Removal of pharmaceuticals by a potassium ferrate(VI) material: from practical implementation to reactivity prediction†

Vanessa Peings *^a, Thierry Pigot ^a, Patrick Baylere ^a, Jean-Marc Sotiropoulos ^a and Jérôme Frayret ^ab
aUniversité de Pau et des Pays de l'Adour, IPREM UMR 5254, 2 Avenue du Président Angot, 64053 Pau Cedex 9, France. E-mail: vanessa.peings@univ-pau.fr; vanessa.peings@gmail.com; Tel: +33 684218733
^bPSI Solutions Environnementales, 570 rue Peyrefitte, 65300 Lannemezan, France

First published on 24th April 2017

---

---

Water impact This work considers ferrate(VI) applications for wastewater treatment as (i) the ferrate material used is synthesized by a cheap and original method producing a highly oxidative material as efficient as pure commercial ferrate and (ii) a combination of experimental measurements and calculations of the ionization energies of the targeted pharmaceuticals is used as a tool to predict organic contaminant oxidation.

1. Introduction

In recent years, an increasing interest in the environmental impact of pharmaceuticals present in the aquatic environment due to industry, hospital and domestic wastewater discharges^2,3 has been observed. Poor knowledge of the chronic effects of exposure to very low concentrations (ng L⁻¹–μg L⁻¹), bioaccumulation, synergy or the combination of several molecules makes it difficult to predict long-term effects on humans and ecosystems. Nevertheless, concerns about potential adverse effects remain and are challenging global water industries to find ways to remove these pollutants. About 80% of the total pharmaceutical load passes through treatment plants using activated sludge and secondary sedimentation.⁴

As the conventional wastewater treatment process is not effective in removing pharmaceuticals, a variety of physical, biological and chemical processes have been investigated and reviewed.⁵ Among the various oxidants used in the wastewater treatment field, iron-based material ferrate(VI) (FeO₄²⁻) has demonstrated its potential to remove a wide range of pharmaceuticals such as antibiotics, lipid regulators, antipyretics, anticonvulsants, and beta-blockers.^6,7 Fe(VI) exhibits many advantages because it acts firstly as an oxidant and disinfectant over the entire pH range, with a redox potential from 0.72 V in alkaline media (pH = 14) to 2.20 V under acidic conditions (pH = 0).⁸ It is then reduced to non-toxic ferric hydroxide Fe(OH)₃ which has coagulant properties.⁹ Because of the oxidation/disinfection–coagulation–precipitation process, Fe(VI) appears to be effective with both organic and inorganic contaminants.^9–11 This multifunctionality and the environmentally benign character of Fe(VI) are substantial advantages over other commonly used oxidants (chlorine, chlorine dioxide, permanganate, hydrogen peroxide and ozone).¹²

The chemistry of Fe(VI) needs to be better understood in order to further its applications in wastewater treatment. In addition, the production cost is the main reason for Fe(VI) not being more widely used for pollution remediation. Among the main synthetic production methods to obtain a stable solid Fe(VI), electrochemical and wet syntheses are long and expensive processes.¹³ Moreover, the efficiency of dry synthesis remains low due to the limited stability of Fe(VI) at the required high temperatures.¹⁴ Fe(VI) is not actually industrially produced by these methods. Since 1990, dry synthesis processes have been developed at room temperature. The resulting material involves a more stable ferrate salt.¹⁵ The FeO₄²⁻ anion is stabilized by replacing a proportion of Fe atoms with an element X existing in the XO₄^m− form (m = integer), where X^(8−m)+ has the electronic structure of a rare gas. The potassium sulfatoferrate K₂(Fe_0.5,S_0.5)O₄ prepared for this study is obtained according to patent WO2008065279.¹ This solid-phase synthesis consists of mixing a powder of Fe₂(SO₄)₃, a powder of Ca(ClO)₂ and KOH pellets. The product of the reaction exists in the form of solid pellets for which no further purification is performed because the other salts present do not alter the quality of the treated water. This method is faster, easier and cheaper than other production methods leading to pure potassium ferrate.

The cost of the Fe(VI) material synthesized according to patent WO2008065279 (ref. 1) allows its application to real water treatment, especially since the efficiency of the sulfatoferrate material for the oxidative removal of phenol has been demonstrated to be similar to the efficiency of a pure commercial ferrate.¹⁶ This study aims to determine the efficiency of the sulfatoferrate material for the oxidative removal of several pharmaceuticals, through the following objectives:

(1) determination of the reaction kinetics of the Fe(VI) material with the β-blocker metoprolol (MET), the antiepileptic drug carbamazepine (CBZ), the antibiotic ciprofloxacin (CIP) and hydroxy-ibuprofen (OH-IBU), a metabolite of the antiphlogistic drug ibuprofen, and comparison with the kinetics obtained with high purity ferrate;

(2) evaluation of the degradation of nine pharmaceuticals including MET, CBZ, CIP and OH-IBU present in a hospital wastewater (HWW);

(3) correlation of the reactivity of Fe(VI) with observed experimental ionization potentials recorded by UV photoelectron spectroscopy (UV-PES) and the first energies of the highest occupied molecular orbitals (HOMOs) obtained from density functional theory (DFT) calculations.

