Assessing of channel structure and magnetic properties on heavy metal ions removal from water

The synthesis of Li1.1Co0.3Fe2.1O4 ferrite nanoparticles has been synthesized by the citrate auto-combustion method. The distribution of cations on A-site and B-site is studied by X-ray diffraction and High-Resolution Transmission Electron Microscopy. The Williamson–Hall analysis and Debye–Scherrer method are used to study the individual contributions of crystallite sizes on the peak broadening of Li1.1Co0.3Fe2.1O4. The estimated crystallite size of the investigated sample obtained from the Williamson–Hall plot and the Scherrer method agrees well with each other. HRTEM analysis confirms the homogeneous formation of the cubic phase. The calculated height and spacing parameters related to roughness are essential to achieve the efficiency of Li1.1Co0.3Fe2.1O4 to be used in microbatteries, smart windows, smart mirrors, gas sensors, and other applications. According to the obtained data, the Li1.1Co0.3Fe2.1O4 has a spiky surface with kurtosis of the line (Rku) equals 5.5. Additionally, the magnetic hysteresis loop has been clarified using the Vibrating Sample Magnetometer. The double peak characteristic in the Switching field distribution reveals the competition between exchange coupling and strong dipolar interactions. Li1.1Co0.3Fe2.1O4 has employed as a sorbent material for the removal of lead (II) ions from wastewater. The main advantages of the synthesized sample are ease of separation, high adsorption, low cost as well as recycled with notable efficiency. Two models of adsorption isotherms (Freundlich and Langmuir) are utilized to recognize the adsorption mechanism.


Introduction
A great properties of spinel ferrites are inexpensive, abundant in nature, environment-friendly and more stable. Spinel Lithium ferrite is a low-cost material that, have enormous applications in various devices of modern microwave technology [1,2]. They also can be used as cathode materials in lithium-ion batteries [3] and as sensing units in gas sensors [4]. It has a high-resistivity, low mobility semiconductor, and low eddy current losses. The properties of lithium ferrite can be enhanced by the substitution of divalent (cobalt) ions.
Removal of Heavy Metals from wastewater is a significant issue because they not only contaminate the water bodies but are also toxic to the ecosystem, as the majority of the heavy metals are non-degradability and highly toxic in nature. Consequently, their concentrations have to be decreased to appropriate levels before discharging into the environment, or else these can cause a threat to human as well as animal health [18,19].
In the present study, Li 1.1 Co 0.3 Fe 2.1 O 4 is synthesized using the citrate precursor method because of its cost-effectiveness, homogeneous formation of cubic ferrites, less time-consuming technique, and uniform particle size. Structural, thermal analysis morphological, and magnetic investigations were performed on the investigated sample. Further, the sample was employed as a sorbent for the removal of heavy metal (lead II) from water. The effect of pH on the adsorption process has been studied.

Experimental work 2.1 Preparation technique
All chemicals were purchased from Sigma-Aldrich with purity 99.9%. LiCoFeO was prepared by mixing non-stoichiometric proportions of iron, cobalt, lithium nitrates with a calculated amount of citric acid by citrate auto-combustion technique as reported previously [20] with some adjustments as shown in Fig. 1a

