Pseudo-superparamagnetic behaviour of barium hexaferrite particles

The effect of hexadecyltrimethylammonium bromide (CTAB) addition on the crystal structure, morphology, and magnetic properties of co-precipitated hexagonal barium ferrite was investigated. For a fixed amount of surfactant, different Fe3+ concentrations and Fe3+/Ba2+ ratios were used to optimize the formation of single-phase barium ferrite particles. The results indicated that the obtained ferrite particles exhibited coercivity changes similar to those of superparamagnetic particles with larger than theoretically calculated particle sizes. This results from the softening of the material due to the size reduction of the grains and incorporation of excess barium, localized on the surface of the particles. Therefore, lowering the energy barrier required to reverse the magnetization was observed, while high magnetization saturation was preserved. The precipitation of barium ferrite particles from a surfactant-rich solution allowed control of BaFe12O19 magnetic properties without introducing any modifications inside the crystal structure.


Introduction
The importance of magnetic compounds has signicantly grown in the past years as their new applications are being intensively developed in the elds of medicine, 1-3 separation technology, 4-7 preparation of smart materials 8,9 and electronics. 10,11 This results in an ongoing challenge to design new materials with the desired properties for a particular application or to develop new methods of their preparation. For the commonly studied ferrite ceramics, their magnetic properties result from interactions between metal ions occupying suitable positions relative to oxygen ions in their crystalline structure. [12][13][14] In this regard, ferrites with hexagonal symmetry are an important class of materials with unique magnetic properties, such as high values of coercivity, magnetization, exchange stiffness, and strong magnetic anisotropy. 15 Recently, their application in photocatalysis, [16][17][18] water treatment processes, 19 and hyperthermia application 20,21 was studied. The most commonly used methods of hexagonal ferrite particles preparation are ball milling, 22,23 thermal treatment, 24,25 hydrothermal treatment, 26,27 sol-gel autocombustion [28][29][30][31] and chemical co-precipitation method. [32][33][34][35][36] At the same time other works also suggest that simple preparation of hexagonal-based magnetic compounds is possible from iron-rich industrial wastes. 37 Compared to spinel ferrites, materials like BaFe 12 O 19 (BaM), being M-type hexagonal ferrite, are usually prone to higher magnetization and exhibit strong uniaxial magnetic anisotropy. The magnetic properties of BaM materials are associated with the changes in the microstructure and ions substitution. Despite the great potential of hexaferrites, methods of their synthesis and possible control over morphology and nal magnetic properties, which are crucial considering their application, are still far less investigated compared to spinel analogues (e.g., ZnFe 2 O 4 , Fe 3 O 4 ). Most of the recent work in this eld is focused on doping of the ferrite structure with transition metals and rare earth elements, [38][39][40][41][42][43][44][45][46] while little attention is given to morphological and size-dependent evolution of BaM's properties.
In this regard, the presented study focused on the preparation of BaFe 12 O 19 by co-precipitation of Fe 3+ and Ba 2+ ions in the presence of a cationic surfactant (CTAB). At present, CTAB addition was found to inuence slightly on the properties of precipitated hexaferrites, however, its' effect was studied only at a limited range of introduced substrates. 32, 33,47 Therefore, in the presented study correlation of BaFe 12 O 19 crystal structure with reagents concentration and reaction dynamics was investigated for the rst time. The structural, textural, and surface characteristics' were performed to understand the structural evolution of the barium ferrite particles. The physical properties measurement system (PPMS) at the temperature of 293 K and in the range of 0-3 T was used to investigate the magnetic properties change as a function of mean particle size and the presence of BaFe 2 O 4 and a-Fe 2 O 3 impurities in the structure of barium hexaferrite.

