MCM-41 - ACS Publications - American Chemical Society

Article pubs.acs.org/JPCC

An Investigation of the Selective Adsorptions of Metals on Mesoporous NH2‑MCM-41 Xinqing Chen,†,§ Wai Kwong Ching,† Koon Fung Lam,† Wei Wei,*,§ and King Lun Yeung*,†,‡ †

Department of Chemical and Biomolecular Engineering and ‡Division of Environment, The Hong Kong University of Science and Technology, Clear Water Bay, Kowloon, Hong Kong, People’s Republic of China § CAS Key Laboratory of Low-carbon Conversion Science and Engineering, Shanghai Advanced Research Institute, Chinese Academy of Sciences, Shanghai 201210, People’s Republic of China S Supporting Information *

ABSTRACT: The adsorption selectivity of NH2-MCM-41 follows the Irving-William order observed in transition metal complexes (i.e., Cu2+ > Zn2+ > Co2+ > Ni2+) that is consistent with X-ray photoelectron spectroscopy (XPS) and adsorption thermodynamics data. The interplays between surface aminopropyl ligands and the transition metals are adequately explained by the Pearson HSAB principle, but a multicomponent Freundlich adsorption model based on the single component adsorption data gives a more accurate prediction and description of the metal adsorptions from binary, ternary and quaternary solutions at different pHs and from complex water sources (i.e., tap water, runoff water, and seawater) that contained myriad humic substances, dissolved minerals, and various anions. preferentially adsorb cesium,14 while NH2-MCM-41 is shown to be selective for Ni2+ and Cd2+ adsorptions.15 Chromate and arsenate oxyanion species were adsorbed on mesoporous silicas containing mono-, di-, and triamino groups through the ionic interactions between the protonated aminos and the negatively charged oxyanions allowing their separation from mixtures containing metal cations.16 Copper(II) ions were separated from Cr(VI) species using COONa-MCM-4117 and magnetic MCM-41. 18 Fryxell et al. 19 employed metal complexes attached on mesoporous silicas to remove arsenate and chromate ions from drinking water. Ferrihydrite was also shown to selectively adsorb arsenite and arsenate.20 These clearly illustrate the possibility of tailoring the sorption properties of mesoporous adsorbent through the judicious choice of surface functional moieties. However, the lack of comprehensive data and suitable predictive model prevented a priori design for selectivity. Cobalt, nickel, copper, and zinc metals are widely used in constructions and manufacturing, and are ubiquitous in commercial products and household items. They are also important environmental pollution that pose a nontrivial health risk, and were reported to be responsible for a number of neurological diseases including Alzheimer’s disease, cancers and even acute toxic reactions.21 These metals do have an economic value, and therefore their removal and possible recovery are

1. INTRODUCTION Selective adsorption offers the possibility of treating metal pollutants in water in a way that allows their separation, removal, and recovery in purified form for recycle and reuse.1−6 A host of adsorbent materials including natural and synthetic, minerals and polymers as well as the staple carbons7−10 have been studied for the adsorptions of heavy metals, including Co2+, Ni2+, Cu2+ and Zn2+. Adsorbents often possess a complex surface populated by a multitude of functional groups of varying chemical properties, making it difficult to reliably predict adsorption behavior and selectivity. Furthermore, the physical form of the adsorbent dictates their accessibility and transport properties, while the water matrix itself can significantly alter metal adsorption.9 Mesoporous silica provides an ideal surface for investigating the role of surface functional group(s) on metal adsorption. MCM-41 with its large accessible surface is amenable to easy modification allowing the tailor-design of surface chemistry to achieve desired adsorption properties. Feng et al.11 achieved high Hg2+ adsorption selectivity by grafting thiopropyl groups on MCM-41, while Kang et al.12 populated the surface of SBA15 with imidazole moieties to obtain selective adsorbents for Pt2+ and Pd2+ ions. The work of Lam and co-workers13 showed that MCM-41 containing RNH2, R2NH, and R3N groups displayed excellent adsorption selectivity for Au3+ from binary solutions that also contained Cu2+ and Ni2+, while their later works2 reported the contrasting adsorption selectivity of SHMCM-41 and NH2-MCM-41 for Ag+ and Cu2+. Copper ferrocyanide immobilized on mesoporous silica was reported to © 2016 American Chemical Society

Received: April 5, 2016 Revised: August 4, 2016 Published: August 15, 2016 18365

DOI: 10.1021/acs.jpcc.6b03480 J. Phys. Chem. C 2016, 120, 18365−18376

Article

The Journal of Physical Chemistry C desirable. Selective adsorption is ideal, but the close similarity between these metals makes it a difficult challenge. The aim of this study is to investigate whether it is possible to predict the adsorptions of these metals, and establish a priori method for designing selective adsorbents. A mesoporous silica, MCM-41 serving as a blank canvas for this study was populated with surface aminopropyls (NH2-MCM-41) and used for the adsorptions of Co2+, Ni2+, Cu2+, and Zn2+ from single, binary, ternary and quaternary solutions. The applicability of the model to complex water systems was also examined for tap water, surface water, and seawater.

qe =

(Co − Ce)V m

(1)

where qe (mmol/g) is the adsorption capacity, Co (mM) and Ce (mM) are the initial and equilibrium metal concentrations, V (L) is the solution volume, and m (g) is the mass of adsorbent. The adsorption thermodynamics can be obtained from the single component adsorption data with the change of enthalpy (ΔH), Gibbs free energy (ΔG), and entropy (ΔS) calculated according to eqs 2−4.22 ΔH = R

