Nano-ZrO₂@C, Nano-(ZrC, ZrO₂)@C and Nano-ZrC@C Composites Prepared by Plasma-Assisted Carbonization of Zr-Loaded Iminodiacetate-Functionalized Styrene-Divinylbenzene Copolymers

Abstract

We have developed an easy route to prepare (nano-ZrO₂,nano-ZrC_x)@C composites with varying ZrO₂/ZrC_x content. The process consists of preparing a zirconium-loaded, iminodiacetate-functionalized styrene-divinylbenzene (STY-DVB) copolymer, and its subsequent carbonization in a tube furnace and/or a thermal plasma reactor. Depending on the zirconium salt used (zirconyl chloride, zirconyl nitrate or zirconium (IV) sulfate) in the Zr loading, the Zr-loaded resins resulted in ZrO₂@C pre-pyrolizates with C to Zr molar ratios of 5.8, 6.8 and 6.60. This carbon surplus is sufficient for the partial or even complete reduction of ZrO₂ into ZrC_0.58 at 1400 °C. The reaction products also contain 5 to 55 mass% residual free carbon. The plasma processing of the ZrO₂@C composite formed at 1000 °C in a tube furnace led to ZrC_0.94@C composites. The transformation of amorphous carbon content during the plasma treatment strongly depended on the atmosphere (He or H₂) in the reactor and the anion type of the Zr salt. In the presence of He, amorphous carbon could be completely transformed into graphite. In the presence of H₂, amorphous carbon and graphite were found at roughly the same ratio. No ZrO₂ could be detected in the plasma-treated samples, whilst different ZrO₂ polymorphs were found in the samples prepared in the tube furnace, depending on the synthesis conditions.

1. Introduction

Composites containing zirconium carbide, amorphous carbon or graphite and some or no ZrO₂ are important strategic materials for the nuclear industry due to their advantageous mechanical properties, neutron absorption cross-section and radiolytic stability [1,2,3,4]. A recently developed method of synthesizing (nano-ZrC,nano-ZrO₂)@C composites is the high-temperature or plasma-assisted carbonization of Zr-loaded functionalized polymers. This can be achieved with polyacrylonitrile-ZrCl₄ composites [3] and quaternary ammonium salt or styrene-divinylbenzene copolymers containing a sulfonate group and loaded with anionic [4] and cationic zirconium species [5], respectively, as precursors. The homogeneous distribution of the functional groups allows for zirconium to be bound at an atomic level, enhancing the distribution of ZrC in the carbon matrix after carbonization.

The chemical nature of functional groups (sulfonate or iminodiacetate) and DVB content in the STY-DVB copolymer have a significant influence on the chemical nature of metal compounds embedded in the carbon matrix during the carbonization of metal-loaded, functionalized, styrene-divinylbenzene (STY-DVB) copolymers, [6,7,8]. In fact, ZrC could be prepared by the carbonization of the quaternary ammonium group-containing and anionic Zr-loaded STY-DVB resins at 1350 °C [3], whereas only ZrO₂@C composites were formed from the sulfonated resins [5]. Due to the moderate reactivity of ZrO₂ in these composites, their subsequent RF plasma treatment in either a He or H₂ atmosphere [5] only resulted in a partial conversion of ZrO₂ into ZrC.

Our goal was to study the effect of functional groups on the formation and composition of zirconium carbide using STY-DVB copolymer precursors functionalized with groups that differed from sulfonate or quaternary ammonium. Since no information is available on the carbonization reactions of Zr-loaded iminodiacetate-functionalized STY-DVB resins, we studied the formation of ZrC in the carbonization process of these resins in both a tubular furnace and/or a thermal plasma-assisted process.

It is known that STY-DVB resins functionalized with iminodiacetate groups and loaded with a polyvalent element such as iron(III) have special properties due to the divalent selective iminodiacetate groups and the presence of a third, outer counter-ion neutralizing the third valence of iron(III) (X=Cl, nitrate or sulfate) [7,9,10]. Therefore, we conducted a detailed study of the influence of counter-ions such as chloride, nitrate, or sulfate on the chemical form of Zr in Zr-loaded iminodiacetate-functionalized STY/DVB copolymers.

Hereinafter, the prepared materials are referred to as L-X, L-Zr-X, L-Zr-X-T-t, where L stands for the Lewatite resin skeleton, X stands for the anion in the mineral acid or Zr salt used in H or Zr loading, respectively. T corresponds to the pyrolysis temperatures (1000, 1200, 1400 °C), whereas t represents reaction time (2 or 8 h). The L-Zr-X-1000-2 samples subjected to subsequent RF plasma processing in a H₂ and He atmosphere were labeled as L-Zr-X-1000-2-P-H₂/He, as summarized in

Table 1. Labels for samples.

Label	Description of Sample
L-X	H-form of iminodiacetate-functionalized STY-DVB copolymers containing 2 mass% DVB.
L-X-Zr	Zirconium-loaded iminodiacetate-functionalized STY-DVB copolymer samples containing 2 mass% DVB with ZrOX2 (X=Cl NO3) and Zr(SO4)2 (X=S) salts.
L-X-Zr-T-t	Samples obtained by carbonizing L-X-Zr resins at the given T temperature and for a carbonization time of t.
L-Zr-X-1000-2-P-H2/He	Plasma-treated samples from L-Zr-X-1000-2 samples containing ZrO2@C in a H2 or He atmosphere.

. Sulfate and nitrate ions were given the abbreviations S and N, respectively, in the L-Zr-X, L-Zr-X-T-t and L-Zr-X-100-2-P-H₂/He formulas.

In this paper, we present the preparation conditions and properties of nano-(ZrC_x,ZrO₂)@C composites made in a tube furnace (1000–1400 °C for 2–8 h) from Zr-loaded iminodiacetate-functionalized STY-DVB resins, and nano-ZrC_x@C samples from L-Zr-X-1000-2 samples containing ZrO₂. The method we used was RF thermal plasma treatment. We also conducted a detailed study of the chemical nature of Zr in the L-Zr-X samples, the number of iminodiacetate groups in the STY-DVB polymer chain, and the influence of the degree of Zr saturation and the anion of the Zr salt used during Zr loading on the properties of the forming nano-(ZrO₂, ZrC)@C.

