Introduction

Among various conducting polymers, polyaniline has been widely studied because of its facile synthesis and its good chemical stability both in air [1, 2] and acidic media over a large pH range. Polyaniline has been used in many applications such as dye-sensitized solar cells [3], hydrogen photoproduction [4], corrosion protection [5], gas sensors [6], glucose biosensor [7], the removal of pharmaceutical drugs [8] as well as the storage and conversion of energy [9].

The polymers family is versatile system with general formula: [(–B–NH–B–NH–)y(–B–N = Q = N–)1–y]n [10]; B and Q denote the rings C6H4 in the benzenoid and the quinoid forms, respectively. The fully reduced leucoemeraldine base polymer (LEB; y = 1), the fully oxidized pernigraniline base polymer (PB; y = 0), the halfoxidized form emeraldine base (EB; y = 0.5) and 75% intrinsically oxidized nigraniline (NA; y = 0.75) are all insulators. On the contrary, the protonation of the EB form produces the polaronic or bipolaronic emeraldine salt (ES) (Fig. 1), with a high conductivity (~ 2–10 S cm−1) [10,11,12].

Fig. 1
figure 1

The protonated emeraldine salt forms (PANI-ES)

Full size image

Many reports have been devoted to polyaniline as an adsorbent for removing heavy metals and chromium (VI) in particular [13,14,15,16,17,18,19]. Nevertheless, to our knowledge, no studies have been reported on the effect of polyaniline particle size on the removal efficiency of chromium (VI) before now.

Chromium belongs to the category of heavy metals that are potentially hazardous for the human health and environment even at very low concentration [20]. In waste aqueous systems, it exists mainly in two oxidation states, trivalent, Cr (III) and hexavalent, Cr(VI) states. The later variety is considered the most toxic by both the World Health Organization (WHO) and the Environmental Protection Agency (EPA) due to its carcinogenicity and mutagenicity on humans [21, 22]. The US Environmental Protection Agency (USEPA) established for the total chromium content (III and VI) that the maximum contaminant level (MCL) was 0.1 mg/L, due to the possible Cr(III)-Cr(VI) inter-conversions, depending on environmental conditions [23]; however, a level of 0.05 mg/L of chromium (VI) has been retained as a provisional guideline value by the WHO [24]. In this work, we investigate a relationship between the polyaniline particles size (PS) and the chromium (VI) sorption efficiency. This is done to enrich the literature on polyaniline where the adsorption properties have not been fully investigated.

Experimental

All products were supplied by the Alfa Aesar Company with purity greater than 99%, and solutions were prepared in deionized water. Polyaniline doped with HCl (PAni-HCl) has been prepared by a typical oxidation method using ammonium peroxodisulfate ((NH4)2S2O8) in aqueous acidic media (HCl) [24].

Initially, 0.055 mmol of aniline (~ 5 mL) was dissolved in 100 mL of HCl solution (1 M) in volumetric flask, then 150 mL of solution containing the oxidized agent (NH4)2S2O8 (0.063 mmol) was added dropwise to the solution containing the aniline monomer. The molar ratio [S2O82−]/[C6H5NH2] was taken to be 1.15. The mixture was left under stirring for 8 h at a temperature in the range (0–3°C), using an ice bath. The dark green precipitate was recovered by vacuum filtration and washed several times with both distilled water and acetone, and finally dried under reduced pressure in an oven for 72 h. The obtained powder was milled using an IKA laboratory mill (M20, Werke Staufen, Germany) for then sieved (30 min) using a FRITSCH vertical vibratory sieve shaker (model Analysette 3 PRO, Germany) into size ranges of < 50, 50–63, 125–160 and > 200 μm. Finally, the powders were placed in desiccators until use.