As far as we know, this is the first study to investigate the application of a non-pure ferrate(VI) material carried out in a non-buffered real HWW, which is an incontestable source of pharmaceuticals in the aquatic environment. The use of a cheaper material than pure ferrate in a non-buffered effluent could be promising for large-scale applications.

2. Materials and methods

2.1. Chemicals

All solutions were prepared in reagent-grade deionized water. Analytical standards: (±)-metoprolol (+)-tartrate salt (≥98.5%), carbamazepine (≥99.0%), ciprofloxacin (≥99.0%) and 2-hydroxyibuprofen (≥99.0%) were purchased from Sigma-Aldrich. Diclofenac, cyclophosphamide, ifosfamide, clarithromycine and oxazepam standards were purchased either from Sigma-Aldrich or Fisher Scientific at above 97% purity. Hospital effluent was collected downstream of the Purpan hospital (Toulouse, France) and upstream of the wastewater treatment plant. The products used in the pH adjustment of the solutions were sodium hydroxide R.P. Normapur from Prolabo (≥98.0%), acetic acid (≥99.0%) and nitric acid (70%) obtained from Fisher Scientific. HPLC-grade acetonitrile solvent was obtained from Sigma-Aldrich.

Potassium sulfatoferrate(VI) was obtained according to the patent WO2008065279.¹ The composition of the resulting pellets is 15–18% K₂FeO₄, 20–26% KOH, and 2–3% Fe(OH)₃. The balance consists of excess chemicals from the synthesis reaction: Cl⁻, SO₄²⁻, Ca²⁺, K⁺, Na⁺, and CO₃²⁻. The cost of this iron(VI) material is estimated at $20 per kg (around $120 per kg of K₂FeO₄) using the current process (including reactants and their preparation, conditioning of the pellets, labour and energy consumption) but the optimization of the synthesis is still being improved. In comparison, the price of commercial pure potassium ferrate(VI) (90%) exceeds $2000 per kg of K₂FeO₄ (Sigma-Aldrich). Fe(VI) stock solutions were prepared in water and used within 20 min of preparation to minimize Fe(VI) autodecomposition.

2.2. Kinetic experiments

Kinetic experiments on MET, CBZ, CIP and OH-IBU oxidation by the potassium sulfatoferrate material were performed at pH 10.3 ± 0.3, at room temperature (20 ± 2 °C) and under rapid stirring. The initial concentration of the pharmaceuticals was 0.05 μM while the initial concentration of Fe(VI) was in the range of 50 μM to 1000 μM. The reactions were initiated by spiking an aliquot of the Fe(VI) stock solution. At a certain time interval, samples of the solution were quenched by decreasing the pH to under 2 because protonated Fe(VI) species are unstable in aqueous media and decompose very quickly.¹⁷ No filtration was performed before acidification so only the oxidation properties of the ferrate anion were observed. Indeed, the chemical reduction of Fe(VI) through its action on the substrate or water led to the formation of insoluble Fe(III) species, which may have the ability to adsorb organic compounds and thereby remove them from solution. In acidic media, the Fe(III) particles are solubilized and the coagulant effect is avoided. The concentrations of residual pharmaceuticals were analyzed using LC/MS. All samples were prepared in triplicate.

2.3. Treatment of the hospital wastewater

The studied HWW was sampled in glass bottles before its discharge into the sewerage network. No pretreatment was performed on this HWW. The pH was close to neutrality (pH = 7.3 ± 0.2). The HWW had a chemical oxygen demand (COD) of 185 mg (O₂) L⁻¹ and contained biological species. Jar-testing experiments were performed under the following protocol: fast mixing for 60 min at 240 rpm; slow mixing for 12 min at 40 rpm; and then sedimentation for 60 min. Preliminary experiments in a range of ferrate doses from 75 to 750 μM were conducted on the HWW. The highest concentration led to more effective abatements of both COD and total concentration of the monitored pharmaceuticals. The results are presented for the ferrate(VI) dose of 750 μM. The final pH of the solution was 10.0 ± 0.4. The experiment was performed in duplicate. The supernatants were filtered through 0.45 μm membrane filters after sedimentation. Samples were extracted on SPE cartridges and further analyzed by LC/MS. Nine selected pharmaceuticals with concentrations ranging from 73 ± 4 ng L⁻¹ to 159 ± 8 μg L⁻¹ were monitored: MET, CBZ, CIP, OH-IBU, diclofenac (DFC), cyclophosphamide (CYC), ifosfamide (IFO), clarithromycine (CLA) and oxazepam (OXA).

2.4. Analytical methods

The composition of the sulfatoferrate product synthesized according to patent WO2008065279 was defined by inductively coupled plasma-optical emission spectrometry (ICP-OES) (ULTIMA, Jobin Yvon). The Fe, K, Ca, Na, and S elements were quantitatively determined from standard curves. To distinguish between the Fe(VI) and Fe(III) forms that constituted the product, the quantitative determination of Fe(VI) is based on the fact that Fe(VI) species are soluble in water but ferric hydroxide Fe(OH)₃ is insoluble. A known weight of the sulfatoferrate material was dissolved in Milli-Q purified water, just before performing the analysis (to avoid Fe(VI) degradation). A proportion of this solution was filtered at 0.45 μm in order to monitor the Fe(VI) concentration, while the other was acidified (1% HNO₃) to obtain the total concentration of iron. A mass balance study provided the other salt concentrations. The decrease in the ferrate(VI) concentration in the solution during the treatments was monitored through measurement of the absorbance at 510 nm using a Shimadzu UV-1800 spectrophotometer.

