The states of water in tryptophan grafted hydroxypropyl methylcellulose hydrogels and their effect on the adsorption of methylene blue and rhodamine B

Tryptophan (Trp) decorated hydroxypropyl methylcellulose (HPMC) cryogels were prepared by a one-step reaction with citric acid. The increase of Trp content in the 3D network from 0 to 2.18 wt% increased the apparent density from 0.0267 g.cm−3 to 0.0381 g.cm−3 and the compression modulus from 94 kPa to 201 kPa, due to hydrophobic interactions between Trp molecules.

Contents lists available atScienceDirect

Carbohydrate Polymers journal homepage:www.elsevier.com/locate/carbpol

The states of water in tryptophan grafted hydroxypropyl methylcellulose

rhodamine B

Paulo V.O Toledoa, Oigres D Bernardinellib, Edvaldo Sabadinib, Denise F.S Petria,*

a Fundamental Chemistry Department, Institute of Chemistry, University of São Paulo, Av Prof Lineu Prestes 748, 05508-000, São Paulo, Brazil

b Department of Physicochemistry, Institute of Chemistry, University of Campinas (UNICAMP), 13083-970, Campinas, São Paulo, Brazil

A R T I C L E I N F O

Chemical compounds studied in the article:

Hydroxypropyl methylcellulose

Citric acid

Tryptophan

Methylene blue

Rhodamine B

Keywords:

Hydroxypropyl methylcellulose

Tryptophan

States of water

Hydrogels

Cryogels

Adsorption

A B S T R A C T Tryptophan (Trp) decorated hydroxypropyl methylcellulose (HPMC) cryogels were prepared by a one-step re-action with citric acid The increase of Trp content in the 3D network from 0 to 2.18 wt% increased the apparent density from 0.0267 g.cm−3to 0.0381 g.cm−3and the compression modulus from 94 kPa to 201 kPa, due to hydrophobic interactions between Trp molecules The increase of Trp content in HPMC-Trp hydrogels increased the amount of non-freezing water, estimated from differential scanning calorimetry, and the amount of freezing water, which was determined by time-domain nuclear magnetic resonance The adsorption capacity of methy-lene blue (MB) and rhodamine B (RB) on HPMC-Trp hydrogels increased with Trp content and the amount of freezing water HPMC-Trp hydrogels could be recycled 6 times keeping the original adsorptive capacity The

diffusional constants of MB and RB tended to increase with Trp content RB adsorbed on HPMC-Trp hydrogels presented a bathochromic shift offluorescence

1 Introduction

Polysaccharide based 3D structures, such as aerogels and cryogels,

are interesting platforms for tissue engineering (Tchobanian, Van

Oosterwyck, & Fardim, 2019), drug delivery (Ulker & Erkey, 2014) and

adsorption of pollutants (Maleki, 2016) Upon contact with aqueous

media, the aerogels or cryogels become hydrogels Understanding the

structure of water around the polysaccharide chains is important

be-cause it might drive the interactions with cells, drugs or pollutants

Water molecules in hydrogels coexist as three different states: (i)

Non-freezing bound water or non-Non-freezing water (Wnf), which results from

tightly bound water molecules, it is not freezable at 0 °C or below 0 °C,

due to the strong interaction with the polymer chain (ii) freezing bound

water or intermediate water (Wfb), which is not freezable at 0 °C, but it

is freezable below 0 °C because the interactions with polymer chains are

not so strong as in the Wnf, and (iii) free water or freezing water (Wf),

which is freezable at 0 °C, in this case, water molecules hardly interact

with the polymer chains, they are just entrapped in the matrix (Tsuruta,

2010) The amount of water molecules in each state depends

fundamentally on the chemical nature of polymer (Hatakeyama & Hatakeyama, 2017)

Hydroxypropyl methylcellulose (HPMC) is a family of water-soluble cellulose ethers widely applied in cosmetics (Lochhead, 2017), phar-maceuticals (Kaur et al., 2018), and food (Burdock, 2007) formulation HPMC chains can be crosslinked via esterification with citric acid (CA),

a nontoxic multifunctional acid, enabling the formation of tridimen-sional structures for drug release (Marani, Bloisi, & Petri, 2015;Reddy

& Yang, 2010) or adsorption of pollutants (Martins, Toledo, & Petri,

2017;Toledo et al., 2019) Cellulose can be modified with amino acids

in order to improve bioaffinity (Kalaskar et al., 2008) or to develop membrane for methanol fuel cells (Zhao et al., 2019) Synthetic poly-mers modified with L-tryptophan (Trp), a hydrophobic essential amino acid (Richard et al., 2009), can be used to produce cryogels with high affinity for proteins (Türkmen et al., 2015) or DNA (Çorman, Bereli, Özkara, Uzun, & Denizli, 2013) Despite the above-mentioned potential applications, systematic investigations about cryogels and hydrogels based on amino acid modified HPMC, the consequences on the me-chanical properties, on the states of water in such materials and on their

https://doi.org/10.1016/j.carbpol.2020.116765

Received 20 February 2020; Received in revised form 27 June 2020; Accepted 11 July 2020

