1 Introduction

Tetracyclines (TCs) constitute a group of antibiotics, some natural and others obtained by semi-synthesis, that cover a wide spectrum of antimicrobial activity. In addition to urinary tract infections, chlamydia, and acne, TCs can also be used to treat other bacterial infections such as respiratory tract infections, skin, soft tissue infections, among others [7, 24, 41, 50]. The wide use of these antibiotics in veterinary medicine and also as an additive in animal feed, have brought as a consequence its presence in food products such as meat, milk, honey and chicken [10, 13, 48]. TCs is not limited to food; they can also enter the environment because sewage treatment plants cannot completely remove them [26]. Moreover, there is a growing environmental concern about antibiotics because their presence in soils and waters leads to the emergence of resistant species. High concentrations of TCs and three degradation products were reported in the effluents (5.28–8.32 μg L−1) and sludges (34.6–49.6 μg kg−1) of three municipal wastewater treatment plants located in Turkey [64] and in aquatic organisms (4.23 to 208.14 ng g−1) [20]. Given these concerns, the development of analytical methodologies that allow the determination of low concentrations of tetracycline in environmental matrices is crucial. Among the analytical methods reported in different works, HPLC, ELISA, capillary electrophoresis, spectrophotometry, chemiluminescence and electrochemical approaches stand out [1, 9, 40, 42, 48, 56, 68]. The main disadvantages of these techniques are that they have tedious sample pretreatment processes and long analysis times [33]. On the other hand, electrochemical sensors are an attractive alternative for the detection of TCs due to their high selectivity, rapid detection and possibility of in situ applications [33]. Graphene materials, as constituent of electrochemical sensors, seem to have adequate properties for these applications. Nanomaterials composed of graphene oxide (GO) have high surface area and high specific capacitance due to its unique structure, which allows it to store a large amount of electrical charge. In addition, GO has oxygenated functional groups, such as epoxide (C–O–C), hydroxyl (C–OH), and carboxyl (-COOH) groups, which are located both in the basal planes and at the edges of the material [8, 54]. These characteristics make them appropriate for the adsorption of different classes of compounds and therefore, for applications such as electrochemical sensors and biosensors for the determination, for example, pesticides, drugs, nitrogenous bases and others [16, 18, 35, 64, 67, 71]. Wong et al. [66] fabricated a carbon paste electrode by combining multiwalled carbon nanotubes and graphene oxide for TC detection. They applied adsorption separation differential pulse voltammetry (AdSDPV) in which tetracycline was firstly electro accumulated and then analyzed. This technique only detects those analytes that are strongly adsorbed on the electrode surface. The sensor was applied to different samples such as artificial urine, river water and pharmaceutical samples. Lorenzetti et al. [34], on the other hand, applied a modification of AdSDPV to determine TC in milk and river samples using disposable screen-printed electrodes. The technique allowed analysts to improve the selectivity of electrochemical sensors by exploiting the adsorption properties of the sensing surface [43].

Within the tetracycline’s family, the so-called tetracycline (TC) is one of the substances that constitute this group of antibiotics. It has different acid groups in its structure and can exist under different ionic species and conformations depending on the pH. The presence of such groups in the TC molecule generates potential sites of interaction with metal ions and surfaces [46].

The adsorption of TC on montmorillonite, mesoporous silice, graphene-based materials and others adsorbents have been previously studied [6, 17, 19, 39, 45, 52]. Nevertheless, there is very little information on the TC adsorption in the presence of a divalent cation such as Ca2+ under different pH conditions. This cation is ubiquitous in natural media and is implicated in many biological and environmental processes. There are studies regarding its influence on the adsorption of different organic and inorganic compounds on different adsorbents [4, 28, 29, 31, 32, 59]. Chowdhury et al., [11] studied the influence of pH, ionic strength, ion valence and the presence of natural organic matter (NOM) on the aggregation and stability of graphene oxide (GO) and three reduced GOs (rGO). They found that the stability depended on pH, ion valence, and concentration of surface functional groups. In the presence of divalent cations (Ca2+, Mg2+), the increase in pH decreased the stability of GO, which was due to Ca2+ adsorption on the surface functional groups of GO. In the presence of NOM and divalent cations (Ca2+, Mg2+), GO aggregates settled from suspension because Ca2+ ions act as bridges between GO functional group and NOM. This indicates that pH and divalent cations can play complex roles in the adsorption properties of rGO and GO and could affect significantly adsorption and detection of third molecules such as tetracyclines.