The remaining pharmaceutical compounds present in the treated test solutions were detected by two devices. The detection of MET, CBZ, CIP and OH-IBU for the kinetics experiments was monitored by a Waters Acquity Ultraperformance LC plus a Waters Xevo TQ MS detector. The separation was achieved using a Waters Acquity BEH C₁₈ column (1.7 μm, 50 mm × 2.1 mm). The mobile phase consisted of 80% water and 20% acetonitrile prepared in 0.5% acetic acid at a flow rate of 0.7 mL min⁻¹. MET, CBZ and CIP were analyzed in electrospray ionization (ESI) positive mode, while OH-IBU was analyzed in ESI negative mode. To measure the target compounds in the HWW, before and after treatments by the sulfatoferrate material, an Agilent 6490 Triple Quadrupole LC/MS was used. The analyses were performed after adding a known quantity of the internal standard compound (isotopically substituted compound).

The photoelectron spectra were recorded on a home-made Phoenix 1 (ref. 18) instrument that was equipped with a 127° cylindrical analyzer and monitored using a microcomputer to which a digital–analog converter was added. The spectra were calibrated to known auto-ionization energies of He at 4.99 eV (He(II)/He(I)) and Xe at 12.13 eV (²P_3/2) and 13.44 eV (²P_1/2). All of the products were vaporized at 1.1 × 10⁻⁵ mbar by the beam heat or by an oven located near the photoionization chamber. The pressure of detection (Channeltron) was below 1.0 × 10⁻⁷ mbar. Measured temperatures were 568 K, 433 K, 443 K, 493 K, 389 K and 401 K for CIP, MET, CBZ, CIP, OH-IBU and DFC, respectively. Spectra were recorded when pressure and temperature were stabilized, preventing intensity variation and guaranteeing good reproducibility.

2.5. Calculations

Calculations were performed by the program Gaussian 09 (ref. 19) using density functional theory.²⁰ The various structures were fully optimized at the CAM-B3LYP level,²¹ with the 6-31G(d,p) basis set.²² All atoms were augmented with a single set of polarization functions. Their second derivatives were calculated to check that the optimized structures were true minima on the potential energy surface. All of the total energies were zero-point energies (ZPEs) and temperature corrected using un-scaled density functional frequencies. For the assignment of the PE spectra, ionization energies were calculated as previously suggested by Stowasser and Hoffmann.‡²³ Molecular orbitals were plotted using GaussView.²⁴

3. Results

3.1. Reaction orders and constants of ferrate with selected pharmaceuticals

Three ferrate concentrations were applied for the treatments of each pharmaceutical: 50 μM, 500 μM and 1000 μM. Although initially present in excess in relation to the solute (0.05 μM), Fe(VI) cannot be assumed to remain constant during the experiments due to its autodecomposition in solution with time and its action on the solute (eqn (1)). The autodecomposition of Fe(VI) is the reaction of a molecule of Fe(VI) with a molecule of water or another molecule of Fe(VI).^25,26 The pH-dependent rate of this reaction has been determined in a previous study.¹⁶ The reactivity of Fe(VI) with a number of organic compounds shows second-order behavior overall, i.e. first order in total concentration [Fe(VI)] and first order in total concentration of the compound.¹⁰ According to the literature, a second-order law is assumed for CBZ and CIP reactions with Fe(VI).^27–29 To our knowledge, no data exists for the oxidation kinetics of MET and OH-IBU by Fe(VI), but a second-order law exists for similar molecules such as propranolol and ibuprofen.^28,30

	(1)

With: [Fe(VI)] = [H₃FeO₄⁺] + [H₂FeO₄] + [HFeO₄⁻] + [FeO₄²⁻] ≈ [FeO₄²⁻] at pH 10.3 ± 0.3 (pK_a = 7.3 ± 0.1).

n = 1 or 2, depending on the pH;¹⁶

[x] the concentration of the solute x (x = MET, CBZ, CIP or OH-IBU);

k ₁ is the second order rate constant associated with the autodecomposition of Fe(VI);

k ₂ is the second order rate constant associated with the reaction of Fe(VI) with x.

Due to the consumption of Fe(VI) during the experiment, a pseudo-first order rate is not appropriate for modeling the pharmaceutical degradations. The procedure described by Lee and co-workers was thus used to determine the values of k₂,³¹ by plotting the natural logarithm of the compound P versus the Fe(VI) exposure, i.e. Fe(VI) concentration integrated over time (eqn (2)).

	(2)

Fig. 1 shows an example of the degradation of both MET and Fe(VI) during the reaction in the molar ratio Fe(VI):MET = 10000:1.

Fe(VI) exposure was measured by determining the gray hatched area (example for 30 min in Fig. 1). The inset in Fig. 1 shows the experimental data obtained for the kinetic studies with the three ratios Fe(VI):MET ([Fe(VI)] = 50 μM, 500 μM and 1000 μM) fitted by a correct linear correlation. The slope representing k₂ yields 3.7 ± 0.3 M⁻¹ s⁻¹. According to the literature, Fe(VI) has a much higher reactivity with the β-blockers atenolol and propranolol, with values of k₂ at pH 8 of ≈7 M⁻¹ s⁻¹ (ref. 6) and 20 M⁻¹ s⁻¹ (ref. 30), respectively. This low reactivity of MET compared to those of atenolol and propranolol has already been demonstrated during treatment in aqueous solution when a ratio of 1:10 [β-blocker]:[Fe(VI)] was used.³²

The same approach has been used to determine the second-order rate constants of the reactions of Fe(VI) with CBZ and CIP (Fig. 2).