2. Results and Discussion

2.1. Preparation and Properties of Zr-Loaded Ion-Exchangers (L-Zr-X, X=Chloride (Cl), Nitrate (N) and Sulfate (S))

The iminodiacetate-functionalized styrene-divinylbenzene copolymer was prepared via chloromethylation of STY-DVB copolymer containing 2% DVB exclusively at the meta-position of STY rings [11] with subsequent coupling with iminodiacetic acid diethyl ester and by alkaline hydrolysis (ESI Scheme S1). The sodium salt of the resin containing the iminodiacetate group was activated with 10% HCl, HNO₃ and H₂SO₄ (Scheme 1a), resulting in L-X resins as starting materials.

The Zr-loading of L-X resins was performed with zirconium salts containing X anions, namely 0.5% zirconyl chloride (ZrOCl₂.8H₂O in a 3 M HCl solution), 0.5% zirconyl nitrate (in 3 M HNO₃) and 0.5% zirconium (IV) sulfate (in 1.5 M H₂SO₄) (Scheme 1a). Due to their acidic pH, the formal chemical species of zirconium (IV) was the “zirconyl” ion, ZrO²⁺ [12,13,14]. Therefore, the CHNS elemental analysis data of L-Zr-X resins, the zirconium content of the solution eluted during Zr loading, and the amounts of L-X resins/salts containing zirconium used provide sufficient information for the calculation of the Zr/iminodiacetate group ratio (loading rate), water content and the Zr/C ratios according to the method given by Dunlop CJ et al. [15] (Supplementary Materials, Table S1).

The ratio of N content (x mole of iminodiacetate is located on the DVB and STY rings) and carbon content (x mole of methylene-iminodiacetic acid groups and the polymer chain consisting of DVB and STY) suggests that almost exactly one -CH₂N(CH₂COO)₂^2- group is located on every aromatic (STY or DVB) ring (Table S1). This means that during chloromethylation of the base polymer (STY-DVB) [16] (and consecutive coupling with iminodiacetic acid ester), only one chloromethyl group (iminodiacetate group) was inserted into every aromatic ring and, according to the synthesis conditions, it was built exclusively into the meta-position (see the IR results below and Supplementary Materials Figure S1) (Supplementary Materials, Scheme S1).

Knowing the ratio of STY/DVB/-CH₂N(CH₂COO)₂^2- groups, we calculated the number of hydrogens belonging to these building elements. The ratio of Zr loading and the amount of iminodiacetate groups in L-Zr-X samples was easy to calculate from the molar ratio of nitrogen and zirconium contents. Zr loading was 34.8%, 27.0% and 22.8% for the samples produced from zirconyl chloride, zirconyl nitrate and zirconium (IV) sulfate, respectively. The difference between the calculated and measured hydrogen content may be attributed to the presence of swelling water, whose amounts were around ~20% (Supplementary Materials Table S1).

2.2. The Chemical Nature of Zirconium in the L-Zr-X Samples

The chelate-forming ability of the iminodiacetate group resulting in the coordination of a nitrogen lonely electron pair to the zirconium d orbitals stabilizes the chelate structure. In this way, the protonation possibility of a tertiary nitrogen atom with HX acids in the chelate ring is lost. Therefore, the presence of foreign anions (chloride, nitrate and sulfate) due to the protonation of tertiary N atoms cannot be expected. The iminodiacetate group is a divalent-ion selective group; thus, one iminodiacetate group neutralizes only two charges of a multivalent cation, and one or more foreign ions neutralize the surplus charges, as was observed in the case of L-Fe-X complexes containing Fe(III) ions [6,7,8,9]. In principle, the counter-ions in the salts used for Zr loading (chloride, nitrate or sulfate) may neutralize foreign anions. However, the elemental analysis data and IR spectroscopy results on L-Zr-X compounds show the absence of any foreign anion; thus, the tertiary amino group of the -CH₂N(CH₂COOH)₂ part must be self-protonated with one of its carboxylic acid group (-CH₂NH(CH₂COO)(CH₂COOH) (Scheme 1b)).

The almost identical N content in the three L-Zr-X samples or the lack of sulfur indicates the absence of nitrate or sulfate anions (Supplementary Materials, Table S1). The matching IR patterns for L-Zr-X samples (X=Cl, N and S, Supplementary Materials, Figure S2) confirm that the chemical nature of these samples is practically the same. Most of the sulfate and nitrate IR bands would coincide with the IR bands of the iminodiacetate group and STY-DVB skeleton. A comparison of the IR spectra of the HCl, HNO₃ and H₂SO₄ salts of the iminodiacetic acid (Supplementary Materials, Figure S3) and L-Zr-Cl (Supplementary Materials, Figure S2) resins showed that the ν_as(S-O) and ν_d(N-O) modes of sulfate and nitrate ions could be identified in the IR spectra of [HN(CH₂COOH)₂].HX (X=NO₃ and ½SO₄) salts as a doublet at 1143/1042 cm⁻¹ and a singlet at 1057 cm⁻¹, respectively. These bands should appear in these spectral ranges to show the presence of sulfate and nitrate ions in the IR spectra of L-X and L-Zr-X resins if the tertiary N atom of the iminodiacetate fragment is protonated with the HX (X=N or S) acids.

However, a comparison of the L-X and L-Zr-X compounds (Supplementary Materials, Figures S4–S6) showed that neither the L-X compounds nor the L-Zr-X compounds contain foreign anions (Cl^-, NO₃^- or SO₄^2-). The presence of the Zr=O group or its hydrated form (Zr(OH)₂) with oxide- or hydroxide-neutralizing anions cannot be confirmed because of the absence of a strong and wide band at 1020–870 cm⁻¹ characterizing the zirconyl ion [14,15,16,17]. Yuchi et al., assumed the presence of Zr species of the type L-Zr(OH)₂ in a Zr-saturated IDA resin, but we did not observe any strong and sharp bands of ionic OH [17,18] above 3500 cm⁻¹ in the IR spectra of our L-Zr-X resins (Figure 1). Therefore, a similar chemical environment of the N atom is expected in all six L-X and L-Zr-X (X=Cl, N and S) compounds.