2. EXPERIMENTAL METHODS 2.1. Mesoporous Adsorbents. Mesoporous MCM-41 was prepared from an alkaline synthesis solution13 containing tetraethyl orthosilicate (TEOS, 98%, Aldrich), cetyltrimethylammonium bromide (CTABr, 99.3%, Aldrich) and ammonium hydroxide (NH4OH, 28−30 wt %, Fisher Scientific). The synthesis mixture of 13.2 SiO2: 1 CTA2O: 584 NH4OH: 5546 H2O was allowed to react at 25 °C for 24 h. The recovered powder was calcined in air at 823 K for 24 h, after drying overnight in an oven at 373 K. Aminopropyls were attached on 2.5 g MCM-41 following 18 h reflux in 250 mL toluene containing 0.1 mol 3-aminopropyltrimethoxysilane (97%, Aldrich). NH2-MCM-41 was recovered by filtration and carefully rinsed with dry toluene, before drying overnight at 383 K. The NH2-MCM-41 and MCM-41 were characterized by Xray diffraction (XRD, Philips 1830), N2 physisorption (Coulter SA 3100) and electron microscopy (HRTEM, JEOL JEM 2010 and SEM, JEOL 6300F) to ascertain the phase structure, crystallinity and textural properties of the materials as well as their particle size and morphology. X-ray photoelectron spectroscopy (XPS, Physical Electronics PHI 5000) and Fourier transform infrared spectroscopy (FTIR, PerkinElmer GX 2000) were used to probe the chemistry of the adsorbent surface before and after modification and adsorption. The number of attached aminopropy groups was determined from thermogravimetric analysis (TGA/DTA, Setaram 31/1190), wherein 20 mg of NH2-MCM-41 was heated in air from 25 to 800 °C at a rate of 5 °C·min−1. The surface zeta-potential of NH2-MCM-41 was measured using Coulter Beckman Delsa 4400SX. 2.2. Adsorption Experiments. The adsorption study was conducted using 50 mg adsorbent in 50 mL metal ion solutions of cobalt nitrate (98%, Aldrich), nickel nitrate hexahydrate (99%, Aldrich), copper nitrate (99%, Nacalai Tesque) and zinc nitrate hexahydrate (95%, Nacalai Tesque). Equimolar metal concentrations were used for the multicomponent adsorption experiments at solution pH of 4.5 to 5.2. The adsorption was carried out in a constant temperature orbital shaker that allows the temperature to be adjusted. The initial and final metal concentrations were measured by inductively coupled plasma− atomic emission spectrometry (ICP-AES, PerkinElmer Optima 3000XL) or inductively coupled plasma−mass spectrometry (ICP-MS, Varian 820). Calibrations were made against standard solutions prepared from 1000 ppm of Co in 2% HNO3, 1000 ppm of Ni in 2% HNO3, 1000 ppm of Cu in 2% HNO3 and 1000 ppm of Cu in 2% HNO3 purchased from High-Purity Standards. Triplicate measurements were made, and samples were remeasured when the RSD was higher than 3%. The equilibrium adsorption capacity was calculated from eq 1.

d(ln Ke) d(1/T )

(2)

ΔG = −n f RT

(3)

T ΔS = ΔH − ΔG

(4)

where Ke is the adsorption equilibrium constant, nf is the model parameter from Freundlich equation, R (8.314 J mol−1 K−1) is the gas constant and T (K) is the adsorption temperature. The ΔH was calculated from the adsorption isotherms from the van’t Hoff equation (eq 2), while ΔG was calculated from eq 3 once the Freundlich parameter, nf is known. Separate experiments were performed to investigate the effects of pH on metal adsorption on NH2-MCM-41. The pH was adjusted with the addition of 0.1 M HNO3 or 0.1 M NH4OH. Besides distilled deionized water, three other water sources were used for the multicomponent adsorption experiments, namely tap water, runoff water and seawater from locations within the campus of the Hong Kong University of Science and Technology. 2.3. Adsorption Models. Langmuir and Freundlich adsorption equations were used to fit the single component adsorption data of Co2+, Ni2+, Cu2+, and Zn2+ on NH2-MCM41. Langmuir adsorption (eq 5) best describes adsorption on homogeneous surface, while Freundlich adsorption (eq 6) takes into account that most surfaces are heterogeneous with adsorption sites of varying energetics.23 The model parameters were determined by minimizing the sum of square error, SSE (eq 7). qcal,L =

kLCe (1 + bLCe)

(5)

qcal,F = KFCenf

(6)

SSE = (qcal − qe)2

(7)

where qcal (mmol/g) is the calculated adsorption capacity, qexp is the experimental value, Ce (mmol/L) is the equilibrium metal concentration, while KL (mmol·g −1 ), bL (L/mmol), K F (mmol1−nf·g−1Lnf), and nf are the model parameters. The Ideal Adsorbed Solution Theory (IAST), LeVan and Vermeulen Model, and Multicomponent Freundlich were used to predict the binary, tertiary, and quaternary adsorption of Co2+, Ni2+, Cu2+, and Zn2+ on NH2-MCM-41. Crittenden et al.24 derived the IAST equation in eq 8 based on a Freundlich adsorption model. ⎛ ∑N n q ⎞ni ⎜ j = 1 j e, i ⎟ Ce, i = N ∑ j = 1 qe, j ⎜⎝ niK i ⎟⎠ qe, i

for species i = 1−N (8)

18366


Article

The Journal of Physical Chemistry C

Figure 1. (a) Scanning electron and (b) transmission electron micrographs of mesoporous silicas along with the (c) X-ray diffraction and (d) nitrogen physisorption isotherms of the MCM-41 and NH2-MCM-41 powders.

where Ki and ni are the model parameters from the Freundlich adsorption equation. The LeVan and Vermeulen equation25 for binary adsorption shown in eq 9 is a simplification of IAST model.

qe,1

1/ n1

( ) C = ⎡ ⎢( ) C + ( ) ⎣ K1 n1

n

e,1

1/ n1

K1 n1

The multicomponent Freundlich equation in eq 10 was introduced by Sheindorf et al.26 to describe competitive adsorption. The model includes a competition coefficient (aij) that accounts for the differing affinities of adsorbates for the adsorbent surface. The competition coefficient of the metal species i and j is a ij , and can be determined from thermodynamic data or from the binary adsorption data.

