Introduction

Synthetic dyes are widely used by different industries such as textile, rubber, leather tanning, paper, printing, cosmetics, food. Around 10,000 types of commercial dyes and more than 0.7 million tons of synthetic dyes are produced annually in worldwide (Mahmoodi et al. 2011). This wide use of dyes in daily life has produced enormous volumes of effluents containing dyestuff and poses considerable management problems in wastewater industry. Around 15% of synthetic dyes used are lost during production and improper processing (Mahmoodi and Arami 2008). Even at low concentrations, these pollutants reduce the sunlight penetration, affecting thereby the process of photosynthesis which destroys the food webs and aquatic life existing in water ecosystem. Human health is threatened not only by the toxicity, carcinogenic, and mutagenic effect of these contaminants, but also by the incorporation of some dyes in food products which may cause asthma, urticarial, eczema and allergic reactions in humans bodies (Wang et al. 2006).

The presence of dyes in environment has become one of the most difficult problems to solve in developing countries. Due to their complex structure and non-biodegradability behavior providing a high stability and resistant to light and oxidative agents complicates the choice of appropriate method for their treatment (Devi et al. 2015). Several wastewater technologies have been used to remove dyes including coagulation/flocculation, advanced oxidation processes, ozonation, membrane separation, electrochemical methods, adsorption on activated carbon and biological treatment (Akar et al. 2009; Belala et al. 2011). However, these technologies are expensive because of capital costs investments, extensive energy and chemicals use (Reddad et al. 2002). In contrast, the adsorption technology has been proven to be advantageous because of less investment cost, easy and simple design operations, no toxic by-products and possibilities of adsorbents regeneration and adsorbate recovery (Mittal et al. 2009).

The selection of adsorbent material is still the main concern in this technique. It is known that the activated carbon is widely used adsorbent due to its high surface capacity and low specificity in respect to properties of pollutants (Chen et al. 2013; Djilani et al. 2015). However, due to higher costs of its synthesis and regeneration activated carbon is difficult to afford by industry in the developing countries (Vimonses et al. 2009a). In recent years, the development of low-cost adsorbents from different locally available materials has attracted considerable attentions of scientific researchers. A wide variety of adsorbents material has been experimented and tested for dyestuff and heavy metals removing from wastewater such as agro waste products (Aziz et al. 2009; Bhatti and Nausheen 2014), sludge from wastewater treatment (Iddou and Ouali 2008), algae biomass (Iddou et al. 2011), biosorbents as: bacteria, yeasts and fungi (Djafer et al. 2014; Kelewou et al. 2014), clay and mineral materials (de Sales et al. 2013; Abidi et al. 2015).

Diatomite is natural aluminosilicate mineral material, which largely consists of fossilized skeletons of unicellular aquatic plants found in sedimentary rocks. It is readily available and low-cost materials of variety of shape and size generally ranging between 10 and 200 µm (Yuan et al. 2010). Many of physical and chemical characteristics, such as highly porous structure, low density, high surface area and low thermal conductivity, lead to use it as filter in water purification, as sound and thermal insulation in building construction, as catalyst support in semiconductors materials, etc. (Wang et al. 2015). Therefore, diatomite has been used successfully as adsorbent in wastewater treatment to remove pollutants such as heavy metals (Ye et al. 2015) and dyes (Sheshdeh et al. 2014).

It was reported that some of these adsorbents require additional pre-treatment in order to enhance their adsorption capacities and achieve the performance of activated carbons; therefore, several treatment methods have been investigated (Johari et al. 2016). The heat treatment is one of the best and simple methods of activation. However, calcination at high-temperature can affect the superficial functional groups and damages the pores structure, and loss of removal efficacy of chemicals by adsorbents after its thermal regeneration has been one of main limitation for wider use of different adsorbents including activated carbons (Wang and Zhu 2006).

In order to overcome this drawback, the objective of this present study was to find, whether thermal treatment of diatomite reduces the adsorption capacities of textile dye and enhances the selectivity and regenerability of the recycled materials in aqueous systems. To determine different parameters affecting the characteristics of regenerated diatomite using heat treatment, batch experiments were done in laboratory scale in normal conditions of temperature and pressure.