The IR spectra of L-X and L-Zr-X resins showed the typical C-H (aromatic and aliphatic) and C-C (aromatic ring) and other vibrational modes of the polystyrene-divinylbenzene copolymers [5,19,20], and the typical normal modes of iminodiacetate groups as well [21]. The evaluation of the spectra imposes a challenge due to the substantial overlaps of the polystyrene network and iminodiacetate group modes. The most critical peaks agree with the reference values [5,14,22,23,24,25,26,27,28] given for polystyrene-DVB samples. There is no peak at 1680 cm⁻¹ associated with the free vinyl groups of the crosslinker. Thus, there are no pendant vinyl groups in the samples. The lack of a peak at 833 cm⁻¹ and the steric hindrance of the ortho positions due to the steric demand of iminodiacetate groups confirm the meta-substitutions on the STY rings [24].

Even when we used a significant excess of zirconium salt, we could not saturate the iminodiacetate groups—the ratio of uncomplexed/chelated iminodiacetate groups in the L-Zr-X samples was ~2:1, ~3:1 and ~4:1 for X=Cl, N and S, respectively. Since the iminodiacetic acid groups forming in the acidic environment during Zr loading rearrange into an N-protonated form with the formation of a carboxylate group, a “free” iminodiacetic group can act as a monovalent acid. It can be considered as only a monovalent anion, [HN(CH₂COO)₂]^-. Thus, the counter-ions of zirconium (IV) used to neutralize the excess of two charges of zirconium (IV) may be two ionic unchelated iminodiacetate groups (Scheme 1b). Due to intramolecular self-protonation, a minimum of two free iminodiacetate groups need to neutralize one chelated iminodiacetate-coordinated zirconium. Thus, the Zr-loading of iminodiacetate groups in the L-Zr-X samples may not be higher than 1/3 due to the 2:1 ratio of monovalent anionic and chelated iminodiacetate groups, respectively, as was found in the case of the zirconyl chloride-loaded L-Zr-Cl sample (Scheme 1b).

To confirm the presence of the protonated/Zr-coordinated nitrogen atoms in the L-Zr-X samples, we used solid-state ¹⁵N and ¹³C CP MAS NMR spectroscopy. Deconvolution of the broad ¹⁵N CP-MAS NMR signal of the iminodiacetate nitrogen resulted in two components for both samples, L-N and L-Zr-N (315/330 and 317/332 ppm, Figure 2a,b, respectively). The signal at 315 ppm may be attributed to the protonated N-atom, whereas the signal at 317 ppm can be attributed to the protonated and Zr-coordinated N-atom. This latter signal is broadened by the quadrupolar magnetic nature of the Zr. A weak ¹⁵N signal of the unprotonated tertiary N atom (due to the equilibrium character of the protonation reaction) also appears in both spectra.

The chemical shifts in the nitrogen atoms in protonated and coordinated forms (Scheme 1b) are similar—the electron-shielding environment is similar. In other words, the strength of the coordination of Zr to the nitrogen atom in the iminodiacetate group of L-Zr-X samples is close to the strength of the proton binding to the protonated nitrogen atoms in -HN(CH₂COOH)(CH₂COO) or -HN(CH₂COO)₂^- groups (Scheme 1b).

When X= sulfate and nitrate, the chelated/unchelated iminodiacetate group ratios were 1:4 and 1:3, respectively. The free iminodiacetate groups are in their acidic form according to the low pH used during Zr loading and, due to self-protonation, these are in the [-CH₂-NH(CH₂COO)(CH₂COOH) (Scheme 1b) form [29]. The asymmetric nature of these iminodiacetate groups results in two C=O carbonyl signals [28] in the ¹³C NMR spectra of the L-X-Zr samples (Supplementary Materials, Figure S7).

Due to the lack of foreign neutralizing ions in the L-Zr-X samples, a question arises: why is Zr loading so different depending on the chemical form of the zirconium salt used in the loading experiments?

The zirconium(IV) ion easily hydrolyzes/polymerizes in aqueous solutions, even under highly acidic conditions [12]. The exclusive form of zirconium(IV) in aq. solutions above pH > 0 are zirconyl ions (ZrO²⁺), but the counter-ions have an enormous influence on the chemical nature of the species containing Zr. For example, solid zirconyl nitrate, ZrO(NO₃)2∙5H₂O, consists of H₂[ZrO(NO₃)₄].H₂O anionic species. It completely dissociates in its aqueous solutions with the formation of a neutral ZrO(OH)NO₃, which could not be loaded on a cationic (Amberlite IR 100) ion exchanger [13].

Zirconium sulfate strongly hydrolyzes in water. Basic zirconium sulfates are formed, including zirconyl sulfate, given the affinity of zirconium to sulfate, which is higher than the affinity to water. Thus, the sulfate ion is bound to zirconium strongly even in aq. solutions. For example, the electrolysis of aqueous Zr(SO₄)₂ solutions results in hydrogen evolution on the cathode, and H₂[ZrO(SO₄)₂∙3H₂O] was identified in the solution [30], which contains the zirconium in an anionic species.

Solid zirconyl chloride (ZrOCl₂.8H₂O) is a tetramer, [Zr₄(OH)₈(H₂O)₁₆]Cl₈ [12], and the zirconyl ion is the dominant form in its aqueous solutions. Zirconium (IV) ions would form only in >12 M HCl solutions and at a very low (0.0001 M) zirconium concentration [12]. We used a 0.055 M zirconyl chloride solution in 10% aq. HCl, thus, this solution contained only ZrO²⁺ and not Zr⁴⁺ ions.

The solution-phase zirconium species form complex equilibriums with the resin-phase iminodiacetate group. The difference between the chemical forms of zirconium in the solutions of the three zirconium salts may be attributed to the differences in Zr loading on the L-X resin under identical Zr loading conditions.

2.3. Thermal Studies on Compounds L-Zr-X

During the heat-induced carbonization of STY-DVB copolymers—with or without functionalization with functional groups—a part of the carbon content turns into volatile compounds such as hydrocarbons and the cracking products of the functional groups [6,9]. The iminodiacetate group is expected to be cracked into fragments of the CH₂N(CH₂COOH)₂ group, including H₂O (from dehydration) or CO₂ (from decarboxylation) and various fragments containing N [31]. The evolution of H₂O, CO₂, STY, DVB, some aromatic hydrocarbons, and the expected fragmentation products of the iminodiacetate group containing N is illustrated as a function of temperature (Figure 3 and Figure 4 and Supplementary Materials, Figures S8–S10).