⁎Corresponding author

E-mail addresses:paulo.vinicius.toledo@usp.br(P.V.O Toledo),oigres.daniel@gmail.com(O.D Bernardinelli),sabadini@unicamp.br(E Sabadini),

dfsp@iq.usp.br(D.F.S Petri)

Carbohydrate Polymers 248 (2020) 116765

Available online 25 July 2020

0144-8617/ © 2020 Elsevier Ltd All rights reserved

T

affinity for dyes, are poorly explored Dyes are environmental

pollu-tants, thus it is important to comprehend the correlation between the

states of water in hydrogels and their adsorption capacity

Preliminary, HPMC cryogels were prepared with four different

amino acids (AA), namely, tryptophan (Trp, hydrophobic), glutamic

acid (Glu, acid), cysteine (Cys, polar) and histidine (His, basic) Among

all AA modified HPMC cryogels, the HPMC-Trp cryogels presented the

highest compression modulus in comparison to pure HPMC cryogels

Similarly, the presence of hydrophobic particles, such as lignin, in

cryogels and hydrogels improved their mechanical properties and

ad-sorptive capacity for dyes (Zhang et al., 2019) Based on this, we

pro-ceed with the systematic investigation on the (i) crosslinking of HPMC

chains and chemical attachment of Trp moieties to HPMC by

ester-ification in one-step reaction, with different CA:Trp ratios, (ii) the

physicochemical properties of HPMC-Trp cryogels, (iii) the states of

water (Wnf and Wfb) in HPMC-Trp hydrogels and (iv) how the Trp

content and states of water affect the adsorption capacity of HPMC-Trp

hydrogels for methylene blue (MB) and rhodamine B (RB) To the best

of our knowledge, it is thefirst systematic study about the correlation

between the states of water and the adsorption behavior of dyes

in-volving AA modified polysaccharides

2 Experimental 2.1 Preparation of Tryptophan decorated cryogels

HPMC E4M (USP HPMC 2910, MS 0.25, DS 1.9) was kindly pro-vided by The Dow Chemical Company (Brazil), GPC measurements (Supplementary Material SM1) indicated Mn 1.1x105g mol−1and

Mw 2.4x105g mol−1 The crosslinking of HPMC chains and the at-tachment of Trp to the HPMC chains were performed with CA in a one-step reaction Briefly, aqueous solutions containing HPMC at 2.0 wt%,

CA at 0.10 wt% or 0.20 wt% (CA, LabSynth, Brazil, 192.13 g.mol−1), sodium hypophosphite monohydrate at 0.05 wt% (HPS, LabSynth, Brazil, 106.14 g.mol−1) and L-tryptophan (Trp, Sigma-Aldrich, 204.23 g.mol−1, T0254) were stored in the refrigerator at 12 °C for 8 h in order

to achieve complete dissolution of HPMC The solutions were prepared

at different CA:Trp molar ratios of 1.0:0.0, 1.0:0.5, 1.0:1.0, 2.0:0.5, 2.0:1.0, and 2.0:2.0 The highest concentration of Trp in solution was 2.0 g.L−1, which is well below its solubility at 25 °C of 11.4 g.L−1 [National Center for Biotechnology Information PubChem Database Tryptophan, CID = 6305, https://pubchem.ncbi.nlm.nih.gov/ compound/Tryptophan (accessed on Feb 20, 2020)] The hydrogels were poured into molds of different sizes, were frozen in a freezer (-40

°C for 2 h) and freeze-dried (-50 °C, 200μmHg, for 8–24 hours) The resulting 3D solid structures were heated at 165 °C for 7 min in order to promote the reaction between HPMC hydroxyl groups and CA carboxylic acid groups and/or with Trp carboxylic acid/amine groups One should notice that in the absence of CA there is no chemical crosslinking among the HPMC chains If CA mass is less than 5 % of polymer mass, the crosslinking efficiency is very low and the hydrogels show no mechanical stability CA mass between 5% and 10 % of polymer mass yields stable HPMC hydrogels (Marani et al., 2015; Martins et al., 2017) Ghorpade and co-workers showed that CA mass of

15 % or 20 % of polysaccharide (carboxymethyl cellulose and tamarind gum) mass is not adequate because the swelling degree of hydrogels tends to decrease (Mali, Dhawale, Dias, Dhane, & Ghorpade, 2018) For compressive strength tests, rectangular samples of 13.0 mmx17.0 mmx

25.0 mm were prepared, whereas for the other analyses the samples were discs of 2 mm thickness and 35 mm of diameter The samples were coded as HPMC-Trp cryogels

Fig 1 Schematic representation of (a) HPMC, (b) CA, (c) Trp, (d) MB and (e) RB

Table 1

CA:Trp molar ratios used in the precursor gels to synthesize the HPMC-Trp

cryogels Nitrogen content determined by elemental analyses (N, %), calculated

Trp content (wt%) and gel content (%)

CA:Trp N (%) Trp content (wt%) Gel content (%)