The oxidation kinetics of CBZ and CIP are higher than that of MET, with k₂ = 13.1 ± 0.8 M⁻¹ s⁻¹ and 89 ± 2 M⁻¹ s⁻¹, respectively, at pH 10.3 ± 0.3. The data found in the literature demonstrate that Fe(VI) reaction rates with CBZ exhibit a strong pH dependence: k₂ has been measured to be 67 M⁻¹ s⁻¹ (ref. 33) and 70 ± 3 M⁻¹ s⁻¹ (ref. 27) at pH 7, 16 M⁻¹ s⁻¹ (ref. 33) or 23.83 ± 0.52 M⁻¹ s⁻¹ (ref. 29) at pH 8 and 1.09 ± 0.11 M⁻¹ s⁻¹ at pH 9.²⁹ The apparent second-order rate constants for CIP have been measured to be 470 M⁻¹ s⁻¹ at pH 7,³³ 113.7 ± 6.3 M⁻¹ s⁻¹ (ref. 28) or 170 M⁻¹ s⁻¹ (ref. 33) at pH 8 and 64.1 ± 1.0 M⁻¹ s⁻¹ at pH 9.²⁸ Finally, there appears to be no reaction between Fe(VI) and OH-IBU. A very low reactivity of the similar molecule ibuprofen has been demonstrated in several works, with a second-order rate constant close to 0.1 M⁻¹ s⁻¹ at pH 8 and below 0.02 M⁻¹ s⁻¹ at pH 9.^6,28,34Table 1 summarizes the determined rate constants k₂ at pH 10.3 ± 0.3 from this study and the data found in the literature at pH 7, 8 and 9. The reaction rates for both CBZ and CIP reported in the literature resulted from the use of a pure potassium ferrate. The ferrate-based material used in this work shows comparable reaction rates, demonstrating that it is as efficient as pure commercial potassium ferrate in the removal of these micropollutants, as it has been shown on the oxidation of phenol in a previous work.¹⁶

---

3.2. Treatment of hospital wastewater

The previous results show that Fe(VI) is a selective oxidant that can reduce the concentration of emerging pollutants in aqueous solutions. This ability needs to be confirmed in complex, actual wastewater. Several pharmaceuticals initially present in the HWW (COD = 185 mg (O₂) L⁻¹) were monitored during the treatment by the sulfatoferrate material added to the concentration of 150 mg L⁻¹ of K₂FeO₄ ([Fe(VI)] = 750 μM): MET, CBZ, CIP, OH-IBU, diclofenac (DFC), cyclophosphamide (CYC), ifosfamide (IFO), clarithromycine (CLA) and oxazepam (OXA). The concentrations of these compounds ([x]₀) and the molar ratio [Fe(VI)]:[x]₀ are given in Table 2. The final pH of the HWW treated by 750 μM Fe(VI) material was 10.0 ± 0.4.

---

Fig. 3 shows the abatement of the targeted pharmaceuticals after 60 min of treatment. The reactivity of Fe(VI) was the best in the following order: CIP > CBZ > DFC > MET > CLA > OXA > OH-IBU > CYC > IFO, and the overall concentration of the compounds decreased by 58%.

Although present at high concentration in the HWW (Table 1), CIP removal is 80 ± 16%, which confirms the good reactivity on CIP observed during the kinetic study of the pure molecule. The removal attributed to the coagulation process initiated by the ferric hydroxides resulting from Fe(VI) decomposition has not been quantified but should be considered. Only the oxidation by Fe(VI) was monitored during the kinetic determinations. The measurement of zeta potential values in CIP solutions treated with Fe(VI) showed that removal due to coagulation is not negligible³⁵ or even high, with an elimination of more than 80% in a wastewater treatment plant using ferric salt.³⁶ On the other hand, ferrate has real oxidation ability. This property is demonstrated by the removal of 77 ± 8% of a non-ionizable compound: CBZ. Indeed, a non-ionizable compound is not affected by a coagulation process.³⁶ The monitoring of MET shows good removal despite the low reactivity observed during the kinetic study (k₂ = 3.7 ± 0.3 M⁻¹ s⁻¹). The 66 ± 27% removal might be the result of the high molar ratio [Fe(VI)]:[MET]₀ and/or a non-negligible coagulation process with ferric salt (<40% according to Vieno et al.³⁶). This coagulation process also involves the OH-IBU removal by 15 ± 1% (the non-oxidability was reported during the kinetic experiment). The other targeted pharmaceuticals: DFC, CYC, IFO, CLA and OXA are removed at 68 ± 18%, 22 ± 2%, 7.8 ± 0.5%, 40 ± 1% and 39 ± 4%, respectively. These results show the selectivity of Fe(VI) and the need to improve our understanding of its action.