The ratio of solid to volatile carbon compounds strongly depends on the decomposition temperature, heating rate, and DVB content and is influenced by the presence of metal ions and the extent of metal loading [6,9]. Generally, the higher the metal or DVB content, the more solid carbon products are formed [6,32,33,34]. Either the 2% DVB content or the incomplete saturation of functional groups by zirconium is a favorable condition for low solid carbon content in the decomposition products [5,6], increasing the Zr/C ratio in the carbonized material. Therefore, first, we studied the mass loss ratio of volatiles (H₂O, CO₂, and organic decomposition products)/solids (ZrC, ZrO₂ and carbon) at 800 and 1400 °C. The measured C/ZrO₂ ratios in the pre-pyrolyzed products of L-Zr-X samples formed at 800 °C (after eliminating all organic volatiles) were ~5.8, ~6.8 and 6.64 for X=Cl, N and S, respectively. Thus, the C/Zr ratio ensures that there is enough carbon content to form ZrC from ZrO₂/C intermediates. The free carbon contents of the decomposition products formed at 1400 °with a 10 °C/min heating rate were ~10, ~5 and ~45 mass% for the composites prepared from L-Zr-X resins (X=Cl, N and S), respectively.

The thermal decomposition of the L-Zr-X samples (X=Cl, N and S) started similarly, by losing the swelling water and some adsorbed organic materials from the pores of the hydrophilic STY-DVB polymer network below 200 °C (Figure 3). The TG-MS curves first showed the elimination of water. Loss of the iminodiacetate functional groups occurred with the evolution of CO₂, H₂O and CH₂COO fragments (Figure 3 and Figure 4, Supplementary Materials, Figures S8 and S9) at around ~220 and ~320 °C. Depolymerization first occurred around ~320 °C, with the formation of STY and DVB. Complete cracking of the organic residues occurred with the formation of aromatic compounds between 400 and 500 °C. The TG-MS curves of decarboxylation, dehydration, depolymerization and cracking are given for L-Zr-X samples in Figure 4 and Supplementary Materials, Figures S8 and S9. The TG-DSC curves of the L-Zr-X systems are attached as Supplementary Materials, Figure S10.

Water elimination, depolymerization and cracking are endothermic, whereas the decomposition of the iminodiacetate functional groups was exothermic for L-Zr-X (X=Cl, N) samples. For X=S, all when the lower amount of Zr is adsorbed on the resin, all the reactions are endothermic (Supplementary Materials, Figure S8a–c). For X=N, zirconium carbide formed as an endothermic process, starting at 1180 and completed at 1321 °C.

2.4. The Composition of Carbonization Products

Carbonization was first studied in a tubular furnace at 1000, 1200 and 1400 °C for 2 h. The effect of reaction time was examined at 1400 °C—holding time was increased up to 8 h. Orthorhombic and tetragonal ZrO₂ were detected as primary decomposition products at 1000 °C. The orthorhombic/tetragonal ratio strongly depended on Zr loading and was 1.36, 0.77 and 0 for the samples with X=Cl, N and S, respectively. Furthermore, the crystallite sizes of ZrO₂ particles ranged between 5 and 10 nm for these samples (

Table 2. Composition and properties of (ZrC_0.58, ZrO₂)@C composites prepared from L-Zr-Cl, L-Zr-N and L-Zr-S resins in a tubular furnace.

Composite	ZrC, (wt%) (Size in nm)	ZrO2 (wt%) (Size in nm)	Carbon (wt%)	BET Surface Area (m2/g)
L-Zr-Cl-1000-2	-	65 (5) Ortho/tetra = 1.36	35	140
L-Zr-Cl-1200-2	75 (10)	15 (11) tetra	10	311
L-Zr-Cl-1400-2	80 (10)	10 (12) tetra	10	310
L-Zr-Cl-1400-8	85 (14)	5 (26) tetra/mono = 1.92	10	245
L-Zr-N-1000-2	-	80(5) ortho/tetra = 0.77	20	278
L-Zr-N-1200-2	65 (10)	25 (7) tetra	10	447
L-Zr-N-1400-2	95 (17)	-	5	430
L-Zr-N-1400-8	95 (18)	-	5	438
L-Zr-S-1000-2	-	10 (8)	90	33
L-Zr-S-1200-2	50 (8)	10 (8) ortho/tetra = 0.26	40	41
L-Zr-S-1400-2	30(13)	15 (8) tetra	55	54
L-Zr-S-1400-8	50 (56)	5 (31) mono/cubic = 1.05	55	35

).

As carbonization temperature was increased to 1200 °C, a cubic ZrC_0.58 (a = 4.6690 Å, Fm3m) appeared as the main reaction product (75, 65 and 50 mass% ZrC and 15, 25 and 10 mass% ZrO₂ content for X=Cl, N and S, with a crystallite size of ~10 nm in every case (Table 2)). The exact composition of ZrC was calculated from the lattice constants determined by XRD (Figure 5 and Supplementary Materials, Figure S11), with the Gusev equation [35], which was valid for ZrC_x samples with 0.5 < x < 1.0. Senczyk’s equation [36] valid for samples 0.65 < x < 01.0 did not give results in this range, which shows that x is out of the validity range of the equation. Residual amorphous carbon content would be enough to form stoichiometric ZrC by incorporating carbon into the ZrC_0.58 structure at high-temperature heat treatment [37,38]. Thus, two sets of heat-treatment tests were carried out at 1400 °C with treatment times of 2 h and 8 h, and an RF plasma treatment was also performed in an H₂ and He atmosphere on nano-ZrO₂@C samples with Zr/C atomic ratios of ~5.8, ~6.8 and 6.64.