Materials and methods

Samples preparation

Diatomite used in this work was provided by the National Company for Non-ferrous Mining Products (ENOF), from Sig deposit, Mascara (West Algeria). After collection, the diatomite was dried at 105 °C until constant mass changes, and then gently crushed and sieved to 250 µm. Thermal treatment in three different temperatures was applied at 300, 600 and 900 °C for 2 h in a programmable muffle furnace, with a heating rate of 3 °C min−1. Obtained materials are named RD (raw diatomite), DH300, DH600 and DH900 and stored in a glass bottle for further use without any additional treatment.

Characterization of adsorbents

The properties of the raw and heated diatomite were determined using various techniques. The particle size distribution was determined from wet dispersion using ultrasound with laser scattering particle size analyzer (Analysette 22 NanoTec, Fritsch, Germany). For particle size calculations Mie theory was used. The procedure consists to add a dry sample into the analyzer cell (filled with distilled water) until the laser beam obstruction was 5–7%. The different crystalline phases were determined using X-ray powder diffractometer (XRD), PANalytical X’Pert PRO MPD diffractometer equipped with an X’Celerator linear detector, using Cu-Kα1 radiation (λ = 1.5406 Å), 40 kV voltages and 30 mA current. The diffraction patterns were recorded between 5 and 80° (2θ) with an angular step interval of 0.0334° and scanning speed of 1°/min. The surface morphology and chemical composition were obtained by field emission scanning electron microscopy (SEM, Tescan Mira/LMU) coupled with X-ray energy-dispersive spectrometry (EDX, Inca Energy 350). The specific surface area of each material was determined at liquid-nitrogen temperature (77 K) using the Brunauer, Emmett and Teller (BET) gas adsorption method using Quadrasorb SI-KR/MP (Quantachrome Instruments, Boynton Beach, FL). The samples were degassed at 100 °C for 24 h under vacuum prior to measurement. The total pore volumes were calculated based on N2 adsorption–desorption, while the pore-size distributions were calculated from the adsorption branch of the isotherm, according to the Barrett–Joyner–Halenda (BJH) model.

Batch adsorption experiments of Red ETL dye

An industrial textile dye appointed Red ETL (R-ETL) was used in this study as target pollution. The dye was provided by SOITEX wet-process textile industry of Tlemcen (West Algeria).

The removal study of R-ETL dye by RD, DH300, DH600 and DH900 consists to optimize on batch mode the physicochemical parameters affecting the adsorption process of R-ETL molecules onto the surface of adsorbents.

Batch experiments were carried out on a magnetic stirrer (Thermo Scientific Cimarec i Multipoint) at 400 rpm using 100 mL Erlenmeyer flask. A desired amount of adsorbent was added to 50 mL of R-ETL dye solution at known initial concentration; the mixture was stirred for desired contact time to ensure equilibrium. Then, the Erlenmeyer’s content was centrifuged at 6500 rpm during 10 min, and the remaining concentration of R-ETL dye in solution was measured by analyzing the supernatant using UV–visible spectrophotometer (HACH DR 4000) at 508 nm wavelength. The adsorbed quantity (Qe) of R-ETL dye adsorbed was calculated according to the following equation:

$$Q_{\text{e}} \left( {{\text{mg}}\;{\text{g}}^{ - 1} } \right) = \frac{{(C_{\text{i}} {-} C_{\text{e}} ) \cdot V}}{W}$$
(1)

where Ci and Ce are, respectively, the initial and equilibrium concentration of R-ETL in solution (mg L−1), V is the volume of solution (L), and W is the amount of adsorbents (g).

In order to determine the effective adsorption conditions, a set of experiments were performed with different pH values, contact time, adsorbents dose (solid/liquid ratio) and initial dye concentrations. The pH value of solution is an important parameter and especially in the adsorption studies of textile dyes, because it can affect not only the adsorption capacity, but also the behavior and solubility of dye molecules in water. The effect of this parameter was studied over the pH range from 2 to 10. The required pH value of solution was adjusted with diluted solutions of HNO3 (0.1 N) and NaOH (0.1 N). The determination of equilibrium state between the liquid and solids phases was studied at different interval time from 5 to 180 min with an initial dye concentration of 50 mg L−1 and optimum pH value. The adsorbent dose can also be very useful for predicting the cost of material per unit of the effluent of dye solution to be treated. Under optimum conditions of pH and contact time, the effect of the ratio solid/liquid was examined from 0.5 to 10 g L−1. Finally, the influence of the initial concentration of R-ETL dye onto the adsorption phenomena was studied at different initial concentrations ranging from 10 to 1000 mg L−1, at optimized above cited parameters.

