High Adsorption of Sulfamethoxazole by an Amine-Modified

Subscriber access provided by University | of Minnesota Libraries

Article

High adsorption of sulfamethoxazole by an amine-modified polystyrene-divinylbenzene resin and its mechanistic insight Chen Ling, Xiaoyun Li, Zhiyun Zhang, Fuqiang Liu, Yingqing Deng, Xiao-Peng Zhang, Aimin Li, Lili He, and Baoshan Xing Environ. Sci. Technol., Just Accepted Manuscript • DOI: 10.1021/acs.est.6b02846 • Publication Date (Web): 30 Aug 2016 Downloaded from http://pubs.acs.org on September 4, 2016

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Environmental Science & Technology is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 32

Environmental Science & Technology

High adsorption of sulfamethoxazole by an amine-modified polystyrene-divinylbenzene resin and its mechanistic insight Chen Ling1,2, Xiaoyun Li2, 3, Zhiyun Zhang4, Fuqiang Liu*,1, Yingqing Deng2, Xiaopeng Zhang1, Aimin Li1, Lili He4, Baoshan Xing*,2 1

State Key Laboratory of Pollution Control and Resource Reuse, School of the Environment,

Nanjing University, Nanjing 210023, P. R. China 2

Stockbridge School of Agriculture, University of Massachusetts, Amherst, Massachusetts 01003,

United States 3

College of Tourism and Environment, Shaanxi Normal University, Xi'an, Shaanxi 710119, P. R.

China 4

Department of Food Science, University of Massachusetts, Amherst, Massachusetts 01003,

United States

1

ACS Paragon Plus Environment


Page 2 of 32

1

ABSTRACT

2

Sulfamethoxazole

3

polystyrene-divinylbenzene resins (PSA/B/C/D) was investigated. All resins showed a

4

similar pH dependent adsorption of SMZ but their capacities were linearly related

5

with the contents of primary amines (-NH2) rather than secondary amines (-NH-).

6

Mechanisms of SMZ adsorption by PSA (highest -NH2 content) were discussed as an

7

example. Due to comparable pKa, H-bonding interactions of -NH20 with SMZ0

8

(regular H-bond) and SMZ- (negative charge-assisted H-bond, (-)CAHB) successively

9

contributed most adsorption (pH 4~9). At weakly acidic pH, -NH20 was partially

10

protonated and electrostatic attraction between -NH3+ and SMZ- occurred

11

concurrently, but could be hindered by increased loading of SMZ0. Hydrophobic/ π-π

12

interactions were not major mechanisms as phenanthrene and nitrobenzenes had little

13

effect on SMZ adsorption. At alkaline pH, where SMZ- and -NH20 prevailed,

14

adsorption was accompanied by the stoichiometric (~1.0) proton exchange with water,

15

leading to OH- release and the formation of (-)CAHB [SO2N-…H…NH2]. The

16

interaction and SMZ spatial distribution in resin-phase were further confirmed by

17

FTIR and Raman spectra. SMZ was uniformly adsorbed on external and interior

18

surfaces. SMZ adsorption by PSA had low-interference from other coexistent matter,

19

but high stability after multiple regenerations. The findings will guide new adsorbent

20

designs for selectively removing target organics.

(SMZ)

adsorption

by

a

series

2


of

amine-modified

Page 3 of 32


21

1. INTRODUCTION

22

Continuous release of antibiotics to natural aquatic systems has raised increasing

23

concern in the last decade because these compounds are not readily biodegradable,

24

impact microbial ecosystems and may result in bacterial resistance proliferation1-4.

25

Based on a survey in 2013, 92,700 tons of 36 commonly used antibiotics were

26

produced for human and animal use in China, of which, approximately 53,800 tons

27

eventually ended up in the environment through various pathways5. Although

28

conventional biological wastewater treatment processes can remove some antibiotics,

29

many of them have been reported to be at concentrations of 10~1000 ng/L upon

30

discharge6, 7. Also, many countries tend to reuse wastewater for soil irrigation,

31

resulting in more antibiotics detected worldwide in surface water as well as ground

32

water8. Therefore, advanced treatment and remediation of water resources are

33

urgently needed9, 10.