Preparation of barium hexaferrite particles
All the reagents were of analytical grade, purchased from Sigma-Aldrich (Poznan, Poland). In order to obtain BaFe 12 O 19 particles, corresponding amounts of Fe(NO 3 ) 3 $9H 2 O and Ba(NO 3 ) 2 were dissolved in distilled water, and 1000 mg dm À3 of hexadecyltrimethylammonium bromide (CTAB) was added. The aqueous solution of metal ions stabilized with the addition of a cationic surfactant was mixed during precipitation reaction. The precipitate was obtained by adding a fresh-made 5 M NaOH solution to pH value above 11 under room conditions. Obtained powders were centrifuged, washed with distilled water, dried at 80 C, ground and then calcined in two steps: with a heating rate of 3 C min À1 to the temperature of 180 C for 45 minutes, then with a heating rate of 10 C min À1 to 1000 C for 2 h.
In contrast to previously reported in the literature, 32,33,36,[48][49][50] NaOH was added rapidly to solution at a rate of 15 cm 3 s À1 . In order to study the effect of reaction dynamic, two control samples with different Fe 3+ /Ba 2+ ratios were prepared by adding NaOH dropwise at a rate of 0.03 cm 3 s À1 . Moreover, the inuence of the precipitating agent was also investigated. In this regard, NH 4 OH/(NH 4 ) 2 CO 3 mixture was used as precipitant, where OH À /CO 3 2À molar ratio was equalled to 2 : 1. All the obtained barium hexaferrite nanoparticles (BaM) are listed in Table 1 and labelled using two numbers, rst one indicating Fe 3+ /Ba 2+ ratio and a second one setting them following the rising concentration within the series. Samples prepared by dropwise precipitation or using carbonate/hydroxide mixture as a precipitating agent are additionally marked with "D" or "C" letters, respectively.

Physicochemical characterization
The crystal structure of the samples was determined by XRD analysis, performed using Rigaku Intelligent X-ray diffraction system SmartLab, equipped with a sealed tube X-ray generator. The scan rate was 1 $min À1 with a step of 0.01 and in the range of 2q from 15 to 75 . Qualitative analysis was performed using an external standard RIR method based on the ICDD database. Morphology and specic surface area of the obtained samples were analysed using Quanta 250 FEG scanning electron microscope and Brunauer-Emmett-Teller (BET) isotherm method, recorded by measuring nitrogen adsorption at the temperature of liquid nitrogen using Micrometrics Gemini V, apparat model 2365.
The surface composition was examined by X-ray photoelectron spectroscopy (XPS) with energy-dispersive X-ray spectroscopy. In this regard, the barium ferrite particles were placed on carbon tape in a copper holder and dried under vacuum. The XPS spectra were recorded on Escalab 250Xi (Thermosher Scientic using Mg K X-rays). Elemental composition was examined by wavelength dispersive X-ray uorescence spectroscopy, recorded using Tiger S8 spectrometer (Bruker).
Magnetic properties were determined by analysing hysteresis loops using Physical Properties Measurement System (PPMS) (Quantum Design, San Diego, CA, USA) at the temperature of 293 K and in the range of 0-3 T.

Crystal structure
The XRD results are presented in Fig. 1a To conrm the crucial effect of ions concentration, additional samples of barium hexaferrite (marked as "C" in Table 1) were precipitated with a mixture of OH À and CO 3 2À from solutions analogical to BaM_10_1 and BaM_12_4. Both of the samples were found to be pure or almost pure BaFe 12 O 19 . Moreover, as shown in Fig. 2b, sample BaM_12_4C was much more crystalline than BaM_12_4, due to the direct formation of BaCO 3 instead of Ba(OH) 2 , which could affect the distribution of Ba 2+ ions inside the precipitate. As seen in Fig. 2b, altering the rate of precipitation from 15 cm 3 s À1 to 0.033 cm 3 s À1 resulted in a lower content of BaFe 2 O 4 phase (see samples BaM_8_3 and BaM_8_3D with 38% and 33% of BaFe 2 O 4 by weight according to the results of RIR analysis) and formation of a-Fe 2 O 3 (see samples BaM_10_1 and BaM_10_1D).

