Introduction

Heat-induced solid-phase quasi-intramolecular redox reactions of [MLn](XO4)m complexes (M = Cu, Zn, Cd, Fe, Co, Cr) containing reducing ligands (NH3, pyridine, or urea. n = 2–6) and oxidizing anions (X = Cl, Mn, Mo, Re, S; m = 0.5–3) ensure an easy and convenient way to prepare simple or mixed transition metal oxides with particle size in the nanometer region [1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17]. These nanosized oxides are important materials in various industrial processes, especially the transition metal oxides with perovskite (ABO3) or spinel (AB2O4) structures (A, B = Mn, Fe, Co, Ni, and Cu) that can replace the noble-metal-based redox catalyst [18]. For example, the cobalt- and copper-manganese mixed spinel oxides are widely studied and used in various processes like nitrate-to-ammonia conversion [19], oxidation of ammonia, propane, and CO [20,21,22,23] or reduction of NO [24, 25]. Co–Mn or Cu–Mn mixed spinel oxides are active catalysts in the oxidation processes of ciprofloxacin or other organic pollutants in wastewater, in the activation of peroxymonosulfate, a promising technology for water treatment [26, 27], or in the degradation of pharmaceuticals [28]. In the reduction reactions of oxygen, it is possible to use not only Co or Cu but also other AMn2O4 (A = Mn(II), Mg(II), Li(I), and Cd(II)) mixed oxides [29]. In our previous work, we dealt with preparation of stoichiometric AgMnO2 [7] from [Ag(NH3)2]MnO4 and from silver manganese oxides with various Ag–Mn ratio via mixed [Ag(NH3)2](ClO4,MnO4) solid solutions [7]. Besides Ag+, it is possible to dope the manganese oxides with Fe3+ (or Zn2+, Cu2+) and this can improve their photocatalytic efficiency [30]. These help us contextualize our previous research about [Fe(urea-O)6](MnO4)3 from which we prepared (Fe,Mn)Ox [1]. These (Fe,Mn)Ox phases contain oxygen vacancies which are important in a catalyst structure since it can help reduce the activation energy of the given process [31] and help the catalytic oxidation of NO with Fe–Mn bimetal oxides [32]. Fe–Mn mixed oxides are widely used (just like Co–Mn mixed spinel oxides) in important technological processes like reduction of nitroaromatic compounds [33] or CO2 [1], gasification of biomass [34], catalytic oxidation of ethylene oxide [35] or formaldehyde [36], and transforming carbon dioxide into jet fuel [37]. Preparation and decomposition of [hexakis(urea-O)iron(III)] peroxydisulfate (compound 1) fit to this systematic work, because the sulfur content of the oxidative peroxydisulfate can be eliminated as SO2, and studies on this compound can serve as a strong basis to prepare various solid solutions of peroxydisulfate with metal containing oxygen bridged anions, e.g., dichromate that facilitates low-temperature synthesis of nanosized Fe,Cr-oxides with adjusted Fe/Cr ratio.

Still, this reaction route to prepare iron-oxides was used only in a few cases, e.g., the decomposition of [Fe(urea-O)6](NO3)3 [16, 17, 38, 39] or [Fe(urea-O)6](MnO4)3 [1] was investigated. Due to the limited information available about hexakis(urea-O)iron(III) compounds [40, 41], we synthesized, characterized, and thermally decomposed a strongly oxidizing anion-containing compound, [hexakis(urea-O)iron(III)] peroxydisulfate (compound 1). In this paper, we present detailed single-crystal X-ray diffraction, IR, Raman, UV, and Mössbauer spectroscopic as well as thermal analysis studies of compound 1, including identification and characterization of the thermal decomposition intermediates and oxide products formed in an inert atmosphere and in air.

Results and discussion

Preparation and properties

[Hexakis(urea-O)iron(III)] peroxydisulfate (compound 1) was first prepared by Barbieri in the metathesis reaction of an aq. [hexakis(urea-O)iron(III)] chloride and a saturated aq. ammonium persulfate solution as a blueish-green crystalline mass, but neither the yield nor the properties of compound 1 were given [41]. A similar reaction of in situ prepared hexakis(urea-O)iron(III) nitrate (from iron(III) nitrate and six equivalents of urea) and sodium persulfate at room temperature resulted in light-blue blocks during a week with 65% yield. Compound 1 is anhydrous, easily soluble in water (57.5 g/100 mL at 25 °C), but insoluble in organic solvents such as benzene, toluene, chloroform, and carbon tetrachloride. Its saturated aqueous solution is yellow and strongly acidic (pH 2.02). Its pycnometric density is 1.94 g cm−3 at 25 °C. No polymorph phase transition of compound 1 was found between 77 K and its decomposition temperature (396 K) (ESI Fig. S1). Its powder XRD diffractogram is given in ESI Fig. S2.

Single-crystal X-ray diffraction

Light-blue blocks of compound 1 crystallized in triclinic system (space group:\(P\overline{1 }\), a = 9.9126 Å, b = 11.917 Å, c = 12.919 Å, α = 63.1748(18)°, β = 88.6049(14)°, γ = 68.4768(18)°, V = 1248.4 Å3), and its calculated powder pattern agrees well with the experimental PXRD results (ESI Fig. S3). Crystal data and details of the structure determination and refinement are listed in ESI Table S1, and selected parameters of the structure are summarized in ESI Tables S2–S8. The asymmetric unit of the crystal contains one Fe(urea)63+ complex cation with six O-coordinated urea ligands and one and a half S2O82− anions [Fig. 1(a)]. The Fe(urea)63+ complex cation has an octahedral arrangement of urea ligands, the rotation direction of the urea propellers is different at the two sides of the complex [Fig. 1(b)] as was found for Fe(urea-O)6](MnO4)3 (compound 2) [1], Fe(urea-O)6]I3 (compound 3) [42], and for Fe(urea-O)6](I3)3 (compound 4) [43]. This propeller-like arrangement is stabilized by intramolecular hydrogen bonds (altogether 6) formed between each coordinated oxygen atom and the neighboring urea NH2 group [Fig. 2(a)]. There are two kinds of peroxydisulfate ions (labeled with B and C) distinguished in the structure by the number and positions of hydrogen bonds to the urea ligands of the complex cation (Figs. 3c, d, and 2b).