Regeneration and reusability

The regeneration of adsorbents is important from an economical point of view; the adsorbents recycling is considered as an important aspect to minimize the costs related to the synthesis of materials. Under the optimum obtained conditions, the thermal treatment was used as regeneration method. 04 g of raw diatomite was mixed with 1 L of R-ETL dye solution at 1 g L−1 until the saturation. Then, the thermal regeneration was performed at 300, 600 and 900 °C during 2 h. Ethanol and deionized water were used as eluent to check the remaining amount of R-ETL after heat treatment. Moreover, in order to evaluate the reusability of regenerated materials, three loading cycles were carried out. The percentage adsorption of dye was obtained using the following equation:

$$R\,\left( \% \right) = \frac{{(C_{\text{i}} {-} C_{\text{e}} )}}{{C_{\text{i}} }} \times 100$$
(2)

Adsorption kinetics

The modeling of kinetics studies allows to describe the mechanisms and phenomena that control the adsorption process using theoretical models. The adsorption kinetics of R-ETL dye on RD, DH300, DH600 and DH900 have been tested on many theoretical models reported in the literature (see Table 1).

Table 1 Lists of kinetic adsorption models and their linearized expressions
Full size table

Adsorption isotherm

The modeling of adsorption isotherms is to describe in a wide range of concentration and temperature, experimental results with mathematical equations. Several isotherm models have been proposed for the adsorption study, and the main models are described in Table 2.

Table 2 Lists of adsorption isotherm models and their linearized expressions
Full size table

Error calculation

In recent decades, the linear regression has been one of the most viable tools used to define the best-fitting relationship and the correlation coefficient was the only indicator confirming the consistency of the theoretical assumptions of the kinetic and isotherm adsorption models (Foo and Hameed 2010). To check the validity of the above-mentioned models, the standard statistics of root-mean-squared error (RMSE) was calculated (Eq. 3) in addition to the correlation coefficient to support the best fit adsorption model.

$${\text{RMSE}} = 100\sqrt { \frac{1}{N} \mathop \sum \limits_{i - 1}^{N} \left( {q_{\text{t}}^{\exp } - q_{\text{t}}^{\text{cal}} } \right)_{i}^{2} }$$
(3)

where q exp t and \(q_{t}^{\text{cal}}\) (mg g−1) are, respectively, the observed and predicted amounts of R-ETL dye adsorbed, and N is the number of experiments.

Results and discussion

Characterizations

Table 3 summarizes the chemical composition and the physical properties of raw and heated diatomite.

Table 3 The chemical compositions of thermal treated diatomite
Full size table

It shows that the main elements for all materials are: O, C, Si, Ca and Al together with other mineral impurities as Fe, Mg, K, Na, etc. The high contents of Si, Ca, O and C correspond, respectively, to the silica frustules of diatoms and carbonate minerals, while the Al content indicates the presence of aluminosilicate minerals (Sun et al. 2013). The decrease in O and C during the heat treatment indicates the elimination of organic matter, while the increase of other elements designates purification and phase transformation of heated materials.

Particle size and pore distribution analysis (Fig. 1) shows that the coefficients of uniformity (Cu = D60/D10) are more then 2 (4.39; 3.60; 3.07; 2.98, respectively, for RD, DH300, DH600 and DH900) indicating a spread granulometry of all materials. As the value of Cu decreases (close to 2) by increasing temperature, it means that the distribution of particle becomes more uniform due to the elimination of impurities.

Fig. 1
figure 1

Dynamic light-scattering size distribution of RD, DH300, DH600 and DH900 in the aqueous suspensions

Full size image

X-ray patterns obtained from raw and heated diatomite are shown in Fig. 2. The broad diffraction at approximately 22° of 2θ is characteristic of non-crystalline opal-A, which could be identified as biogenic hydrated amorphous silica (Sun et al. 2013). Some mineral impurities as montmorillonite and ancrite are also found in the diatomite samples. After thermal treatment, the mineral structure of diatomite was not dramatically changed up to 600 °C, except a slight decrease in the pattern intensity of Calcite and Quartz. However, a new crystalline phase wollastonite appears on the XRD pattern of DH900. The appurtenance of wollastonite patterns can be explained by the reaction between the lime liberated from the decomposition of calcite and silica of diatoms skeletons which is found in abundance (Ibrahim and Selim 2012); this reaction is also followed by the desperation of calcite and quarts patterns.