1/ n2

K2 n2

e,1

⎤1 − n̅ Ce,2 ⎥ ⎦

+ ΔF2

n

qe, i = K iCe, i(∑ aijCe , i)ni − 1

(9a)

j=1

n

qe,2 =

e,2

1/ n1

K1 n1

1/ n2

K2 n2

e,1

⎤1 − n̅ Ce,2 ⎥ ⎦

3. RESULTS AND DISCUSSION 3.1. Adsorbent Characterization. The free-flowing powders of MCM-41 and NH2-MCM-41 were made of 0.8 μm particles as shown in the scanning electron micrograph in Figure 1a. Both MCM-41 and NH2-MCM-41 prepared in this study display highly ordered pores with the characteristic p6 mm hexagonal symmetry as shown in Figure 1b. This is confirmed by the X-ray diffraction in Figure 1c. Populating the surface with 2.3 mmol·g−1 aminopropyls (Supporting Information (SI) Figure S1) caused a concomitant increase in disorder as observed in the decrease in the intensity of ⟨110⟩ and ⟨200⟩ peaks. The surface modification according to FTIR study (SI Figure S2) also resulted in a decrease in surface silanols (i.e., 3745 cm−1) and the appearance of the characteristic signals at 3288 and 3360 cm−1 belonging to −NH2 stretching as well as the bands at 2920 and 2850 cm−1 from the asymmetric and symmetric stretching of n-propyl chains. The slight shift in the ⟨100⟩ peak of NH2-MCM-41 in Figure 1b toward higher θ indicates a narrowing of the pores following aminopropyls attachment. The nitrogen physisorption isotherms of MCM-41 and NH2-MCM-41 in Figure 1d displayed a Type IV isotherm characteristic of mesoporous materials.22 The prepared MCM41 had a specific surface area of 1080 m2g−1, while the denser NH2-MCM-41 measured 750 m2g−1. 3.2. Single Component Adsorption. The single component adsorptions of the metal ions by NH2-MCM-41

+ ΔF2 (9b)

where

n̅ =

1/ n1

( ) ( )

n1

K1 n1

K1 n1

1/ n1

1/ n2

( ) C +( ) C K2 n2

Ce,1 + n2 Ce,1

1/ n1

K2 n2

e,2

1/ n2

e,2

1/ n2

( ) C( ) ΔF = ⎡ ⎢( ) C + ( ) ⎣ K1 n1

2

K1 n1

1/ n1

e,1

(10)

1/ n2

( ) C ⎡ ⎢( ) C + ( ) ⎣ K2 n2

K2 e ,1 n 2

K2 n2

⎡⎛ ⎞1/ n1 K × ln⎢⎜ 1 ⎟ Ce,1 + ⎢⎣⎝ n1 ⎠

1/ n2

(9c)

Ce,2 ⎤2 − n ̅ Ce,2 ⎥ ⎦

⎛ K 2 ⎞1/ n2 ⎤ ⎜ ⎟ Ce,2 ⎥ ⎥⎦ ⎝ n2 ⎠

(9d)

qe,1 and qe,2 (mmol/g) are the equilibrium adsorption capacity of the adsorbent for metals 1 and 2, Ce,1 and Ce,2 (mM) are the equilibrium concentrations of metals 1 and 2, while K1 (mmol1−n1·g−1Ln1), K2 (mmol1−n2·g−1Ln2), and n1 and n2 are the Freundlich constants from the single component adsorptions of metals 1 and 2, respectively. 18367


Article


Figure 2. Single component adsorption of (a) Co2+, (b) Ni2+, (c) Cu2+ and (d) Zn2+ by NH2-MCM-41 at room temperature. Note: symbols − experimental data; dotted line − Langmuir model calculation; solid line − Freundlich model calculation.

Table 1. Summary of Single Component Adsorption of Metal Ions on NH2-MCM-41 Langmuir model Co2+ Ni2+ Cu2+ Zn2+ a

qe,maxa (mmol/g)

KL (mmol/g)

bL (L.mmol−1)

0.64 0.70 0.76 0.73

78.8 279 304 525

130 430 431 782

Freundlich model SSE 1.1 1.5 7.9 8.0

× × × ×

10−2 10−2 10−3 10−3

KF (mmol1−nf·g−1Lnf)

nf

0.630 0.680 0.747 0.721

0.105 0.114 0.094 0.102

SSE 8.1 4.8 4.1 7.9

× × × ×

10−3 10−3 10−3 10−3

The maximum adsorption capacities are the average values of last three points in single component isotherm.

be approximated within the definition of a “Langmuir adsorbate” for which the configurational term for entropy of the surface is assumed to be zero. The standard state of the adsorbate is taken as θA° = 0.5, and the standard state of the ion is taken as [M2+] = 1 mM. It can be seen that adsorption was spontaneous, and the calculated Gibbs free energy indicated that Co2+ > Ni2+ > Zn2+ > Cu2+ with Cu2+ adsorption being most favorable among the four divalent metal cations. The adsorption of the metal ions on NH 2 -MCM-41 was endothermic, and this could be related to the energies required for the dehydration of the solvation sheath around the metal ions prior to the adsorption. Amzel30 observed that loss of a water molecule from the solvation sheath corresponds roughly to ΔS = 40 J/mol·K and the large ΔS observed in this study is consistent with the dehydration of the solvation sheath following adsorption. Adsorption of the metal ions also disrupted and introduced disorder to the surface aminopropyl groups on the surface, altering their motional freedom and energetics as reflected by the positive ΔS. Table S2 summarizes the ΔH and ΔS values reported in the literature. ΔH values of Cu2+ and Zn2+ are larger than that on Co2+ and Ni2+, which is the same trend in our work. ΔH values of Cu2+ is about 0.17 kJ/mol·K is comparable with that reported for SBA-15 (ΔH = 0.16 kJ/mol·K in Table S2), both being of similar materials.