Even though there was a large excess of Fe(VI) compared to the targeted compounds, the HWW contained other organic and inorganic molecules, leading to competition between several compounds in the solution. Fe(VI) had an impact on organic matter, as the initial COD of the effluent of 185 mg(O₂) L⁻¹ was reduced by 69%.

3.3. Photoelectron spectra and DFT calculations

Ferrate(VI) tends to have a great capacity to oxidize molecules containing electron-rich moieties (ERMs) rather than those without ERMs.⁶ Indeed, the Fe(VI) oxidation mechanism is a capture of electrons from the target molecules by Fe(VI).^10,11 The first ionization potentials of the target pharmaceuticals could directly influence the reactivity of ferrate. These potentials represent the energy that an atom in the gaseous state requires in order to remove the least-bound electrons (valence electrons) and form positive ions. These experimental data can be correlated with the calculated molecular orbitals (vide infra). Although the Fe(VI) reactions take place in solution, the IP is a good indicator of the ease of removing an electron in a molecule. This approach does not take account of solvation effects. Nevertheless, it could lead to a satisfactory rationalization of the results.

To confirm this hypothesis, the relationship between the constant oxidation rate of certain organic compounds by Fe(VI) and the first IP of that compound was plotted from the data in Table 3 (Fig. 4). The first IPs are taken from the Handbook of HeI photoelectron spectra of fundamental organic molecules.³⁷

---

According to these data found in the literature, the molecules with the lowest IP are the most reactive with Fe(VI), while those with an IP above 10 eV are much less reactive. Experimental measurements and DFT calculations have been conducted to verify the correlation of the reactivity of Fe(VI) with the values of the first ionization potentials of each pharmaceutical studied.

The UV photoelectron spectra of the target molecules were successfully recorded following their ionization (see Fig. A in the ESI†). Table 4 summarizes the experimental data obtained during this study, including the first UV-PES bands for each pharmaceutical and the first ionization energy calculations with the corresponding schematic molecular orbitals (HOMOs: highest occupied molecular orbitals).

---

We can observe a good correlation between the experimental UV-PES bands and calculated values of the ionization energies. Table 2 shows that the lower the first ionization band, the higher the oxidation kinetic by Fe(VI). The experimental measurements of the first ionization potential can thus be used to predict the reactivity of the chemical oxidant Fe(VI) in relation to molecules. Nevertheless, the only position of the HOMOs does not explain the reactivity of Fe(VI) on organic molecules. A difference in reactivity for molecules with similar HOMO energies exists, for example for CBZ and CIP. Also, Fe(VI) does not oxidize OH-IBU, while in Table 2 reaction occurs when the first ionization potential is below 10 eV. Thus, the nature of the HOMOs as well as the position of the HOMOs must also be analyzed.

4. Discussion

OH-IBU is the compound which exhibits the highest first IP (IP₁) and the lowest reactivity towards Fe(VI), in both the kinetic study and HWW treatment. The OH-IBU molecule has a single benzene ring which could be conducive to electrophilic attack due to the electronic density in the aromatic ring. Both the carboxylic group and alcohol group are electron-withdrawing functional groups (inductive effect acceptor −I) that weaken the reactivity of the aromatic system with Fe(VI). The structure of OH-IBU does not contain electron-rich moieties (ERMs) so it resists degradation by Fe(VI). This can explain the very low rate constant measured during the kinetic study. The slight removal of OH-IBU by Fe(VI) in the hospital effluent can probably be attributed to the coagulation effect of reduced ferric ions from Fe(VI).

MET has a lower IP₁, 8.2 eV, and some oxidation of the molecule by Fe(VI) is observed and not negligible. The electronic density in the HOMO is mainly located on the π-bond of the phenolic moiety. The electrophilic attack on this moiety could lead to the formation of a phenoxy radical and then generate a mechanism similar to the oxidation of phenol.¹⁶ The electrophilic attack of amine II is also possible, as the presence of a lone pair on the nitrogen atom makes the amines nucleophilic (Lewis base). Its electronic density is found in HOMO-1 according to DFT calculations.

The IP₁ of DFC is 7.8 eV. DFC has been found to be highly reactive towards Fe(VI) in the HWW. The aniline function has been found to be very easily oxidizable.⁶ It corresponds closely to the HOMO. HOMO-1 corresponds to the π-bond of the phenyl ring. Oxidation products of DFC were found in the study of Zhou and Jiang.²⁹ No opening of the phenyl ring appears, which shows that the reaction occurs more easily on amine II of the aniline function. It results in a mono-aromatic compound formed by the cleavage of the N–C bond between two aromatic rings.

In CBZ, the HOMO corresponds mainly to the π-bond of the central ring. HOMO-1 exhibits a representative weight of the electronic density on the cyclic nitrogen urea. The free lone pair of this nitrogen is less accessible than lone pairs on unhindered nitrogen and less reactive due to the attractive adjacent carbonyl. An identification of some oxidation products by LC/MS/MS has been performed during the kinetic experiment on CBZ oxidation by Fe(VI). Two intermediates have been identified, and their structures are proposed based upon the molecular ions masses and MS/MS fragmentation patterns (Fig. 5, CBZ-1 and CBZ-2). Fe(VI) attacks the central urea function to form CBZ-1. The intermediate CBZ-2 confirmed that an electrophilic attack of Fe(VI) occurs on the central π-bond, which is an ERM reported as the main attacking site during chemical oxidations.⁴⁰ These attack mechanisms are consistent with data of previous studies.^27,29 CBZ-2 was detected as a by-product only in the case of the oxidation of CBZ by Mn(VII).²⁷ Its detection in our study can be explained by the higher molar ratio oxidant/target applied in this work than those applied in previous works.^27,29 After 30 min of reaction, CBZ-1 and CBZ-2 are no longer detected and no other products are identified by LC/MS/MS analyses, suggesting that further oxidation occurred.