34

Adsorption is one of the most reported technologies for the removal of antibiotics

35

from water because it is simple, green, and the adsorbents can be reused. Most

36

antibiotics are polar compounds with both hydrophobic and hydrophilic parts. Thus,

37

various adsorbents have been developed to remove antibiotics that rely on van der

38

Waals, hydrophobic, and electrostatic interactions11-14. Many of these materials

39

encountered lower selectivity and their adsorption decreased significantly in the

40

presence of coexisting substances such as natural organic compounds and inorganic

41

ions. Therefore, to achieve high adsorption capacity and selectivity, it is critical to

42

design more specific interaction sites tailored to remove specific antibiotics, which 3



43

has to date been rarely reported15-17.

44

Hydrogen (H-) bonding is one highly selective interaction contributing to the

45

adsorption of weak organic acids (WOAs, including many antibiotics) from water to

46

adsorbents when WOA-adsorbent H-bonds overcome the competition from

47

WOA-water and water-adsorbent H-bonds18, 19. Recently several studies reported

48

that amines enriched materials showed superior adsorption to WOAs from aqueous

49

phase and all proposed that H-bonding interaction played an important role in the

50

adsorption. Suriyanon et al. studied the adsorption of acetaminophen (ACT) onto an

51

amine-functionalized silica-based material (A-HMS). They speculated H-bonding

52

between positively charged amine (-NH3+) and hydroxyl of acetaminophen20.

53

However, partial deprotonation of -NH3+ and distribution of amine species were not

54

considered and the functional groups participating in H-bonding interaction were not

55

specified, as both ACT and A-HMS contained various groups with H-bonding ability.

56

Similarly, Zhou et al. investigated the role of H-bonding sites in the adsorption of

57

bisphenol A, but the adsorption sites (hydroxyl and/or secondary amines) were not

58

confirmed21. Chen et al. examined the mechanisms of p-nitrophenol (PNP)

59

adsorption onto a tetraethylenepentamine tailed resin at pH 5 through the

60

competition of Cu(II) adsorption, which was expected to coordinate with amine

61

groups, however whether all amine groups were available as sites for PNP

62

adsorption as well as the adsorption mechanisms at pH>5 are still unknown22. At

63

environmentally relevant pHs, many antibiotics are partly dissociated and present as

64

neutral molecules and anions. Several studies have showed an exceptionally strong 4


Page 4 of 32

Page 5 of 32


65

H-bond (e.g. negative charge-assisted H-bond, (-)CAHB) formed between anionic

66

sulfamethazine, phthalic acid, or naphthenic acids and surface carboxyl/ hydroxyl

67

groups23-25, however, little is known about (-)CAHB in regard to surface amine

68

groups with H acceptor ability. Therefore, the preferred amine groups (i.e. among

69

primary, secondary and tertiary amines), the structure-affinity relationship for

70

antibiotic adsorption, and the contribution of H-bonding to the overall adsorption at

71

different pH still need to be explored as the information is vital to guide how to

72

design adsorbents and their applications in the removal of specific antibiotic

73

families.

74

In this work, we focused on the adsorption of sulfamethoxazole (SMZ) onto

75

resins containing abundant amines (primary and secondary amines). SMZ was

76

selected as a model antibiotic because it is a representative sulfonamide antibiotic

77

widely used and frequently detected in groundwater in China, USA and Europe.5, 9, 26.

78

Moreover, SMZ in water exists as a cation (SMZ+), neutral molecule (SMZ0) and as

79

an anion (SMZ-) due to the protonation of the aromatic amine (-NH3+-, pKa≈1.7)

80

and the dissociation of sulfonamide group (-SO2NH-, pKa≈5.7)23. Specifically, we

81

hypothesize that H-bonding interactions between the neutral -SO2NH- group

82

(H-donor) and the surface neutral primary amine (-NH2, H-acceptor) contribute to

83

the majority of SMZ adsorption at pH≤5.7. Also, at higher pH, (-)CAHB interactions

84

between the anionic -SO2N-- group and the surface neutral -NH2 (pKa-NH3+→-NH2

85

=3.5~7.1 for different linear unsubstituted polyamines) may occur as their pKas are

86

comparable. The results of this study will provide new information for adsorbent 5



87

designs and their applications in the efficient removal of antibiotics from waters.