2.2 HPMC-Trp cryogels characterization

For the characterization, all cryogels were rinsed with MilliQ water

until the rinsing water achieved conductivity of ≈ 5 μS cm−1 This

procedure removed the unreacted molecules, which could be only

physically attached to the samples After that, they were freeze-dried

and weighed again The gel content (Gel %) was calculated according to

Eq.(1):

= ⎡

⎣

⎢ −⎛

⎝

⎠

⎤

⎦

⎥ ×

m

where mpolis the initial mass of HPMC and mdriedis the mass of the

freeze-dried sample

The swelling degree (SD) was determined with a precision

tensi-ometer Krüss K100 at (24 ± 1) °C as the mass of sorbed MilliQ water

(pH 5.5) at equilibrium divided by the mass of dried adsorbent:

=

SD m

m

water

The apparent density of HPMC-Trp cryogels was determined at

(24 ± 1)oC by dividing the mass of freeze-dried cryogels by the

cor-responding volume, which was estimated by their dimensions The

di-mensions of seven different samples of the same chemical composition

were measured using a pachometer; the same seven samples were

weighed in an analytical balance The mean mass divided by the mean

volume determined for seven samples yielded the mean apparent

den-sity (ρap) value SEM analyses were performed for gold-coated (by

sputtering) samples in a Jeol Neoscope microscope JCM 5000,

oper-ating at 5 kV voltage The compressive tests were performed for 10

cryogels (rectangular) samples using an Impac, Digital Dynamometer

IP-90DI, with a 10 N load cell, at the strain rate of 0.01 s−1and at

(24 ± 1)oC and (70 ± 5) % relative air humidity The samples HPMC

and HPMC-Trp 2.18 wt% were compressed and decompressed in MilliQ water up to 5 times; in thefirst cycle of compression/decompression, the materials presented larger hysteresis and larger mechanical re-sistance than in the following cycles due to the presence of air en-trapped inside the cryogels, which after thefirst cycle was expelled by water Fourier transform infrared spectroscopy analyses in the atte-nuated total reflectance mode (FTIR-ATR) were performed in a Perkin Elmer Frontier equipped with ZnSe crystal, resolution of 4 cm-1and in the wavenumber range of 600 cm−1to 4000 cm−1 Elemental analysis (CHN) was performed with a Perkin Elmer 2400 Series II equipment For the differential scanning calorimetry (DSC TA Instruments Q10), the samples were swollen in MilliQ water (≈ 5 μS cm−1) at 20 °C for

12 h, degassed under vacuum pump (10 min at 100 mmHg) and subject

to 2 cycles from– 40 °C to 40 °C, at 5 °C.min−1rate Measurements of TDNMR were performed in a Minispec 20 MHz at 33 °C T2relaxation time measurement was carried out with a standard Carr-Purcell-Meiboom-Grill (CPMG) pulse sequence with 30000 echoes and an echo time of 16μs (Carr & Purcell, 1954) SKL Neo MultiExp program for inverse Laplace transform (ILT) was used and Log-Normal distribution integration aided by an user-guided program called Peakfit 4.00 (Jandel scientific software); the dried cryogels (≈ 25 mg) were swollen in MilliQ water (1.0 mL) at 20 °C inside the NMR probe and the results were acquired every 3 min from the completely dry sample to the ab-solute swelling that lasted a total of 45 min

2.3 Adsorption studies Prior to the adsorption experiments, all cryogels were rinsed with MilliQ water until the rinsing water achieved conductivity of≈ 5 μS

cm−1 After that, they were freeze-dried and weighed again For the adsorption studies, methylene blue (MB, Sigma-Aldrich, 319.81 g.mol

-1) dissolved in Tris-HCl 0.05 mol.L-1buffer at pH 7.0 (Trizma Base, Sigma-Aldrich, 121.14 g.mol-1) and rhodamine B (RB, Sigma-Aldrich,

Fig 2 (a) Dehydration of citric acid and anhydride formation upon heating (b) Esterification and crosslinking between two HPMC chains (c) Esterification of anhydride and HPMC hydroxyl group followed by attachment of Trp amino group to form an amide linkage

479.01 g.mol-1) dissolved in Tris-HCl 0.05 mol.L-1buffer at pH 2.5 were

used as model molecules Noteworthy, these pH conditions were chosen

because at pH 7 the adsorption of MB on the walls of the vials and

self-association by π stacking was avoided At pH higher than three, the

adsorption of RB molecules on the HPMC or HPMC-Trp hydrogels was

too low All adsorption experiments were performed at (24 ± 1) °C and

contact time of 24 h, in order to assure equilibrium conditions

(Kaewprasit, Hequet, Abidi, & Gourlot, 1998)

The equilibrium adsorption capacity (qe, mg g−1) of MB or RB was

calculated dividing the concentration of adsorbed MB or RB by the mass

of dried cryogels (m) and multiplying by the solution volume (v):