Fig. 2
figure 2

XRD patterns of RD, DH300, DH600 and DH900

Full size image

As seen in Fig. 3, the SEM micrographs (b) and (c) still show the original form of diatoms genus Coscinodiscus Ehrenberg (Centrales). The disc-shaped diatomite treated at 600 °C has highly developed porous network ranging between macroporous and mesoporous structure. In addition, the concave and convex aspects of the surface were almost disappeared and smoother surface was formed due to the removal of impurities and organic matter. As reported by Ediz et al. (2010) that the heat treatment is the suitable method to increase the permeability of diatomite. The deconstruction of the shape and pores collapsing of diatom skeleton were occurred at 900 °C. The same finding was reported by Rangsriwatananon et al. (2008), regarding the form of diatoms at different thermal and acid treatments.

Fig. 3
figure 3

Scanning electron micrograph of heated diatomite a RD, b DH300 c DH600 and d DH900

Full size image

Nitrogen adsorption/desorption isotherms (inset in Fig. 4) and corresponding Barrett, Joyner and Halenda (BJH) pore-size distribution curves of RD, DH300, DH600 and DH900 samples are shown in Fig. 4. The N2 adsorption/desorption diatomites exhibit type II adsorption isotherms with H3 hysteresis loop according to the International Union of Pure and Applied Chemistry (IUPAC) classification (Gregg and Sing 1982). The hysteresis loop of this isotherm is associated with the filling and emptying of the mesopores by capillary condensation, implying that the pores are mainly narrow and slit-like (Sing and Williams 2004). In addition, increasing in the sharp of N2 adsorbed quantity near the relative pressure of 1 reflecting the existence of macropores in all samples, except DH900 sample.

Fig. 4
figure 4

Particle size distributions curves and nitrogen, adsorption–desorption isotherms (the inset) of heated diatomite samples

Full size image

The textural parameters of the raw and heated diatomite samples such as the specific surface area (SBET), pore volume (Vpore) and average pore size (W) are summarized in Table 4.

Table 4 Textural parameters of the heated diatomite samples at different temperatures
Full size table

It evident to say that up to 600 °C the isotherms shape of the heated samples was not greatly affected, and the average pore size was improved due to the elimination of natural organic impurities. However, the treatment of diatomite at higher temperatures (900 °C) decreases not only the nitrogen adsorption, which is indicated by a weakening of the hysteresis, but also the specific surface area and pores volume. The average pore size was slightly decreased reflecting the deconstruction of the porous structure of diatoms frustule.

Adsorption of R-ETL dye

Effect of pH solution

As shown in Fig. 5, the pH of solution affects significantly the adsorption of R-ETL dye. A decrease in the adsorption capacity with increasing pH medium was noted. The maximum quantities of R-ETL adsorbed were obtained at pH 2 with 12.7, 13.5, 15.2 and 8.8 mg g−1, respectively, for RD, DH300, DH600 and DH900, while the lowest quantities were recorded at pH 10 with 0.43, 0.1, 2.3 and 1.2 mg g−1 for RD, DH300, DH600 and DH900, respectively.

Fig. 5
figure 5

Influence of pH solution on the adsorption capacity of R-ETL dye onto RD, DH300, DH600 and DH900

Full size image

The highest adsorption capacities obtained at acidic condition might be due to the protonation effect of superficial functional groups of adsorbents, providing a strong electrostatic attractive forces between the surface of adsorbents and the negatively charged R-ETL dye molecules. The lower adsorption capacities observed at alkaline condition are probably due to the competition between the excess hydroxyl ions and the negative R-ETL dye molecules to occupy the same binding sites (Ncibi et al. 2009).

Effect of contact time

The results illustrated in Fig. 6a show a rapid increase in adsorption capacity during the first 20 min, where an adsorption amount of 11.3, 12.5, 13, 5 and 6.2 mg g−1 was recorded, respectively, for RD, DH300, DH600 and DH900. The time required to reach equilibrium between RD, DH300, DH600 and DH900 and the R-ETL dye solution is 30 min. The extension of the contact time up to 180 min (3 h) leads to an insignificant improvement, and the adsorption capacity remains almost constant.

Fig. 6
figure 6

a Effect of contact time on the adsorption process of R-ELT dye; b kinetic adsorption using pseudo-second-order model for modeling the adsorption of R-ETL dye onto RD, DH300, DH600 and DH900

Full size image

The rapid increase in the adsorbed quantity at the initial stage is probably due to the abundance of active sites easily accessible by the R-ETL dye molecules. The slowdown of the adsorption rate may be due either to the reduced of the available active sites on the surface or to the saturation stat of the adsorbents, which causes repulsive forces between the adsorbed dye molecules onto the surface and the remaining dye molecules in the solution (El Haddad et al. 2012).