Figure 1
figure 1

(a) The asymmetric unit of compound 1 and (b) propeller-like arrangement of the urea ligands.

Full size image
Figure 2
figure 2

(a) Intramolecular hydrogen bonds in the Fe(urea)63+ cation. (b) Arrangement of the hydrogen-bonded peroxydisulfate anions around the Fe(urea)63+ cation.

Full size image
Figure 3
figure 3

Intermolecular hydrogen bonds in compound 1. The intermolecular hydrogen bonds formed by the (a) upper and (b) lower part of the Fe(urea)63+ ion. The hydrogen bond network of (c) the “B” and (d) the “C” peroxydisulfate anions.

Full size image

The Fe(urea)63+ cation of compound 1 contains altogether 24 N–H bonds which form both intramolecular (within the cation, N–H⋯O bonds between the neighboring ligands) and intermolecular (between the cation ligand and peroxydisulfate anions) hydrogen bonds in the structure. Six of these hydrogen bonds, just discussed, are monodentate intramolecular hydrogen bonds [Fig. 2(a)]. Due to the differences between the axial and equatorial Fe⋯O=C(NH2)2 distances, the O⋯H–N distances and the strength of these hydrogen bonds are not equal (ESI Table S2).

Intermolecular hydrogen bonds between the Fe(urea)63+ cations and the peroxydisulfate anions set up a complex hydrogen bond network. Every urea ligand has two NH2 groups each having 2 potential H-bond-forming sites, thus the number of hydrogen bonds per urea ligands can vary between 0 and 4. The complexity of the system is further increased by the formation of bifurcated H-bonds, and by the presence of two different peroxydisulfate anions in the asymmetric unit with different hydrogen bond patterns. The intermolecular H-bond interactions in compound 1 are visualized in Fig. 3.

Two different peroxydisulfate anions (labeled as anions B and C) could be found in the asymmetric unit of the crystal structure of compound 1 with different hydrogen bond patterns. The hydrogen bond network formed by anion “C” is symmetrical, that is, the two halves of the anion form the same number of hydrogen bonds with the same number of cations [Fig. 3(d)]. In contrast, H−bond interactions of anion “B” are arranged asymmetrically [Fig. 3(c)].

Anion “B” connects 6 Fe(urea)63+ cations with a total of 20 intermolecular hydrogen bonds. The two O3S−O− half−units of the anion form different interactions, thus the hydrogen bond network of this anion is strongly asymmetric. The first half unit of the anion (with oxygen atoms O1B, O2B, O3B, and O4B) forms 11 hydrogen bonds: O1B and O2B each form 4 hydrogen bonds with 3 urea ligands, while O3B and O4B form 2 and 1 hydrogen bonds, respectively. In contrast, the other half unit (with oxygen atoms O5B, O6B, O7B, and O8B) forms only 9 hydrogen bonds: the 3 terminal oxygens (O6B, O7B, and O8B) each form 3 hydrogen bonds, while the bridging peroxy oxygen (O5B) does not form hydrogen bonds. While O7B forms hydrogen bonds with 3 different urea ligands, O6B and O7B each form hydrogen bonds with 2 ligands, thus these oxygen atoms each interact with 2 nitrogens from the same ligands.

Interestingly, while anion “C” also connects 6 Fe(urea)63+ cations with a total of 20 hydrogen bonds, it forms a symmetric hydrogen bond network. Two of the three terminal oxygens (O1C and O3C) each form 3 hydrogen bonds with 3 and 2 urea ligands, respectively, while O2C forms only two hydrogen bonds with two different nitrogen atoms of the same ligand. The bridging peroxy oxygen (O4C) also forms 2 hydrogen bonds with two different urea ligands.

The six urea ligands of a Fe(urea)63+ cation (numbered from 1 to 6 in Fig. 3) form different number of hydrogen bonds with different number of peroxydisulfate anions [Fig. 3(a), (b)]. Ligand 1 forms 5 hydrogen bonds with 3 “B” anions: N1B is connected to the terminal oxygens of two different anions with 1 and 2 hydrogen bonds, while N1A forms 1 hydrogen bond with a terminal oxygen atom and 1 hydrogen bond with a bridging peroxy oxygen atom of a different anion. Ligand 2 forms a total of 6 hydrogen bonds with 4 “B” and 1 “C” anions: N2A forms 2 hydrogen bonds with 2 different “B” ligands, while N2B forms 4 hydrogen bonds to terminal oxygens of 2 “B” and a terminal and a peroxy oxygen of 1 “C” anions. Ligand 3 forms the smallest number of hydrogen bonds: it makes only 4 hydrogen bonds with terminal oxygens of 2 “B” and 1 “C” anions. Ligand 4 and 5 both make 5–5 hydrogen bonds with 2 “B” and 1 “C” anions but in a different arrangement. In the case of ligand 4, N4A makes 4 hydrogen bonds with 2 “B” and 1 “C” anions, while N4B forms 1 hydrogen bond with a “B” anion. For ligand 5, N5A forms 1 hydrogen bond with a “C” anion, while N5B forms 4 hydrogen bonds with 2 “B” anions and the same “C” anion that interacted with N5A. Ligand 6 makes 6 hydrogen bonds with 1 “B” and 1 “C” anions: N6A forms 3 hydrogen bonds with terminal oxygens of 1 “B” and 1 “C” anions, while N6B makes 3 hydrogen bonds with terminal and peroxy oxygens of a single “C” anion.

The packing of the molecules is presented in Fig. 4. Layers of Fe(urea)63+ and S2O82− alternate with each other in the crystal [Fig. 4(b), (c)]. The double−propeller−shaped complex cations form columns along with the unit cell axis c. The peroxydisulfate ions, filling the channels between the rods, interconnect the Fe(urea)63+ ions within and between the columns by forming hydrogen bonds.

Figure 4
figure 4

View of the unit cell along (a) the a, (b) b, and (c) c axes.