Morphology and surface characterization
As presented in Table 1, the BET surface area varied from 0.3 m 2 g À1 to 8.6 m 2 g À1 for BaM_12_1 and BaM_10_1C, respectively. The observed development of the specic surface area for samples within series 10 and barium hexaferrite particles precipitated with the mixture of CO 3 2À /OH À may result from the stabilization of precipitated particles and inhibition of their secondary growth. The presence of carbonates in the precipitation environment affected the morphology of the nal particles. The introduction of carbonates increased the content of BaCO 3 /Fe 2 (CO 3 ) 3 in the precipitate, which may result in a relatively higher surface area. The main XPS results for the selected, pure BaFe 12 O 19 samples with different Fe 3+ /Ba 2+ ratios are illustrated in Fig. 3 and listed in Table 2. Aer subtracting baseline, in which C 1s peak at 285 eV was used for charge correction, the peaks at $795 eV and $779 eV are ascribed to Ba 3d 3/2 and Ba 3d 5/2 peaks, respectively. The XPS spectrum of Fe 2p can be resolved in two peaks, which are ascribed to Fe 2+ at $710 eV and Fe 3+ at $712 eV. For all samples, the percentage amount of Fe 2+ was higher than Fe 3+ ions at the BaFe 12 O 19 surface, which could result from oxygen vacancy inside the structure, which caused Fe 3+ reduction to compensate charge distribution. 52,53 The highest content of Fe 2+ compared to Fe 3+ was observed for the sample BaM_12_3, which can contribute to the lower M S value of this sample. However, for all samples, Fe 3+ content was quite low, comparing to other reported results (approx. 72% of Fe 2+ in this study vs. 27% in ref. 54), suggesting that surface composition plays a minor role in overall sample magnetization.
Moreover, the presence of Ba 2+ intermediates such as BaCO 3 , and Ba(OH) 2 at the surface of obtained BaM was analysed based on the energy difference between observed metal-oxygen signals. The DBa-O differences were higher than 248.9 eV for all samples, which is characteristic for BaM. For BaM_10_1 and BaM_12_2 their overall spectra were similar to those obtained by Atuchin et al. in their study on BaFe 12 O 19 electronic structure. 55 For sample BaM_12_3 no other barium compounds were observed on the surface of BaFe 12 O 19 . However, its surface was visibly enriched in iron, which was followed by a higher DFe-O energy difference. It suggests that the BaM_12_3 surface structure was more similar to iron oxide rather than barium ferrite. [58][59][60] The presence of amorphous, iron-rich structure could additionally explain the decrease of measured magnetization.
The morphology of BaFe 12 O 19 was studied by SEM, and exemplary images of the selected samples BaM_10_1,  BaM_12_2, and BaM_12_3 are presented in Fig. 4 (all were found to be single-phase BaFe 12 O 19 ). As expected, the changes in the reaction conditions resulted in clearly visible microstructure differences of the obtained materials. For sample BaM_10_1 formation of ultrane, irregular-shaped grains sintered into larger aggregates were observed. For both samples BaM_12_2 and BaM_12_3, the grains formed regular hexagonal platelets or bifrustums crystals. Among the samples, BaM_12_3 ferrite particles had a larger grain size, which could be a direct result of the increased substrates concentration and additional sintering required to form the single-phase BaFe 12 O 19 . For a detailed study of obtained particles, the distribution of BaM grains' width and height of at least 200 grains for each sample is presented in Fig. 7. An increase of the CTAB/Fe 3+ ratio resulted in both decreasing grains' size and narrowing the size distribution. Since the introduced CTAB at the preparation step should be removed at approx. 400-500 C during the sintering, 61 its presence is mostly assumed to affect precipitation process, as the BaFe 12 O 19 nucleation was found to start only in the range of 600 to 700 C (see ESI for XRD patterns †). Therefore, observed differences are expected to result from the evolution of precipitated, amorphous precursor. In general, surfactant presence inuence the formation of primary particles and their secondary growth, e.g., through their stabilization and inhibition of Ostwald ripening. 27,62 Since the same process can be discussed for amorphous precipitates, 63 a similar stabilization is expected to take place inside the investigated system. The effectiveness of such a process should depend on the CTAB amount that adsorbed on the particles surface. Therefore increase in the Fe 3+ and Ba 2+ concentration, followed by the formation of larger amounts of the precipitate, could result in not effective inhibition of the particle growth in the presence of CTAB. The BET analysis of precipitated precursor conrmed the changes for samples BaM_10_1 and BaM_12_4 with the specic surface area of 124 and 36 m 2 g À1 , respectively. Therefore, the change in CTAB/Fe 3+ ratio is responsible for the size evolution of the obtained precursor, and is further re-ected in the observed grains size of the nal ferrite particles. An increase of the precursors' size could also inuence on the longer calcination time of sample BaM_12_4, which could additionally intensify observed grains growth. Discussed mechanism is presented schematically in Fig. 5. A crucial role of surfactant on the grains size was also conrmed by additional SEM analysis of sample BaM_10_4, which was found to form polydisperse mixture of platelet particles, mostly in the range of 1 to 20 micrometres size (see ESI for exemplary images †).