=C −C ×

(3) The concentration of adsorbed MB or RB onto the cryogels was

determined as the difference between the initial concentration (C0) of

MB or RB and the concentration of MB or RB in the supernatants after

24 h contact, or the equilibrium concentration (Ce) First, calibration

curves of absorbance intensity as a function of MB and RB

concentra-tion were determined by means of spectrophotometry in a Beckmann

Coulter DU640 spectrophotometer, respectively at 664 nm and 553 nm

(Supplementary Material SM2)

For the adsorption isotherms of MB onto Trp decorated HPMC

cryogels (∼30 mg dry mass), the initial concentration (Ci) of MB

ranged from 0.5 mg.L−1to 5.0 mg.L−1 For the adsorption isotherms of

RB onto HPMC-Trp cryogels, the initial concentration (Ci) of RB ranged

from 0.5 mg.L−1 to 4.0 mg.L−1 Adsorption/desorption cycles were

carried out by immersing the RB coated samples in MilliQ water (10

mL) for 15 min After this period, aliquots of supernatant were with-drawn, and the concentration of released RB was determined by pho-tometry at 553 nm Then the hydrogels were immersed in the RB so-lution for 10 min to proceed with next adsorption/desorption cycle After this period, aliquots of supernatant were withdrawn, and the concentration of remaining RB was detected Then the next desorption process was conducted

In order to get insight about the interactions between MB or RB molecules and Trp moieties bound to HPMC chains,fluorescence was measured with a Shimadzu RF6000 spectrofluorometer, at (24 ± 1) °C,

at 600 nm.min−1, excitation and emission bandwidth of 5 nm, re-solution of 1.0 nm and reproducibility of 0.2 nm After 48 h adsorption

of MB or RB (C0= 1.0 mg.L-1) on HPMC-Trp hydrogels, a holder po-sitioned at 45° with respect to the optical axis suspended the swollen hydrogels in the air Samples carrying MB or RB were excited at 650 nm

or 540 nm, respectively; the emission spectra were measured in the range of 670 nm–820 nm or 550 nm–700 nm, respectively Fluorescence of MB and RB solutions at 0.33 mg.L-1 in the corre-sponding buffers was measured in a 10 mmx10 mm quartz cuvette All measurements were performed for at least two samples of the same composition For comparison, the intensity values of fluorescence spectra were normalized with respect to the intensity at the maximum wavelength Fig 1 represents the chemical structures of the main chemical compounds used in the experiments

Fig 3 (a) Apparent density (ρap), (b) compression modulus (ε) as a function of Trp content in the HPMC-Trp porous materials, (c) Dependence of ε on ρap, the red line corresponds to thefit ε = k ρapm, R² = 0.9051, m = 1.8195, and (d) swelling degree (SD) (For interpretation of the references to colour in thisfigure legend, the reader is referred to the web version of this article)

3 Results and discussion

3.1 Characterization of HPMC-Trp cryogels

Preliminarily, HPMC cryogels were prepared with four different

amino acids, namely, tryptophan (Trp, hydrophobic), glutamic acid

(Glu, acid), cysteine (Cys, polar) and histidine (His, basic) at CA:AA

molar ratio of 1.0:1.0 The corresponding compression moduli (ε)

va-lues followed the sequence: Trp > His >

HPMC-Cys > HPMC-Glu > pure HPMC; the data are provided as

Supplementary Material SM3 Theε values followed the relative

hy-drophobicity of amino acids given as Trp > > Cys > His > > Glu

(Wimley & White, 1996), indicating that the inclusion of hydrophobic

amino acid improved the mechanical properties of HPMC cryogels

Based on this experimental observation and on the lack of reports about

the physicochemical properties of Trp modified HPMC cryogels, Trp

was chosen to proceed with the systematic modification of HPMC

cryogels

The synthesis of HPMC-Trp cryogels with different CA:Trp molar

ratios led to different contents of N in the HPMC cryogels, as revealed

by CHN analyses (Table 1) The N% content in the samples stems

ex-clusively from the chemical attachment of the Trp to the cryogels, since

HPMC and CA carry no N atom in their structures Considering the N contents determined from elemental analyses (Table 1) and the N% content in Trp of 11.75 %, the Trp content in the cryogels was calcu-lated (Table 1) The chemical attachment of Trp to the cryogels was favored by the increase of CA and Trp concentrations in the precursor gel; the highest Trp content of 2.18 ± 0.07 wt% was achieved for the CA:Trp molar ratio of 2.0:2.0 On the other hand, the gel content (Gel

%) was the smallest at the CA:Trp molar ratio of 2.0:2.0, indicating competition between HPMC hydroxyl groups and Trp amine groups for the CA carboxylic acid CA carries three carboxylic acid groups, which

in the presence of HPS and under heating undergoes dehydration and anhydride formation (Peng, Yang, & Wang, 2012) (Fig 2a) The an-hydride might react with HPMC hydroxyl groups, such esterification can take place with hydroxyl groups of a second HPMC chain, pro-moting the crosslinking between HPMC chains (Fig 2b.) The anhydride bound to an HPMC chain can also react with Trp amino groups to form amide groups (Fig 2c), decreasing the crosslinking between two HPMC chains Thus, the increase of CA:Trp ratio increases the competition between these reactions, decreasing the crosslinking (Gel %) and in-creasing the attachment of Trp to HPMC chains