Full size image

The shortest distances between the iron ions in the structure are 6.433 Å and 6.532 Å connecting neighboring complexes along unit cell axis a. The iron atoms along the other two axes have greater separation by the urea molecules and the channels between them are filled by peroxydisulfate ions. Their distances are 9.9–11.8 Å (ESI Fig. S4).

Vibrational spectra of compound 1

There are 3 crystallographically different persulfate anions, 2 of which are at positions of trivial symmetry and there are 6 crystallographically different urea ligands. Altogether 74 atoms build the unit cell. Since there are 3 × 12 and 6 × 27 internal normal modes in 3 sets of persulfate and 7 sets of urea bands, respectively, are expected, together with external modes and overtone/combination bands, it is not worth doing correlation analysis. Instead of that, deuteration experiments were done on compound 1 to separate the heavily overlapping modes belonging to the urea NH2 other groups. The IR and Raman spectra of compound 1 and compound 1-D were recorded at room temperature in the range of 4000–100 cm−1 (Fig. 5 and ESI Fig. S5). Due to the deuteration, the modes that belong to the N–D bonds shifted to lower wavenumbers as compared to the modes of N–H bonds (υ(NH)/υ(ND)≈1.35, ESI Fig. S6). Substantial shifts in the positions of the persulfate normal modes were also observed due to differences in the strength of O…H and O…D interactions (Table 1, Fig. 5). Low-temperature (263, 193 and 93 K) Raman spectra were also recorded (ESI Fig. S7). The assignments of the vibrational modes of the anionic (Table 1) and the cationic (Table 1 and ESI Table S9) parts of compound 1 are given. Out of the 24 normal modes of the peroxydisulfate ion [44, 45], five belong to the bridging S–O–O–S structure. The vibrational modes of the formal S–O–O–S linkage are shown in ESI Fig. S8. All of them are expected to be IR and Raman active. Due to the presence of two different peroxydisulfate anions, two series of bands belonging to the normal modes may be expected to appear. Most of the peroxydisulfate ion modes coincide with some vibrational modes of the urea ligands (Table 1). To handle this coincidence, compound 1 was deuterated with D2O, and the IR spectra of the perdeuterated compound 1 (compound 1-D) [Fig. 5(b)] were evaluated to separate the bands belonging to urea and peroxydisulfate.

Figure 5
figure 5

The room-temperature IR (a) and Raman spectra (b) of compounds 1 and 1-D.

Full size image
TABLE 1 Assignment of peroxydisulfate ion vibrational modes (cm−1) in the IR and Raman spectra of compound 1.
Full size table

The deformation modes of S–O–O–S linkage are expected in the far-IR range (< 400 cm−1) together with the υ(Fe–O) modes [44,45,46,47]. The presence of two kinds of peroxydisulfate ion and the hydrogen bond induced asymmetry of the “B” peroxydisulfate ion results in more complicated experimental IR and Raman spectra than that of alkali peroxydisulfates [45] (ESI Fig. S8). The IR and Raman bands located around 400 cm−1 were assigned as the δSOOS modes by Cleaver [45] and as the terminal δSO3 mode by Skogareva [48]. Comparing the available IR and Raman data of the alkali, alkali metal and some [M(NH3)4]2+ (M = Cu, Zn) complex peroxydisulfates [45, 49], increasing the polarizing effect of cation (decreasing the ionic size or increasing the charge) increases the wavenumber of δ(SOOS) and υ(SOOS) modes. Due to the presence of the trivalent iron cation in compound 1, the bands at 409 and 391 cm−1 might be assigned to δSOOS mode of the two different peroxydisulfate ions. The size of the complex cation, however, increased due to the presence of six urea ligands, resulting in a strong decrease in the polarizing effect of the complex cation. Comparison of the υas(SOOS) IR wavenumbers found in the IR spectra of alkali, alkaline earth metal and divalent metal (Cu, Zn) ammine complex peroxydisulfates, and compound 1, it is obvious that the complex cations ([M(NH3)4]2+ (M = Zn, Cu, (υas(SOOS) = 1048 cm−1) and [Fe(urea)6]3+ (1045 an 1035 cm−1) have less polarizing effects than the less polarizing alkali (cesium) cation (υas(SOOS) = 1052 cm−1). Based on this relationship, the bands at 391 and 409 cm−1 in the IR spectrum of compound 1 belong to the terminal S–O bonds and not to the deformation mode of SOOS linkage. The peroxy-bond is not IR active in peroxydisulfates [44]. However, due to the strong distortion (e.g., the antisymmetric hydrogen bond in the “B” peroxydisulfate anion) of symmetries, the bands appearing at 796 cm−1 and 830 cm−1 in the IR and at 810 and 823 cm−1 in the Raman spectra of compound 1 might be attributed both to the υ2 (O–O) [44] and to υs(SOOS) modes. The symmetric and antisymmetric stretching and deformation modes of the terminal SO3 group were also assigned (Table 1). There was no pyrosulfate ion (S2O72−) as a decomposition product detected.

The vibrational modes assigned to the urea ligands and iron–oxygen interactions are given in ESI Table S9. The coordination of urea through its oxygen atom influences on the C=O and the C–N bond lengths. The C=O bond length elongates, whereas the C–N bond length shortens compared to the gaseous urea. Therefore, the corresponding νCO and νCN stretching modes shift to lower and higher wavenumbers, respectively (ESI Table S9). Due to the hydrogen bonds formed by the –NH2 groups, the N–H bonds become longer; thus, the NH2 stretching vibrations will shift toward lower wavenumbers as was observed in the spectra of solid urea [49] and hexakis(urea-O)iron(III) permanganate [1]. The IR spectrum of compound 1 contains two high- and two low-intensity bands in the range above 3000 cm−1. The overlapping bands were separated via deuteration of compound 1, by comparing the IR spectra of the deuterated and non-deuterated samples. Furthermore, the combination/overtone bands could also be used to determine some original band positions as well. Following the method given in [50], together with the deuteration experiments, ν(C=O) components in the complex band systems were located at 1504 cm−1 and 1503 cm−1 in the IR and Raman spectra, respectively. A wide band appears at 1625 cm−1 in the IR spectrum of compound 1 and splits into two well-defined peaks (1693 cm−1 and 1620 cm−1) in the Raman spectra. The δs(N–H) band in the Raman spectra could be assigned as sharp peaks at 1692 cm−1 and at 1693 cm−1 at room temperature and at 100 K, respectively. The δs(N–H) position in the IR spectrum was calculated to be 1685 cm−1. The δas(N–H) mode is at 1625 cm−1 and at 1620 cm−1 in the IR and Raman spectra, respectively.