Elemental analysis
The XRF results showed relative amounts of Fe and Ba for the selected samples, see in Fig. 6 55 It could be important for samples created in barium rich environment as suggest that overall Ba excess could tend to localize on the grains surface and possible their boundary inside the material.

Magnetic properties and microstructure
The PPMS analysis results are presented in Table 1, while obtained hysteresis loops for the most different samples in each series are shown in Fig. 8a- However, M S value did not decrease linearly with increasing BaFe 2 O 4 content, suggesting that there were synergic interactions between both barium phases. As presented in Fig. 9, the highest divergence between theoretical and experimental M S     O 19 formation in the samples, as mentioned above. Presented magnetic measurements results implied that approximately 20% of their content was not prone to magnetization, probably due to the creation of an amorphous phase, right next to the highly crystalline region. For other BaM samples, the paramagnetic phase was about 4-6%, which could be a net result of both amorphous content presence, spin canting on the surface of the material, and the iron deciency conrmed by XRF analyses. In most of the presented hysteresis, there were also visible steps, which indicated that obtained samples did not behave uniformly. 66 Because this phenomenon also applied to pure BaM samples, it could result from their polydispersity and interparticle interactions rather than differences in crystal structure itself. 68,69 A smooth hysteresis was obtained by lowering the precipitation rate, as shown for sample BaM_8_3D in Fig. 8d.
Finally, the most noticeable changes were in magnetic coercivity (H C ) and remanence (M R ) of the obtained samples, between which a direct connection was found, as shown in Fig. 10a. Remanence values were normalized with M S to exclude the effect of non-magnetic phases. In literature, BaFe 12 O 19 coercivity usually t between 200-500 kA m À1 , while it is seen that most of the presented samples had signicantly lower H C , ultimately reaching the difference between 6 and 215 kA m À1 (samples BaM_10_1 and BaM_12_3, respectively, both being single-phase BaM). The overall trend of H C growth together with applied Fe 3+ concentration and higher Fe 3+ /Ba 2+ ratio was quite well observed and is shown in Fig. 10b. The only sample that was not tting this relation was BaM_12_1, mostly due to approximately 30% of a-Fe 2 O 3 in its' structure, which, compared to BaFe 2 O 4 , is strictly paramagnetic. The obtained results are in agreement with the literature showing that the presence of the non-magnetic layer between misoriented nanograins/particles resulted in the enhancement of coercivity. 70,71 The analogical statement could be made for samples 12.3 and 12.4, considering their high amorphous/non-magnetic phase content. The simultaneous decrease of both M R and H C suggested that material could tend toward its superparamagnetic state. 72 It is well known that both coercivity and magnetic remanence depend on particle size, exhibiting maximum approximately at the point of single to multidomain transition. 73,74 It results from magnetic domain motion as well as magnetization vector's coherent rotation for particles larger and smaller than a critical value, respectively. For BaFe 12 O 19 , this point is not strictly dened. However, as previously reported, it is between 500-1000 nm. 15 Following this, most of the grains observed for BaM_10_1 and BaM_12_2 should behave as single-domain particles, while sample BaM_12_3 could be seen as multidomain. Therefore coercivity of BaM_12_3 should depend mostly on wall motion, that especially will become pinned at grain boundaries as reported by Dho et al. 25 On the other hand, for a single domain particles, the energy barrier preventing  This journal is © The Royal Society of Chemistry 2020 RSC Adv., 2020, 10, 18784-18796 | 18791 magnetization reversal is proportional to magnetic anisotropy, following the relation DE ¼ KV, where K is an anisotropy constant and V is particle's volume.