The X-ray diffraction (XRD) data were recorded with a PANalytical X’Pert PRO diffractometer (CuKα radiation,λ = 1.54056 Å) over 2θ range (3–60°) with an increment of 0.008° and a scanning rate of 0.1° min−1. The morphology and particle size were determined with a QUANTA 250 Scanning Electron Microscope (LV-SEM). The Fourier transform infrared (FTIR) analysis was performed using a Shimadzu sperctrophotometer equipped with a DTGM KBr detector. The particle size analysis was performed by dry mode (at 1 bar), according to Mie theory, using a CILAB 1190DL Laser Particle Size Analyzer.

The sorption experiments were performed in a double-walled Pyrex reactor (500 cm3 capacity). A 250 mg of PAni-HCl powder was added to 250 mL of Cr2O72− solution (250 mg L−1, pH ~ 4.5). The sorbent powder was dispersed by vigorous stirring (500 rpm), and the temperature was set at 25 °C thanks to a thermostat (LAUDA E10 S). Samples of 1 mL were drawn at 0, 0.5, 1, 2, 5, 10, 20, 40 and 60 min using a syringe with FTFE filter and then were diluted to 10 mL with deionized water in appropriate volumetric flasks.

The residual chromium concentration was evaluated in the λ-range (200–370 nm), using a UV-2401 Shimadzu spectrophotometer. The calibration curve (slope = 0.015 mg−1 L, R2 = 0.998) was performed using samples with Cr(VI) concentrations in the range (1−50 mg/L). The adsorption capacity (Qt, mg g−1) and the removal efficiency (τ, %) are calculated using Eqs. (1) and (2):

$$Q_{t} \left( {{\text{mg }}\,{\text{g}}^{ - 1} } \right) = \{ [{\text{Cr}}\left( {{\text{VI}}} \right)]_{0} - [{\text{Cr}}\left( {{\text{VI}}} \right)]_{t} /{\text{m }}\} \times {\text{V}}$$
(1)
$$\tau \left( \% \right) \, = \{ [{\text{Cr}}\left( {{\text{VI}}} \right)]_{0} - [{\text{Cr}}\left( {{\text{VI}}} \right)]_{e} /[{\text{Cr}}\left( {{\text{VI}}} \right)]_{0} \} \times 100$$
(2)

where [Cr(VI)]0, [Cr(VI)]t and [Cr(VI)]e are the initial, at time t and equilibrium concentrations (mg L−1) of the chromium (VI); V and m are the volume of the solution and the mass of polyaniline adsorbent (g).

Results and interpretation

Material characterization

Figure 2 shows a typical X-ray diffraction pattern of a semi-crystalline PAni-HCl structure. All peaks are assigned to the monoclinic unit cell (space group: P21), in agreement with the literature data [25, 26]. Furthermore, the broad peaks at ~ 20 and 25° are attributed, respectively, to the periodicity parallel and perpendicular to the ES chains [27]. The main vibration peaks identifying the chemical structure of polyaniline are shown in the FTIR-ATR spectra (Fig. 3). The peaks at 794 and 1101 cm−1 are attributed to the C–H out-of-plane deformation in the 1,4-disubstituted benzene ring, and aromatic C–H in-plane stretching vibration, respectively [28, 29]. The absorption peaks 1240 and 1292 cm−1 are ascribed to the C–N+ frequency mode of the polaronic lattice and to the C–N vibration in benzenoid ring, respectively [30]. Moreover, the peaks 1468 and 1651 cm−1 are assigned to the C=C stretching deformation in the sequencing of benzenoid and quinoid units, respectively [31]. Furthermore, the peak at 3440 cm−1 belongs to the stretching vibration of the CN groups.

Fig. 2
figure 2

XRD pattern of prepared material

Full size image
Fig. 3
figure 3

FTIR infrared spectra of the elaborated material

Full size image

The SEM image (Fig. 4) shows a non-uniform and porous microstructure based on highly agglomerate particles. This morphology offers more favorable adsorption sites compared to a compact surface and seems to have an appreciable potential for retaining Cr(VI) ions.