UV spectroscopy

Compound 1 has three well-defined absorption bands in its solid-phase UV–Vis spectrum (ESI Fig. S9). The UV spectra of compound 1, in principle, are built up from the increments of FeIII crystal field bands, the bands belonging to the urea ligand and to the persulfate anion.

Compound 1 contains a high-spin FeIII state with a 6S ground term. The 6S term is not split by any crystal field, and the electronic transitions are spin and Laporte forbidden. Thus, among the expected 4T1g(G) ← 6S, 4T2g(G) ← 6S, and the 4Ag, 4Eg(G) ← 6S transitions of the [Fe(urea-O)6]3+ cation, the first does not appear, the second one appears at 590 nm as a very weak and wide band, and the third transition was found at 432 nm. The 4Ag,4 Eg(G) ← 6S transition is degenerate in octahedral symmetry.

The urea n-π* and π − π* bands in the UV spectrum of compound 1 show up as a wide band in the UV range centered at 212 and 228 nm, due to various strength and bond lengths of N–H and C=O linkages, including possible FeIII–O=C LMCT bands. The uncoordinated urea has bands in the UV range at 211, 230–260, 299, 337, 361, and 380 nm due to various tautomer and hydrogen-bonded clusters present in solid urea [50]. The absorption belonging to the excitation of the peroxide bond in the persulfate anion [51] is located around 265 nm, near to the band found in the UV spectrum of sodium persulfate at 260 nm (ESI Fig. S10a, b).

Room-temperature and low-temperature (liq. N2) Mössbauer spectrum of compound 1

The Mössbauer measurements of compound 1 were done both at room temperature and at liquid nitrogen temperature. At both temperatures, the Mössbauer spectra show only broad singlet-like envelopes attributable to only one iron(III) site with magnetic relaxation (Fig. 6). This line shape is quite common for this type of complexes [1, 52,53,54]. The spectra were evaluated by the Blume-Tjon two-state magnetic relaxation model (B-T model) provided by the Mosswinn code [55]. The Mössbauer parameters calculated by this model are as follows: the isomer shift (δ), the amplitude of the relaxing hyperfine magnetic field (H), the line width (Γ), the component Vzz of the electric field gradient (EFG), the asymmetry parameter of the EFG (ETA) (these latter two are not calculated in a Lorentzian fit, only the quadrupole splitting), and, finally, the most specific Mössbauer parameter of this model is the jump up the rate of the relaxing magnetic field (forced to be equal to the jump down rate because of the two-state relaxation). This parameter is the base ten logarithm of the jump-up frequency (W), from which the spin relaxation time may be calculated as π /W.

Figure 6
figure 6

The (a) room and (b) liquid nitrogen temperature Mössbauer spectrum of compound 1.

Full size image

At the first attempt, we used B-T model for the evaluation of our experimental Mössbauer data without any fixed values. In this case, a very broad distribution of fitted parameters was the result of the evaluation. In some cases, even unphysical parameters were obtained. We had a similar experience in the case of hexakis(urea-o)-iron(III)-permanganate complex that we have discussed in ref. [1]. At the second attempt, we used the B-T model again, but now the previously published ETA = 0.232 and Vzz = -0.762 × 1021 V m−2 parameters obtained by density functional theory calculations for the complex cation only (ESI Table S10) were used as fixed values. The Mössbauer parameters that we obtained in this way (ESI Table S11) met the Mössbauer parameters expected for other high-spin iron(III) urea complexes [1, 52,53,54].

The experimental Mössbauer spectra were recorded at 295 K and liquid N2 (80 K) temperatures. The isomer shifts obtained were 0.416 mm s−1 and 0.542 mm s−1, respectively. This confirms that compound 1 is a high-spin iron(III) complex. The isomer shifts are similar to those we have found for hexakis(urea-o)-iron(III)-permanganate (0.412 mm s−1 and 0.501 mm s−1) [1] and are smaller than those for the previously published hexakis(urea-o)-iron(III)-chloride (0.58 mm s−1 and 0.60 mm s−1) [53]. Since single-crystal X-ray diffraction proved the presence of hydrogen bonds between [Fe(urea)6]3+ cations and (S2O8)2− anions, the isomer shifts of compound 1 are in good agreement with our previously published finding [1]. We can conclude that the hydrogen bonds reduce the 3d electrons density on the iron(III) central ion and thereby the isomer shift decreases. The line widths are 0.608 mm s−1 and 0.966 mm s−1 at 295 K and 80 K, respectively, and it is similar to the [Fe(urea)6](MnO4)3 line widths (0.626 mm s−1 and 0.875 mm s−1) which is consistent with the spin relaxation effect [56]. The shortest distance between the Fe–Fe atoms is rather high, 6.433 Å, which supports this type of relaxation.

Solid-phase decomposition reactions of compound 1

The thermal decomposition of compound 1 was followed by TG-MS and DSC methods in inert (argon or N2) and oxidative (air) atmospheres. The overall weight losses until 800 °C were found to be 75% and 85% in the inert and in the oxidative atmosphere, respectively (ESI Figs. S11–S14). The final product of the decomposition was α-Fe2O3 in every case as was confirmed by powder XRD and Mössbauer spectroscopy as well (ESI Figs. S15, S16).