The superparamagnetic state is observed when this barrier can be overcome by thermal uctuations, leading to spontaneous switching of the magnetization vector. For the obtained samples, an effective anisotropy constant was found by analysing the high-eld region of hysteresis loops. In this region, magnetization changes reversibly due to the alignment of magnetic moments within the material with a magnetic eld vector. Therefore, it causes them to dri away from their randomly-distributed easy magnetization axes. According to the law of approach saturation to magnetization (LAMS) this region could be described by equation: 75,76 MðHÞ where M is magnetization, H is magnetic eld strength, c is high eld magnetic susceptibility and a and b are numerical parameters corresponding to materials defects and anisotropy, respectively. While the direct interpretation of a is not well known, for a compound with hexagonal symmetry b is found to be: from which, effective anisotropy constant (K) could be easily calculated. Approximated values of best-tting a and b parameters, together with obtained K values for selected samples are presented in Table 3.
For samples BaM_10_1 and BaM_12_2 calculated K values were very close to a known value of 3.3 Â 10 5 J m À3 , 15 and the parameter a was 0 or had an extremely low value. As expected for BaFe 12 O 19 , calculated anisotropy values were high, and sizes of grains observed by SEM analysis were signicantly larger than theoretically calculated for superparamagnetic particles. On the other hand, an alternative explanation of coercivity loss through possible wall motion, similar to sample BaM_12_3, could not be fully proven. In this case, grain growth should result in a smaller area of grain boundary and further H C decrease through inhibition of wall pinning, 25 which was not observed. It shows that other factors, outside of simple size reduction of BaM grains, are responsible for enhanced coercivity decrease. Observed nonstoichiometry of obtained samples could be an inuencing factor on the overall magnetic properties of BaFe 12 O 19 , and the hypothetical barium/iron excess inside the ferrite structure is rst to be considered. Prathap et al. have studied in detail the effect of iron deciency on analogical lead hexaferrite's properties. 64 Their results suggest that the formation of PbFe 12Àx -O 19Ày could indeed inuence coercivity loss. However, it is accompanied by a signicant loss in saturation magnetization value. Although a similar relation between H C and measured Fe/ Ba ratio could be observed in this study (see Fig. 11a), the M S value appears to be independent of the material's composition. On the other hand, lowering of Fe/Ba ratio could be understood as the incorporation of barium surplus to the BaM lattice. Zhao et al. studied the effect of Ba surplus on the properties of BaCoTiFe 10 O 19 . 78 Despite the natural differences in coercivity between BaM and its Co + Ti modied analog (being so ferromagnetic), their results indicate that changing the ratio of (Fe + Co + Ti)/Ba from 12 to 10 should result in a signicant M S decrease and a unit cell extension. As shown in Fig. 11b, both a and c lattice parameters of sample BaM_10_1 are very close to a known value for BaFe 12 O 19 , 15 indicated as empty points in Fig. 11b. However, possible barium incorporation could be observed for sample BaM_8_1 and is quite reasonable with increased barium content during the preparation. The possibility of Ba excess inside BaM structure for sample BaM_8_1 could also explain the decrease of its magnetization value, comparing to other samples. However, no visible increase in H C was observed, that could have been expected. 78 On the other hand, XPS analysis suggested that Ba tends to accumulate at the grains' surface. Composition change at the grain boundary can be an inuencing factor on the coercivity of sintered material due to affecting the intergrain coupling and the properties of the boundary phase. 79,80 Since no evidence of other crystalline-phases was observed based on the XRD patterns, the barium excess could be present as an amorphous, possibly thin layer on the grains' surface. Analogical layers were observed for modied Mn-Zn 81 and SrFe 12 O 19 (ref. 82) ferrites and were studied in detail for alloy magnets. 83,84 The effect of such boundary-phase could vary signicantly depending on its specic conditions and the morphology of the grains. Existing studies suggest that the coercivity of the hard magnetic phase could be signicantly decreased if the distribution of the magnetic moments became misaligned at the boundary. 