are summarized in Figure 2 along with the Langmuir and Freundlich model calculations. The amount of metals adsorbed on unmodified MCM-41 (Table S1) was negligible in comparison. The adsorption capacities of NH2-MCM-41 for Co2+, Ni2+, Cu2+ and Zn2+ are 0.64, 0.70, 0.76, and 0.73 mmol.g−1, respectively. It is clear from the figure that Freundlich model provides a better description of the adsorption data (cf. Table 1). This implies that the adsorbent is far from homogeneous despite having aminopropyls as the main adsorption site. Indeed, interactions of aminopropyls with their neighbors and the unreacted hydroxyl groups on MCM41 can change their adsorption energetics, while the adsorption of charged metal ions perturbs the neighboring sites as a results of charge repulsion. Also, the adsorption of the divalent metal ions requires cooperation between neighboring aminopropyls making their spatial distribution an important factor in metal adsorption as shown in a prior study.27 The adsorption thermodynamics was measured from the adsorption isotherm data in Figure 3, and the calculated adsorption enthalpy, Gibbs free energy and entropy28 are listed in Table 2. Savara et al.29 proposed a possible framework for comparing thermodynamic data for the adsorption of “molecules” on “solids”. Although some of the underlying assumption may not be completely valid for ionic species that have more specific and energetic interactions, our system can 18368


Article


Figure 3. Adsorption isotherms of (a) Co2+, (b) Ni2+, (c) Cu2+, and (d) Zn2+ on NH2-MCM-41. Note: symbols − experimental data; solid line − Freundlich model calculation.

Table 2. Thermodynamic Properties of Metal Ion Adsorption on NH2-MCM-41 ΔG(kJ/mol)

ΔS (kJ/mol.K)

adsorbate

θA

ΔH(kJ/mol)

4 °C

10 °C

23 °C

4 °C

10 °C

23 °C

Co2+

0.2 0.5 0.8 0.2 0.5 0.8 0.2 0.5 0.8 0.2 0.5 0.8

29.1 23.2 13.1 33.9 25.8 14.9 43.1 33.1 19.9 36.6 26.9 14.3

−8.23 −8.23 −8.23 −9.23 −9.23 −9.23 −15.6 −15.6 −15.6 −14.9 −14.9 −14.9

−11.7 −11.7 −11.7 −13.5 −13.5 −13.5 −21.5 −21.5 −21.5 −21.9 −21.9 −21.9

−21.7 −21.7 −21.7 −22.7 −22.7 −22.7 −30.9 −30.9 −30.9 −30.2 −30.2 −30.2

0.134 0.113 0.076 0.155 0.125 0.087 0.211 0.175 0.127 0.185 0.151 0.104

0.144 0.123 0.087 0.167 0.138 0.100 0.228 0.192 0.146 0.206 0.172 0.127

0.168 0.148 0.114 0.187 0.160 0.124 0.244 0.211 0.168 0.220 0.189 0.147

Ni2+

Cu2+ Zn2+

decreases from Cu2+ > Zn2+ > Co2+ > Ni2+, while the calculated R values follows a similar trend with Cu2+ > Zn2+ > Co2+ ∼ Ni2+. This implies that Cu2+ had a stronger interaction with the amino groups on NH2-MCM-41 compared to the other metal cations, and the observed trend is consistent with the adsorption thermodynamics data in Table 2. The XPS results are also consistent with the Pearson hard− soft, acid−base (HSAB) principle, which states that the interactions between Lewis acids and bases of similar “hardness” are faster and stronger. The HSAB principle has proven particularly useful in explaining the stability, mechanisms, and pathways of chemical interactions between metal ions and organic ligands.32 The aminopropyls of NH2-MCM-41 are Lewis base of intermediate hardness, and were selected for their good affinity for Co2+, Ni2+, Cu2+, and Zn2+ that are Lewis acids of moderate hardness. XPS is able to provide a more refined measure of this interaction (cf. Figure 4 and Table 3),

Figure 4 shows the binding energy of nitrogen (N 1s) in NH2-MCM-41 before and after adsorption of the respective metal cations. Samples with comparable adsorbed amounts of metal cations were dried in vacuum and analyzed by X-ray photoelectron spectroscopy (XPS). The nitrogen in aminopropyls had N 1s binding energy of 399.5 ± 0.5 eV28 and the adsorption of metals caused a shift to higher energies (i.e., 400−402 eV). The formation of dative bonds between the metal ions and the surface amino groups involved the free electron pair in the nitrogen atom resulting in a positive polarization that led to the higher N 1s binding energy.15 It is possible to deconvolute N 1s into two main component peaks centered at 400 eV (Peak I) from the amino group and at 402 eV (Peak II) belonging to a charged ammonium (i.e., −NH3+ or NH2···M2+).31 The binding energy of Peak II and its relative ratio to Peak I (R) were observed to change with the adsorbed metal ions as shown in Table 3. The Peak II binding energies 18369


Article


The pH does have an effect on NH2-MCM-41 and on the metal adsorption as shown in Figure 5. Figure 5a plots the surface zeta potential of NH2-MCM-41 at different pH. It can be seen from the figure that, at low pH, the surface is positively charged due to the protonation of the aminopropyls on NH2MCM-41. This repels the metal cations and prevents their adsorption at pH below 2.5 (Figure 5b). Above pH 2.5, the NH2-MCM-41 surface becomes progressively less positively charged, and metal cation adsorption increases, reaching a maximum capacity at pH 4.0 to 5.5. Precipitation generally occurred above pH 5.5. Figure 5b shows that all four metal cations, Co2+, Ni2+, Cu2+ and Zn2+, follow this general trend. 3.3. Multicomponent Adsorptions. Binary Component Adsorptions. Binary component adsorptions of Co2+, Ni2+, Cu2+ and Zn2+ on NH2-MCM-41 are plotted in Figure 6, and the adsorption capacities are summarized in Table 4. The overall adsorption capacity of NH2-MCM-41 for the metal ions were on the average between 0.7−0.9 mmol/g, which is comparable to its single component adsorption capacity (Table 1). Figure 6a shows that NH2-MCM-41 adsorbs both Co2+ and Ni2+ to about similar extent from the Co2+-Ni2+ binary solution. It is clear from the data that Cu2+ is preferentially adsorbed from all binary solutions that contained the metal ion as shown in Figure 6b for Cu2+−Co2+, Figure 6d for Cu2+−Ni2+, and Figure 6f for Cu2+−Zn2+, while Zn2+ is mainly adsorbed from Zn2+−Co2+ (Figure 6c) and Zn2+−Ni2+ solutions (Figure 6e). A prediction of the adsorption behavior of binary solutions was made from the single component adsorption data (Figure 2). Using the single component Freundlich isotherm data in Table 1, three multicomponent adsorption models including IAST, LeVan and Vermeulen, and multicomponent Freundlich were evaluated, and the calculated sums of square errors (SSE) for each model are summarized in Table 4. The data did not fit the IAST model very well (Table 4) and this may be due to the underlying assumptions that the spreading pressures of the mixture and pure component are the same.24 The LeVan and Vermeulen equation is adequate in predicting the adsorption capacities and selectivities at high concentrations. However, it failed to predict metal adsorption at low concentrations resulting in the relatively large calculated SSE in Table 4. The IAST model and LeVan and Vermeulen equation ignored competitions between the adsorbate species, while the multicomponent Freundlich equation specifically accounts for this using the competition coefficient in eq 10. The calculated competition coefficients are listed in Table 5 with Cu2+ being the highest and Ni2+ the lowest. The model lines in the binary