The thermal decomposition is a multistep process with well-separated steps and starts at around 130 °C with exothermic character in both atmospheres. Thus, the aerial oxygen does not play a role in the initiation of the decomposition reaction (ESI Figs. S11–S14). The peak temperatures of the first decomposition steps are 151 °C and 159 °C(DTG) with ∆Hr = 386.45 and 375.93 kJ/mol (DSC peak temperatures are 154 and 155 °C) in inert and oxidative atmospheres, respectively (ESI Figs. S13, S14). The thermal decomposition reaction of compound 1 has a complicated character. First, instead of an endothermic ligand loss, an exothermic process takes place as was observed in the case of [Fe(urea)6](MnO4)3 [1], where the permanganate ions oxidized the urea in the first stage of the decomposition reaction. The TG-MS study of the decomposition of compound 1 confirmed the presence of such urea oxidation products as H2O (m/z = 18), CO2 (m/z = 44), and N2 (m/z = 28). NO (m/z = 30) and SO2 (m/z = 64) as a reduction product of the persulfate ion could also be detected in small amounts (Fig. 7). There was no gaseous ammonia found (the m/z = 17 peak found belongs to the OH fragment of water based on the m/z = 18 and m/z = 17 peak intensity ratio [Fig. 7(a)], which agrees well with the value for H2O → OH fragmentation [57].

Figure 7
figure 7

The TG–MS ion curves of compound 1 in argon atmosphere. The numbers in the figure represents the m/z values of the formed ions.

Full size image

In order to study the phase relations in the decomposition intermediates formed in the thermal decomposition step, isotherm heatings of compound 1 at some particular temperatures (T) were done for 0.5 h both in air and in inert atmospheres. The samples are denoted as I–T–O2(N2), where T means the temperature of isotherm (I) heating in the indicated inert (N2) atmosphere or in air (O2). The Mössbauer spectroscopic results of the solid decomposition intermediate I-160-N2 of compound 1 indicate only iron(III) environments (Fig. 8 and ESI Table S12). Thus, there was no redox reaction occurring between the iron(III) center and the urea ligand. Due to the lack of oxygen, the only oxidizing agent could only be the persulfate ion of compound 1. In accordance with this, the bands belonging to the persulfate ion completely disappeared from the IR and Raman spectra of I-160-N2 [Fig. 9(a) and ESI Figs. S17, S18]. The IR and Raman results of the intermediate I-160-N2 confirmed the presence of sulfate ion (ν3(F2) = 1104 cm−1, ν1(A1) = 986 cm−1 (strongest Raman peak of SO42−), ν4(F2) = 597 cm−1, and ν2(E) = 466 cm−1 [58]) as a persulfate reduction product. Surprisingly, bands belonging to ammonium ions appeared in the IR and Raman spectra of I-160- N2 [ν1, ν2, ν3, and ν4 = 3201, 3077, 2834, 1419 cm−1 Fig. 9(a)] (and in the IR spectra of I-160-O2 too (ESI Fig. S18a)). The ammonia source during the decomposition of compound 1 is the direct decomposition or hydrolysis, shown by equations E1, E2, E3, (water forms as oxidation products from urea and persulfate detected by TG-MS) as was observed in the case of [Fe(urea)6](MnO4)3 [1]

$${\text{2 H}}_{{2}} {\text{NCONH}}_{{2}} = {\text{H}}_{{2}} {\text{NCONHCONH}}_{{2}} + {\text{NH}}_{{3}}$$
(E1)
$${\text{H}}_{{2}} {\text{NCONH}}_{{2}} { = }\,{\text{NH}}_{{3}} { + }\;{\text{HNCO}}$$
(E2)
$${\text{H}}_{{2}} {\text{NCONH}}_{{2}} + {\text{H}}_{{2}} {\text{O}}\,{ = }\,{\text{2NH}}_{{3}} { + }\,{\text{CO}}_{{2}}$$
(E3)
Figure 8
figure 8

The Room-temperature Mössbauer spectra of decomposition intermediates formed at various temperatures in the isothermal decomposition of compound 1 in (a) air and (b) in N2 atmosphere. Please mind the different velocity scales.

Full size image
Figure 9
figure 9

The IR spectra (a) and the XRD (b) of decomposition intermediates formed at various temperatures in the isothermal decomposition of compound 1 under N2.

Full size image

The presence of biuret and isocyanate could be detected with IR [Fig. 9(a) at 160 °C], where the ν(NH) bands and mixed amide ν(CONH) bands do not change much on heating up to 300 °C. An IR band at 2213 cm−1 belongs to an intermediate with carbon–nitrogen multiple bond (e.g., coordinated or non-coordinated isocyanate) and appears as a weak band in the IR spectrum obtained for sample I-160-N2. The formation of isocyanate intermediates was detected in the oxidative degradation of hexakis(urea-O)iron(III) nitrate as well [59].

The main iron-containing phase of the decomposition intermediates formed at 160 °C either in air or in inert atmosphere was assigned to pyracmonite, (NH4)3Fe(SO4)3 [60] [Fig. 9(b) and ESI Fig. S19]. The Mössbauer spectrum obtained for sample I-160-N2 shows that ca. 73% of the iron-containing products is (NH4)3Fe(SO4)3, whereas the remaining 27% of iron is in the trivalent state located in octahedral (71%) and tetrahedral (29%) coordination environment [Fig. 8(b) and ESI Table S12] [61]. The same phases are present in the intermediate I-160-O2 [Fig. 8(a) and ESI Table S13] in the same amount and ratio. Thus, external oxygen does not play any role in this process.

The formation of (NH4)3Fe(SO4)3 in huge amounts shows that the persulfate ion is mainly reduced into sulfate ion, and the urea ligand is decomposed mainly into biuret and ammonia (without redox reaction). The persulfate-driven oxidative urea degradation with the formation of water, CO2, and SO2 also occurs, as was detected in the first decomposition step by TG-MS [Fig. 7(c), (d)]. This exothermic degradation reaction can serve as a source of activation energy for the decomposition of urea with NH3 elimination; however, the majority of urea decomposes without a redox reaction; thus, the persulfate has to be consumed by another reducing agent. The large amount of (NH4)3Fe(SO4)3 formed (71%) (ESI Tables S12, S13) shows that the main redox reaction does not involve oxygen transfer from the persulfate ion (no SO2 formation), because the Fe-to-S ratio in (NH4)3Fe(SO4)3 is the same as that in compound 1 (2:6); thus the huge SO2 loss would prevent the formation of (NH4)3Fe(SO4)3 in such a large amount (71%) as we found. The (NH4)3Fe(SO4)3 formation cannot be 100%, of course, because a fraction of the urea ligands is oxidized by persulfate consuming actually oxygen from the persulfate and resulting in the formation of H2O, CO2, and SO2 as detected by TG-MS [Fig. 7(a), (c), (d)]. The relatively low mass loss found (9.2 and 14.6% in air and argon, respectively) confirms that the complete urea degradation may only be a minor side-reaction (theoretical mass loss related to one urea is 9.9%) and the reduction of persulfate results mainly in sulfate ions.