69 By assuming that this misaligned region became enlarged by the existence of a non-stoichiometric surface phase, the overall material could be effectively soened. This reasoning could explain the observed drop of coercivity for materials obtained at barium rich conditions, as a mixed effect of size reduction and modication of boundary properties through the formation of the barium-rich layer. On the other hand, samples characterized by the iron excess are in agreement with the study by Zhao et al. focused on the formation of BaFe 12+x O 19+1.5x ferrites. 65 Their results suggest that incorporation of iron surplus inside BaM lattice both expands its unit cell and increase observed M S and H C values. In this study, both coercivity increase and unit cell enlargement were observed for samples BaM_12_2 and BaM_12_3, as shown in Fig. 11a and b. Moreover, the XPS analysis conrmed that no non-stoichiometric iron excess is not present at the surface. On the other hand, both samples possessed rather small values of M S , comparing to increased magnetization reported by Zhao and co-workers. However, it could be a result of structure deciency, which was already revealed for sample BaM_12_3.
In order to gain a better insight into magnetic interactions occurring inside synthesized materials, a differential dM/dH curves were analysed for selected single-phase BaM samples. Obtained results are presented in Fig. 12b showing changes between III to I quadrants of the hysteresis, with a range limited from À1 to 1 T (outside this, no other peaks were observed). It was found that samples obtained at different conditions were characterized by different dM/dH character. Sample BaM_12_3 exhibits typical behaviour for hard ferromagnetics, with a single peak dominating in a differential curve at H z 0.3 T and a smooth, broad M(H) hysteresis. On the other hand, sample BaM_12_2 possesses two visible peaks around H z 0 and H z 0.47 T, which is characteristic for a weakly coupled magnetic systems. 66,85 It was changed for samples BaM_8_1 and BaM_10_1 for which the H z 0.5 T peak disappears almost completely (especially for BaM_10_1). The systematic disappearance of the second peak on the dM/dH curve could indicate an enhancement of the coupling behaviour inside the material. For a BaFe 12 O 19 , coupling with a so magnetic phase could lead to a coercivity loss. 86,87 However, the nal properties should heavily depend on the fraction of both phases. As throughout the Fe 3+ /Ba 2+ ¼ 8 and 10 series, no strong dependence between magnetic properties and synthesis conditions was observed, and therefore between possible Ba content, it seems unlikely that similar coupling is mostly responsible for observed H C loss. Moreover, no enhancement of the remanence was observed, which should be characteristic for exchange-coupled BaM composites regardless of the simultaneous H C change. [86][87][88][89] Final remanence value was always proportional to the coercivity, as shown before, and the highest H C was observed for sample BaM_10_4, which was also visibly less coupled than BaM_10_1 (see in the ESI for the comparison of the dM/dH curves †). On the other hand, the same sample also possessed visibly larger grains than BaM_10_1, suggesting that observed particles size was still signicant for the nal properties. Alternatively, dipolar interactions could also be responsible for observed hysteresis behaviour. In this regard, simultaneous H C and M R decrease was characteristic for more interacting particles. 90,91 A rough estimation of the energy of such interactions can be made through relation: 92 where m 0 is magnetic constant, m 1 and m 2 are magnetic moments of interacting particles (grains) and r is distance between them. It can be further multiplied by the number of neighbouring particles (n) and for sample BaM_10_1 calculated energy indeed could be an important factor. Especially, considering relatively big particle interacting with a smaller one, anisotropy of the later one could be overcomed (for a 115 Â 50 nm ellipsoid interacting with a 245 Â 110 nm one, at the distance of 90 nm and with n ¼ 5, calculated E dipolar /E anizo ¼ 1.073). Importance of the dipolar interactions can somehow explain differences observed between samples BaM_10_1 and its resynthesized version. The new sample was characterized with similar XRD, morphology, size distribution, BET, M S , and K values as the original one and a slightly higher H C and M R , which could be connected to an increase in distance between the particles and the number of a possible neighbours. It was shown in Fig. 13, where obtained relationship between particles size and H C was presented (V was calculated as a mean, weighted with a size distribution for every sample (Fig. 7), treated as ellipsoids of revolution).