The kinetic study of R-ETL dye onto RD, DH300, DH600 and DH900 is important to understand the dynamics of the adsorption reaction in terms of order of the rate constant. Thus, the kinetics parameters provide information for designing and modeling of the adsorption process. Four kinetic models: pseudo-first order, pseudo-second order, intra-particle diffusion and extra-particle diffusion were used to examine the experimental data.

Apart from the correlation coefficient (R2), RMSE was used to measure the goodness-of-fit. Table 5 regroups the values of kinetic parameters, correlation coefficients (R2), and the calculated errors for each model.

Table 5 Kinetic parameters for the adsorption of R-ETL dye onto RD, DH300, DH600 and DH900
Full size table

Except the pseudo-second-order kinetic model, the low correlation coefficients and the high RMSE errors confirm that the other kinetic models do not give a good regression, indicating a poor description of the removal phenomena of R-ETL molecules by RD, DH300, DH600 and DH900. However, the best correlations were obtained for the pseudo-second-order kinetic model with highest R2 (0.99) and lowest RMSE errors. Furthermore, the calculated Qe values through this model are close to the experimental values (see Fig. 6b). This strongly suggests that the adsorption of R-ETL molecules onto raw and heated diatomite is most appropriately represented by the pseudo-second-order kinetic model, supposing that two reactions either in series or in parallel are occurring, the first one is fast and reaches equilibrium quickly and the second is a slower reaction that can continue for a long period of time.

Effect of adsorbent dosage

Figure 7 depicts the effect of adsorbents dosage on the adsorption capacity of R-ETL dye. As shown in Fig. 8, the increase in the adsorbent dosage leads to the increase in the adsorption capacity. For instance, by increasing the adsorbents concentration from 0.5 to 10 g L−1 the adsorption increased from 0.97 to 5.33; 1.24 to 5.84; 1.63 to 8.57; 0.88 to 3.00, respectively, for RD, DH300, DH600 and DH900. This fact is mainly due to a substantial increase in the number of active sites available on the surface of the tested adsorbents. Further, by increasing the adsorbent concentration up to 10 g L−1 the adsorbed amount remains constant. This may due to the saturation of the available active sites or to the equilibrium stat established between the adsorbed dye molecules and those remain in the solution at the studied initial dye concentration (Ncibi et al. 2009).

Fig. 7
figure 7

Effect of adsorbent dosage on the adsorption capacities of R-ETL dye onto RD, DH300, DH600 and DH900

Full size image
Fig. 8
figure 8

a Influence of the initial R-ETL dye concentration onto the adsorption process; b adsorption isotherm using BET model for modeling the adsorption of R-ETL dye onto RD, DH300, DH600 and DH900

Full size image

Effect of initial dye concentration

It can be seen from Fig. 8a that at lower concentration values (C0 < 200 mg L−1) the uptake capacity increased from 0.19 to 34.86; 2.68 to 52.23; 3.3 to 51.64 and 0.1 to 39.38 mg g−1, respectively, for RD, DH300, DH600 and DH900 as the initial dye concentration increased from 10 to 200 mg L−1. Afterward, by further increasing the R-ETL dye concentration (up to 1000 mg L−1) the adsorption behavior of R-ETL dye displays a point of inflexion and the adsorption capacities reach 291.15; 359.29; 390.64 and 322.36 mg g−1, respectively, for RD, DH300, DH600 and DH900. This same phenomenological was reported by Bisinella Scheufele (Bisinella et al. 2016) for the adsorption of RB5G dye on sugarcane bagasse.

This effect could be explained as: (1) At lower initial dye concentrations, all R-ETL dye molecules could interact with the binding sites on the adsorbents surface and high sorption rates occur which lead to the formation of monolayer (one molecule is adsorbed per site). After that, the binding sites on the adsorbents surface became saturated and no further adsorption occurs; (2) At high initial R-ETL dye concentrations the driving force increases due to the gradient of concentration at solid/liquid interface generating a strong adsorbate–adsorbate interaction, which promote the formation of other layers (multilayer adsorption) involving some intermolecular interactions (as van der Waals forces) supplemented by various electrostatic contributions (as polarization, field–dipole and field gradient–quadrupole interactions) between the molecules of R-ETL dye (Naderi 2014).