The heating time of 2 h at 1400 °C resulted in ZrC_0.58@C (X=N) and (ZrC_0.58,ZrO₂)@C composites with a ZrC_0.58/ZrO₂ molar ratio of 8 and 2 (Table 2). ZrO₂ was only detected as its tetragonal modification in these samples. Increasing the heating time to 8 h resulted in an increasing ZrC_0.58 content and a partial transformation of ZrO₂ into its monoclinic form (the tetragonal/monoclinic ratio was 2 and 5 for X=Cl and S, respectively). The crystallite sizes of ZrC_0.58 ranged between 10 and 20 nm. Neither graphite nor Zr₃C₂ was found in these samples. Even under prolonged heating, additional carbon atoms were not incorporated into the ZrC0.58 structure. The reactivity of ZrO₂ and carbon species in composites prepared from various Zr-loaded resins with various zirconium content was different in the carbonization reactions, despite the lack of foreign anions in the L-Zr-X samples. Therefore, we studied the behavior of all three L-Zr-X-1000-2 composites (containing ZrO₂) (X=Cl, N and S) in RF plasma under an inert (He) and a reducing (H₂) atmosphere.

Due to the high temperature that prevailed in the RF thermal plasma, the carbonization reaction between the ZrO₂ and C species took place rapidly and zirconium carbide crystallized with a composition of ZrC_0.94 (a = 4.6930, Fm3m) could be prepared. The ZrC stoichiometry in ZrC_x was determined from the lattice constants of the particular samples (Figure 5) with the Senczyk equation [36]. The composition of samples did not depend on the counter-ion of the loading Zr salt, but the amount of ZrC/amorphous carbon/graphite ratio strongly depended on this.

The main features of ZrC_0.94@C samples are the complete transformation of ZrO₂ into ZrC. The amorphous carbon is expected to transform into graphite during the plasma treatment [37]. The average size of ZrC crystallites ranged between 35 and 62 nm, which is higher than that of the samples made in a tube furnace. This may be attributed to the effect of high temperature, resulting in fast diffusion and growing crystallites.

In general, the low DVB content (2%) of L-Zr-X samples resulted in a relatively high ratio of organic volatiles during carbonization, resulting in high ZrO₂/C and high ZrC/C ratios in the precursor and product samples. The incorporation of iminodiacetate functional groups for Zr loading instead of sulfonyl groups [37] increased ZrC content in the ZrC@C composites. The composition of ZrC in the RF plasma-treated samples was almost the same, Zr_0.93 and Zr_0.94, for the samples prepared from sulfonated [37] and iminodiacetate-loaded resins, respectively. However, the ZrC_0.94 content and ZrC_0.94/ZrO₂ ratio were far higher in the case of iminodiacetate-loaded than the sulfonated resins [37].

2.5. Surface Characterization of the Carbon Composites Containing Nano-ZrO₂ and Nano-ZrC

We characterized the surface of the carbonized samples containing carbon balls decorated with nano-ZrO₂ polymorphs, nano-ZrC or their mixture by N₂ BET surface area analysis (Table 2 and

Table 3. Composition and properties of ZrC_0.94@C composites prepared in RF plasma.

Sample	ZrC, (wt%) (Size in nm)	Carbon (wt%)	BET Surface Area (m2/g)	Total Pore Volume (cm3/g)
L-Cl-1000-2-P-He	95 (49)	5	50	0.21
L-Cl-1000-2-P-H2	90 (40)	10	62	0.26
L-N-1000-2-P-He	95 (62)	5	52	0.22
L-N-1000-2-P-H2	90 (35)	15	82	0.22
L-S-1000-2-P-He	80 (40)	20	52	0.20
L-S-1000-2-P-H2	75 (42)	20	58	0.22

), X-ray photoelectron spectroscopy (XPS) and high-resolution transmission electron microscopy (HRTEM). The surface area of nano-ZrO₂@C composites prepared at 1000 °C for 2 h was 140, 278 and 33 m²/g for the samples with X=Cl, N and S, respectively. We found the same tendency for the samples containing ZrC and ZrO₂ + ZrC prepared in a tubular furnace as the ZrO₂@C samples. The highest values were found for X=N and the lowest values for X=S. Increasing the temperature from 1200 to 1400 °C had no critical effect on the surface area, whereas heating time (8 h instead of 2 h at 1400 °C) only had a considerable influence in the case of the X=Cl resin. The plasma treatment brought about significant sintering, resulting in a decreased surface area (Table 3).

The surface composition of some nano-ZrO₂@C, nano-(ZrC_0.58, ZrO₂)@C and nano-ZrC_0.94@C composites were determined by XPS, and the Zr_3d, C_1s and O_1s spectra are given in Figure 6, Figure 7, Figure 8 and Figure 9 and Supplementary Materials, Figures S12 and S13. The calculated surface concentrations are given in Supplementary Materials, Tables S2–S5.

The surface of the L-Zr-X-T-t samples with X=Cl and S contains approximately 85–95% graphite. Zr_0.58C (Zr_3d5 BE ≈ 179.9 ± 0.5 eV) was found only in samples heated above 1200 °C in low concentration, but its presence was undoubtedly detected based on the zirconium spectra (Figure 6). Increasing the temperature to 1400 °C or applying prolonged heating had no significant effect on the concentration of ZrC on the surface. The surface concentration of ZrC in the L-Zr-Cl samples is approximately double that of the L-Zr-S samples. Zirconium was present in the samples in two other forms: ZrO₂ (Zr_3d5 BE ≈ 183.8 ± 0.5 eV) and presumably some kind of zirconium oxycarbide (Zr_3d5 BE ≈ 182.5 ± 0.5 eV) (Tables S2 and S3, Figure 6, Figure 7, Figure 8 and Figure 9, Supplementary Materials, Figures S12 and S13). Oxidized carbon derivatives are also detectable on the surface in a few percentages. The line positions of the corresponding oxygen, O_1s 531.2 eV and carbon, C_1s 182.0 eV agreed with the literature data [39]. The appearance of the ZrO₂ minor phase can be attributed to the high affinity between Zr and O [37,38]. Some of the oxygen is chemically bound to carbon (O_1s 533.0 ± 0.2 eV, C_1s 286.7 ± 0.2 eV). In the carbon spectra, the quantity of the oxidized carbon species can only be determined with an error due to the presence of the high graphite peak. The graphite peak was fitted with an asymmetric peak shape with AS(40;0.6) parameters. π-π^* satellites were also present in the spectra at a binding energy of approximately 290 eV. Chlorine or sulfur could not be detected in the L-Zr-Cl-T-t and L-Zr-S-T-t samples.