The overall results showed that the observed coercivity decrease results mostly from the size reduction of BaM grains and possible increase in their interactions. Especially, dipolar interactions could be an important factor, considering ultrane particles obtained in barium-rich conditions. It was followed by the change in elemental composition and particularly incorporation of barium surplus tends to localize on the boundary of the particles. This behaviour may further affect interactions between the grains and therefore resulting properties. For ultrane particles, this could lead to a more collective state of a material, where magnetization reversal is a continuous process of succeeding switches of neighbouring grains rather than anisotropy overcoming for isolated particles. Ultimately, by changing the size and composition of obtained materials, they became effectively soened, mimicking the superparamagnetic behaviour, despite being far from their calculated critical size and remaining ferrimagnetic (none of it reached H ¼ 0 and M ¼ 0 points during measurements).
Previously, in the literature, the observed grains had similar dimensions and signicantly larger H C values. 25,62 In this study, both phase and elemental composition of BaFe 12 O 19 was found to vary depending on the applied Fe 3+ /Ba 2+ ratio and the relative amount of surfactant. Therefore, both of these parameters are important for the preparation of a single-phase barium hexaferrite particles with different coercivity. Ultimately, signicant changes in morphology and magnetic properties of BaM could be observed with a high suitable CTAB/Fe 3+ ratio.

Conclusions
In this study, based on the structural, textural, and elemental characteristics' a possible magnetic microstructure of barium hexaferrite particles was discussed. Barium hexaferrite properties highly depended on the synthesis conditions, especially during precipitation with the addition of CTAB surfactant. For a xed amount of surfactant and Fe 3+ /Ba 2+ ratio, an increase in ions concentration resulted in the formation of bigger grains and higher content of the barium-rich phase in a nal product. Obtained series of pure BaM samples exhibited signicant changes in elemental composition, allowing for the formation of both iron and barium excessive single-phase materials. Meanwhile, the surplus iron can be incorporated into the BaFe 12 O 19 structure, while barium tends to accumulate on the grain surface despite the overall Fe/Ba ratio. It suggests that the grain boundary of as-synthesized ferrite could possess properties different than bulk BaM. For the samples obtained at a barium rich environment and with a high CTAB/Fe 3+ ratio, a signicant soening of the nal material was observed, connected to both decreased grain size and the possible effect of dipolar interactions, together with a formation of Ba-rich layer at the surface. The overall approach allowed us to synthesize a series of BaFe 12 O 19 materials with behaviour similar to superparamagnetic transition, despite being far from the theoretical point of superparamagnetic critical size. This study provides a new method for tailoring magnetic properties of barium hexaferrite, where high magnetization is preserved, and additional elements are not required to modify the crystal structure of the material.

Conflicts of interest
There are no conicts to declare.