3.2 Characterisation of GO/HPAMAM/DAC adsorbent
Figure 2a shows the FT-IR spectra of GO, DAC, HPAMAM, and GO/HPAMAM/DAC. For GO, the stretching vibration peak of C = C, C = O and –OH appeared at 1623, 1718 and 3419 cm− 1, respectively (Qi et al. 2017). For HPAMAM, the bending vibration peak of –CH2–appeared at 1417 cm− 1. The –NH–flexural vibration peak and –C = O tensile vibration peak of the amide bond appeared at 1500 and 1585 cm− 1, respectively. The symmetric and asymmetric tensile vibration peaks of –CH2– appeared at 2843 and 2936 cm− 1. The tensile vibration peak of –NH2 appeared at 3084 cm− 1. The broad peak at 3287 cm− 1 was attributed to the superposition of the –NH–stretching vibration peak of the amide bond and the –NH2 stretching vibration peak. For DAC, the C = O stretching vibration peak of the aldehyde group appeared at 1750 cm− 1 (Zhu et al. 2016), confirming the introduction of the aldehyde group on the cellulose molecular chain after oxidation. In contrast, some new peaks appeared in GO/HPAMAM/DAC. The peaks at 1450, 1650 and 3380 cm− 1 correspond to the vibrations of –CH2–, –CONH– and –NH–, respectively (Zeng et al. 2007), confirming the graft of HPAMAM on GO through an amide bond. On the other hand, the C = N vibration peak appeared at 1558 cm− 1 and the C = O stretching vibration peak of the aldehyde group at 1750 cm− 1 disappeared, indicating that DAC was grafted onto GO/HPAMAM by forming the Schiff-based structure between the amino group of HPAMAM and the aldehyde group of DAC. Thus, each building unit of GO/HPAMAM/DAC adsorbent was connected by a covalent bond to achieve structural stability.
The XRD spectra of GO, GO/HPAMAM and GO/HPAMAM/DAC are shown in Fig. 2b. A strong peak appeared in the XRD spectra of GO, which was caused by the interlayer spacing of GO sheets (Lv et al. 2018). After grafting HPAMAM, the characteristic diffraction peak of GO was reduced, which may be due to the shielding of GO by HPAMAM. In addition, the characteristic diffraction peak of GO was further reduced after grafting DAC, and the characteristic diffraction peak corresponding to the typical crystalline structure of cellulose Ⅰ appeared at about 2θ = 23°. The decrease in the crystallinity of GO may be due to the grafting of cellulose molecular chains, which resulted in the non-uniform content of GO (Liu et al. 2017).
The surface chemical structure of GO/HPAMAM/DAC adsorbent, which affects its absorption performance, was analysed using XPS. As shown in Fig. 2c, a distinct characteristic peak corresponding to N 1s appeared at about 400 eV, which implies the grafting of HPAMAM in the adsorbent. This sharp peak indicates the relatively high nitrogen content in GO/HPAMAM/DAC, which might increase the adsorption capacity. The high-resolution spectrum of N 1s is shown in Fig. 2d. The characteristic peaks corresponding to C–N–C, C–N, N(C)3, C = N and N–C = O appeared at 399.2, 399.8, 400.1, 400.8 and 401.4 eV, respectively (He et al. 2016). The appearance of N–C = O and C = N proves the grafting of HPAMAM and DAC on GO, which is consistent with the results of FT-IR spectra. The primary, secondary and tertiary amines presented in GO/HPAMAM/DAC could promote the adsorption of heavy metal ions.
The SEM images of GO, GO/HPAMAM and GO/HPAMAM/DAC are shown in Fig. 3. The GO (Fig. 3a, b, c) displayed a smooth sheet-layer structure with a certain degree of stacking. After grafting HPAMAM (Fig. 3d, e, f), GO showed a rough surface. This may be due to the branched molecular structure and the nano-cavities of HPAMAM. The abundant active groups in HPAMAM could improve the affinity between adsorbents and heavy metal ions. In addition, the stacking degree of GO was reduced. After the grafting of DAC onto GO/HPAMAM (Fig. 3g, h, i), the surface of GO became rougher and micro/nano 3D bump appeared, which could increase the contact area between the adsorbent and pollutants for an efficient adsorption process. The large number of active groups in HPAMAM and the micro/nano 3D structures fabricated by cellulose would improve the adsorption performance of GO.