Fig. 4
figure 4

SEM micrograph of the recovered powder. Inset: The particle size distribution of the powder with small size range (< 50 µm)

Full size image

Particle size analysis

The particle size distributions (PSD) are obtained in terms of cumulative and frequency semi-logarithmic plots except for particles sizes smaller than 50 μm (Fig. 4, Inset), the distributions are multimodal (multiple peaks). All the most intense peaks are wide and non-symmetric (left and right-skewed) indicating polydispersed particles. For a non-symmetrical distribution and large particles, its more accurate to report the volume weighted mean diameter \({D}_{\text{4,3}}\), given by equation (3) [32, 33]:

$${D}_{\text{4,3}}=\left(\sum {n}_{i}{D}_{i}^{4}\right)/\left(\sum {n}_{i}{D}_{i}^{3}\right)$$
(3)

where ni is the number-based frequency of particles in the same class i, having the mean diameter Di. All \({D}_{\text{4,3}}\) values are presented in Table 1.

Table 1 Particle size specifications and removal properties
Full size table

Chromium (VI) speciation

According to the Cr(VI) predominance diagram (at 25 °C) [34, 35], the pH affects the significantly the chemical equilibrium involving both the H2CrO4−Cr2O72−−CrO4− or H2CrO4−HCrO4−−CrO4− systems, depending on the total chromium concentration (TCC); all species are linked by the following acid–base equilibriums [13, 34]:

$${\text{H}}_{2} {\text{CrO}}_{4} {\text{HCrO}}_{4}^{ - } + {\text{H}}^{ + } \quad (pKa = 0.8)$$
(4)
$${\text{HCrO}}_{4}^{ - } \rightleftarrows {\text{CrO}}_{4}^{2 - } + {\text{ H}}^{ + } \quad (pKa = \, 6.5)$$
(5)
$$2{\text{HCrO}}_{4}^{ - } \rightleftarrows {\text{Cr}}_{2} {\text{O}}_{7}^{2 - } + {\text{H}}_{2} {\text{O}}\quad \left( {pK = - 1.52} \right)$$
(6)

In the pH range (2–6), Cr(VI) exists as dichromate Cr2O72− or as bichromate HCrO4−, depending on TCC less or higher than 10−1.68 mol/L (pCr = 1.68). Nevertheless, at pH 4.5 and TCC of 2.157 × 10–3 M (pCr = 2.66), the Cr(VI) species are mainly present in the form of HCrO4−, and therefore, the H2CrO4−HCrO4−−CrO4− system is considered in our study with TCC given by (7):

$$[{\text{Cr}}\left( {{\text{VI}}} \right)]_{{{\text{TCC}}}} = [{\text{H}}_{2} {\text{CrO}}_{4} ] \, + [{\text{HCrO}}_{4}^{ - } ] \, + [{\text{CrO}}_{4}^{ - } ]$$
(7)

Spectrophotometry

The residual chromate Cr(VI) concentration versus time was evaluated by UV–visible spectrophotometry. Figure 5 gives the absorption spectra of residual Cr(VI) samples. The two intense peaks at 350 nm (28,571 cm−1) and 274 nm (36,496 cm−1) are attributed to 2t1u–eg and 2t1u–t2g transitions, respectively. It is well known that the inter-bands electronic charge transfer in both tetrahedral complexes Cr2O72− and CrO42− occurs between the higher occupied oxygen anti-bonding 2p*(O2−) molecular orbital and the unoccupied chromium non-bonding eg and t2g of 3d (Cr7+) molecular orbital. The 2p*-3d transitions are classified as ligand-to-metal charge transfer (LMCT) and are spin and Laporte allowed [36].

Fig. 5
figure 5

UV–Vis absorption spectra of residual Cr(VI) solutions versus the contact time with particles sizes of: a < 50 µm and b > 200 µm

Full size image

The spectrograms clearly exhibit the effect of the polyaniline particles size on the amount of Cr(VI) adsorbed. Indeed, the optical density abruptly decreases for grain sizes less than 50 μm, while it gradually decreases for particle sizes > 200 μm (Fig. 5, Inset).