The present work focuses on obtaining GO nanoparticles with TC adsorption properties for a future application as electrochemical sensors or as TC sorbents for decontamination. The synthesis and structural characterization of GO is firstly presented, followed by an evaluation of the TC adsorption properties. Special attention is paid to the effects of Ca2+ on the adsorption at different pH, which could be used to optimize the performance of GO as a sensor or as a TC adsorbent.

2 Materials and methods

2.1 Materials

Graphite powder (Gr) with a particle size < 50 µm was supplied by Sigma-Aldrich. KMnO4 (99.5%), H2SO4 (95–98%), H3PO4 (85%) and HCl (36–38%) were supplied by Cicarelli. Hydrogen peroxide (30% v/v) was supplied by Anedra and ethanol (96%) was supplied by Porta. Tetracycline hydrochloride (TC, purity 99%) was obtained from Parafarm and was used without further purification. TC stock solutions were prepared just before use to avoid degradation caused by oxygen and light. CaCl2 and 0.01 M KCl solutions were used for Ca2+ effect experiments and as electrolyte for zeta potential measurements. NaOH and HCl solutions were used for pH adjustment.

2.2 Synthesis of graphene oxide

GO was synthesized using Tours’ method [37]. A 9:1 mixture of H2SO4:H3PO4 (360:40 mL) was added to Gr powder (3.0 g) and KMnO4 (18.0 g), producing a slightly exothermic reaction, which increased the temperature of the reaction vessel to 35–40 °C. The product of the reaction was then heated to 50 °C and stirred for 12 h, cooled at room temperature and poured afterwards onto ice (~ 400 mL) with 30% H2O2 (3 mL). This last procedure produced important bubbling, a yellow coloration on the product, which finally became brownish. The resulting mixture was centrifuged (10,000 rpm for 20 min), and the supernatant was discarded. The remaining solid material was washed first with 200 mL water, then with 200 mL 30% HCl, and finally with two washings of 200 mL ethanol. The washed material was dried overnight at 42 ºC in air. The obtained brown material will be referred as GO hereafter.