Supplementary Material SM4 provides FTIR-ATR spectra of pure HPMC and HPMC-Trp in the 4000 to 600 cm−1 range All samples

Fig 4 SEM images for (a) pure HPMC cryogels and HPMC-Trp cryogels with (b) 0.73 wt%, (c) 1.09 wt%, (d) 1.17 wt%, (e) 1.38 wt% and (f) 2.18 wt% Trp Scale bars correspond to 50μm

Table 2

Experimental values of HΔ endo determined for the HPMC-Trp hydrogels by DSC, the W t , W nfand(W f+W fb)values calculated with Eqs.(4)and(5) W fbfraction was determined by TD-NMR measurements (detailed in Section3.3)

Trp

(wt%)

H

(%)

W nf

(%)

+

( f fb) (%) W fb(%) (TD-NMR)

presented the characteristic HPMC bands: in the 3500–3200 cm−1

re-gion (OH vibrational stretching); at 2930 cm−1and 2850 cm−1

(sym-metrical and asym(sym-metrical CH stretching); and in the 1200–850 cm−1

region (CO and CCee stretching vibrations of the glucopyranose ring)

(Silverstein, Webster, Kiemle, & Bryce, 2014) The esterification

be-tween CA and HPMC was identified by the bands at ≈1730 cm−1and

1640 cm−1, which were assigned to C]O stretching of ester and acidic

forms, respectively (Bueno, Bentini, Catalani, & Petri, 2013), the bands

at 1458 cm−1, 1408 cm-1,and 1322 cm−1were assigned to CH2scissor,

symmetric axial deformation of C]O of esters and stretching of COe of

carboxylate groups HMPC-Trp presented characteristic amide

vibra-tional bands at 1642 cm−1and 1312 cm−1, assigned toνC=Oamide I

andνC-Namide, respectively (Liu, Shen, Zhou, Wang, & Deng, 2016),

which overlapped other characteristic bands already observed in pure

HPMC cryogels, impairing the identification of Trp by these vibrational

bands However, HPMC-Trp with 1.38 wt% and 2.18 wt% Trp

pre-sented two weak bands at 748 cm−1and 707 cm−1, which are

char-acteristic of Trp indole ring; these bands did not appear in the spectra of

pure HPMC cryogels and, therefore, indicated the chemical attachment

to the HPMC hydrogels

Fig 3a and b shows that the apparent density (ρap) and the

compression modulus (ε) increased considerably with the increase of Trp content in the HPMC-Trp cryogels, indicating that hydrophobic interaction between Trp molecules might contributed to the cell wall structuring HPMC-Trp 2.18 wt% presented ε value twofold of that determined for pure HPMC cryogels For comparison, theε values of HPMC cryogels modified with 15 wt% of cellulose nanocrystals in-creased only 20 % in comparison to pure HPMC cryogels (Toledo et al.,

2019), but the addition of 2 wt% hydrophobic lignin particles to hy-drophilic poly(vinyl alcohol), PVA, hydrogels increased fourfold theε value over that of pure PVA hydrogels (Bian et al., 2018) Thus, self-assembling of hydrophobic moieties added to hydrophilic cryogels might improve the mechanical properties of cryogels

Theε values increased with ρap1.82(Fig 3c); the index value of 1.82

is typical for isotropic open-cell structures (Scotti & Dunand, 2018) Fig 4shows the SEM images of pure HPMC and HPMC-Trp with dif-ferent Trp contents Regardless of the Trp content, all cryogels pre-sented isotropic open cellular structure, in agreement with the index value of 1.82 The swelling degree (SD) values (Fig 3d) presented no significant dependence on the Trp content On average, the SD for water amounted to≈ 47 g per g of cryogel This value is ≈ 10 units smaller than those found for cryogels made of negatively charged

Fig 5 Representation of the T2decay-time distribution influenced by the pore size schematically indicated in the SEM image, (a) small pores and free “bulk” water, (b) large pores and free water and (c) small and large pores and free water (d) T2decay-time distribution of the hydrogen atoms of water measured as soon as the HPMC-Trp cryogels were put in contact with water (virtually at 0 min) and after 45 min contact The Trp content in HPMC-Trp samples varied from 0 wt% to 2.18 wt

%

polysaccharides, like xanthan gum (Toledo, Marques, & Petri, 2019) or

carboxymethyl cellulose (Toledo, Limeira, Siqueira, & Petri, 2019), due

to the absence of charges in the HPMC structure Fig.Supplementary

Material SM5 shows the compressive stress-strain curves measured in

MilliQ water for HPMC and HPMC-Trp 2.18 wt% The samples were

compressed and decompressed up to 5 times, but for the sake of clarity,

only the 2nd cycles were presented Both HPMC and HPMC-Trp 2.18 wt

% hydrogels presented high resilience, small hysteresis and similar

stiffness This trend is in agreement with the independence of SD values

on Trp contents (Fig 3d)