3.3 Adsorption performance of GO/HPAMAM/DAC adsorbent
As shown in Fig. 4a, the surface charge of GO/HPAMAM/DAC adsorbent in the pH range of 3.0–12.0 was analysed. It was easy to find that the zeta potential of the adsorbent decreased with the increase of pH. When the pH was 8.24, the zeta potential was zero. Therefore, the isoelectric point (pHip) of GO/HPAMAM/DAC adsorbent was 8.24. The surface charge of GO/HPAMAM/DAC adsorbent was positive at pH < 8.24 because the primary and secondary amines in HPAMAM interacted with H+ to form cationic groups (Zhu et al. 2016).
Figure 4b shows the adsorption capacity of the adsorbents for Pb(Ⅱ), Cd(Ⅱ) and Cu(Ⅱ). The unmodified GO had the worst adsorption capacity (42.4, 28.7 and 10.8 mg/g for Pb(Ⅱ), Cd(Ⅱ) and Cu(Ⅱ), respectively). After grafting HPAMAM, the adsorption capacity of GO/HPAMAM was improved (112.8, 58.9 and 24.3 mg/g for Pb(Ⅱ), Cd(Ⅱ) and Cu(Ⅱ), respectively). This was mainly due to the introduction of more active groups on GO, which enhanced the affinity between the adsorbent and heavy metal ions. After grafting DAC, the adsorption capacity of GO/HPAMAM/DAC showed more improvement (401.5, 242.8 and 131.4 mg/g for Pb(Ⅱ), Cd(Ⅱ) and Cu(Ⅱ), respectively). This was mainly because the micro/nano 3D structures fabricated by cellulose on GO increased the contact area between the adsorbent and the pollutants. Based on the above results, it can be concluded that introducing more active groups using HPAMAM and fabricating the micro/nano 3D structure using cellulose is an effective strategy to improve the adsorption capacity of GO. Subsequently, the adsorption kinetics and adsorption isotherms of Pb(Ⅱ), Cd(Ⅱ) and Cu(Ⅱ) onto GO/HPAMAM/DAC were investigated.
The adsorption kinetics of Pb(Ⅱ), Cd(Ⅱ) and Cu(Ⅱ) onto GO/HPAMAM/DAC adsorbent at 298 K are shown in Fig. 5a. There were two adsorption stages for these three adsorbates. First, the adsorption capacity increased rapidly in the first 150 min. Then, the adsorption process tended to balance, and the adsorption capacity reached a stable level. A large number of active groups and micro/nano 3D bumps in GO/HPAMAM/DAC adsorbent increased its interaction with heavy metal ions. Pseudo-first-order and pseudo-second-order kinetics were fitted to measure the data to investigate the adsorption process.
The pseudo-first-order kinetics is expressed as follows:
\(-\text{ln}\left(1-F\right)={k}_{1}t+C\) \(F=\frac{{Q}_{\text{t}}}{{Q}_{\text{e}}}\) (1)
where Qt (mg/g) is the amount of heavy metal ions adsorbed at the contact time t (min), Qe (mg/g) is the amount of heavy metal ions adsorbed at equilibrium, and k1 (min− 1) is the kinetic rate constant. The curves of -ln(1།F) vs. t are shown in Fig. 5b, and the k1 values are displayed in Table 1.
Table 1
Adsorption kinetic parameters of two kinetic models for different adsorbates.
Adsorbate
|
Pseudo-first-order
|
|
Pseudo-second-order
|
Qe(mg/g)
|
k1(1/min)
|
R2
|
k2(g/mg.min)
|
Qe(mg/g)
|
R2
|
Pb(Ⅱ)
|
148.8
|
0.01522
|
0.9940
|
|
0.00017
|
169.8
|
0.9977
|
Cd(Ⅱ)
|
131.6
|
0.00924
|
0.9857
|
|
0.00021
|
138.7
|
0.9970
|
Cu(Ⅱ)
|
118.2
|
0.01349
|
0.9749
|
|
0.00016
|
134.2
|
0.9978
|
The pseudo-second-order kinetics is expressed as follows:
$$\frac{t}{{Q}_{\text{t}}}=\frac{1}{{Q}_{\text{e}}}t+\frac{1}{{k}_{2}{{Q}_{\text{e}}}^{2}}$$
2
where k2 (g mg− 1 min− 1) is the kinetic rate constant. The curves of t/Qt vs. t are shown in Fig. 5c, and the k2 values are displayed in Table 1.