2.3 Characterization techniques

Textural characterization of Gr and GO were conducted with N2 adsorption/desorption isotherms at 77 K using a Micromeritic Gemini V2.0 2380 equipment. Samples were degassed at 323 K for approximately 12 h. Surface area values were calculated from the linear adjustment of the Brunauer–Emmett–Teller (BET) equation with adsorption data obtained in the relative pressure (p/p◦) range between 0.01 and 0.33. The total pore volume and the average pore size were obtained by the Barrett-Joyner-Halenda (BJH) method. X-ray diffraction (XRD) patterns were obtained with a Rigaku D-Max II-C diffractometer, with CuKα radiation (λ = 1.54 Å) of 40 kV and 20 mA. Scans were recorded between 5 and 60º 2θ, with a step size of 0.02º 2θ and a scanning rate of 2º min−1. Crystal dimensions were calculated using the Scherrer equation. The morphology of GO was characterized using transmission electron microscope (TEM, JEOL 100 CX) images. The TEM micrographs of the GO sample were collected at a maximum accelerating voltage of 100 kV. The sample was prepared by dispersing a small amount of GO in ethanol via sonication for about 15 min. A few drops of this suspension were placed in 200 mesh grids provided with a Formvar film. FT-IR spectra were obtained with a IRTracer-100 Shimadzu spectrophotometer. The sample was prepared in KBr pellets with 0.3% w/w of GO or Gr. The spectra were recorded between 400 and 4000 cm−1 with a 4 cm−1 resolution and a 32 min−1 acquisition rate. X-ray Photoelectron Spectra (XPS) were obtained with a Thermo Scientific K-Alpha + X-ray Photoelectron Spectrometer. Spectra were recorded at room temperature, using Al-K (1200W) radiation for excitation and a 180° double focus hemispherical analyzer in a vacuum chamber of 1–10 mbar. Spectra were collected for oxygen (O1s), carbon (C1s) and a survey to identify the different species. The deconvolution of the spectra was obtained by means of Igor software using Voigt functions. Thermogravimetric analyses (TGA) data were recorded on a Mettler Toledo TGA/DSC1 instrument, the samples were heated from room temperature to 1000 ºC (10 ºC min−1) in N2 atmosphere. UV–Vis spectra were collected in the range of 200–700 nm using a Genesis 10S UV–Vis Spectrophotometer (Thermo Scentific). A zetasizer (Nano-ZS) Malvern equipment was used to determine the zeta potential of Gr and GO particles at different pH values. Gr and GO suspensions were prepared in 0.01 M KCl solutions, the pH was adjusted to the desired value and after 15 min equilibration, the electrophoretic mobility was measured. Zeta potential was calculated using the Smoluchowski equation. The effects of Ca2+ on the electrophoretic mobility of GO at different pH values were also evaluated. For this, a stable GO dispersion was prepared in 0.01 M KCl as electrolyte and in presence of different CaCl2 concentration (0.001 M and 0.01 M). The pH was modified gradually in all cases, and after equilibration the mobility was measured.

2.4 Adsorption studies

Adsorption studies under equilibrium conditions were carried out to investigate the removal capacity of tetracycline (TC) and the effects of pH and Ca2+ on it. Batch adsorption experiments were carried out in 10 mL glass flasks continuously shaken with an orbital shaker (200 rpm) at 22 ± 1 °C in duplicate. The concentration of TC in the equilibrium solution was quantified by UV–Vis spectrophotometry. Since the wavelength of the absorption maximum of TC changes with pH and presence of Ca2+, calibration curves in the concentration range of 0–48 mg L−1 were performed at the corresponding pH values, with and without 0.01 M CaCl2.

The equilibrium adsorption capacity (Qe, mg g−1) of GO was calculated as

$${Q}_{e}=\frac{\left({c}_{i}-{c}_{e}\right)V}{m}$$

where ci and ce are the initial and equilibrium TC concentrations, V is the volume and m the mass of adsorbent used in the experiments. The ratio m/V is known as the adsorbent dose.

To acquire optimal adsorption experimental condition, the effects of some operation parameters (i.e., adsorbent dose, TC initial concentration and pH solution) on TC adsorption were explored. The effect of adsorbent mass was evaluated at an initial concentration of 30 mg L−1 and pH 5.

Adsorption isotherms at 22 °C were performed in 10 mL capped glass vessels varying TC initial concentration from 20 to 90 mg L−1 with the same adsorbent dose (250 mg L−1). Two pH values were investigated (5 and 9.5). Once equilibrated, the dispersions were filtered with 0.45 µm filter, and the supernatant separated for TC quantification. Similar adsorption experiments were performed in the presence of 0.01 M Ca2+, in order to evaluate the effects of calcium in the adsorption behavior of GO.

The effect of pH on TC adsorption was investigated in the pH range 2.5–11.0. The initial concentration of TC was 30 mg L−1, and the adsorbent dose was 250 mg L−1. Capped glass vessels with a total volume of 10 mL were used. The pH effect was also analyzed in the presence of 10–2 and 10–3 M Ca2+.