Adsorption test
The batch equilibrium technique was employed as mentioned in the previous work [21]. The Pb 2? -ion concentration was measured by atomic absorption spectrophotometry ICP (spectrometry Prodigy 7). The adsorption efficiency (g) and the equilibrium adsorption capacity (q) of metal ions are calculated according to the following equations [22][23][24]: where C i and C e are the initial and equilibrium (final) concentrations (mg L -1 ) of metal ion solution, respectively, while V and m are the volume of Pb (II) solution, and the mass of adsorbent, respectively. The nanoferrite sample is attracted to the magnet and accordingly can be easily separated as illustrated in Fig. 1b Figure 2a illustrates the XRD pattern of the Li 1.1 Co 0.3 Fe 2.1 O 4 sample revealing the formation of cubic spinel phase with space group (Fd-3 m). The XRD data are compared with JCPDS card 04-022-8066. There are no appreciable impurity peaks in the spectrum of the sample. The indexed peaks are (220), (311), (222), (400), (422), (511), (440), (620), and (533). The more intense peak is observed at 2h = 35.73°with hkl (311) which agrees well with the reported literature [25].
Certainly, the stoichiometric ratio for the tetra/octa is 0.50 causing a single spinel phase. In the present study, this ratio equals 0.66 yielding an excess amount of metals. In fact, the XRD pattern does not detect any secondary phases. This can be attributed to the large surface-to-volume ratio leading to the flexibility of the structure as well as the flexibility of the composition [26]. Li-ferrite nanoparticles show a partially inverted spinel structure [27]. The incorporation of the two cations at both cationic lattice sites of the spinel structure allows a broad compositional range of stability, even with rich compositions (cations in A/cations in B [ 0.5). Additionally, this nonstoichiometry can be structurally accommodated by the formation of point defects.
The average crystallite size of the investigated samples can be calculated using the following Scherrer equation [28]: where k is the wavelength of X-ray line (k = 0.154056 nm), b signifies the FWHM (full width of the diffraction peak) and h is the Bragg angle of the corresponding diffraction peak. However, we do not know the exact value of K for the present material system so a generic value of K = 0.9 is used, which means that all D values in this work are estimates only.
The estimated average crystallite size (D) of the prepared nanoferrite sample is 35 nm. On the other side, the crystallize size (D W-H ) can be also estimated from the Williamson-Hall equation in the manner prescribed below. Due to lattice imperfection and distortion, the induced strain can be calculated using the following Wilson equation [29]: The Williamson-Hall equation varies with tanh only, instead of 1=cos h as Debye-Scherrer equation [30,31]. The addition of Eq. 3 and Eq. 4 gives the observed broadening assuming the contribution of particle size and strain: Equation 6 shows the Williamson-Hall equation for estimating the average crystallite size and the average lattice strain (e). Figure 2b shows the fitting plots of b cosh versus 4 sinh. The slope gives 4e and the intercept Kk/D. The calculated values of e and D W-H are listed in Table 1. The strain is considered to be uniform in all the crystallographic directions. The W-H plot for the investigated sample shows the negative value of the strain indicates the compressive nature of it which leads to lattice shrinkage.
The dislocation density d represents the total defects in the sample. It is defined as the length of dislocation lines per unit volume of the crystal and it is calculated using the following equation [32]: Other important parameters that reveal the crystallization of the sample crystals are the distortion parameter (g), inter-chain distance (r), and theoretical density (D x ). These parameters can be calculated from XRD data using the following equations [33][34][35]: where Z (8) is the number of molecules per unit cell, M is the molecular weight, N is Avogadro's number and V is the unit cell volume. All the calculated parameters are tabulated in Table 1.
The prepared sample is of high quality with good crystallinity and fine crystallite size as confirmed from the calculated parameters in Table 1. The shape of Li 1.1 Co 0.3 Fe 2.1 O 4 nanoparticle is roughly spherical and highly agglomerated due to the magnetic nature of the sample. The SAED consists of concentric rings with spots over the rings. This characteristic specifies that the sample is good nanocrystalline in nature [36]. The rings with a dotty pattern in SAED ratify the varied size distribution of the nanoferrite sample. The inset in Fig. 3b shows the size distribution of the sample and the mean particle size is 70 nm.

Surface topography using AFM
Atomic force microscopy (AFM) is a crucial tool to scrutinize the surface topography of the material. Figure 4 shows the 2D-AFM image of the Li 1.1 Co 0.3 Fe 2.1 O 4 nanoferrite particle. The images show that Li 1.1 Co 0.3 Fe 2.1 O 4 exhibits waviness surface texture. The hill region is comparatively rough compared to the valley region. In other words, the typical Table 1 Experimental lattice parameter (a), Crystallite size (D W-H ), dislocation density (d), Distortion parameter (g), micro-strain (e), Inter chain distance (r), Unit cell volume (V) and X-ray density (D x ) roughness is the same for surfaces with roughness profiles completely different, because it takes into consideration the average profile of heights only. For this reason, additional sophisticated parameters can be clarified to fully describe a surface when more significant data are required. The profile peak height (R p ) is the measure of the highest peak around the surface profile from the baseline. Similarly, the valley depth (R v ) is the measure of the deepest valley across the surface profile analyzed from the baseline. Consequently, the maximum height of the profile (R T ) can be identified as the vertical distance between the minimum valley and the maximum peak as shown in Fig. 4c.
The roughness parameters are estimated by analyzing the topography scans of the sample's surface  Table 2. As shown from the table, the surface profile parameters include average roughness (R a ), root mean square roughness (R q ), maximum peak to valley height (R t ), skewness of the line (R sk ), and kurtosis of the line (R ku ). These parameters are important for obtaining specified data, like a scratch or an unusual crack on the material.
R sk is used to measure the symmetry and is more sensitive to occasional deep valleys or high peaks. Usually, R sk is used to distinguish two profiles of the same R a or R q values but of different shapes.
R ku is utilized to estimate the distribution of the spikes above and below the mean plane. For spiky surfaces, R ku [ 3; for bumpy surfaces, R ku \ 3. According to the value of R ku the Li 1.1 Co 0.3 Fe 2.1 O 4 has a spiky surface (R ku = 5.5).
Finally, the characterization of materials through their roughness allows one to attain data on the efficiency of samples in many application areas. The height and spacing parameters associated with roughness are essential to achieve the efficiency of Li 1.1 Co 0.3 Fe 2.1 O 4 to be used in microbatteries, smart mirrors, gas sensors, and other applications.