As shown in Table 1, the R2 of the pseudo-second-order kinetic model for Pb(Ⅱ), Cd(Ⅱ) and Cu(Ⅱ) adsorption were all larger than those of the pseudo-first-order kinetic model. It confirmed that the pseudo-second-order kinetic model can describe the Pb(Ⅱ), Cd(Ⅱ) and Cu(Ⅱ) adsorption onto GO/HPAMAM/DAC adsorbent, and chemical adsorption is the rate-determining step (Xu et al. 2017).
Figure 6a presents the adsorption isotherms of Pb(Ⅱ), Cd(Ⅱ) and Cu(Ⅱ) onto GO/HPAMAM/DAC adsorbent at 298 K. The adsorption capacities of GO/HPAMAM/DAC adsorbent for Pb(Ⅱ), Cd(Ⅱ) and Cu(Ⅱ) increased with the increase of initial concentration of heavy metal ions. The GO/HPAMAM/DAC adsorbent exhibited the strongest adsorption capacity for Pb(Ⅱ). The Langmuir and Freundlich isotherm models were used to demonstrate the adsorption process of GO/HPAMAM/DAC adsorbent for Pb(Ⅱ), Cd(Ⅱ) and Cu(Ⅱ).
The Langmuir isotherm is expressed as follows:
$$\frac{{C}_{\text{e}}}{{Q}_{\text{e}}}=\frac{{C}_{\text{e}}}{{Q}_{\text{m}}}+\frac{1}{b{Q}_{\text{m}}}$$
3
where Qm (mg/g) represents the maximum adsorption capacity of unit weight adsorbent for heavy metal ions, and b (L/mg) is the Langmuir adsorption constant. The curves of Ce/Qe vs. 1/Ce are shown in Fig. 6b, and the Qm and b values are displayed in Table 2.
Table 2
Adsorption isotherm parameters of two isotherm models for different adsorbates.
Adsorbate
|
Langmuir model
|
|
Freundlich model
|
Qm(mg/g)
|
b(L/mg)
|
R2
|
k(mg/g)
|
n
|
R2
|
Pb(Ⅱ)
|
680.3
|
0.0284
|
0.9927
|
|
29.91
|
2.3737
|
0.9123
|
Cd(Ⅱ)
|
418.4
|
0.0032
|
0.9909
|
|
2.74
|
1.5183
|
0.9648
|
Cu(Ⅱ)
|
280.1
|
0.0028
|
0.9907
|
|
19.11
|
4.3397
|
0.8174
|
The Freundlich isotherm is expressed as follows:
$$\text{ln}{Q}_{\text{e}}=\text{ln}k+\frac{1}{n}\text{ln}{C}_{\text{e}}$$
4
where k and n are Freundlich adsorption constants, k (mg/g) is a rough index for the adsorption capacity, 1/n represents the empirical parameter of adsorption intensity. The curves of ln Qe vs. ln Ce are shown in Fig. 6c, and the k and n values are displayed in Table 2.
As shown in Table 2, the R2 of the Langmuir isotherm model for Pb(Ⅱ), Cd(Ⅱ) and Cu(Ⅱ) adsorption were larger than those of the Freundlich isotherm model. Therefore, it can be concluded that the Langmuir isotherm model can describe the Pb(Ⅱ), Cd(Ⅱ) and Cu(Ⅱ) adsorption onto GO/HPAMAM/DAC adsorbent. In addition, the Qm of Pb(Ⅱ), Cd(Ⅱ) and Cu(Ⅱ) were 680.3, 418.4 and 280.1 mg/g at 298 K. Compared to most of the adsorbents listed in Table 3, GO/HPAMAM/DAC adsorbent exhibits a higher Qm for Pb(Ⅱ), Cd(Ⅱ) and Cu(Ⅱ). All the above results confirmed that GO/HPAMAM/DAC has the potential to adsorb heavy metal ions in wastewater as a promising adsorbent.