Figure 4. N 1s X-ray photoelectron spectra of NH2-MCM-41 before and after metal ion adsorption.

Table 3. XPS N 1s Binding Energies of Fresh and Spent NH2-MCM-41 peak area % sample NH2-MCM41 Co-NH2MCM-41 Ni-NH2MCM-41 Cu-NH2MCM-41 Zn-NH2MCM-41

peak I (−NH3+)

peak II (−NH2)

100.0

0.0

85.2

binding energy (eV) R

(NH3+/ NH2)

peak I (−NH3+)

peak II (−NH2)

0

399.78

N/A

14.8

0.17

399.86

401.75

86.5

13.5

0.16

399.85

401.72

78.7

21.3

0.27

399.89

401.88

82.3

17.7

0.22

399.88

401.78

and offers a convenient tool for the design of selective adsorbents.

Figure 5. (a) Surface zeta potential of NH2-MCM-41 at different pH and (b) the corresponding single component adsorption of Co2+ (◊), Ni2+ (×), Cu2+ (△), and Zn2+ (□) at room temperature. 18370


Article


Figure 6. Binary component adsorptions of Co2+, Ni2+, Cu2+, and Zn2+ on NH2-MCM-41 (◊ Co2+; × Ni2+; △ Cu2+; □ Zn2+) at room temperature. Note: symbols − experimental data; lines − multicomponent Freundlich model calculation.

Table 4. Binary Metal Adsorptions on NH2-MCM-41 and Model Comparison equilibrium concentration (mM) metal 1 metal 2 metal 1 Co2+ Co2+ Co2+ Ni2+ Ni2+ Cu2+

Ni2+ Cu2+ Zn2+ Cu2+ Zn2+ Zn2+

2.57 3.21 3.14 3.07 3.17 2.20

adsorption capacity (mmol/g)

sum of square error

metal 2 metal 1 metal 2 2.68 2.28 2.32 2.12 2.40 2.72

0.46 0.02 0.00 0.00 0.00 0.76

0.30 0.78 0.73 0.92 0.70 0.14

total adsorption capacity (mmol/g)

IAST model

LeVan and Vermeulen model

multicomponent Freundlich model

0.76 0.80 0.73 0.92 0.70 0.90

0.11 0.07 0.12 0.07 0.29 0.22

0.77 0.10 0.50 0.13 0.13 0.18

0.02 0.02 0.05 0.02 0.03 0.03

Table 5. Competition Coefficients, aij for Multicomponent Freundlich Equations

component adsorption plots in Figure 6 were calculated from the multicomponent Freundlich eq (eq 10) using the single component adsorption data (cf. Figure 2, Table 1). It is clear that the model calculation is in excellent agreement with the experimental data, and can predict adsorption behavior over the entire range of concentrations. Furthermore, the same model can be extended to predict the effect of pH on the binary component adsorption as shown in Figure 7. The model assumes that aminopropyls are the primary adsorption site for the metal cations, and the hydronium ions (H3O+) compete

metal 2 Co2+ Metal 1

2+

Co Ni2+ Cu2+ Zn2+

Ni2+

Cu2+

Zn2+

2.29 0.006 0.007

0.437 0.001 0.003

175 1000 11.2

140 400 0.089 -

18371


Article


Figure 7. Effects of solution pH on the binary component adsorptions of Co2+, Ni2+, Cu2+ and Zn2+ by NH2-MCM-41 at room temperature (◊ Co2+; × Ni2+; △ Cu2+; □ Zn2+). Note: symbols − experimental data; lines − multicomponent Freundlich model calculation.

Figure 8. Ternary component adsorptions of Co2+, Ni2+, Cu2+, and Zn2+ on NH2-MCM-41 (◊ Co2+; × Ni2+; △ Cu2+; □ Zn2+) at room temperature. Note: symbols − experimental data; lines − multicomponent Freundlich model calculation.

18372


Article


Figure 9. Effect of solution pH on the ternary component adsorptions of Co2+, Ni2+, Cu2+, and Zn2+ by NH2-MCM-41 at room temperature (◊ Co2+; × Ni2+; △ Cu2+; □ Zn2+). Note: symbols − experimental data; lines − multicomponent Freundlich model calculation.

Figure 10. Quaternary component adsorptions of Co2+, Ni2+, Cu2+, and Zn2+ on NH2-MCM-41 (◊ Co2+; × Ni2+; △ Cu2+; □ Zn2+) as a function of (a) Ce and (b) pH at room temperature. Note: symbols − experimental data; lines − multicomponent Freundlich model calculation.

with their adsorptions. The H3O+ adsorption can be estimated from the Zeta potential measurements in Figure 5 and used to calculate the competition coefficients. It is clear from Figure 7 that the model calculations are in close agreement with the experimental data. Ternary and Quaternary Component Adsorptions. Ternary and quaternary component adsorptions were carried out on solutions containing equimolar concentrations of metal cations. The multicomponent Freundlich adsorption model perfectly predicted the adsorption behavior of the ternary solutions as shown in Figure 8. In all solutions containing Cu2+ as shown in Figure 8a,b,d, Cu2+ was preferentially adsorbed by NH2-MCM-41. Figure 8b plots the adsorption from Cu2+, Co2+, and Ni2+ solution where the selectivity for Cu2+ ranged from 60% at low metal concentrations (i.e., Ce ≤ 0.5 mM) to better than 99% at Ce above 1 mM when NH2-MCM-41 is fully saturated. Zinc(II) ions were also adsorbed to a lesser extent from Cu2+, Zn2+, and Ni2+ and Cu2+, Zn2+, and Co2+ solutions as shown in Figure 8a and 8d. The adsorbent’s selectivity for Cu2+ was between 40 to 85% and that of Zn2+ was lower at 15 to 65% for these solutions. Similarly, Figure 8c shows that Zn2+ was adsorbed to a greater extent than both Co2+ and Ni2+ by NH2-MCM-41.