BET analysis
The adsorption of molecules on surfaces as a function of the relative pressures (P/P o ) of the adsorbate is elucidated by Brunauer, Emmett, and Teller theory. Figure 5a illustrates the adsorption/desorption isotherm of nitrogen at 77 K for Li 1.1 Co 0.3 Fe 2.1 O 4 ferrite nanoparticles. As shown from the plot the desorption branch does not follow the adsorption branch. It forms an obvious hysteresis loop. This loop is due to the amount of adsorbed along the desorption branch and the adsorption branch is quite different. The sorption isotherm is categorized as type IV and the hysteresis loop is of type H1. This agrees well with the IUPAC (International Union of Pure and Applied Chemistry) classification.
The distribution of pore radius versus pore volume of the studied sample is shown in Fig. 5b. Adsorption is dependent on the surface area, pore sizes as well as pore volume. The BET surface area of Li 1.1 Co 0.3 Fe 2.1-O 4 is about 127 m 2 g -1 . The pore volume and pore sizes are 0.436 cm 3 g -1 and 2.034 nm, respectively. According to the value of the pore size, the investigated sample can be classified as mesoporous materials [37]. In mesoporous materials, the sorption efficiency depends on two issues the first one is fluid wall interaction strength. The second is the attractive interactions between the fluid molecules. Consequently, multilayer adsorption and capillary condensation can occur in the pore.

Magnetic properties measurements
Measurement of magnetization at room temperature for Li 1.1 Co 0.3 Fe 2.1 O 4 nanocrystalline spinel powder is carried out using the VSM technique. Figure 6a nanoparticle also can be detected by an approximation of Stoner-Wohlfarth theory by extrapolating the plot of magnetization versus 1/H 2 to approach zero [38][39][40]. This theory is applied for non-interacting particles as confirmed from the squareness values in Table 3. In this way, the M s value is equal to 40.10 emu g -1 . The obtained value is very comparable to the experimental value signifying that an applied field of ± 20 kOe is appropriate to saturate the investigated samples.
The Li 1? ion is a nonmagnetic cation (0 lB) that does not respond to the magnetization of the sublattice. Consequently, it does not impart to the net magnetic moment. Li 1.1 Co 0.3 Fe 2.1 O 4 ferrite has a partially inverse structure so most of the lithium ions occupy octahedral sites (B sites) where as Fe 3? (5 lB) Table 2 Maximum profile valley depth (R v ), Maximum profile peak height (R p ), Maximum height of the profile (R T ), average roughness (R a ), root mean square roughness (R q ), skewness of the line (R sk ), and kurtosis of the line (R k ) The obtained value of saturation magnetization is lower than that found in previous studies of Li 0.5-Fe 2.5 O 4 nanoparticles [45]. The less magnetization of the samples is due to the fact that the magnetization of A-sublattice becomes much diluted and the A-B exchange interactions become weaker or comparable with the B-B exchange interactions. Consequently, the canted spins and the Yafet-Kittel (Y-K) angle are increased [46]. Additionally, it can be also owed to the formation of an inactive magnetic layer [47,48] and the disordered cation distribution [49].
From the hysteresis loops, the coercivity (H C ), M S , M r , squareness (M r /M s ), exchange bias [50], and magnetic moment (n B ) [51] are calculated and tabulated in Table 3.
Coercivity is often considered an extrinsic property of materials. It is sensitive to defects as grain boundaries, and dislocations. The value of the squareness ratio (M r /M s ) is an indication of the soft character of the nanoparticles and its value varies from 0 to 1 [52]. In the present case, the   [53]. The anisotropy constant is determined from the Stoner-Wohlfarth equation as follows: where H c is the coercivity, M s saturation magnetization and K is magnetic anisotropy constant. The weak anisotropy of the investigated samples is due to the non-collinearity (canting) of spins on their surface. A noticeable shift of the loop is detected and is ascribed to exchange bias phenomena. The exchange bias field are detected from the following equation and tabulated in Table 3.
This is referred to the existence of different spin configurations. In nanostructure sample, the magnetic structure of the core is different from that of the surface where spin frustration predominates [54,55].
The Switching Field Distribution (SFD) for the materials is one of the significant criteria for magnetic recording. It has a high impact on high-density recording performance [56]. Figure 7 shows the SFD value for the Li 1.1 Co 0.3-Fe 2.1 O 4 as a function of coercivity. Here, the SFD value is calculated from Eq. (14) as the ratio of the half-width of the peak in the derivative M-H and coercivity The derivative curve shows a double peak typical of a two-step reversal. Generally, the double peak characteristic in the derivative reveals the competition between exchange coupling and strong dipolar interactions.
Tuning the dipolar field energy to a value of the same order as the exchange coupling energy leads to a double peak derivative whose first peak usually occurs before reaching zero field coming from the saturation field [57][58][59]. Therefore, for the investigated sample, the dipolar interactions are more effective, which implies a strong reduction of the inter bump exchange coupling in our bump array.
From the above arguments, it is concluded that the small SFD value (0.54) of the investigated sample is due to an intrinsic magnetic nature such as uniformity of substitution and crystallinity as confirmed from XRD data.