88

2. MATERIALS AND METHODS

89

2.1. Materials

90

Four resins with amine

91

polystyrene-divinylbenzene resin by reacting the resin with either ethylenediamine,

92

diethylenetriamine, triethylenetetramine or tetraethylenepentamine (the details are

93

shown in Figure S1 in the Supporting Information, SI), and were named as PSA, PSB,

94

PSC and PSD, respectively, based on their incremental amine chain length. Four

95

sulfonamides including sulfamethoxazole (SMZ), sulfadiazine (SFZ), sulfamethazine

96

(SMT) and sulfapyridine (SFD) were purchased from Sigma-Aldrich with a purity of

97

99% and used as received.

98

from American Radiolabeled Chemicals, Inc. and the other reagents were all

99

analytical grade purchased from Fisher Scientific. The typical properties of all

100

adsorbents and adsorbates used in this study are listed in Table S1 and Table S2 in the

101

SI.

102

2.2. Adsorption experiments

103

Adsorption was conducted using the established batch techniques detailed in the SI.

104

Briefly, in order to avoid pH buffering by the resin, the desired amount of resin (10.0

105

mg) was first pre-wetted using 20 mL of ultrapure water (pre-adjusted to a desired pH)

106

for 24h at 298±1 K followed by the addition of 20 mL of the adsorbate solution at the

107

same pH as the equilibrium solution pH in the pre-wetting stage. The adsorption was

108

processed for another 72 h, which was the reaction time found to be required to reach

groups were

14

synthesized from chloromethylated

C labeled (S. A. 77 mCi/mmol) SMZ was purchased

6


Page 6 of 32

Page 7 of 32


109

adsorption equilibrium from the prior kinetic experiments (Figure S2). Liquid-phase

110

composition is specified in each figure. After phase separation, SMZ concentration

111

with a 14C-labeled was determined in the supernatant by liquid scintillation counting.

112

Other solutes of interest (organic and inorganic compounds) were detected by UV

113

spectrophotometer or ion chromatography (details in the SI, Table S3). The

114

equilibrium pH was also measured. There was no significant loss of solutes to

115

degradation or adsorption onto vial walls.

116

2.3. Data Analysis

117

The equilibrium adsorption amount (Qe µmol/g) was calculated using Eq.1. Three

118

typical isotherm models were used to fit the adsorption data. Mean weighted squared

119

error (MWSE) was calculated by Eq.2. ( − ) ⋯ ( . 1)

!"# (): = 1 + =

)

%&ℎ (%): = ( * Dubinin − Ashtakhow model (DAM): log = − (>/ )@ O

1 AB = D[(FGHIJK − FLKM )N / FGHIJK N ] … ( . 2) C PQ)

120

where, C0 and Ce (µmol/L) are the initial and equilibrium dissolved concentrations,

121

respectively. The m (g) is the mass of resin, V (L) is the volume of the adsorbate

122

solution. Qmax and Q0 (µmol/g) are the maximum adsorption capacities for LM and

123

DAM, respectively. KL (L/µmol) and Kf represent the adsorption affinity parameters

124

of LM and FM, respectively. The n is the Freundlich isotherm constant. The >(kJ/mol)

125

= RT ln (Cs/Ce) is the effective adsorption potential, where Cs (µmol/L) is the water 7



126

solubility of the compound at the tested pH, R is the universal gas constant

127

[8.314×10-3 kJ/(mol·K)] and T (K) is the absolute temperature. E [kJ/mol] is the

128

“correlating divisor” of >. The b is the DAM fitting parameter. Both E and b are

129

related to the adsorption affinity of organic compounds. MWSE and the correlation

130

coefficient (r2) were used to evaluate the goodness of fit on the data. X is the number

131

of observations and C is the degree of freedom (X-2 for LM and FM, X-3 for DAM).