The surface of RF-plasma-treated samples also contained 80–90 atomic% graphite and 1–8 atomic% ZrC_0.94. The quantity of ZrC0.94 is the highest in the L-Zr-N samples and the lowest in the L-Zr-S samples. In the case of H₂ plasma treatment, the surface ZrC_0.94 contents are always higher than in the case of He plasma treatment. The ZrC_0.94 content vs. the zirconium oxide content on the surface is the highest in the L-Zr-N samples. ZrO₂ and zirconium oxycarbide concentrations varied between 2 and 11 atomic% (Supplementary Materials, Table S5), and a low percentage of oxidized carbon derivatives were also detected.

2.6. TEM Analysis of the L-Zr-S-1400-8 and L-Zr-N-1000-P-H2 Samples

We studied HRTEM samples containing ZrC_0.58 and ZrC_0.94 prepared from the L-Zr-N and L-Zr-S copolymers in RF plasma and tubular furnace. The L-Zr-S-1400-8 sample shows a crystalline ZrO₂ monoclinic phase (ESI Figure S14) together with the ZrC cubic phase. (Figure 10).

The HAADF image and elemental maps show the distribution of Zr and O to be very similar. This may be partly attributed to the presence of ZrO₂ and partly to the ZrC_0.58 phase embedded in a carbon matrix containing a homogeneous oxidized surface. The zirconium-rich ZrC_0.58 phase surface may contain some amount of oxygen, as oxycarbide species were detected by XPS as well.

A cubic ZrC_0.94 (Fm3m) main phase was identified in L-Zr-N-1000-P-H₂ sample by the selected area electron diffraction (SAED) pattern (Figure 11b). Oxygen mainly appears in the smallest particles, which might be attributed to the oxidation of the carbon surface.

The plasma-treated L-Zr-N-1000-P-H₂ sample mainly contains ZrC_0.94 (Figure 5), which may be attributed to the reaction of zirconium dioxide particles passing through the plasma flame, evaporating, and reacting with carbon. The appropriate residence time for particles in the hottest zone of the plasma varied depending on the particle sizes and the actual thermal trajectory of the individual particles.

2.7. Raman Studies on ZrO₂@C, ZrC@C and (ZrC,ZrO₂)@C Composites

Powder XRD analysis is inadequate when exactly determining the ratio of amorphous carbon/graphite components in the nano-(ZrO_2, ZrC)@C composites [40]. However, the Raman spectroscopy is suitable for following the quantitative changes in amorphous carbon, graphite and distorted graphite structures and the graphitization processes [27,28,29,30,31,32]. Highly ordered monocrystalline graphite (ideal graphite lattice) has two first-order Raman bands, G (E_2g, degenerated optical mode at the center of the Brillouin zone) and G1 (2D) (harmonic in-plane transverse optical mode close to the zone boundary K point) at around 1580 cm⁻¹ and 2687 cm⁻¹, respectively [33,34]. The disordered graphitic lattice has a band D1 (D) (A_1g, phonon of the in-plane longitudinal optical branch close to the zone center) and D2 (D’) (E_2g, in-plane acoustic branch close to the K point), or graphene layer edges and surfaces, at around 1350 and 1620 cm⁻¹, respectively [37,40,41,42].

The amorphous carbon D3 band (D” or A) is at ca. 1500 cm⁻¹, whereas the D4 (D*) band at ca. 1200 cm⁻¹ contains mixed modes, including a mode of the disordered graphite lattice (A_1g), and polyenes or ionic impurities induced modes [27,30,40,41]. Moreover, the Raman spectra of carbonaceous materials contain bands belonging to weak combination (D + D”) and harmonic overtone (2D’) modes as well [42,43,44].

The Raman spectra of composites prepared from L-Zr-X samples between 1000 and 1400 °C for various times for X=Cl, N and S are given in Supplementary Materials, Table S6 and illustrated in Supplementary Materials, Figures S15–S17. The spectra and peak positions/intensities for plasma-treated samples are given in Figure 12 and Figure 13 and

Table 4. Raman shifts peak positions (in cm⁻¹) and their intensities (in arbitrary units) for the plasma-treated samples.

Sample		D *	D	D”	G	D’	2D	D + D’	ID/IG	I2D/IG
L-Zr-Cl-1000-2-P-H2	R shift [cm−1]	1100	1340	1572	1592	1614	2676	2929
	Intensity	0	2650	1552	784	585	1313	173	3.38	1.68
	FWHM	0	53.53	44.38	29.40	23.47	71.72	40.98
L-Zr-Cl-1000-2-P-He	R shift [cm−1]	1100	1339	1488	1574	1603	2675	2930
	Intensity	0	1647	259	1012	705	305	110	1.63	0.30
	FWHM	0	81.88	115.5	52.17	35.59	79.04	54.11
L-Zr-N-1000-2-P-H2	R shift [cm−1]	1100	1340	1480	1574	1606	2680	2933
	Intensity	0	2762	206	2907	802	2070	157	0.95	0.71
	FWHM	0	57.92	88.54	40.51	29.73	60.06	35.56
L-Zr-N-1000-2-P-He	R shift [cm−1]	1100	1339	1481	1574	1607	2679	2933
	Intensity	0	2581	475	2670	1004	1343	178	0.97	0.50
	FWHM	0	80.57	109.4	46.94	30.00	62.47	57.14
L-Zr-S-1000-2-P-H2	R shift [cm−1]	1100	1339	1517	1582	2078	2671	2925
	Intensity	0	1943	60.55	1349	0	387	127	1.44	0.29
	FWHM	0	65.45	0.113	64.27	0	79,9	59.33
L-Zr-S-1000-2-P-He	R shift [cm−1]	1100	1340	1494	1575	1604	2676	2934
	Intensity	0	2128	392	1511	1068	558	143	1.41	0.37
	FWHM	0	97.33	109.5	51.36	35.97	80.13	75.83

, respectively.