Additional experiments to evaluate the effect of Ca2+ on TC adsorption were performed at constant pH and varying calcium concentration. The initial concentration of TC was 30 mg L−1, the adsorbent dose was 250 mg L−1 and the final volume was 10 mL. Used Ca2+ concentrations were 0.0001 M, 0.001 M, 0.01 M, 0.1 M and 1.0 M, at two different pH values, 2.5 and 9.5.

For the adsorption − desorption cycle test, 62.5 mg of the adsorbent was added into a 250 mL TC solution (30 mg/L) in absence (pH 5) and presence of Ca2+ (0.01 M) at pH 9 for 24 h. The first adsorption capacity was calculated as previously described. The adsorbent with tetracycline-adsorbed was regenerated by immersing it into water at pH 5 and water containing 0.01 M Ca 2+ at pH 9. Then, the adsorbent was reused in adsorption experiments. The removal efficiency was obtained on the basis of the initial adsorption capacity. The extra adsorption − desorption processes were repeated four times.

3 Results and discussion

3.1 Characterization

The results obtained by N2 adsorption/desorption experiments are listed in Table S1, in the Supporting Information. Gr has a BET area of 14 m2 g−1, which increased to 30 m2 g−1 after the oxidation process. The average pore size of GO is 5.6 nm, indicating that it is a mesoporous material. In comparison with others graphene oxide obtained from natural graphite and oxidized by a modified Hummers method [28, 29], the studied sample has intermediate specific surface area and pore volume [5, 28, 29, 58, 61]. The pH values of Gr and GO dispersions (Table S1, in the Supporting Information) also indicate that GO exhibits strong acidity while Gr is almost neutral (slightly acidic).

The XRD patterns and structural parameters of Gr and GO are shown in Fig. 1A and B, and listed in Table S2 in the Supporting Information, respectively. Gr shows three main reflections around 26.6°, 44.7°, and 54.8° 2θ corresponding to (002), (101), and (004) crystallographic planes, respectively. The intense and sharp peak of the (002) plane suggests a highly ordered material, with multiple graphene layers [38]. After the oxidation, the peak at 26.6° 2θ disappeared as a consequence of the complete oxidation of Gr [21] and a typical peak associated to GO ((001) plane) appeared at around 9.26º 2θ. These findings are indicative of a significant reduction of the close-packed hexagonal structure and an increase of stacking disorder due to the existence of intercalated oxygenous groups [70]. The increase in the interlayer spacing (from 0.34 nm to 0.95 nm) was attributed to the conversion of sp2 carbon to sp3 carbon, evidencing GO oxidation and exfoliation [53, 62].

Fig. 1
figure 1

(A) XRD pattern of Gr. (B) XRD pattern of GO. (C) TEM micrograph of GO

Full size image

The crystallites height (Lc) decreased after oxidation from 21.49 nm in Gr to 12.06 nm in GO (Table S2, in the Supporting Information). The layer separation in graphene oxide enhanced by oxygenated functionalities is also evidenced by TEM analysis (Fig. 1C). Dark areas indicate the thick stacking nanostructure of several graphene oxide sheets. More transparent areas indicate thinner films of a few layers of GO resulting from exfoliation [60]. As it was reported in the literature, the wrinkled morphology is attributed to the presence of a large number of functional groups such as hydroxyl and carboxyl groups on the edge, and carboxyl and epoxide groups in the inner part of GO [12].

The FT-IR spectra of Gr and GO, displayed in Fig. 2A and B, show characteristic vibration bands of these materials. Gr spectrum resulted to be relatively flat, with weak signals. The band at 1580 cm−1 is attributed to the C = C vibration of graphene sheets. The spectrum of GO was very different, with stronger signals. The most intense band appears at 3400 cm−1 and corresponds to the O–H stretching vibration of hydroxyl groups and water molecules. The deformation vibration mode of O–H groups appears at around 1410 cm−1. The band at 1740 cm−1 was associated to the C = O stretching vibration of a carbonyl group; another band at approximately 1620 cm−1 was attributed to the C = C skeletal vibration of the graphene sheets (unoxidized graphitic domains) [55]; a band appearing at 1220 cm−1 was in turn attributed to the stretching vibration of epoxy C–O–C group; and the band at 1050 cm−1 was finally assigned to the alkoxy C-O stretching vibration (carboxyl group) [2, 15, 65]. A comparison of Gr and GO spectra clearly indicates that GO is the material with the highest proportion of oxygenated groups.