132

Qmeasured and Qmodel are the measured and model-estimated equilibrium adsorption

133

amount, respectively. Statistical analysis was conducted using the commercial

134

software programs (Microsoft excel 2013, Origin 8.0 and R version 3.2.4).

135

2.4. Characterization techniques

136

The elemental content of the whole resin (WR) and the resin external surface (ES)

137

were examined by elemental analyzer (CHN-O-Rapid, Germany) and X-ray

138

photoelectron spectroscopy (XPS, PHI 5000 VersaProbe, Japan), respectively. The

139

speciation of surface amines and SMZ was analyzed by theoretical calculation from

140

Visual MINTEQ (ver. 3.0, USA). Solid-state IR spectra (KBr pellets) in the region of

141

2000~500 cm-1 were recorded on a VERTEX 80 v fourier transform infrared

142

spectrometer (Bruker, Germany). A DXR Raman microscope (Thermo Fisher

143

Scientific, Madison, WI., U.S.A.) with a 633 nm laser and 10×, 20× confocal

144

microscope objectives were used in this study. Each spectrum was scanned from 3400

145

to 400 cm-1 with 8 mW laser power and 2 s exposure time. Raman maps were

146

integrated based on the characteristic peaks in the Raman spectra using OMINC

147

software (Thermo Fisher Scientific). Raman mapping was completed with a 50 µm 8


Page 8 of 32

Page 9 of 32


148

pinhole aperture to control the confocal depths. The step size in the X and Z directions

149

was both 10 µm, and each image contained 64 spots.

150

3. RESULTS AND DISCUSSION

151

3.1. Influence of pH and the adsorption isotherms at the optimal pH

152

SMZ adsorption onto the four resins reached the maximum at pH 4.5-6.0 and

153

decreased on either side of this range (Figure1A). Generally, the order in adsorption

154

amount was PSA>PSB>PSC>PSD. The adsorption isotherm of SMZ at an optimal pH

155

(5.5) was conducted to compare the adsorption capacities of different resins (Figure

156

1B). DAM fit better than LM or FM with lower MWSE values and higher correlation

157

coefficients (r2). The fitting parameters are presented in Table S4. Consistently, From

158

PSA to PSD, their Q0 values gradually decreased by 25.8%. However, the affinity

159

indexes (E and b values) were relatively stable ( pHe1 at pHe>5 10


Page 10 of 32

Page 11 of 32


192

(meaning OH- was released or H+ was adsorbed due to the interaction between SMZ

193

and resin at pHe>5). The two stages indicate different adsorption mechanisms.

194

Considering the environmentally-relevant pH range of water is 4.0~9.0, the

195

adsorption of SMZ at pH around 4.5, 6.0 and 8.5 was particularly studied.

196

3.2.1. pH around 4.5

197

At pH 4.5, SMZ0 (95%) and SMZ- (5%) make up the total SMZ in solution, while

198

the amine groups of PSA are comprised of -NH20 (25%) and -NH3+ (75%). Possible

199

dominating interactions between PSA and SMZ are listed below20, 27, 28. UVWXY + BZ [ = UVWXY ∙ BZ [ (1) Electrostatic attraction UVWXY + BZ = UVWXY ⋯ BZ (2) H − Bonding interaction UVWN + BZ = UVWN ⋯ BZ (3) H − Bonding interaction U + SMZ = U∎BZ (4) Non − specific interaction

200

M2 is excluded from the predominant mechanisms because the adsorption amount

201

of SMZ clearly decreased with the increase of both -NH3+ and SMZ0 at a pH range

202

from 5 to 3 (see Figure 1A and Figure 2). Moreover, if M1 is thought to be a

203

significant mechanism for SMZ adsorption at pH 4.5, part of the SMZ0 in the solution

204

was expected to dissociate to form SMZ- with the release of H+, as more than 70%

205

SMZ was adsorbed onto the resin-phase. However, Figure 2C shows no pH decrease

206

due to the adsorption of SMZ at a pH lower than 5, arguing against M1 as a leading

207

contribution of SMZ adsorption onto PSA.