In CIP, a saturated heterocyclic ring contains two nitrogen atoms in opposite positions (piperazinyl ring). The HOMO corresponds to the electronic density on amine III of the piperazine, and the deeper HOMO-2 and -3 to amine II. The electronic density of HOMO-1 is mainly located on the alkene conjugate of the heterocycle containing the cyclopropylamine group. A recent study on the oxidation of CIP by Fe(VI) confirms the presence of these two reactive sites on the molecule.²⁸ The attack on the piperazinyl ring and the alkene conjugate of the heterocycle appeared to lead to the cleavage or hydroxylation of the rings.²⁸

Regarding the reductions in the other compounds monitored in the HWW, we could assume that the 40% reduction in CLA is the result of the amine III attack by Fe(VI), since there are no other ERMs in the macrolide CLA. The reactivity of several macrolides has been determined with ClO₂ and O₃, and has shown that the oxidants selectively attack amine III.^41–43 The OXA removal corresponds closely to the CLA removal. This reactivity can be attributed to the presence of amine II and amine III. CYC and IFO show very low reactivity, as with OH-IBU. The low reactivity of the nitrogen atoms in CYC is probably due to the electronic density on the nitrogen atoms being diminished due to the presence of the phosphamide group. The lower reactivity of IFO may be attributed to the presence of both the phosphamide group and the chloro group (−I strong and mesomeric effect +M low) close to the nitrogen.

5. Conclusion

The ferrate-based material used in this study, synthesized according to patent WO2008065279, shows comparable oxidative capacities to pure commercial ferrate for a reduced cost. The benefits of Fe(VI) as a tertiary treatment on a wide range of micropollutants are demonstrated. Nevertheless, the effectiveness of Fe(VI) oxidation is very sensitive to the structure of the targeted molecules. The monitoring of the pharmaceutical abatement in pure water and in a real hospital wastewater has shown a close correlation with both the observed experimental ionization potentials (IPs) and the ionization energies (IEs) of the HOMOs given by DFT calculations. The lower the first IE of the pharmaceutical, the higher its oxidation by Fe(VI). The experimental measurements of the first IP can thus be used as a tool to predict organic contaminant abatement. The nature of the HOMOs must also be analyzed in Fe(VI) treatments resulting in the selective oxidation of electron-rich moieties (ERMs) such as the lone pair on the nitrogen atom (secondary and tertiary amines, aniline, etc.) or π-bonds of heterocyclic rings. The lack of ERMs leads to a poor or non-reactivity, as found for hydroxyibuprofen, cyclophosphamide and ifosfamide.

Acknowledgements

The authors wish to thank Dr. Lobinski Ryszard, Dr. Vacchina Véronique and Dr. Albasi Claire for their contribution to this work. The French ANR is acknowledged for financial support through project ANR-13-ECOT-0010.