Fig. 2
figure 2

(A) FTIR spectra of Gr in KBr pellets. (B) FTIR spectra of GO in KBr pellets. (C) XPS spectrum of GO. (D) Relative proportions of C1s signals of GO

Full size image

The presence of oxygenated groups on GO was also detected by XPS. The elemental analysis obtained by XPS indicated C: 64.3 atom%, O: 32.1 atom%, N: 1.9 atom% and S: 1.6 atom%. The C/O ratio equal to 2 indicates a successful introduction of oxygen atoms into the ordered arrangement of graphite structure. The C1s XPS signal shows different types of carbon components. The C1s spectrum show four peaks that correspond to the following functional groups: carbon sp2 (C = C, 284.2 eV), carbon sp3 (C–C, 284.9 eV), epoxy/hydroxyls (C-O, 286.8 eV), and carbonyl (C = O, 288.6 eV) (Fig. 2C) [49, 69]. The relative proportion of these groups is shown in Fig. 2D.

Figure 3A shows the TGA of Gr and GO and Fig. 3B shows the DSC of GO, all performed in N2 atmosphere. Gr was stable in the whole range of temperature analyzed. On the contrary, GO begun to decompose at approximately 50ºC, and lost up to 85% of its total weight when heated to 200ºC. The first endothermic peak is due to water desorption, with around 10% of mass loss. The higher mass loss occurred between 150–190 °C, with two marked exothermic peaks at 162 and 181 °C. This high thermal reactivity of GO can be attributed to the decomposition of the labile oxygen-containing moieties, promoted by the disruption of the multilayered stacks structure of GO [16, 18, 23].

Fig. 3
figure 3

(A) TGA analysis corresponding to Gr and GO under N2 atmosphere. (B) DSC plot of GO

Full size image

The UV–vis spectrum of GO is shown in Fig. S1, in the Supporting Information. It presented a characteristic absorption band at 230 nm and a broad shoulder at 300 nm, which can be assigned to the π-π* transition of the C = C bonds and n-π* transition of the C = O bonds (carbonyl or carboxyl group), respectively. The presence of the shoulder at 300 nm is a good evidence of oxidation [25, 27, 47]. On the contrary, Gr particles settled so quickly that only a clear supernatant remained, and that is what is observed in the UV–visible absorption spectrum.

The electrokinetic properties of Gr and GO are shown in Fig. 4A. Gr had a negative zeta potential of around -22 mV at high and intermediate pH, and then the potential decreased as the pH decreased, reaching a nearly isoelectric point at pH 2.1. GO, on the contrary, exhibited a more negative zeta potential in the whole range of pH investigated. It only decreased from -35 mV to -30 mV from pH 10.5 to pH 2.1. The negative zeta potential is in line with the fact that oxidation treatment introduced oxygen-containing functional groups on the surfaces of GO [28,29,30, 57]. Consequently, GO suspensions resist more aggregation and tend to disperse better than Gr. Figure 4B shows that calcium produces significant changes to the electrokinetic behavior of GO. The zeta potential becomes less negative as calcium concentration increases, and the isoelectric point shifts to pH around 6 in 10−2 M calcium, evidencing interaction of Ca2+ with the surface groups of GO.