208

Some non-specific interactions (M4), especially hydrophobic effects (HPO) and

209

π-π electron donor-acceptor (EDA) interactions between carbonaceous sorbents and 11



Page 12 of 32

210

aromatic compounds have been frequently mentioned in previous literature29-31. A

211

large amount of alkane/ benzene carbon (>60%) in the matrix and the amine chain of

212

PSA is regarded as hydrophobic sites. The contribution of HPO between SMZ and

213

PSA at pH 4.5 was tested with the competition of phenanthrene (PHE) which is much

214

more hydrophobic than SMZ as reflected in its logKow (log-transformed octanol-water

215

distribution coefficient, Table S2). π-π EDA interactions occurs when two oppositely

216

polarized quadrapoles of benzene rings approach each other23, 32. SMZ could be a

217

π-acceptor due to the high electronegativity of the -SO2NH- group, while the

218

polystyrene-divinylbenzene matrix has a π-donor ability27,

219

interactions between SMZ and PSA, the competitive effect of two strong π-acceptor

220

compounds: 1,3-dinitrobenzene (DNB) and 2,4-dinitrotoluene (DNT) on SMZ

221

adsorption was studied

222

co-presence of the three uncharged competitors exerted very little inhibition (smaller

223

than 3%) on SMZ adsorption by PSA. Moreover, the adsorption strength based on

224

Ce-normalized adsorption amount was in the order PHE>SMZ>>DNT>DNB in both

225

single-solute and bi-solutes adsorption systems. Hence, the role of HPO and π-π EDA

226

interactions for SMZ adsorption within this pH range was not important. Compared

227

with the high adsorption potentials of carbon nanotubes (CNTs), biochars and

228

hypercrosslinked resin (hundreds m2/g)23, 36, 37, PSA (maximum 26.73 m2/g) probably

229

does not have enough HPO area to make sufficient contact with SMZ (polar: nonpolar

230

surface area ratio=1:3, see Table S1 and S2), yielding a much lower adsorption

231

capacity.

33

. To test π-π EDA

34, 35

. As shown in Table S6 and Figure S7 in the SI, the

12


Page 13 of 32


232

Based on the above discussion, we propose that H-bonding interaction (M3) is the

233

predominant mechanism for SMZ adsorption at pH 4.5 and occurred between

234

sulfonamide (-SO2NH-, with H-bond donor ability αm=1.15) and the neutral primary

235

amine (-NH2, with H-bond acceptor ability βm=0.6)38, 39. H-bond strength gradually

236

increases as the difference of donor-acceptor proton affinities (∆PA) or acidic

237

constants (∆pKa) decreases40. We compared the adsorption of SMZ along with three

238

other sulfonamide compounds with different pKa values (see Table S2) because of the

239

different heterocyclic rings. Note that >95% of each compound is in its neutral form

240

at pH 4.5. If M3 predominantly contributed to the adsorption mechanisms, the affinity

241

of the four compounds to PSA would increase from SMZ, SFZ, SMT to SFD due to a

242

smaller ∆pKa between resin primary amines (pKa,resin ~5.0) and -SO2NH- groups (pKa,

243

solute,

244

initial concentration, the adsorption amount was as expected and in the order of SMZ>

245

SFZ> SMT> SFD. Moreover, the affinity indexes (KL in LM, Kf in FM and E, b in

246

DAM, Table S7) all show a linear relationship with the ratio of acid dissociation

247

constants of the solute -SO2NH- groups and the resin primary amines (Ka, solute/ Ka, resin,

248

equal to 10-∆pKa) (r2>0.94), which strongly demonstrates the predominant role of M3

249

in SMZ adsorption.

SMZ 5.7< SFZ 6.3< SMT 7.4< SFD 8.6). Figure S8 shows that for the same

250

3.2.2. pH around 6.0

251

PSA had the highest adsorption strength of SMZ at a pH around 6.0 where SMZ is

252

~60% anionic while ~80% of the primary amines in PSA are in neutral form. Similar

253

to the case at an initial pH of 4.5, M1 and M3 could be concurrent mechanisms. The 13



254

apparent adsorption of SMZ0 and SMZ- was calculated using Visual MINTEQ based

255

on the concentration of total SMZ and solution pH (initial and equilibrium).