Adsorption mechanism
Physicochemical interactions of the adsorbate-adsorbent play an important role in an efficient adsorption system. Therefore, it is essential to understand the adsorption mechanism. The adsorption of heavy metals onto spinel ferrite and their composites is due to (i) chemisorption, (ii) surface

Effect of pH on heavy metal removal
The removal of Pb 2? heavy metal cations from an aqueous solution is detected through batch experiments using nano-Li 1.1 Co 0.3 Fe 2.1 O 4 as an adsorbent. The interaction between adsorbent and adsorbate can be adjusted by experimental parameters as pH values [58]. The concentration of heavy metal nitrates is measured by the ICP technique. The removal efficiency of investigated sample is determined using Eqs. (1,2) as mentioned in the experimental section. Figure 8 illustrates the effect of pH on the adsorption efficiency of Pb 2? ions in the aqueous solution. The pH can influence the sorption capacity of Pb 2? ions by changing the surface charges of Li 1.1 Co 0.3 Fe 2.1 O 4 , the Pb 2? complex type, and the number of active sites available for the adsorption process [59].
The results indicate that Pb 2? removal increases to the maximum with increasing pH from 4 to 8 at room temperature (30°C) as shown in Table 4. In the present study, the maximum efficiency of Pb 2? removal is about 99.8% at pH 8. This is elucidated in view of a competition between H ? and Pb (II) ions to be adsorbed on Li 1 adsorption process [58]. However, at high pH values, the H ? ions competition nearly vanishes and the positively charged Pb 2? and Pb (OH) ? ions can be attached to the free binding sites, increasing the uptake of the Pb 2? ions [62].
The most recurrently utilized models to designate the adsorption results are the Freundlich and Langmuir isotherms. In the case study, these models are applied to scrutinize the adsorption of the heavy metals on Li 1.1 Co 0.3 Fe 2.1 O 4 at different Pb 2? -ion concentrations. This study is carried out for equilibrium metal ion concentrations varying from 11.94 to 0 mg/ L. The constants of the two models are calculated using the following relations and the results are presented in the inset of Fig. 9a, b.
The Langmuir equation is given by [21] C e q e ¼ 1 While the Freundlich logarithmic form of equation is specified as where Ce is the equilibrium metal concentration (mg L -1 ), q m and K L are the Langmuir constants accompanying with maximum adsorption capacity  (mg g -1 ) and the relative energy of adsorption (L mg -1 ), respectively, q e is the equilibrium sorption capacity (mg/g), and K f and n are Freundlich equilibrium constants. It indicates that the experimental data fitted well to both isotherm models. The Langmuir model, considering a monolayer adsorption process of molecules on solid surfaces, gives a good model for the adsorption system. According to this model, see Eq. (15) and Fig. 9a, the maximum adsorption capacity, q m , is estimated as 102 mg g -1 . The Langmuir constant, K L , equals 0.25 L mg -1 which designates high sorption energy between Pb II and Li 1 Furthermore, K f and 1/n are calculated from the intercept and slope of Fig. 9b. The obtained data are given in the inset of the figure. The small value of 1/n (less than 1) ratifies that the adsorption of lead Pb 2? on the Li 1.1 Co 0.3 Fe 2.1 O 4 is favorable.
The removal of heavy metals using magnetic materials has been studied by many authors as illustrated in nanosample was prepared in singlephase cubic spinel with space group (Fd-3 m). XRD assure that the sample was prepared in nanoscale with crystallite size 31 nm. The shape of nanoparticles is roughly spherical and highly agglomerated. M r /M s ratio assure the sample Li 1.1 Co 0.3 Fe 2.1 O 4 has magnetostatic interaction. The sample has weak ferrimagnetic behavior with Curie temperature near 715 K. The adsorption data are fitted to the Freundlich and Langmuir isotherm models. The best   [66] fitting is obtained by utilized the Freundlich model for Pb ?2 metal, i.e., Pb II with R 2 value 0.884.