Table 3
The Qm for GO/HPAMAM/DAC compared with other adsorbents reported.
Adsorbent
|
Qm (mg/g)
|
Temperature
|
Reference
|
Pb(Ⅱ)
|
Cd(Ⅱ)
|
Cu(Ⅱ)
|
Graphene oxide/cellulose membranes
|
107.9
|
16.7
|
14.3
|
298 K
|
Sitko et al. (2016)
|
Modified biochar
|
-
|
-
|
16.1
|
293 K
|
Yang et al. (2014)
|
GO/CMC
|
76.7
|
46.1
|
82.9
|
293 K
|
Zhang et al. (2014)
|
Lignosulfonate-graphene oxide-polyaniline
|
216.4
|
-
|
-
|
303 K
|
Yang et al. (2014)
|
Multi-metal binding biosorbent
|
63.4
|
38.3
|
108.1
|
296 K
|
Abdolali et al. (2017)
|
Cross-linked grapheme oxide sheets via modified extracted cellulose
|
186.5
|
-
|
46.4
|
298 K
|
Yakout et al. (2017)
|
GO-HBP-NH2-CMC
|
150.3
|
-
|
137.5
|
298 K
|
Kong et al. (2020)
|
GO/HPAMAM/DAC
|
680.3
|
418.4
|
280.1
|
298 K
|
This work
|
In addition, cycling stability is an important factor that restricts the practical application of adsorbents. The adsorption capacity of GO/HPAMAM/DAC adsorbent for Pb(Ⅱ), Cd(Ⅱ) and Cu(Ⅱ) in five cycles is shown in Fig. 7. Even after five cycles, more than 90% of Pb(Ⅱ), Cd(Ⅱ) and Cu(Ⅱ) could be re-adsorbed. The good recycling performance of GO/HPAMAM/DAC adsorbent was attributed to the structural stability contributed by covalent grafting between GO, hyperbranched polymer and cellulose.
3.4 Adsorption mechanism of GO/HPAMAM/DAC adsorbent
The GO/HPAMAM/DAC adsorbent with a large number of active groups and micro/nano 3D structure showed strong adsorption capacity for heavy metal ions in wastewater. To study the adsorption mechanism of GO/HPAMAM/DAC adsorbent, FT-IR and XPS were conducted to analyse the adsorption sites of GO/HPAMAM/DAC adsorbent. The FT-IR spectra of GO/HPAMAM/DAC before and after heavy metal-ion adsorption are shown in Fig. 8a. For GO/HPAMAM/DAC, the broad peak at 3000–3500 cm− 1 was attributed to the superposition of O–H and N–H vibration, and the peaks at 1650 and 1558 cm− 1 were ascribed to O = C–NH and C = N vibration. After the adsorption of Pb(Ⅱ), Cd(Ⅱ) and Cu(Ⅱ), the peaks at 3000–3500 cm− 1 tended to broaden. This confirms that the hydroxyl and amino groups contributed to the adsorption of heavy metal ions. In addition, the peaks at 1650 and 1558 cm− 1 were weakened, suggesting that the nitrogen- and oxygen-containing groups on GO/HAPAMAM/DAC adsorbent could chelate and/or compound heavy metal ions by acting as electron donors.
The XPS spectra of GO/HPAMAM/DAC before and after heavy metal-ion adsorption are shown in Fig. 8b. The Pb, Cd and Cu appeared in XPS spectra after adsorption, confirming the pollutant adsorption on GO/HPAMAM/DAC. Figure 8c, d show the O 1s and N 1s XPS spectra of GO/HPAMAM/DAC before and after Pb(Ⅱ), Cd(Ⅱ) and Cu(Ⅱ) adsorption. For GO/HPAMAM/DAC, the peaks of O 1s and N 1s appeared at 531.5 and 398.6 eV. After adsorption, these two peaks moved towards high binding energy, indicating the interaction between oxygen-/nitrogen-containing groups and heavy metal ions in the adsorption process. This is consistent with the results of FT-IR.