References

1. C. Boesinger, N. Costarramone and A. Castetbon, Method for preparing alkali metal ferrates, WO2008065279, 2008 .
2. D. J. Caldwell, F. Mastrocco, L. Margiotta-Casaluci and B. W. Brooks, An integrated approach for prioritizing pharmaceuticals found in the environment for risk assessment, monitoring and advanced research, Chemosphere, 2014, 115, 4–12 CrossRef CAS PubMed .
3. O. Cardoso, J.-M. Porcher and W. Sanchez, Factory-discharged pharmaceuticals could be a relevant source of aquatic environment contamination: Review of evidence and need for knowledge, Chemosphere, 2014, 115, 20–30 CrossRef CAS PubMed .
4. B. Roig, Pharmaceuticals in the Environment: Current Knowledge and Need Assessment to Reduce Presence and Impact, IWA Publishing, 2010 Search PubMed .
5. J. Wang and S. Wang, Removal of pharmaceuticals and personal care products (PPCPs) from wastewater: A review, J. Environ. Manage., 2016, 182, 620–640 CrossRef CAS PubMed .
6. Y. Lee and U. von Gunten, Oxidative transformation of micropollutants during municipal wastewater treatment: Comparison of kinetic aspects of selective (chlorine, chlorine dioxide, ferrate(VI), and ozone) and non-selective oxidants (hydroxyl radical), Water Res., 2010, 44, 555–566 CrossRef CAS PubMed .
7. V. K. Sharma, Oxidative transformations of environmental pharmaceuticals by Cl2, ClO2, O3, and Fe(VI): Kinetics assessment, Chemosphere, 2008, 73, 1379–1386 CrossRef CAS PubMed .
8. R. H. Wood, The Heat, Free Energy and Entropy of the Ferrate(VI) Ion, J. Am. Chem. Soc., 1958, 80, 2038–2041 CrossRef CAS .
9. J.-Q. Jiang, S. Wang and A. Panagoulopoulos, The exploration of potassium ferrate(VI) as a disinfectant/coagulant in water and wastewater treatment, Chemosphere, 2006, 63, 212–219 CrossRef CAS PubMed .
10. V. K. Sharma, Ferrate(VI) and ferrate(V) oxidation of organic compounds: Kinetics and mechanism, Coord. Chem. Rev., 2013, 257, 495–510 CrossRef CAS .
11. V. K. Sharma, Oxidation of inorganic contaminants by ferrates (VI, V, and IV)–kinetics and mechanisms: A review, J. Environ. Manage., 2011, 92, 1051–1073 CrossRef CAS PubMed .
12. V. K. Sharma, R. Zboril and R. S. Varma, Ferrates: Greener Oxidants with Multimodal Action in Water Treatment Technologies, Acc. Chem. Res., 2015, 48, 182–191 CrossRef CAS PubMed .
13. J. M. Schreyer and L. T. Ockerman, Stability of Ferrate(VI) Ion in Aqueous Solution, Anal. Chem., 1951, 23, 1312–1314 CrossRef CAS .
14. L. Machala, R. Zboril, V. K. Sharma, J. Filip, O. Schneeweiss, J. Madarász, Z. Homonnay, G. Pokol and R. Yngard, in Ferrates, ed. V. K. Sharma, American Chemical Society, 2008, vol. 985, pp. 124–144 Search PubMed .
15. O. Evrard, R. Gerardin, N. Schmitt and J. L. Evrard, Ferrates de metaux alcalins ou alcalino-terreux, leur preparation et leurs applications industrielles, 1994 Search PubMed .
16. V. Peings, J. Frayret and T. Pigot, Mechanism for the oxidation of phenol by sulfatoferrate(VI): Comparison with various oxidants, J. Environ. Manage., 2015, 157, 287–296 CrossRef CAS PubMed .
17. W. F. Wagner, J. R. Gump and E. N. Hart, Factors Affecting Stability of Aqueous Potassium Ferrate(VI) Solutions, Anal. Chem., 1952, 24, 1497–1498 CrossRef CAS .
18. K. Miqueu, J.-M. Sotiropoulos, P. Baylère, S. Joantéguy, G. Pfister-Guillouzo, H. Ranaivonjatovo, J. Escudié and M. Bouslikhane, The electronic structure of a diarsaallene –As=C=As– and a phosphaarsaallene –P=C=As–: UV photoelectron spectroscopy and theoretical studies, J. Mol. Struct., 2004, 690, 53–61 CrossRef CAS .
19. M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A. Robb, J. R. Cheeseman, G. Scalmani, V. Barone, B. Mennucci, G. A. Petersson, H. Nakatsuji, M. Caricato, X. Li, H. P. Hratchian, A. F. Izmaylov, J. Bloino, G. Zheng, J. L. Sonnenberg, M. Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O. Kitao, H. Nakai, T. Vreven, J. A. Montgomery, J. E. Peralta, F. Ogliaro, M. Bearpark, J. J. Heyd, E. Brothers, K. N. Kudin, V. N. Staroverov, R. Kobayashi, J. Normand, K. Raghavachari, A. Rendell, J. C. Burant, S. S. Iyengar, J. Tomasi, M. Cossi, N. Rega, J. M. Millam, M. Klene, J. E. Knox, J. B. Cross, V. Bakken, C. Adamo, J. Jaramillo, R. Gomperts, R. E. Stratmann, O. Yazyev, A. J. Austin, R. Cammi, C. Pomelli, J. W. Ochterski, R. L. Martin, K. Morokuma, V. G. Zakrzewski, G. A. Voth, P. Salvador, J. J. Dannenberg, S. Dapprich, A. D. Daniels, Ö. Farkas, J. B. Foresman, J. V. Ortiz, J. Cioslowski and D. J. Fox, Gaussian 09, Revision E.01, Gaussian, Inc., 2009, at <http://www.gaussian.com/g_tech/g_ur/m_citation.htm%3e> Search PubMed .
20. R. G. Parr and Y. Weitao, Density-Functional Theory of Atoms and Molecules, Oxford University Press, 1989 Search PubMed .
21. T. Yanai, D. P. Tew and N. C. Handy, A new hybrid exchange–correlation functional using the Coulomb-attenuating method (CAM-B3LYP), Chem. Phys. Lett., 2004, 393, 51–57 CrossRef CAS .