256

Adsorption at initial pH 5.3 and pH 6.1 are compared in Figure 3. DAM fits the

257

adsorption isotherm much better than LM and FM with the lowest MWSE values and

258

highest correlation coefficient (r2) (Table S8). The affinity indexes (E and b in DAM)

259

for SMZ0 increased while those for SMZ- decreased with rising initial pH, which

260

agrees with the speculation that the available -NH20 sites increased while the available

261

-NH3+ sites reduced at a higher pH. This was also examined by the higher adsorption

262

amount of the structural analogue, 4-aminobenzenesulfonic ions (ABS-, pKa，ABS→

263

ABS-=-3.39)

264

that SMZ- apparent adsorption at the initial pH of 5.3 approached saturation at its

265

equilibrium concentration of 200 µmol/L, while SMZ- adsorption at the initial pH of

266

6.1 kept increasing with the increase in SMZ- equilibrium concentration. The earlier

267

saturation of SMZ- loading at a lower pH is probably caused by the steric hindrance of

268

the loading of SMZ0. Overall, at a total SMZ concentration lower than 50 µmol/L,

269

both SMZ0 and SMZ- were almost completely adsorbed by PSA at the same ratio of

270

their concentrations in solution (1.7 at pH 5.3 and 0.45 at pH 6.1). With total SMZ

271

concentration increasing, the ratio of the SMZ0 and SMZ- adsorption amount kept

272

rising to 3.3 at the initial pH 5.3 or 1.3 for the initial pH 6.1. The preference of SMZ0

273

adsorption onto PSA indicates that H-bonding (M3) is likely the predominant

274

mechanism occurring rather than electrostatic attractions (M1).

275

at pH 5.3 than at pH 6.1 (Figure S9). On the other hand, it is noteworthy

3.2.3. pH around 8.5 14


Page 14 of 32

Page 15 of 32


276

At the initial pH of 8.5, SMZ- reaches the maximum percentage of total SMZ

277

adsorption (>99.5) while the primary amines in the resin are totally -NH20. Since the

278

hydrophobic and π-π EDA interactions (M4) are very weak at low pH where SMZ0

279

(logKow ~0.79) is the predominant species, M4 will be further weakened with the

280

decrease of solute hydrophobicity (logKow ~0.06) at a much higher pH36. The clear

281

adsorption of SMZ- at this pH range as shown in Figure 1A is expected to be

282

contributed to electrostatic interactions or anion exchange (M1) because most of the

283

secondary amines are still positively charged (>80%) at pH 8~9. SMZ- adsorption was

284

conducted under various solute conditions with one or more cosolutes including

285

H2PO4- and SO42- (the molar ratio cosolute to SMZ was 103~104), assuming that high

286

concentrations of inorganic anions (especially divalent SO42-) could strongly compete

287

for most of the M1 sites30. As shown in Figure 4A, SMZ adsorption at pH 8.5

288

declined

289

SMZ+H2PO4-+SO42- three-solutes system. The adsorption amount was basically stable

290

when more SO42- was added to the adsorption system, suggesting that only around 12%

291

SMZ adsorption is attributed to M1. On the other hand, the weaker adsorption of

292

H2PO4-, SO42- and the structural analogue, ABS- than SMZ- in the single-solute system

293

at the same initial pH 8.5 again demonstrates M1 is present, but not critical at this pH

294

range. More importantly, since M1~M4 were all ruled out, the major adsorption

295

(~88 %) of SMZ- must be only driven by a different mechanism.

with

increasing

coexistent

inorganic

anions,

by

11.7%

in

296

A special H-bond called negative charge-assisted H-bond, (-)CAHB, is a subset

297

of low barrier hydrogen bond and generally recognized as occurring in the 15



298

liquid-phase when the H-donor and acceptor have similar acidic constants (∆pKa=

299

pKa, H-donor−pKa, H-donor≤ ~4) 40. Moreover, the strength of (-)CAHB increases with