The mechanism of the Cr(VI) removal was described by many authors [16, 17, 37] as chelate-type interactions between the imine and amine functional groups of the polyaniline particles and the negatively charged chromium species, which is pH-dependent. In parallel to the electrostatic interactions, a fraction of Cr(VI) is reduced to Cr(III).

The adsorption of the predominating form HCrO4− on protonated PAni-HCl particles occurs simultaneously with releasing Cl− ions, through an ion-exchange process. Therefore, we suggest mechanism according to Eqs. (8) and (9), in agreement with ref. [38, 39]:

$$\left( { - {\text{B}} - {\text{N}}^{ + } {\text{H}} - {\text{B}}^{ - } ,{\text{ Cl}}^{ - } } \right) + {\text{ HCrO}}_{4}^{ - } \leftrightarrow \left( { - {\text{B}} - {\text{N}}^{ + } {\text{H}} - {\text{B}} - ,{\text{ HCrO}}_{4}^{ - } } \right) + {\text{Cl}}^{ - }$$
(8)
$$\left( { - {\text{B}} - {\text{N}}^{ + } {\text{H}} = {\text{Q}} = ,{\text{ Cl}}^{ - } } \right) + {\text{ HCrO}}_{4}^{ - } \leftrightarrow \left( { - {\text{B}} - {\text{N}}^{ + } {\text{H}} = {\text{B}} = ,{\text{ HCrO}}_{4}^{ - } } \right) + {\text{Cl}}^{ - }$$
(9)

Effect of contact time and particle size

The evolutions of [Cr(VI)] versus contact time for different adsorbent particle sizes (< 50, 50–63, 125–160 and > 200 μm) are illustrated in Fig. 6. The batch adsorption experiments were conducted at 25 °C for 60 min, with an initial concentration of 250 mg L−1 and polyaniline adsorbent dose of 1 g/1000 mL. The figure reveals that the uptake is significantly influenced by both the particle size and contact time; the latter clearly enhanced the percent removal of Cr(VI).

Fig. 6
figure 6

Effect of the particle size and contact time on the Cr(VI) removal. Inset: Kinetic of pseudo-second order. Initial Cr(VI) concentration = 250 mg L−1; pH 4.5; T = 298 K

Full size image

In addition, under the same experimental conditions, powder with small particle size allows both rapid and significantly improved absorption. Indeed, for the first 30 s, more than 80% of Cr(VI) was removed using small sizes particles (< 50 μm), in contrast with those of > 200 μm where only 14% was eliminated.

For both particle sizes < 50 μm and > 200 μm, the equilibrium was reached after 40 min, and the calculated uptake efficiencies were found to be 99 and 77%, respectively. Such results are due to the enhanced external surface absorption, and hence to increase the number of adsorption sites. It is well established now that both physical and chemical phenomena are strongly influenced by the size and geometry of the materials [40, 41]; they become more pronounced as the particles size decreases, and the surface area-to-volume ratio increases [42].

Adsorption kinetic study

Throughout the studied particle sizes range, the linear plots t/Qt versus time (t) illustrated in Fig. 6, Inset, indicate that the Cr(VI) adsorption follows a pseudo-second-order (PSO) model, given by [43]:

$$t/Q_{t} = \left\{ {1/K_{2} Qe^{2} + t/Q_{e} } \right\}$$
(10)

where k2 (mg g−1 min−1) is the rate constant of pseudo-second-order adsorption, Qe (mg g−1) and Qt (mg g−1) the adsorbed amounts at equilibrium at time t. The values of Qe and k2 are evaluated, respectively, from the slope and the intercept of the straight lines.

According to Ho and Mckay, the adsorption kinetics obeying to the PSO model mainly occurs by means of chemisorption, through an electronic exchange between the adsorbate species Cr(VI) and the adsorption sites [44].