Fig. 4
figure 4

(A) Zeta potential (ζ) as a function of pH in 0.01 M KCl: Gr (●) and GO (○). (B) Zeta potential (ζ) of GO in the presence of CaCl2: GO with CaCl2 10−2 M (Δ) and GO with CaCl2 10–3 M (▲)

Full size image

3.2 Adsorption studies

Figure S2 exhibits that TC removal capacity (%) on GO increases with the adsorbent dose from 50 to 500 mg L−1 at pH 5. A dose of 250 mg L−1 was selected to study the effect of concentration, pH and Ca on the adsorption of TC on GO. A dose of 250 mg L−1 results in a removal of 60%, which allows the effect of the experimental variables on the adsorption of CT to be analyzed.

TC adsorption isotherms on GO at pH 5 and 9 are presented with Langmuir settings in Fig. 5 and Table 1. TC adsorption decreased with increasing pH from 5 to 9.5 in absence of CaCl2. In the literature, the pH was found to impact similarly the adsorption of TC on different solids [36, 44],Topal and Arslan, 2020). The results suggest that electrostatics interactions are playing a role on the adsorption, because the adsorbent is negatively charged in the analyzed pH range (Fig. 4A) and TC species increase their negative charge as pH increases [44]. In fact, at pH 5 TC is present in the zwitterionic (TCH±) and anionic (TCH−) forms, whereas at pH 9.5 it is in its negative and doubly negative forms (TCH− and TC2−). Even though the net charge of TCH± and TCH− is 0 and -1, respectively, both species contain a positively charged group in their structure, the dimethylammonium group and thus, the molecules can arrange at the surface locating the positively charged group close to the surface and the negatively charged group(s) as far as possible from the surface [45], resulting in adsorption at pH 5. At pH 9.5, repulsive forces seem to prevail, and adsorption is nearly zero. These results were also corroborated by the pH effect adsorption assay (see below).

Fig. 5
figure 5

Adsorption isotherms and Langmuir settings of TC on GO: at pH 5 with 0.01 M CaCl2 ( ) and without 0.01 M CaCl2 ( ); at pH 9.5 with 0.01 M CaCl2 (▲) and without 0.01 M CaCl2(●)

Full size image
Table 1 Comparison of the Adsorption Capacity of GO with other graphenic adsorbents for TC
Full size table

The effect of Ca2+ on TC adsorption isotherms at pH 5 and 9.5 is also shown in Fig. 5. The isotherms practically did not change by changing the pH in the presence of calcium. At pH 5 the presence of calcium decreased the removal capacity of GO, whereas at pH 9.5 the divalent cation significantly increased the removal capacity. The result was that isotherms practically coincided at pH 5 and 9.5, when calcium was present. It is possible that at pH 5 Ca2+ competes with the positively charged group of TC species, decreasing the adsorption. This competition is no longer operative at pH 9.5, where an opposite Ca2+ effect was observed, probably as a consequence of the interaction of the divalent cation with negatively charged groups of the GO surface and negatively charge groups of TC species. Parolo et al., [46] reported the same effects of calcium on TC adsorption on the negatively charged montmorillonite surface, postulating the formation of Ca2+ bridges between the surface and TC. This seems to be also the case with GO, with Ca2+ forming GO-Ca2+-TC bridges and thus inducing the adsorption of negatively charged TC species on a negatively charged surface.

As it is shown in Table 1, the TC adsorption capacity of GO is comparable to other graphenic adsorbents reported in the literature. There are results informed for pure GO and for functionalized GOs. The performance is better than that of other GO with no functionalization. The performance is worse than that of some functionalized GO or composites containing GO.

The effect of calcium on TC adsorption was explored in more detail by investigating the removal at eight different pH values and at three different calcium concentrations. The results are shown in Fig. 6, and can be understood as an interplay between the two main roles played by calcium on TC adsorption, i.e., competitor or bridging cation. At pH lower than around 6, the competitor role of calcium prevails and thus TC removal decreased by increasing calcium concentration to 0.001 M or 0.01 M. At pH higher than around 6, on the contrary, calcium seems to act mainly as a bridge between the surface and TC, increasing TC removal by increasing calcium concentration to 0.001 M or 0.01 M. The general effect of calcium is, then, to flatten the adsorption vs pH curves, making the removal only weakly dependent on pH.