Figure 9 plots the ternary metal adsorption from solutions of different pH. It is clear from the plots that there is a close agreement between the experimental data and model calculations. The results suggest that high selectivity for Cu2+ adsorption can be attained by properly adjusting the solution pH. Indeed, it is possible to adsorb only Cu2+ from all four ternary solutions such that Cu2+ can be recovered at high purity for reuse. Lowering the pH to 2.5 changes the surface charge from negative to positive with the protonation of the aminopropyls that alters the “hardness” of the Lewis base and therefore their affinity for metal adsorption. The quaternary metal adsorption results are summarized in Figure 10. The model calculations from the multicomponent Freundlich equation using single-component adsorption data still provide a good description of the ternary adsorption on NH2-MCM-41, but tends to overestimate the adsorption of Cu2+. This may be due to a greater complexity of interactions between the ions and an increase in the ionic strength of the solution. The plots show that Cu2+ and Zn2+ were adsorbed exclusive of Co2+ and Ni2+ over the entire range of solution concentrations and pH studied. The Cu2+ adsorption selectivity of NH2-MCM-41 for equimolar quaternary solution ranges from 60 to 99%. 18373


Article

The Journal of Physical Chemistry C Table 6. Metal Adsorption from Different Water Systemsa percent removal (%) deionized distilled H2O tap water run-off water seawater a

pH

Ca2+ (ppm)

TOC (ppm)

COD (ppm)

Co2+

Ni2+

Cu2+

Zn2+

6.0 6.7 7.5 8.0

0 32 53 358

< 0.1 0.2 2.5 22

< 0.1 < 0.1 1.0 n.a.

75.5 (64) 70.6 68.1 54.7

52.0 (51) 35.2 26.0 13.2

99.8 (100) 99.8 99.6 99.0

98.4 (96) 98.5 98.2 98.1

Note: numbers in parentheses are the calculated percent removal of metal ions according to multicomponent Freundlich equation.

expected to interfere with the adsorptions of Co2+, Ni2+, Cu2+, and Zn2+. High partition coefficients KCu and KZn of 1.0 × 107 and 1.2 × 106 were maintained, indicating excellent specificity for Cu2+ and Zn2+ adsorptions. The humic substances on the other hand are mostly “soft” Lewis acids and are likely to form complexes with Ni2+ and Co2+ stabilizing these metal ions in the solution. This could explain the lower adsorption of these metal ions with increasing TOC content of the water. The percent removal of metal ions from the solutions was also calculated using the multicomponent Freundlich equation and listed in Table 6. There is good agreement between the calculated values and the experimental measurements illustrating the usefulness of the model calculation in predicting adsorption behavior from complex water systems.

The multicomponent adsorption experiments clearly show that adsorption selectivity of NH2-MCM-41 follows the IrvingWilliam order seen in transition metal complexes (i.e., Co2+ > Ni2+ < Cu2+ > Zn2+) that is often attributed to the ligand field effects, but can be very well explained by Pearson HSAB principle.32 Ni2+ (η = 8.50 eV) and Zn2+ (η = 10.80 eV)33 are observed to have greater hardness (η) and therefore interact more strongly with the solvating water molecules, making them more stable as dissolved ions,15,32 while the similarity in the hardness of Cu2+ (η = 8.27 eV) and aminopropyl resulted in the preferred adsorption of Cu2+. The observed adsorption selectivity is also consistent with the results of the adsorption thermodynamics study and XPS measurements. Indeed, model calculations based on the single-component adsorption thermodynamics data can very well predict the multicomponent adsorption behavior of the NH 2 -MCM-41 adsorbent. 3.4. Water Matrix. Studies have shown that water matrix can have a dramatic effect on adsorption.34 Water is complex and contains dissolved minerals (e.g., Na+, Ca2+, Mg2+, K+), various anions (i.e., Cl−, Br− HCO3−, NO32−, SO42−), and myriad humic substances. A full water analysis is not possible and Table 6 lists only the main properties (i.e., pH, [Ca2+], TOC and COD) of the raw waters used in the study. The waters were collected from various sources within the university campus and were filtered through a 0.45 μm Millipore filter to remove suspended particulates and microorganisms. The water from the laboratory tap had a pH of 6.4, a residual Cl2 of 0.4 mg/L, a hardness of 80 mg·L−1 and a total organic content (TOC) of 0.2 ppmC. The runoff water was collected from a hillside stream and the water had a pH of 7.5, a hardness of 185 mg·L−1 and a TOC of 2.5 ppmC. The seawater was collected from a depth of 1.5 m from the campus pier. Analysis showed that the seawater had a pH of 8.0, a TOC of 22 ppmC and contained 9643 ppm of Na+, 526 ppm K+, 358 ppm of Ca2+, 1040 ppm Mg2+, 15,413 ppm of Cl− and 2017 SO42− ppm. The tap water, surface runoff water and seawater were contaminated with select metal pollutants, specifically 1.0 ppm of Co2+, 1.0 ppm of Ni2+, 1.0 ppm of Cu2+ and 1.0 ppm Zn2+. The prepared water solutions were allowed to equilibrate and any precipitates that are formed were removed by centrifugation before measuring the dissolved metal concentration using ICP-MS. Adsorption was carried out without pH adjustment at room temperature over 24 h. The multicomponent adsorption from the different water matrix is summarized in Table 6. It is clear that NH 2 -MCM-41 remained selective for Cu 2+ adsorption followed by Zn2+, Co2+, and Ni2+, a trend similar to that of deionized distilled water. This can be explained by the fact that most of the minerals (i.e., Na+, Ca2+, Mg2+) found in these waters are “hard” Lewis acids, while the anions (i.e., Cl−, SO42−) are “hard” Lewis bases that, according to the Pearson HSAB principle, have poor interactions with the aminopropyls, a Lewis base of moderate “hardness” and are therefore not