For the L-Zr-Cl-T-t samples, the intensity ratio of the peaks belonging to amorphous/graphitic components (D”/(D+G)) gradually decreases with an increasing reaction temperature, suggesting an advance in graphitization. The D and G bands become sharper due to the formation of more crystalline phases. The intensities of D* and D’ bands decrease with increasing temperature, showing the partial decomposition of organic polyene structures and surface graphene layers, respectively. Increasing the reaction time does not have a considerable effect. The I_2D/I_G ratio, related to the thickness or separation of the monolayer carbon sheets, slightly and irregularly changes with increasing reaction temperature and time, indicating that Zr atoms are not incorporated into the carbon network to a great extent [45]. The formation of graphite may also be attributed to the increased temperature and the catalytic effect of ZrC [46] forming above 1200 °C. The more ZrC is formed, the more graphite/disordered graphite was detected.

For the samples L-Zr-N-T-t, the catalytic effect of ZrC on graphitization might be more pronounced than the effect of temperature, whereas, for the L-Zr-S-t-t samples, the increase in temperature and reaction time resulted in regular improvements in the graphitization process (Supplementary Materials, Table S6). For the samples made from zirconyl nitrate, the ratio of amorphous carbon to graphite decreases with increasing temperature. The amorphous phase and polyenes completely disappear at 1400 °C in 2 h. Amorphous carbon remains in another experiment conducted at the same temperature but with a longer time (8 h). This shows that there are some differences between the two processes in terms of graphitization. This effect does not directly depend on heating time, but some other unknown factor(s), e.g., activation of some catalytic processes to increase nucleation or crystal growth, might play a role in its appearance.

The Raman spectra of all plasma-treated samples show the complete disappearance of the D* bands due to the thermal decomposition of polyene structures and the transformation of surface graphene layers. The amount of defective graphite and amorphous carbon is the highest in the X=Cl sample made in H₂, and when the atmosphere was changed to He, the ID/IG and ID”/(ID+G) ratios decreased from 0.45 to 0.097 and from 3.38 to 1.63, respectively. For the samples with X=N, the ID/IG and ID”/(ID+G) ratios are close to each other. For the samples with X=N, less amorphous carbon is found, and the ratio of defective/regular graphitic structures is almost the same (ID/IG = 0.95 and 0.97 in H₂ and He, respectively). A considerable influence of the atmosphere in ABAB stacking and the separation of the monolayer sheets (I_2D/I_G ratio) was only found for the X=Cl sample [47,48,49].

3. Materials and Methods

Chemical-grade resins and other chemicals (HCl, sulfuric acid, nitric acid, zirconyl nitrate, zirconyl chloride, zirconium(IV) sulfate, sodium hydroxide, calcium oxide and other analytical reagents), were supplied by Deuton-X Ltd., Hungary.

3.1. Experimental

3.1.1. Preparation of L-X-Zr Samples

The iminodiacetate functionalized STY-DVB copolymer containing 2% DVB (Lewatit TP-207, BASF, Lanxess, Germany) was soaked in a 10% aq. acid (HCl, H₂SO₄ and HNO₃) solution for 24 h and the beads that swelled were filled into a 20-cm-long glass column. Activation was conducted using a 10% aq. HCl, H₂SO₄ and HNO₃ solutions, respectively, with a flow rate of 5 mL min⁻¹, and Zr loading was performed using a zirconyl chloride solution containing 0.5 mass% Zr in 3 M HCl until the maximum possible loading with zirconium was reached [5]. Similarly, 0.5 mass% of zirconyl nitrate in 3 M HNO₃ and 0.5 M Zr(SO₄)₂ in 1.5 M H₂SO₄ were used under the same conditions as ZrOCl2.8H₂O.

The L-X-Zr samples that swelled during Zr loading were dried at 90 and 110 °C for 3 h and 2 h, respectively, in ambient air at atmospheric pressure with a heating rate of 10 °C/min. The beads were left to cool in a desiccator filled with freshly prepared CaO.

3.1.2. Carbonization Experiments

The Zr-loaded L-Zr-X resins were subsequently ground in a planetary ball mill (225 rpm, 30 min; Idar-Oberstein, Germany) until their size was reduced to ~63 µm. In each case, ca. 1.6 g of the sample was put into a quartz boat, and the carbonization process (pyrolysis) was performed in a tubular furnace under an Ar atmosphere (alumina tube) at temperatures of 1000, 1200 and 1400 °C for 2 h and 8 h. The heating rate (15 °C min⁻¹) and the volume flow rate of the gas (1.5 L⋅min⁻¹) were kept the same for all the experiments.

3.1.3. RF Plasma Processing

Due to the ~20% swelling water content in the resins (Table S1) and the short residence time of the particles within the RF plasma, the L-Zr-X samples could not be sufficiently directly carbonized. Therefore, we exposed the sample L-Zr-1000-2 prepared with the carbonization of the L-Zr-X samples at 1000 °C for 2 h and containing only ZrO₂ and elemental carbon (Table 2). These samples were exposed to in-flight thermal plasma treatment both in an inert (He) and a redutive (H₂) atmosphere. The experimental setup is described in [5,28,37]. This consisted of a reactor chamber supplied with a radio-frequency, inductively coupled plasma torch (PL-35, TEKNA Ltd., Sherbrooke, Canada, mounted on the top) and an RF generator (LEPEL, Waukesha, USA) with a maximum power of 30 kW at 4–5 MHz, together with a cyclone, a filter unit and a vacuum pump [5]. The solid samples were delivered into the plasma as described in [5] via a Praxair feeder (Danbury, Connecticut, USA) with He as carrier gas (flow rate of 5 L⋅min⁻¹). The feeding probe was coaxially in the middle of the induction coil. The feeding rate of the powdered resins was 3 g⋅min⁻¹. The plasma gas was a mixture of 11 L⋅min⁻¹ Ar and 6 L⋅min⁻¹ He, whereas the sheath gas was a mixture of 35 L⋅min⁻¹ Ar and 5 L⋅min⁻¹ He. The same experiments were performed with hydrogen in sheath gas instead of He with a flow rate of 5 L⋅min⁻¹. The most important operating conditions for the particular tests were adapted from [5] (70 kPa and 25 kW of power).

3.1.4. Elemental Analysis

The CHS analysis of the L-Zr-X samples were carried out after drying them in glass vials in a vacuum oven (Thermo Electron, Karlsruhe, Germany) at 140 °C for 19 h (tightly capped under N₂ and sealed with a plastic film) was performed with a Carlo Elba 1106 instrument.