Fig. 6
figure 6

Effect of pH in presence of Ca2+ on the adsorption of TC in GO. GO dose 250 mg L−1 and initial TC concentration of 30 mg L−1

Full size image

The effects of calcium were further investigated by working at two extreme pH values (2.5 and 9.5), with six different calcium concentrations. The results are shown in Fig. 7, which presents TC removal at different Ca2+/TC ratios, for a constant TC concentration. At pH 2.5, where TC is fully in its cationic form, the results in Fig. 7 are typical of competition between calcium and TC for adsorption sites, with a monotonous decrease in TC adsorption as the Ca2+/TC ratio increases. At pH 9.5, instead, the trend is no so simple. TC removal increases by increasing calcium concentration up to a Ca2+/TC ratio of 11, passes through a maximum, and then decreases at higher Ca2+/TC ratios. Bridging calcium may be still responsible for the adsorption under these conditions. However, it is very difficult to completely understand the whole behavior, especially why the removal decreased at very high calcium concentrations. At such high Ca2+ concentrations the surface probably becomes positively charged, as suggested by zeta potential data. In addition, the fact that the molar Ca2+ concentration is 11,000 times higher than that of TC may induce the formation of Ca2+-TC species with an excess of calcium in the structure and low affinity for the surface. In fact, speciation calculations by Parolo et al. [45] showed that at high calcium concentrations and pH > 6 the dominant species in solution in a Ca2+-TC mixture is Ca2TC2+, which may have low affinity for the positively charged GO surface.

Fig. 7
figure 7

Effect of CaCl2 concentration on TC adsorption on GO at pH 9.5. GO dose 250 mg L−1 and initial concentration of TC 30 mg L−1. Ca2+/TC: ratio of molar concentrations

Full size image

Adsorption–desorption experiments using water at two different pH as eluent were conducted to check the reusability of GO for TC adsorption. As illustrated in Fig. 8, after three cycles of adsorption–desorption, there were similar decreases in the uptake performance for both pH used (32.68% at pH 5 and 29.56% at pH 9) that might be ascribed to partially irreversible adsorption that occupied some of the adsorption sites or the loss of adsorbent during the desorption process. Thus, due to the fact that the adsorbent regeneration can be affected by various parameters such as the type and amount of the used reagent [14], the decrease in efficiency is acceptable, and further study on the use of a more efficient regeneration solution should be explored.

Fig. 8
figure 8

Reusability of GO for TC adsorption. At pH 5 the adsorption step was conducted in pure water, since this is a situation where high adsorption is attained. At pH 9, the adsorption step was conducted in the presence of 0.01 M CaCl2, for the same reason

Full size image

4 Conclusions

GO nanoparticles were prepared and characterized for TC adsorption. Adsorption varies widely with pH and Ca2+ concentration. In the absence of Ca2+, adsorption was high at low pH and decreased as pH increased. It can be seen that in the presence of Ca2+ decreased TC adsorption at low pH but increased it at high pH. It is postulated that at least two different adsorption processes take place in the presence of Ca2+: at pH ≤ 5 there is cation exchange, and Ca2+ behaves as a competitor of the positively charged TC species. At pH ≥ 5, the formation of Ca bridges (GO-Ca2+-TC) is favored, increasing the adsorption of the antibiotic. Consequently, in the presence of Ca2+, the pH does not produce significant variations in the adsorption of TC. The combination of Ca2+ and pH effects are important to optimize the behavior of GO and the use of Ca2+ or not will depend on the application needed. For example, not using calcium in the adsorbing media is important to reversibly adsorb TC at low pH and desorb it at high pH, which may be useful for decontamination and further regeneration of the adsorbent. Using Ca2+, on the contrary, may be very advantageous for using GO in an electrochemical sensor, which will have a pH-independent response, facilitating the use of the device.