4. CONCLUSIONS This work used mesoporous MCM-41 to create selective adsorbent by attaching surface aminopropyl moieties on its large accessible surface. Adsorption selectivity was investigated for Co2+, Ni2+, Cu2+, and Zn2+ that have very similar chemistry and properties. The results of the adsorption study suggests that Pearson HSAB principle can serve as a rational framework for the heuristic design of a selective adsorbent. XPS provides a fast and convenient method for rapid screening of candidate surface ligands for a given target metal by measuring the strength of their interactions. Adsorption thermodynamics data provides a more accurate measure of these interactions and can form the basis for an accurate model for the adsorption. Indeed, a multicomponent Freundlich model that takes into account competitions gave good prediction of binary, ternary, and quaternary adsorptions of Co2+, Ni2+, Cu2+, and Zn2+ on NH2MCM-41.

■

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.jpcc.6b03480. Additional experimental resutls including TGA and FTIR analysis; Additional single component adsorption capacity on MCM-41 and NH2-MCM-41; Summary of the ΔH and ΔS values reported in the literature; and the method for calculating multicomponent Freundlich model at different pH (PDF)

■

AUTHOR INFORMATION

Corresponding Author

*Tel: +852 2358 7123; Fax: +852 2358 0054; e-mail: [email protected] (K.L.Y.). Notes

The authors declare no competing financial interest. 18374


Article


■

(11) Feng, X.; Fryxell, G. E.; Wang, L.-Q.; Kim, A. Y.; Liu, J.; Kemner, K. M. Functionalized Monolayers on Ordered Mesoporous Supports. Science 1997, 276, 923−926. (12) Kang, T.; Park, Y.; Choi, K.; Lee, J. S.; Yi, L. Ordered Mesoporous Silica (SBA-15) Derivatized With Imidazole-Containing Functionalities as a Selective Adsorbent of Precious Metal Ions. J. Mater. Chem. 2004, 14, 1043−1049. (13) Lam, K. F.; Yeung, K. L.; McKay, G. An Investigation of Gold Adsorption from Binary Mixture using Selective Mesoporous Silica Adsorbents. J. Phys. Chem. B 2006, 110, 2187−2194. (14) Lin, Y.; Fryxell, G. E.; Wu, H.; Engelhard, M. Selective Sorption of Cesium Using Self-Assembled Monolayers on Mesoporous Supports. Environ. Sci. Technol. 2001, 35 (19), 3962−3966. (15) Lam, K. F.; Yeung, K. L.; McKay, G. A New Approach for Cd2+ and Ni2+ Removal and Recovery using Mesoporous Adsorbent with Tunable Selectivity. Environ. Sci. Technol. 2007, 41, 3329−3334. (16) Yoshitake, H.; Yokoi, T.; Tatsumi, T. Adsorption Behaviour of Arsenate at Transition Metal Cations Captured by Amino-Functionalized Mesoporous Silicas. Chem. Mater. 2003, 15, 1713−1721. (17) Lam, K. F.; Yeung, K. L.; McKay, G. Selective Mesoporous Adsorbents for Cr2O72‑ and Cu2+ Separation. Microporous Mesoporous Mater. 2007, 100, 191−201. (18) Chen, X.; Lam, K. F.; Zhang, Q.; Pan, B.; Arruebo, M.; Yeung, K. L. Synthesis of Highly Selective Magnetic Mesoporous Adsorbent. J. Phys. Chem. C 2009, 113, 9804−9813. (19) Fryxell, G. E.; Liu, J.; Hauser, T. A.; Nie, Z.; Ferris, K. F.; Mattigod, S.; Gong, M.; Hallen, R. T. Design and Synthesis of Selective Mesoporous Anion Traps. Chem. Mater. 1999, 11, 2148− 2154. (20) Couture, R. M.; Rose, J.; Kumar, N.; Mitchell, K.; Wallschlager, D.; Van Cappellen, P. Sorption of Arsenite, Arsenate and Thioarsenates to Iron Oxides and Iron Sulfides: A Kinetic and Spectroscopic Investigation. Environ. Sci. Technol. 2013, 47, 5652− 5659. (21) Quinn, J. F.; Crane, S.; Harris, C.; Wadsworth, T. L. Copper in Alzheimer’s Disease: Too Much or Too Little? Expert Rev. Neurother. 2009, 9, 631−637. (22) Pan, B. C.; Zhang, Q. X.; Meng, F. W.; Li, X. T.; Zhang, X.; Zheng, J. Z.; Zhang, W. M.; Pan, B. J.; Chen, J. L. Sorption Enhancement of Aromatic Sulfonates onto An Aminated Hyper-crosslinked Polymer. Environ. Sci. Technol. 2005, 39, 3308−3313. (23) Do, D. D. Adsorption Analysis: Equilibria and Kinetics; Imperial College Press: London, 1998. (24) McGinley, P. M.; Katz, L. E.; Weber, W. J., Jr. A Distributed Reactivity Model for Sorption by Soils and Sediments. 2. Multicomponent Systems and Competitive Effects. Environ. Sci. Technol. 1993, 27 (8), 1524−1531. (25) LeVan, M. D.; Vermeulen, T. Binary Langmuir and Freundlich Isotherms for Ideal Adsorbed Solutions. J. Phys. Chem. 1981, 85, 3247−3250. (26) Sheindorf, C.; Rebhun, M.; Sheintuch, M. A freundlich-type Multicomponent Isotherm. J. Colloid Interface Sci. 1981, 79 (1), 136− 142. (27) Mehandzhiyski, A. Y.; Riccardi, E.; van Erp, T. S.; Trinh, T. T.; Grimes, B. A. Ab Initio Molecular Dynamics Study on The Interactions Between Carboxylate Ions and Metal Ions in Water. J. Phys. Chem. B 2015, 119, 10710−10719. (28) Zhang, Q.; Teng, J.; Zou, G.; Peng, Q.; Du, Q.; Jiao, T.; Xiang, J. Efficient Phosphate Sequestration for Water Purification by Unique Sandwich-like MXene/Magnetic Iron Oxide Nanocomposites. Nanoscale 2016, 8, 7085−7093. (29) Savara, A. Standard States for Adsorption on Solid Surfaces: 2D Gases, Surface Liquids, and Langmuir adsorbates. J. Phys. Chem. C 2013, 117, 15710−15715. (30) Amzel, L. M. Loss of Translational Entropy in Binding, Folding, and Catalysis. Proteins: Struct., Funct., Genet. 1997, 28, 144−149. (31) Hooper, A. E.; Werho, D.; Hopson, T.; Palmer, O. Evaluation of Amine- and Amide-terminated Self-assembled Monolayers as ‘Molec-