The formation of pyracmonite with NH4+ to FeIII to SO42− stoichiometry of 3:1:3 can be explained if we assume that the main persulfate-driven redox reaction is the oxidation of ammonia and not directly of urea.

$${\text{3S}}_{{2}} {\text{O}}_{8}^{2 - } + {\text{8NH}}_{{3}} = \,{3}\left( {{\text{NH}}_{{4}} } \right)_{{2}} {\text{SO}}_{{4}} + {\text{N}}_{{2}} + {\text{3SO}}_{4}^{2 - }$$
(E4)

Since two iron(III) belong to three persulfate ions in compound 1, the ammonium and sulfate stoichiometry for pyracmonite can be provided by the ammonium sulfate formed in reaction E4. If one assumes an oxygen transfer reaction, e.g.,

$${\text{3S}}_{{2}} {\text{O}}_{8}^{2 - } + {1}0{\text{NH}}_{{3}} = \,{3}\left( {{\text{NH}}_{{4}} } \right)_{{2}} {\text{SO}}_{{4}} + {\text{ 3SO}}_{{2}} + {\text{ 6OH}}^{ - } + \,{\text{2N}}_{{2}} ,$$
(E5)

then the amount of sulfate formed decreases and could not fulfill the stoichiometry needs for the formation of two iron(III) since the FeIII-to-SO42− ratio would only be 2:3 at maximum.

The second decomposition step in an inert atmosphere is a two-step endothermic process (overall reaction heat is 264.5 kJ/mol) with 201 and 207 °C DSC peak temperatures (ESI Figs. S13, S14). In the intermediate I-220-N2, two different iron environments were found by Mössbauer spectroscopy: one iron(II) similar to (NH4)2Fe(SO4)2 with 15% relative fraction [61], and another one, iron(III) environment (85%) [Fig. 8(b) and ESI Table S12] identified as (NH4)3Fe(SO4)3 [62]. The presence of (NH4)3Fe(SO4)3 could be confirmed, but the peaks for (NH4)2Fe(SO4)2 could not be detected by powder XRD [Fig. 9(b) and ESI Fig. S18]. The reduction of iron(III) by the biuret-like components is confirmed by the evolution of H2O and CO2 around 200 °C [Fig. 7(a), (c)]. The formation of iron(II) as (NH4)2Fe(SO4)2 detected by Mössbauer Spectroscopy [Fig. 8(b) and ESI Table S12] probably takes place within the lattice of (NH4)3Fe(SO4)3.

The intermediate I-220-O2 contains 93% (NH4)3Fe(SO4)3 and 7% iron(III) in octahedral environment [Fig. 8(a) and ESI Table S13]. The reaction is exothermic with 207 °C peak temperature, ∆Hr = − 289.2 kJ/mol. The intermediate containing C–N multiple bond (isocyanate) disappears from the IR spectrum recorded for sample I-220-O2 [Fig. 9(a)]. The bands belonging to biuret-like components (ν(C=O) = 1669 cm−1) are present with even lower intensity than in the IR spectrum of I-160-O2 [Fig. 9(a)].

The further heat treatment of I-220 samples prepared either in air or in inert atmosphere shows the main decomposition characteristics of (NH4)3Fe(SO4)3 and (NH4)2Fe(SO4)2 [62,63,64,65,66,67,68].

The main decomposition route of (NH4)3Fe(SO4)3 is

$$\left( {{\text{NH}}_{{4}} } \right)_{{3}} {\text{Fe}}\left( {{\text{SO}}_{{4}} } \right)_{{3}} \to {\text{NH}}_{{4}} {\text{Fe}}\left( {{\text{SO}}_{{4}} } \right)_{{2}} \to {\text{Fe}}_{{2}} \left( {{\text{SO}}_{{4}} } \right)_{{3}} \to \alpha{\text{-Fe}}_{{2}} {\text{O}}_{{3}} \left( {{8}00\,^\circ {\text{C}}} \right)$$

The primary thermal decomposition product of (NH4)3Fe(SO4)3 was identified as NH4Fe(SO4)2 by XRD [Fig. 9(b)] and confirmed by the Mössbauer spectrum as the only product in air [Fig. 8(a)]. In the samples made in inert atmosphere, three different iron coordination environments could be found: the main component NH4Fe(SO4)2 (69%), and two iron environments belonging to transition states, one contains iron(III) (17%), whereas the other iron(II)-containing species has an environment similar to that in (NH4)2Fe(SO4)2 (14%) [Fig. 8(b) and ESI Table S12].

The amount of biuret-like condensation products decreases (ν(C=O = 1669 cm−1), and the most important change in the IR and Raman spectra is shifting the νs(S–O) stretching mode from ν1 = 986 cm−1 and 996 cm-1 (IR) to 1036 and 1024 cm−1, respectively [Fig. 9(a)]. It may be attributed to the change in the sulfate coordination environment (formation of a polymer-like sulfate-bridged structure) [47].