3.2 Determination of non-freezing water (Wnf) in HPMC-Trp hydrogels by

means of DSC

HPMC-Trp cryogels become hydrogels upon immersion in water

The amount of Wnfwas calculated with Eq.(4)(Kim, Lee, & Kim, 2004):

⎜ ⎟

⎝

⎞

⎠

×

H

whereW t is the total water in the hydrogel, the sum W f and W fb

cor-responds to the fraction of free water One should notice that by DSC it

was not possible to discriminate the W f and W fbdue to the low polymer

content

W tcan be determined by Eq.(5):

= ⎛

⎝

⎠

×

m

where m swis the mass of swollen hydrogel at equilibrium

The ratio of the melting enthalpy values of free water in the

hy-drogel ( HΔ endo)and of ice ( HΔ m o = 334 J.g−1) (Bouwstra, Salomon-de Vries, & van Miltenburg, 1995) was calculated to estimate the fraction

of free water The DSC curves were provided asSupplementary Ma-terial SM6.Table 2shows the experimental values of HΔ endodetermined for the HPMC-Trp hydrogels by DSC, theW t , W nf and(W f +W fb)values calculated with Eqs (4)and(5) All HPMC-Trp hydrogels presented

lower W nf values than bare HPMC hydrogels (17.72 %), because the amount of bound water depends strongly on the polymer hydro-philicity For instance, the increase of the degree of substitution (DS) of

carboxymethyl cellulose (CMC) from 0.7 to 1.8 increased the W nfvalues from≈ 75 % to ≈ 90 % (Hatakeyama & Hatakeyama, 2017) However,

among the HPMC-Trp samples, the W nf values tended to increase with the Trp content Higher contents of Trp attached to HPMC were achieved by higher CA:Trp ratios (2.0:0.5, 2.0:1.0 and 2.0:2.0), so that not only the Trp content increased but also the number of carboxylate groups For those samples, the increase of hydrophobic Trp content from 1.17 wt% to 2.18 wt% was accompanied by the increase of hy-drophilic carboxylate groups stemming from CA

3.3 Investigation of water in HPMC-Trp hydrogels through TD-NMR

Fig 5indicates a schematic representation of T2 decay-time dis-tribution of the hydrogen atoms corresponding to the water molecules that arefilling the pores of the cryogels (SEM image) and the excess of water (free water) The intensity of the signal relates to the number of

Fig 6 Adsorption capacity (qe) of (a) MB at pH 7 and (b) RB at pH 2.5, respectively, onto HPMC-Trp hydrogels (30 ± 5 mg dried basis and 10 mL solution) as a function of the equilibrium concentration of (Ce), at (24 ± 1)oC KFvalues determined for MB and RB as a function of (c) Trp content and (d) freezing bound water fraction (Wfb)

hydrogen nuclei of water molecules inside each pore population, while

the rate of decay is associated with the mobility of the molecules

Surface relaxation of the wetting phase is strongly dependent on the

wettability environment within the pores space (Câmara et al., 2020;

Schmidt-Rohr & Spiess, 1994;Vidal, Bernardinelli, Paglarini, Sabadini,

& Pollonio, 2019) For this reason, water in small pores provides short

relaxation time (Fig 5a), while water residing in large pores is related

to longer relaxation times (Fig 5b) Fig 5c represents the expected

decay for water in a matrix containing the two pores populations The

longest relaxation time corresponds to the population of free “bulk”

water Fig 5d shows the distribution of T2 obtained for HPMC-Trp

hydrogels with Trp content ranging from 0 wt% to 2.18 wt%

The corresponding decay curves were previously treated by the ILT,

as shown in theSupplementary Material SM7 For each sample, plots

for water distribution were recorded as soon as water was added to the

cryogel (0’), and after 45 min The characteristic signal at around

2,500−3,000 ms corresponds to the free water (excess of water)

pre-sent in the NMR tube and in the hydrogels Signals at a smaller time

scale were attributed to Wfb The mean T2values and the percentage of

water in each population (area of the bands) were analyzed by

con-sidering the T2distribution at 0’ and 45’, as shown in Supplementary

Material Table SM1 The areas corresponding to each band were

ob-tained through the deconvolution of plots by using Log-Normal

dis-tribution integration, representing the fraction of water free and present

in the pores (Supplementary Material SM8) The signals below 1 ms,

which might be related to bound water (Wnf) (Wang et al., 2020;Wei

et al., 2018), were too weak to be accurately treated with ILT and

ap-peared only in two samples (Supplementary Material SM9)

Except for HPMC-Trp 2.18 wt%, the population of free water

remained practically constant, indicating that the majority of the water was incorporated into the pores of the cryogels very fast (shorter than 3 min) The pattern of the signals for the HPMC-Trp samples differed from the one for pure HPMC hydrogels Without HPMC-Trp 0.73 wt%, the general trend was an increase of Wnfwith the increase of Trp content in the HPMC-Trp hydrogels, as shown inTable 2 The samples with the highest Trp contents (1.38 wt% and 2.18 wt%) presented the shortest