3.1.5. Specific Surface Area Measurements

We calculated the surface area using the Brunauer–Emmett–Teller (BET) equation from nitrogen adsorption data collected at −196 °C (Autosorb 1C, Quantachrome, Boynton Beach, Florida, USA). The samples were evacuated for 24 h at 100 °C prior to measurements.

3.1.6. Transmission Electron Microscopy

Transmission Electron Microscopy (TEM) was carried out with a 200 keV (Talos Thermo Scientific, Santa Clara, CA, USA) transmission electron microscope. Grains were crushed in EtOH and deposited onto Cu grids covered by Lacey carbon. We obtained bright-field TEM (BFTEM), high-resolution TEM (HRTEM), and high-angle annular dark-field (HAADF) scanning TEM images as well as selected-area electron diffraction (SAED) patterns. In TEM mode, the current was ~600 pA and the SAED patterns were obtained with a camera length of 520 mm [5]. In addition, we performed elemental mapping in STEM mode to study the grains’ chemical composition and heterogeneity. In STEM mode, the current was ~200 pA. Beam convergence was 10.5 mrad. The camera length was 125 mm. Dwell time was 10 µs. We used a “Super-X” detector system with 4 silicon drift detectors built into the microscope column. Fast Fourier transforms (FFTs) were calculated from the HRTEM images with the Digital Micrograph 3.6.1 software (Gatan Inc., Pleasanton, CA, USA). The details of TEM spectroscopy are given in [32].

3.1.7. Thermal Studies

We performed simultaneous thermogravimetric analysis (TGA), differential scanning calorimetry (DSC) and mass spectrometric evolved gas analysis (MS-EGA) using a SETARAM Labsys Evo and a Pfeiffer Vacuum OmniStar instrument, under a high-purity He gas stream (6.0) with a heating rate of 10 °C/min from 25 °C to 1600 °C. The details of thermal measurements are given in [31,32]

3.1.8. Powder X-ray Diffraction

A Philips Bragg–Brentano parafocusing goniometer was used to acquire powder XRD patterns (Cu weighted Kα radiation, 1.5406/1.5444 Å) in the range of 4–70° 2θ (with a step size of 0.02° and an interval time of 1 s). Although the unit cell dimensions of the phases were well-determined, the composition (in mass%) of phases should be considered an estimate rather than an accurate value. Due to the amorphous carbon content, problems occurred when determining the background. Thus, no goodness-of-fit parameters are given. The details of the XRD measurements are given in [32].

3.1.9. Vibrational Spectroscopy

The analytical IR and far-IR range spectra were collected on a Bruker Alpha FT-IR (Bruker, Ettingen, Germany) and a Biorad-Digilab FTS-30-FIR far-IR instrument (Hercules, CA, USA) with a Smart Attenuated Total Reflectance (ATR) accessory, respectively. All spectra were recorded from 16 scans and acquired with a resolution of 4 cm⁻¹ [33,34,35].

A Horiba Jobin–Yvon LabRAM micro-spectrometer (Longjumeau, France) was used, with an external Nd-YAG laser source of 532 nm for Raman spectroscopy. We focused the laser beam on an objective (20X, numerical aperture = 0.4) using a D0.3 or D0.6 optical filter, which reduced the intensity of the laser to avoid degradation. A total of 1000 µm of the confocal hole was used, with a grating monochromator of 1800 mm⁻¹ for light dispersion. The wavenumbers were scanned with a resolution of 3 cm⁻¹_, with an accumulation time of 90–120 s per point for all the spectra in a range of 200–3400 cm⁻¹. After baseline correction, the first- and second-order regions of the spectra were deconvoluted into Lorentzian functions using the fitting tool available in the Origin software.

3.1.10. X-ray Photoelectron Spectroscopy

The X-ray photoelectron spectra (XPS) of the RF plasma-treated samples were recorded with a Kratos XSAM 800 instrument (Manchester, UK) in fixed analyzer transmission mode, with a pass energy of 40 eV, Mg K_α1,2 (1253.6 eV) excitation and an analysis chamber pressure of < 1·10⁻⁷ Pa. Survey spectra were recorded for both samples in the 100–1300 eV kinetic energy range. High-resolution spectra of the characteristic photoelectron lines of zirconium, oxygen, and carbon were recorded with 0.1 eV steps and a dwell time of 1 s. Quantitative analysis (after removal of the Shirley background) was performed with the XPS MultiQuant program (Version 7.8, M. Mohai, Budapest, Hungary) with experimentally determined photoionization cross-section data and asymmetry parameters [39].

3.1.11. Solid-State NMR Measurements

¹⁵N NMR magic angle spinning (MAS) spectra were recorded on a Varian System spectrometer operating at a ¹H frequency of 600 MHz with a Chemagnetics 6.0-mm narrow-bore triple resonance T3 probe in double resonance mode. ¹⁵N cross-polarization spectra were recorded with 5 ms contact time at 18 °C with a spinning rate of 1.6 kHz. ¹³C magic angle spinning (MAS) spectra were recorded on a Varian System spectrometer operating at a ¹H frequency of 400 MHz with a Chemagnetics 4.0 mm narrow-bore double resonance T3 probe. ¹³C cross-polarization spectra were recorded with 2 ms contact time at 18 °C with a spinning rate of 8.0 kHz. A repetition delay of 20 s and TPPM decoupling were used in all the measurements. For the ¹⁵N and ¹³C measurements, glycine and adamantane were used as a chemical shift reference.

4. Conclusions

(1)

Due to the low (2%) divinylbenzene content of the Zr-loaded iminodiacetate functional groups containing styrene-divinylbenzene (L-Zr-X) copolymers samples, a significant loss of the organic volatiles was observed during the carbonization reaction of the L-Zr-X samples, which resulted in high ZrO₂/C and ZrC/C ratios in the precursor and product samples, respectively.

(2)

Changing the anion in the zirconium salt used to load the chelate-forming resin resulted in ZrO₂@C precursors with varying reactivity levels towards carbonization and alterable Zr content.

(3)

ZrC_0.54@C samples can be prepared with or without ZrO₂ in a tubular furnace, with nanosize ZrC and ZrO₂ content.

(4)

The RF plasma treatment of ZrO₂@C samples led to ZrC_0.94@C samples with high (75–95%) nanosized ZrC_0.94 content and a pore volume of 0.20–0.26 cm³/g, with or without graphite content.