation of sunlight, h⁺, • OH and • O₂ ⁻ jointly attack glyphosate molecules to break them up.

As Supplementary Figure S4 shows the bond breaking mode of glyphosate degradation, but after analysis in this experiment mainly C-N bond breaking was the main focus. In conjunction with SN/T 1923–200 Appendix A, as can be seen in Supplementary Figure S5A, the relative molecular mass of the glyphosate derivative is 391 and the presence of a cation peak at m/z 392 ([M + H]⁺) is clearly representative of the presence of the glyphosate molecule and the presence of fragment ion peaks at m/z 213.95 ([M-C₁₄H₁₀ + H]⁺) and m/z 88.04 ([C₃H₈NO₅P-H₃PO₃+H]⁺) in the full scan mass spectra of the daughter ions. For Supplementary Figure S5B, it can be seen that after photodegradation of glyphosate to small molecules, the main AMPA derivative has an ion peak at m/z 334 ([M + H]⁺), while its fragment ion peaks at m/z 179.01 ([C₁₄ + H₁₀]⁺) and m/z 112.06 ([CH₆NO₃+H]⁺) are present. It was shown that the main pathways of degradation of glyphosate in the samples were • OH and • O₂ ⁻ attacking the C-N bond of glyphosate to generate intermediates such as AMPA and acetic acid, or directly skipping acetic acid to generate AMPA and glyoxalate intermediates, for the nitrogen in AMPA, NH₄ ⁺-N was the main mineralisation product (Negishi et al., 2012), therefore, eventually AMPA degraded to inorganic substances such as PO₄ ^3-, NH₄ ⁺-N, H₂O and CO₂, and acetic acid or glyoxalate gradually degraded to CO₂ and H₂O. Unlike the analysis by Jaisi et al. (2016) and Echavia et al. (2009), its degradation intermediate is sarcosine, which then further breaks the bond and cleaves to aliphatic structures such as formic acid, oxalic acid and acetic acid, eventually generating inorganic substances such as phosphate and CO₂.