T2values and the distributions of the two populations (small and large pores) became closer The sample HPMC-Trp 2.18 wt% presented an interesting redistribution of the free water population after 45 min of exposition, the band area of free water decreased 2.5 times and the band area related to water into the small pores increased 1.7 times, in comparison to the initial contact

3.4 Adsorption of MB and RB on HPMC-Trp hydrogels

Fig 6a and b shows the equilibrium adsorption capacity (qe) of MB

at pH 7 and RB at pH 2.5, respectively, onto HPMC-Trp hydrogels (30 ± 5 mg dried basis and 10 mL solution) as a function of the equi-librium concentration of (Ce), at (24 ± 1)oC At pH 7, MB has one negative charge in the conjugated nitrogen (Flury & Wai, 2003), whereas at pH 2.5, RB is positively charged (Ramette & Sandell, 1956) (Supplementary Material SM10) The experimental qe values in-creased linearly with Cevalues Fittings with Langmuir and Freundlich models (Supplementary Material SM11) indicated that the adsorption behavior of MB and RBfitted better the Freundlich model Moreover, desorption experiments showed no desorption of MB or RB after 24 h immersion in the respective solvents, impairing the fitting with the Langmuir model The Freundlich model is an empirical model, which

Fig 7 Adsorption capacity qt(mg g−1) as a function of time t (min) determined for MB at (a) 3.0 mg.L−1and (b) 5.0 mg.L−1, and for RB at 2.0 mg.L−1and 4.0 mg.L−1

yields parameters KFand n, the former is related to the adsorptive

ca-pacity and the latter to the surface homogeneity (Foo & Hameed, 2010)

Fig 6c and d shows the KFvalues determined for MB and RB as a

function of Trp content and freezing bound water fraction (Wfb),

re-spectively The general trend for MB and RB was an increase of KF

values with Trp content (Fig 6c) or Wfb(Fig 6d), indicating that the

adsorption capacity correlated well with intermediate water fraction in

hydrogels On the other hand, the dependence of KF values on Wnf

showed a maximum at ≈ 8% for both adsorbates (Supplementary

Material SM12), which corresponds to the highest Trp content

Therefore, the increase of Trp content caused an increase of Wfb, Wnf

and KFvalues The n values showed no dependence on Trp content

(Supplementary Material SM13); for MB and RB, the n values

amounted to 1.01 ± 0.05 and 0.92 ± 0.03, indicating similar chemical

homogeneity of HPMC-Trp hydrogels used for MB and RB adsorption

The qevalue of≈ 1.0 mg.g−1for MB or RB on HPMC-Trp 2.18 wt%

was similar to those determined for MB on tannin-immobilized

cellu-lose hydrogel (qe= 1,1 mg/g) (Pei et al., 2017) and for RB on carbon

nanospheres (qe= 1,15 mg/g) (Qu, Zhang, Xia, Cong, & Luo, 2015)

The qevalue of≈ 1.0 mg.g−1for MB or RB on HPMC-Trp 2.18 wt% was

used for estimating the number of adsorbed MB or RB molecules per Trp

molecules chemically attached to the HPMC hydrogels In one gram of

HPMC-Trp 2.18 wt%, there are≈ 1×10-4moles Trp, whereas in 1.0 mg

of MB and RB there are≈ 3×10-6moles MB and≈ 2×10-6moles MB

Thus, Trp moieties are in excess in comparison to the adsorbate

mole-cules In order to achieve complete saturation of Trp adsorbing sites

with MB and RB molecules, the qe maximal values should be∼ 33

mg.g−1 and 50 mg.g−1, respectively These values are similar to

lit-erature values determined for lignin-based adsorbents (hydrophobic

surfaces) under similar pH and dye dilute range For instance, the qe

maximal values for MB on organosol lignin (Zhang, Wang, Zhang, Pan,

& Tao, 2016) or blends of lignin and chitosan (Albadarin et al., 2017)

amounted to (∼ 20.6 mg.g−1) and (∼ 36 mg.g−1), respectively

Con-sidering dilute range of RB (> 10 mg/L), the adsorption capacity of

HPMC-Trp (50 mg.g−1) was on the same order of magnitude of that

determined for activated carbons obtained from lignocellulosic waste

(39 mg/g) (da Silva Lacerda et al., 2015)

The reusability of the Trp-HPMC hydrogels for RB was evaluated

Supplementary Material SM14 shows the removal capacity of RB (Ci

= 1.0 mg/L, 10 mL) by HPMC-Trp 2.18 wt% after six adsorption cycles

The adsorption time was 10 min, after that the absorbance of

super-natant was measured and the hydrogels were immersed in 10 mL MilliQ

water for 15 min For more concentrated RB solutions (< 2.0 mg/L) the

complete removal was obtained after four consecutive times in contact

with 10 mL of fresh MilliQ water Regardless the initial RB

concentra-tion, the hydrogels were reused six times, keeping the removal

effi-ciency at the original level The reusability of the HPMC hydrogels for

MB adsorption is also feasible by rinsing with HCl 1.0 mol/L; even after

10 recycles, the adsorption capacity was kept at original level (Toledo,

Martins et al., 2019) However, HPMC-Trp hydrogels are advantageous

over pure HPMC because MB molecules undergo pronounced

photo-fading in the presence of Trp molecules (Knowles and Gurnani, 1972)