22. P. C. Hariharan and J. A. Pople, The influence of polarization functions on molecular orbital hydrogenation energies, Theor. Chim. Acta, 1973, 28, 213–222 CrossRef CAS .
23. R. Stowasser and R. Hoffmann, What Do the Kohn−Sham Orbitals and Eigenvalues Mean?, J. Am. Chem. Soc., 1999, 121, 3414–3420 CrossRef CAS .
24. R. Dennington, T. Keith and J. Millam, GaussView, Semichem Inc., 2009 Search PubMed .
25. J. D. Carr, in Ferrates, ed. V. K. Sharma, American Chemical Society, 2008, vol. 985, pp. 189–196 Search PubMed .
26. Y. Lee, R. Kissner and U. Gunten, Reaction of Ferrate(VI) with ABTS and Self-decay of Ferrate(VI): Kinetics and Mechanisms, Environ. Sci. Technol., 2014, 48, 5154–5162 CrossRef CAS PubMed .
27. L. Hu, H. M. Martin, O. Arce-Bulted, M. N. Sugihara, K. A. Keating and T. J. Strathmann, Oxidation of Carbamazepine by Mn(VII) and Fe(VI): Reaction Kinetics and Mechanism, Environ. Sci. Technol., 2009, 43, 509–515 CrossRef CAS PubMed .
28. Z. Zhou and J.-Q. Jiang, Reaction kinetics and oxidation products formation in the degradation of ciprofloxacin and ibuprofen by ferrate(VI), Chemosphere, 2015, 119, S95–S100 CrossRef CAS PubMed .
29. Z. Zhou and J.-Q. Jiang, Treatment of selected pharmaceuticals by ferrate(VI): Performance, kinetic studies and identification of oxidation products, J. Pharm. Biomed. Anal., 2015, 106, 37–45 CrossRef CAS PubMed .
30. G. A. K. Anquandah, V. K. Sharma, V. R. Panditi, P. R. Gardinali, H. Kim and M. A. Oturan, Ferrate(VI) oxidation of propranolol: Kinetics and products, Chemosphere, 2013, 91, 105–109 CrossRef CAS PubMed .
31. Y. Lee, J. Yoon and U. von Gunten, Kinetics of the Oxidation of Phenols and Phenolic Endocrine Disruptors during Water Treatment with Ferrate (Fe(VI)), Environ. Sci. Technol., 2005, 39, 8978–8984 CrossRef CAS PubMed .
32. M. L. Wilde, W. M. M. Mahmoud, K. Kümmerer and A. F. Martins, Oxidation–coagulation of β-blockers by K2FeVIO4 in hospital wastewater: Assessment of degradation products and biodegradability, Sci. Total Environ., 2013, 452–453, 137–147 CrossRef CAS PubMed .
33. Y. Lee, S. G. Zimmermann, A. T. Kieu and U. von Gunten, Ferrate (Fe(VI)) Application for Municipal Wastewater Treatment: A Novel Process for Simultaneous Micropollutant Oxidation and Phosphate Removal, Environ. Sci. Technol., 2009, 43, 3831–3838 CrossRef CAS PubMed .
34. V. K. Sharma and S. K. Mishra, Ferrate(VI) oxidation of ibuprofen: A kinetic study, Environ. Chem. Lett., 2006, 3, 182–185 CrossRef CAS .
35. S. Barışçı, F. Ulu, M. Sillanpää and A. Dimoglo, The usage of different forms of ferrate (VI) ion for amoxicillin and ciprofloxacin removal: density functional theory based modelling of redox decomposition: The usage of different forms of ferrate (VI) ion, J. Chem. Technol. Biotechnol., 2016, 91, 257–266 CrossRef .
36. N. Vieno, T. Tuhkanen and L. Kronberg, Removal of Pharmaceuticals in Drinking Water Treatment: Effect of Chemical Coagulation, Environ. Technol., 2006, 27, 183–192 CrossRef CAS PubMed .
37. K. Kimura, Handbook of HeI photoelectron spectra of fundamental organic molecules: ionization energies, ab initio assignments, and valence electronic structure for 200 molecules, Japan Scientific Societies Press, 1981 Search PubMed .
38. J. D. Carr, P. B. Kelter, A. Tabatabai, D. Splichal, J. Erickson and C. W. McLaughlin in Water Chlorination Chem. Environ. Impact Health Eff, Lewis Publishers, 1985, vol. 5, pp. 1285–1298 Search PubMed .
39. M. D. Johnson and B. J. Hornstein, Kinetics and mechanism of the ferrate oxidation of hydrazine and monomethylhydrazine, Inorg. Chim. Acta, 1994, 225, 145–150 CrossRef CAS .
40. D. C. McDowell, M. M. Huber, M. Wagner, U. von Gunten and T. A. Ternes, Ozonation of Carbamazepine in Drinking Water: Identification and Kinetic Study of Major Oxidation Products, Environ. Sci. Technol., 2005, 39, 8014–8022 CrossRef CAS PubMed .
41. M. C. Dodd, M.-O. Buffle and U. von Gunten, Oxidation of Antibacterial Molecules by Aqueous Ozone: Moiety-Specific Reaction Kinetics and Application to Ozone-Based Wastewater Treatment, Environ. Sci. Technol., 2006, 40, 1969–1977 CrossRef CAS PubMed .
42. M. M. Huber, S. Canonica, G.-Y. Park and U. von Gunten, Oxidation of pharmaceuticals during ozonation and advanced oxidation processes, Environ. Sci. Technol., 2003, 37, 1016–1024 CrossRef CAS PubMed .
43. M. M. Huber, S. Korhonen, T. A. Ternes and U. von Gunten, Oxidation of pharmaceuticals during water treatment with chlorine dioxide, Water Res., 2005, 39, 3607–3617 CrossRef CAS PubMed .

---

Footnotes

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c7ew00038c

‡ Estimated IPs (except for IP1) were calculated by applying a uniform shift at different Kohn–Sham energies (shift: x = IP1exptl + eKS (HOMO)), where eKS (HOMO) is the highest-occupied DFT/6-31G(d,p) Kohn–Sham MO energy of the ground-state molecule and IP1exptl is the lowest experimental IP of the molecule.

This journal is © The Royal Society of Chemistry 2017