On further heating, NH4Fe(SO4)2 decomposes [69] in air in a redox reaction with the formation of water ad SO2 around 400 °C according to equation E6 [66]:

$${\text{6NH}}_{{4}} {\text{Fe}}\left( {{\text{SO}}_{{4}} } \right)_{{2}} = \, {\text{3Fe}}_{{2}} \left( {{\text{SO}}_{{4}} } \right)_{{3}} + {\text{ 3SO}}_{{2}} + {\text{4NH}}_{{3}} + {\text{N}}_{{2}} + {\text{6H}}_{{2}} {\text{O}}$$
(E6)

Both H2O an SO2 could be detected by TG-MS [Fig. 7(a), (d)], the reaction product at 565 °C is solely Fe2(SO4)3 (ESI Fig. S19). Further decomposition of iron(III) sulfate into iron(III) oxide with SO2 evolution above 500 °C [Fig. 7(d)] can also be detected [1, 67, 68]. By the end of the heat treatment in air, the final product at 800 °C is hematite based on the powder XRD and Mössbauer spectroscopy (ESI Fig. S15). In the case of I-640-O2, the sample contains 6% of wüstite. This phase could form during the decomposition of the Fe2(SO4)3. The release of the SO2 results in the formation of vacancies in the microenvironment of iron that can produce wüstite-like parameters for this metastable phase.

The sample made in an inert atmosphere at 420 °C consists of unreacted NH4Fe(SO4)2 and Fe2(SO4)3 as decomposition product, and Mössbauer spectroscopy shows the presence of an iron(II) compounds as well (58%) [Figs. 8(b), 9(b) and ESI Table S12]. The low-temperature Mössbauer measurement showed that there are two iron(II) signals with the same quadrupole splitting but different isomer shifts (ESI Fig. S20 and ESI Table S14). These two phases could be assigned to iron coordination environments found in (NH4)2Fe(SO4)2 and FeSO4. At room temperature, the Mössbauer parameters of the (NH4)2Fe(SO4)2 and FeSO4 cannot be distinguished, but at 86 K, they can easily be resolved. The appearance of the second doublet only at 86 K is due to the different temperature dependence of the Mössbauer parameters of the different iron(II) microenvironments.

XRD of the sample I-420-N2 contains the peaks belonging to NH4Fe(SO4)2, Fe2(SO4)3, and FeSO4, and heating until 490 °C resulted in decreasing peak intensities of NH4Fe(SO4)2 and increasing peak intensities of FeSO4 [Fig. 9(b)]. It means that there is a substantial reduction rate driven by the ammonium ion of NH4Fe(SO4)2 and the parent of FeSO4 is the NH4Fe(SO4)2 itself. The transformation of NH4Fe(SO4)2 into Fe2(SO4)3 around 400 °C is a well-known process [63], but neither FeSO4 nor (NH4)2Fe(SO4)2 was found. Our experiments were done at 420 °C. Therefore, we checked the effect of temperature on the quasi-intramolecular redox reaction of NH4Fe(SO4)2. The Mössbauer data showed the increase of the iron(II) phase content up to 75% [Fig. 8(b) and ESI Table S12] on heating the sample I-420-N2 until 490 °C in inert atmosphere. The iron(II) compound is FeSO4, according to the powder XRD results [Fig. 9(b)]. Its formation is summarized in equation E7.

$${\text{2NH}}_{{4}} {\text{Fe}}\left( {{\text{SO}}_{{4}} } \right)_{{2}} { = }\,{\text{2FeSO}}_{{4}} + {\text{ 2SO}}_{{2}} + {\text{4H}}_{{2}} {\text{O}} + {\text{N}}_{{2}}$$
(E7)

Two peaks in the TG-MS belonging to SO2 (m/z = 64) appear around 540 °C and 590 °C, and this is consistent with the DTG peaks found at 526 °C and 579 °C [Fig. 7(d)]. The first peak belongs to the redox reaction of NH4Fe(SO4)2 leading to FeSO4 formation, whereas the second peak belongs to the SO2 evolution from Fe2(SO4)3 and FeSO4 with the formation of α-Fe2O3 (E8) [65,66,67,68].

$${\text{2FeSO}}_{{4}} + \, = \alpha {\text{{-}Fe}}_{{2}} {\text{O}}_{{3}} + {\text{SO}}_{{2}} + {\text{`SO}}_{{3}}{\text{'}}$$
(E8)

The iron(III) content consists of two phases. Regular hematite and defect hematite (supposedly with oxygen vacancies) phase formed through the decomposition of Fe2(SO4)3 [64,65,66,67]. Further increasing the temperature until 590 °C in inert atmosphere, the hematite content increased to 95% and the amount of the defect phase decreased to 5% [Fig. 8(b) and ESI Table S12]. In the powder XRD of the sample, a shoulder can be seen on a few reflections that can indicate the defect phase [Fig. 9(b)].

Both I-590-N2 and I-640-O2 contain small amount of FeOOH, 5 and 12%, respectively (Fig. 8 and ESI Tables S12, 13). This relatively small amount shows that this phase may form only at the surface of the samples during the removal from the furnace due to the humidity of air.

Conclusion

[Hexakis(urea-O)iron(III)] peroxydisulfate and its perdeuterated form were synthesized and spectroscopically characterized. The IR and Raman bands of [hexakis(urea-O)iron(III)] peroxydisulfate were assigned using low-temperature Raman measurements and spectroscopical study of the perdeuterated compound 1. The main structural features of [hexakis(urea-O)iron(III)]peroxydisulfate are as follows:

  1. (a)

    Compound 1 has a cation octahedrally coordinated by six different urea ligands via their oxygen in a propeller-like arrangement, as well as two kinds of persulfate anions which are involved in different intermolecular hydrogen bond networks with the urea ligands of neighboring complex cations.

  2. (b)

    One of the persulfate ions has a symmetric connection to the hydrogen bond system via its two peroxy and the terminal oxygens, whereas the other kind of persulfate ion has asymmetrical connection with only one of its peroxy and the terminal oxygens involved.

The thermal studies of compound 1 showed that its thermal degradation reaction is initiated by oxidation of ~ 1 molecule of urea with persulfate ion, together with the releasing of ammonia from the remaining urea molecules while biuret and isocyanate form. The remaining persulfate ions are reduced into sulfate by the liberated ammonia through SO2 and ammonium ion formation. The solid phase contains mainly (NH4)3Fe(SO4)3, which decomposes stepwise into NH4Fe(SO4)2, Fe2(SO4)3, and α-Fe2O3 upon increasing the temperature. In inert atmosphere, some amount of iron(II) compound formed even at 220 °C. The thermal decomposition of the NH4Fe(SO4)2 intermediate also depends on the atmosphere and the temperature. Basically, Fe2(SO4)3 formed together with N2 and SO2, but at 420 °C and 490 °C, in inert atmosphere, FeSO4 also formed in an amount of 27 and 75%, respectively. Finally, both FeSO4 and Fe2(SO4)3 decomposed into α-Fe2O3 and SO2, but the temperature of α-Fe2O3 formation was not the same, 590 and 800 °C, in inert atmosphere and in air, respectively.