The kinetic parameters of PSO are listed in Table 1. The correlation coefficient (R2) was found to be close to unity, revealing the good correlation for the Cr(VI) HCrO4− adsorption of on polyaniline particles. The PSO model is verified for a linear regression coefficient R2 ≥ 0.9 [44]. In addition, for particles with small sizes, the experimental and theoretical values of Qe,exp and Qe,cal were found to be close to each other.

Figures 7 and 8 show, respectively, the variation of the adsorption capacity and the removal efficiency versus the inverse of the diameter ratio \({{D}_{\text{4,3}}}^{-1}\), normalized to \({D}_{\text{4,3}}\) of particles with size range > 200 μm. The experimental data are well fitted by using, respectively, the trend lines:

$$Q_{e,exp} \left( {mg \, g^{ - 1} } \right) = - 59.83 \, x \, \left( {D_{4,3} } \right)^{ - 1} + 252.3 \, \left( {R^{2} = 0.965} \right)$$
(11)
$$\tau \, \left( \% \right) = - 23.57 \times \left( {D_{4,3} } \right)^{ - 1} + \, 100.5\quad \left( {R^{2} = 0.996} \right)$$
(12)
Fig. 7
figure 7

Adsorption capacity versus the mean diameter D4,3

Full size image
Fig. 8
figure 8

The Cr(VI) removal efficiency versus the mean diameter D4,3. Insert: the rate constant versus the inverse of the mean diameter

Full size image

Figure 8, insert, reveals a clear linear relationship between the rate constant and the (normalized) diameter ratio. In fact, the obtained data are well-fitted using a linear model according to the trend line:

$$K_{2} \left( {{\text{mg}}\,{\text{ g}}^{ - 1} {\text{min}}^{ - 1} } \right) = 0.15 \, (D_{4,3} ) - 0.128 \left( {R^{2} = \, 0.981} \right)$$
(13)

The regression coefficients indicate a strong correlation between all the properties obtained and the average diameter ratio D4.3, over a grain size range (20–300 µm). Therefore, it can be seen that the smaller the particles of the adsorbent, the faster the adsorption process and the greater the amount of Cr(VI) removed.

The relationship of the proportionality was highlighted in the previous works [45,46,47,48]. In fact, Krishna [46] has studied the effect of the particles size (PS = 0.6, 0.8 and 1.7 mm) on the adsorption kinetics for removing Cr(VI) species media using calcined brick powder as adsorbent. He has attributed the increase in the adsorption capacity using smaller particles to both the greater accessibility to pores and a larger surface area for the bulk adsorption. Shanmugam et al. [48] have given the same explanation after studying the sorption of Cibacron blue F3GA using a spirulina platensis biomass as adsorbent, with an increases in particle sizes from < 75 to > 600 μm. They indicate that the particles with smaller sizes improve the availability of the surface area and consequently the number of binding sites.

The uptake capacities have been compared with those of other adsorbents for the removal of Cr(VI) (Table 2), and it is clearly seen that PAni-HCl exhibits an interesting potential as adsorbent for removing heavy metals, particularly when the uptake is undertaken using small size particles.

Table 2 Cr(VI) adsorption capacities of the previous works
Full size table

Conclusion

In summary, HCl-doped polyaniline was prepared by chemical route at low temperature (0–3° C). The XRD pattern reveals a typical semi-crystalline structure. The SEM characterization showed a non-uniform and porous microstructure based on highly agglomerate particles, offering more adsorbing sites and higher uptake performance. The recovered powder was crushed and sieved into size ranges < 50, 50–63, 125–160 and > 200 µm. The particle sizes analysis indicated a multimodal distributions, except for < 50 μm. The residual [Cr(VI)] versus time plots revealed that the small particle sizes give both fast and efficient metal removal, approaching ~ 100% efficiency. The adsorption follows a pseudo-second-order model. Inverse linear relationships were demonstrated between adsorption capacity and removal efficiency relative to the mean (normalized) diameter ratio, unlike the rate constant k2 for which the experimental data are well fitted using a linear model.