300

decreasing ∆pKa. Here, ∆pKa between H-donor (-SO2N--) and acceptor (-NH20) is

301

only ~0.6. Therefore, we assume that (-)CAHB may occur between SMZ- and the

302

surface –NH20 sites accompanying with the proton exchange of SMZ- with water, as

303

described by the following equations: BZ [ + WN j = BZ + jW[ (kl m&ℎ) UVWN + BZ = UVWN ⋯ BZ (W − n l&l)

304

That is, UVWN + BZ [ + WN j = [UVWN ⋯ W Y ⋯ BZ [ ] + jW[ ((−)oWp, 5)

305

(-)CAHB was previously reported to form between anionic sulfamethazine or

306

carboxylates and the surface carboxyl or hydroxyl groups of biochars and CNTs,

307

accompanied by the release of OH- in a stoichiometry of up to 1.0 for the monobasic

308

acid19, 23, 41.

309

To examine our hypothesis, the adsorption isotherm of SMZ (SMZ-) and the

310

buffering capability of PSA were studied at the same initial pH 9.0 (Figure 4B). It

311

revealed that OH- was released due to SMZ- adsorption: as the total loading of SMZ

312

(all species) went up from 0 to 550 µmol/g, the solution equilibrium pH clearly

313

increased from 7.02 to 8.68, while the pH in the corresponding controls without PSA

314

remained unchanged. Considering PSA would keep buffering the solution pH during

315

the entire adsorption process, the stoichiometry of OH- release was estimated by

316

summing the moles of OH- that appeared in the solution and the moles of OH-

317

consumed by PSA at the same pH due to its buffering capability. The stoichiometry 16


Page 16 of 32

Page 17 of 32


318

reaches a maximum of 0.96 and decreases with the increase of SMZ loading. The

319

decline in stoichiometry probably resulted from the increase in SMZ adsorption that

320

was contributed by other mechanisms. In the case of the energy conservation, the

321

standard free energy of proton exchange is estimated by ∆tu v #wx = UylnG,HLMIz −

322

UylnG,{GzJ = −32.52 − (−79.88) = +47.35 kJ/mol, where Ka,

323

are the acid dissociation constants of SMZ and water, respectively. To account for the

324

adsorption of SMZ- onto the PSA surface, the energy demand of proton exchange has

325

to be compensated for by the free energy (∆t([)u ) of (-)CAHB [SO2N-…H…NH2]

326

formation. Although to date the latter can be only estimated as the range from about

327

-56 kJ/mol for hydrogen dicarboxylate conjugate pairs, [RCO2⋯H⋯O2CR]- (∆pKa

328

~0), to -50.2 kJ/mol for the acetate/phenol pair, [C6H5O⋯H⋯O2CCH3]- (∆pKa ~5.2)19,

329

23

330

stoichiometry and energy compensation strongly support the formation of (-)CAHB as

331

contributing to the majority of SMZ- adsorption at alkaline pH. In essence, the

332

apparent acidity constant of the adsorbed SMZ (pKa, ads) onto PSA increased to 7.5

333

compared with that of aqueous SMZ (pKa2, 5.7) (see Figure 4B), which facilitates the

334

H-bonding between SMZ and PSA at a wider pH range. As the initial pH increased,

335

proton exchange of SMZ- became energetically costlier at a faster pace due to the

336

suppression by the high OH- concentration. Thus, SMZ- adsorption and OH- release

337

declined sharply to ~zero at initial pH >11 (Table S9).

solute

and Ka,

water

, ∆t([)u is enough to compensate for ∆tu v #wx . Therefore, OH- release

338

3.2.4. Confirmation of adsorption sites and SMZ spatial distribution in PSA

339

FTIR spectra of fresh PSA, SMZ loaded PSA (PSA+SMZ) and pure SMZ are 17



340

shown in Figure 5A. The peak at 1571 cm-1 in the spectra of PSA is assigned to the

341

stretching vibrations of -NH242-44. The peak almost disappears in the PSA+SMZ

342

spectrum, confirming that -NH2 participated in the adsorption. Spectra comparison of

343

SMZ and PSA+SMZ showed an obvious shift in the vibration of S-N bond (SMZ:

344

928 cm-1; PSA+SMZ: 943 cm-1), while the stretching vibration of aromatic –NH2 at

345

1500 cm-1 and 1595 cm-1 remained unchanged45, 46. All the information supports the

346

H-bonding interaction between the sulfonamide group and the primary amine.