Materials and methods

Synthesis of compound 1

Iron(III) nitrate nonahydrate (8.08 g, 0.02 mol) and urea 7.20 g, (0.12 mol) were dissolved in 9 mL of water, and then, the orange-colored solution obtained was mixed with an aq. solution of Na2S2O8 (7.86 g Na2S2O8 in15 mL of water) under stirring. The resulted dark orange-colored solution was left to crystallize at room temperature for a week when light-blue block single crystals were formed. The solid crystalline mass was separated by filtration on a G3 glass filter, washed with a copious amount of cold (0 °C) water, and dried with abs. ethanol and then with diethyl ether. The yield was 65% (18.27 g).

Material characterization

All used chemicals (iron(III) nitrate nonahydrate, sodium peroxydisulfate, ethanol, diethyl ether, nitric acid, potassium permanganate, ammonium iron(II) sulfate (Mohr salt)) were supplied by Deuton-X Ltd., Érd, Hungary.

X-ray powder diffraction patterns were recorded by using a Bragg–Brentano parafocusing goniometer manufactured by Philips (PW-1050) equipped with a Cu anode (40 kV, 35 mA tube power), a secondary beam graphite monochromator, and a proportional counter. Every scan was recorded in a step mode and the diffraction patterns were evaluated by full-profile fitting techniques.

A light-blue, block-like single crystal of 0.3 × 0.2 × 0.1 mm was selected for single -crystal X-ray diffraction measurement. The crystal was mounted on a loop and measured on an XtaLAB Synergy R diffractometer equipped with PhotonJet-R rotating anode source [Cu-K-alpha radiation (λ = 1.54184 Å)], confocal mirrors as a monochromator, and Hypix-6000HE detector. Data collection and data reduction were carried out using CrysAlisPro v.1.171.40.68a program [70]. Data were collected at 100 K using 104 ω scans. A total of 17,580 frames were collected (0.5° rotation, 0.05 s exposure time/frame). The structure was solved using direct methods as implemented in the SHELXS program [71]. Refinement was carried out using SHELXL [71] and Olex2 [72] with full matrix least-squares method on F2. All non-hydrogen atoms were refined anisotropically. Hydrogen atoms were generated based on geometric evidence and their positions were refined by the riding model. Hydrogen atoms were found in difference Fourier maps. Atomic displacement parameters of the S2O82− ions were refined with rigid bonds restraints applied. Olex2 [72] and Mercury [73] were used for molecular graphics and analyzing crystal packing. Crystal data and details of the structure determination and refinement are listed in ESI Table S1, and selected parameters of the structure are summarized in ESI Tables S2–S8. The X-ray crystallographic data have been deposited in the Cambridge Crystallographic Data Centre with CCDC ID 2,192,412 and can be obtained free of charge from the CCDC via www.ccdc.cam.ac.uk/getstructures

The analytical range FT-IR and far-IR spectra of compound 1 were recorded in an attenuated total reflection (ATR) mode at room temperature using a BioRad-Digilab FTS-30-FIR and a Bruker Alpha IR spectrometer for the 4000–400 and 400–100 cm−1 range, respectively.

The Raman measurements of compound 1 were carried out at 93 K, 193 K, 263 K, and 298 K on a Horiba Jobin–Yvon LabRAM microspectrometer. External (532 nm and 785 nm) Nd-laser sources operated at ~ 40 mW and an Olympus BX-40 optical with a temperature-controllable microscope stage (Linkam THMS600) were used in the 4000–100 cm–1 spectral range with 3 cm−1 resolution. The laser beam (50 × objective) was focused and a D0 intensity filter was applied. A confocal hole of 1000 µm and 1800 groove mm−1 grating monochromator was used for light dispersion resolution. The exposure time was 15 s.

The solid-phase UV–Vis (diffuse reflectance) spectrum of compound 1 was measured at room temperature (Jasco V-670 UV–Vis spectrophotometer, equipped with a NV-470 integrating sphere, BaSO4 standard).

57Fe Mössbauer spectroscopy measurements were performed at room temperature and at T = 80 K with a conventional WissEl Mössbauer spectrometer (Starnberg, Germany) operating in constant acceleration mode (57Co source in Rh matrix). Low-temperature measurements were done using a SVT-400-MOSS cryostat (Janis, Woburn, MA, USA) filled with liquid nitrogen. The powdered samples were mixed with polyethylene powder to ensure a perfectly random orientation of the crystallites. The Mössbauer spectra recorded were evaluated by standard computer-based statistical analysis methods, including fitting the experimental data by a sum of Lorentzians or a magnetic relaxation model using a least-squares minimization procedure (MossWinn 4.0 program [55]). The isomer shifts are given relative to α-Fe standard at room temperature.

The TG-MS measurements were carried out in both argon and air by a simultaneous thermal analyzer SDT Q600 from TA Instruments, online coupled with an HPR20/QIC mass spectrometer from Hiden Analytical. The sample holder was an alumina crucible, and an empty alumina crucible was used as reference. The sample mass was 1–3 mg. The carrier gas flow rate was 50 mL min−1. The decomposition was followed from room temperature to 650 °C by a heating rate of 10 °C min−1. During evolved gas analysis, ions selected between m/z = 1–98 were monitored in multiple ion detection modes (MIDs).

The non-isothermal DSC curve of compound 1 was recorded up to 400 °C with a Perkin Elmer DSC 7 apparatus. Another measurement was done between − 130 °C and room temperature to check polymorphic phase transitions. Sample masses varied between 3 and 6 mg. The heating rate was selected to be 5 °C/min in both cases under a continuous nitrogen flow (20 cm3 min−1). The aluminum pans used as sample holders were unsealed.