ACKNOWLEDGMENTS The authors gratefully acknowledge the financial support of the Hong Kong Research Grant Council (605009), the Environment and Conservation Fund (ECWW11EG02) of the HKSAR Government and NSFC (No. 21507141). We also thank the Materials Characterization and Preparation Facility (MCPF) and the Advanced Engineering Material Facility (AEMF) of the Hong Kong University of Science and Technology for the use of their equipment.

■

NOMENCLATURE Co mM, initial metal concentration Ce mM, equilibrium metal concentration V L, volume of solution m g, the mass of adsorbent qe mmol/g, the adsorption capacity ΔH kJ/mol, enthalpy of adsroption ΔG kJ/mol, Gibbs free energy ΔS kJ/mol.K, entropy of adsorption KL and bL mmol/g and L/mmol, Langmuir constants KF and nf mmol1−nf·g−1Lnf, Freundlich model parameters Ki and ni mmol1−ni·g−1L ni, model parameters from the Freundlich equations T K, temperature aij competition coefficients from multicomponent Freundlich model equaitons θA relative adsorbate coverage SSE sum of squares for error

■

REFERENCES

(1) Dabrowski, A.; Hubicki, Z.; Podkoscielny, P.; Robens, E. Seletive Removal of the Heavy Metals Ions from Waters and Industrial Wastewater by Ion-Exchange Method. Chemosphere 2004, 56, 91−106. (2) Lam, K. F.; Chen, X. Q.; Fong, C. M.; Yeung, K. L. Selective Mesoporous Adsorbents for Ag+/Cu2+ Separation. Chem. Commun. 2008, 2034−2036. (3) Gupta, V. K.; Carrott, P. J. M.; Ribeiro Carrott, M. M. L.; Suhas. Low-cost Adsorbents: Growing Approach to Wastewater Treatment: Review. Crit. Rev. Environ. Sci. Technol. 2009, 39, 783−842. (4) Wan Ngah, W. S.; Teong, L. C.; Hanafiah, M. A. K. M. Adsorption of Dyes and Heavy Metal Ions by Chitosan Composites: A Review. Carbohydr. Polym. 2011, 83, 1446−1456. (5) Chen, X. Q.; Lam, K. F.; Mak, S. F.; Yeung, K. L. Precious Metal Recovery by Selective Adsorption using Biosorbents. J. Hazard. Mater. 2011, 186, 902−910. (6) Chen, X. Q.; Lam, K. F.; Yeung, K. L. Selective Removal of Chromium from Different Aqueous Systems using Magnetic MCM-41 Nanosorbents. Chem. Eng. J. 2011, 172, 728−734. (7) Zhang, Q. R.; Du, Q.; Hua, M.; Jiao, T. F.; Gao, F. M.; Pan, B. C. Sorption Enhancement of Lead Ions from Water by Surface Charged Polystyrene-supported Nano-zirconium Oxide Composites. Environ. Sci. Technol. 2013, 47, 6536−6544. (8) Cimen, A.; Torun, M.; Bilgic, A. Immobilization of 4-amino-2hydroxyacetophenone onto Silica Gel Surface and Sorption Studies of Cu(II), Ni(II), and Co(II) Ions. Desalin. Water Treat. 2015, 53, 2106− 2116. (9) Kanungo, S. B.; Tripathy, S. S.; Mishra, S. K.; Sahoo, B.; Rajeev. Adsorption of Co2+, Ni2+, Cu2+, and Zn2+ onto Amorphous Hydrous Manganese Dioxide from Simple (1−1) Electrolyte Solutions. J. Colloid Interface Sci. 2004, 269, 11−21. (10) Wang, W.; Jiao, T.; Zhang, Q.; Luo, X.; Hu, J.; Chen, Y.; Peng, Q.; Yan, X.; Li, B. Hydrothermal Synthesis of Hierarchical Core−shell Manganese Oxide Nanocomposites as Efficient Dye Adsorbents for Wastewater Treatment. RSC Adv. 2015, 5, 56279−56285. 18375


Article

The Journal of Physical Chemistry C ular Glues’ for Au and SiO2 Substrates. Surf. Interface Anal. 2001, 31, 809−814. (32) Pearson, R. G. The Concept of the Chemical Bond. In Absolute Electronegativity and Absolute Hardness; Springer-Verlag: Berlin, 1990, Chapter 2. (33) Pearson, R. G.; Parr, R. G. Absolute Hardness: Companion Parameter to Absolute Electronegativity. J. Am. Chem. Soc. 1983, 105, 7512−7516. (34) Chouyyok, W.; Shin, Y.; Davidson, J.; Samuels, W. D.; Lafemina, N. H.; Rutledge, R. D.; Fryxell, G. E.; Sangvanich, T.; Yantasee, W. Selective Removal of Copper(II) from Natural Waters by Nanoporous Sorbents Functionalized with Chelating Diamines. Environ. Sci. Technol. 2010, 44, 6390−6395.

18376