347

To further investigate the spatial distribution of SMZ in the whole resin, the

348

Raman spectra of PSA and PSA+SMZ cross-sections at 0, 10, 50, 100 µm from the

349

external surface were obtained in Figure 5B. For PSA, peaks at 1610 cm-1 and 1575

350

cm-1 were assigned to the bending stretching of –NH2 scissors which were strongly

351

detected at all inner cross-sections of PSA47. As to PSA+SMZ spectra, there were

352

many new peaks at 1600 cm-1, 1026 cm-1, 1122 cm-1 and 950 cm-1, specifically

353

corresponding to the vibration of amine (at the benzene ring), isoxazole ring and two

354

symmetric stretching of the sulphonyl group, respectively45-47. Their presence at each

355

cross-section directly proves that SMZ has diffused into the interior surface. Also,

356

similar to the results in FTIR, the amine peak at 1575 cm-1 disappeared in all spectra

357

of PSA+SMZ, which is consistent with the involvement of primary amines in SMZ

358

adsorption.

359

We remain curious as to whether SMZ adsorption would be hindered during the

360

diffusion. Raman signal intensity is reduced with the increasing access depth of the

361

laser. For every 64 detection spots in each cross-section, the intensity of guest -SO2 18


Page 18 of 32

Page 19 of 32


362

peak at 1122 cm-1 was normalized by the intensity of the resin-intrinsic benzene peak

363

at 641 cm-1, alkane peak at 1455 cm-1 and –NH2 peak at 1610 cm-1, with the averaged

364

values reported in Table S10. Values of I-SO2 (S) /I-alkane (R) and I-SO2 (S) /I-benzene (R) showed

365

a slight increase with increasing depth, indicating that the concentration of SMZ in the

366

resin-phase is basically uniform from the external surface towards the interior surface.

367

The unhindered diffusion is probably attributed to the large pore structure of PSA as

368

shown in the SEM image (Figure S10A). More importantly, values of I-SO2 (S)/ I-NH2(R)

369

were relatively stable around at 0.9 for all detection spots, suggesting that more SMZ

370

was detected where more primary amines were grafted onto the resin (Figure S10B,

371

C), again supporting that primary amines are the predominant sites for SMZ

372

adsorption.

373

All the data were consistent with our hypothesis. Primary amines were the most

374

important sites for SMZ adsorption. The proposed adsorption mechanisms of SMZ

375

onto PSA surface at different pH ranges are illustrated in Figure 6. For a given initial

376

concentration (e.g. 250 µmol/L), SMZ could be progressively adsorbed first through

377

regular H-bond (M3) and charge pairing with primary amines (M1a) in a mole ratio of

378

46:6 at pHe~4.5, and 60:35 at pHe~6.0. Then adsorption through (-)CAHB with

379

primary amines (M5) and charge pairing with secondary amines (M1b) in a ratio of

380

34:4 at pHe~8.5 likely occurs, accompanied by proton exchange with water in the

381

aqueous phase. Not shown are non-specific interactions which may operate in all

382

cases, but not dominantly. All of the interactions could freely occur at the external and

383

interior surfaces where the corresponding sites are present. 19



384

3.3. Practical application and environmental implication

385

As SMZ adsorption onto PSA predominantly relies on the H-bonding interactions,

386

the low susceptibility of SMZ adsorption onto PSA resin is expected in the aqueous

387

phase where other concurrent substances such as inorganic ions and natural organic

388

matters (NOM) occur. Figure S11 shows the impact of coexisting substances on the

389

adsorption of SMZ by PSA along with another two commonly used adsorbents (anion

390

exchange resin, D201 and activated carbon, AC, see Table S1)48, 49. A small decrease

391

(

High Adsorption of Sulfamethoxazole by an Amine-Modified

Recommend Documents