In order to better understand the mechanisms involved in the adsorption process of R-ETL dye onto RD, DH300, DH600 and DH900, it is important to establish a more appropriate relationship between the adsorption capacity (Qe) and the equilibrium concentration (Ce). Various adsorption isotherm models: Freundlich, Langmuir, Dubinin–Radushkevich and BET were employed. The calculated isotherm constants and error values obtained for each model are given in Table 6.

Table 6 Adsorption isotherm constants for the adsorption of R-ETL dye onto RD, DH300, DH600 and DH900
Full size table

The negative obtained values of Langmuir constants (Qmax and KL) describe that the adsorption process of R-ETL dye did not follow Langmuir isotherm model, since KL indicate the adsorption binding energy on the surface. Furthermore, and except the isotherm data of DH900 which is best-fitted by Freundlich isotherm model with a correlation coefficient of 0.94 and RMSE of 11.58, the BET isotherm model is the most appropriate to describe the adsorption of R-ETL dye on RD, DH300 and DH600 (see Fig. 8b) with a correlation coefficient of 0.99, 0.98, 0.98, respectively. These best correlations confirming the assumptions of mono- and multi-layers phenomenological involved during the adsorption process of R-ETL onto the surface of adsorbents.

Effect of calcination temperature on the adsorption capacity

The structure of diatomite from the one hand was affected by the calcination process but form the other hand the adsorption capacities were improved. Based on the chemical compositions, the RD show some natural organic impurities sticks on the surface which may intervene in the adsorption of N2 gas and give more surface area than the heated samples. These organics matter probably competes the molecule of R-ETL dye to be absorbed into the surface. After calcination, at 300 °C the adsorption capacity improved due to transformation of this organic matter to coal (12.67%). Afterward, the heating up at 600 °C burn out all the organic impurities from the pores and makes the surface smoother (see Fig. 3c) which facilitate the interaction between the R-ETL dye molecules and the functional groups onto the surface of adsorbent that explain the highest adsorption quantity of dye on DH600. The calcination at 900 °C is not suitable as method of treatment, because the most structural characteristics including the surface area of diatoms frustule were damaged.

Material recycling

Figure 9a, b shows, respectively, the number of adsorption reusability cycles and UV–visible absorption spectrum of the released amount of R-ETL dye after thermal regeneration. It is clear that the recovery of R-ETL dye from aqueous solution slightly decreases and around 10% of removal efficiency was lost after each regeneration cycle. The recycled samples can restore up to 66.4, 73.35 and 67.85%, respectively, for DH300, DH 600 and DH900 after three thermal regeneration cycles. Even, the absorbance of the released amount was decreased in both UV and Visible ranges with increasing of thermal regeneration. As well illustrated in Fig. 9b, during thermal regeneration the organic dye substances adsorbed onto the surface or in pores decomposes to carbon. Thereafter, by increasing temperature up to 600 °C, the carbon formed oxidizes in air and disappears completely from the surface, making the pores free and fresh for further adsorption cycles (Vimonses et al. 2009b).

Fig. 9
figure 9

a Reusability cycles of regenerated samples to remove R-ETL dye; b UV–visible absorption spectra of R-ETL dye obtained from different samples after thermal regeneration

Full size image

Conclusion

In this study, the heat treatment of diatomite was used for the one hand to improve its adsorption capacities for textile dye removal, and for the other hand as regeneration approach in water purification field. We showed that the specific surface area and pore volume are the main affected textural characteristics, but the macro-structure pores network of heated diatomites remains intact after calcination. At low temperatures (less than 600 °C), the mineral phases remain unchangeable, and the disk shape of treated samples still shows the original form of diatoms skeletons. The maximum adsorption capacities were obtained at low pH value (pH 2), and high adsorbent dose (4 g L−1). Based on the effect of contact time, the adsorption process was fast and the equilibrium was rather short which reached after 30 min. The kinetic studies were well represented by the pseudo-second-order kinetic model. The isotherm adsorption revealed that the BET model describes satisfactorily the adsorption process of R-ETL dye onto RD, DH300 and DH600, while Freundlich model is more appropriate to the adsorption of R-ETL onto DH900. The mono- and multilayer adsorption phenomenological of R-ETL dye is the main assumptions for the present adsorption process. Our findings allow to conclude that the thermal treatment is an effective method of treatment which lead to enhance the adsorption capacities of diatomite and promote the possibilities for material recovery and recycling.