Reactions involving singlet oxygen and Trp and triplet state of MB and

Trp cause the photofading of MB (Smith, 1978).Supplementary

Ma-terial SM15 shows that upon increasing the Trp content in the HPMC

hydrogels the photofading effect becomes more evident

The adsorption kinetics of MB and RB onto HPMC-Trp hydrogels

was systematically investigated for MB at 3.0 mg.L−1and 5.0 mg.L−1,

and for RB at 2.0 mg.L−1and 4.0 mg.L−1, as shown inFig 7 The data

were fitted with the intraparticle diffusion model (Weber & Morris,

1963), which has been widely applied for the analysis of mass transfer

from solution to the solid-liquid interface and the diffusion of the

ad-sorbate into the porous media:

=

where kintrais the diffusion rate, which is proportional to the diffusion

coefficient of the adsorbate Supplementary Material SM16 provides typical dependence of qton t0.5for MB and RB at 5.0 mg.L−1and 4.0 mg.L−1, respectively; data for more dilute solutions are available as In general, two different slopes were obtained by piecewise linear re-gression (dot lines), which correspond to fast and slow stages The change from the transport rate regime took place after approximately

10 min Supplementary Material Table SM2 shows that the kintra

values decreased one order of magnitude from the 1st to the 2nd ad-sorption stage and increased with the increase of adsorbate con-centration There was a general tendency for the increase of kintravalues with the increase of Trp content in the HPMC-Trp hydrogels The ad-sorption kinetics was also quantitatively evaluated using the pseudo-1st order and pseudo-2nd order models, as shown in theSupplementary Material Table SM3 The adsorption kinetics of MB and RB on HPMC-Trp hydrogelsfitted better with the pseudo-2nd order equation than with the pseudo-1st order equation because thefittings were over a whole time range (30 min) and the calculated qevalues were similar to the experimental data However, the rate constants presented no gen-eral trend regarding the Trp content

Supplementary Material SM17 shows the normalizedfluorescence spectra obtained for RB (solution), MB (solution), HPMC-Trp hydrogels with different Trp contents after 48 h RB or MB adsorption, with the corresponding qevalues The bathochromic shifts of emission maximum

of 8 nm (0 wt% and 0.73 wt% Trp) and 11 nm (1.09 wt% Trp, 1.38 wt% and 2.18 wt%) indicated specific interactions between RB and Trp molecules Trp and dyes might form complexes by van der Waals forces, leading to bathochromic shifts and quenching effects (Doose, Neuweiler, & Sauer, 2005) On the other hand, no significant shift in the emission maximum was observed for MB adsorbed on HPMC-Trp 2.18 wt% hydrogels, only the shoulder at≈ 750 nm presented redshift to ≈

770 nm

4 Conclusions The present study presented the chemical crosslinking and

mod-ification of HPMC chains with citric acid (CA) and Trp by one-step reaction, creating new functional polysaccharides Upon increasing the CA:Trp ratio to 2.0:2.0 in the synthesis, the Trp content in the HPMC cryogels increased to 2.18 wt%, enhancing the compression modulus from 94 ± 5 kPa (pure HPMC cryogel) to 201 ± 9 kPa The insertion of hydrophobic moieties increased the Wnf fraction and the adsorption capacity for RB and MB Trp-HPMC hydrogels presented resilience and reusability Thefluorescence experiments evidenced specific interac-tions between RB and Trp, but they were absent (or very weak) in the case of MB Therefore, the adsorption capacity oh Trp-HPMC hydrogels was favored not only byπ-π interaction among the aromatic rings of Trp and MB or RB, but also by the increase of Wnfportion, or by the

“cage” water molecules This finding is of high relevance because it demonstrated for thefirst time the interdependence between the ad-sorption of water-soluble adsorbates (dyes, drugs, cells, DNA, etc.) and the intermediate water fraction (Wnf), which can be tuned by the degree

of hydrogel modification with hydrophobic moieties (Trp)

CRediT authorship contribution statement

Paulo V.O Toledo: Methodology, Investigation, Data curation Oigres D Bernardinelli: Methodology, Investigation, Data curation Edvaldo Sabadini: Conceptualization, Writing - review & editing Denise F.S Petri: Conceptualization, Writing - review & editing, Supervision, Funding acquisition, Project administration

Acknowledgments Authors gratefully acknowledge financial support from Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq Grant 306848/2017 and 421014/2018) and São Paulo Research Foundation

(FAPESP, Grant 2018/13492-2) This study wasfinanced in part by the

Coordenação de Aperfeiçoamento de Pessoal de Nível Superior - Brasil

(CAPES) - Finance Code 001

Appendix A Supplementary data

Supplementary material related to this article can be found, in the

online version, at doi:https://doi.org/10.1016/j.carbpol.2020.116765

References

Albadarin, A B., Collins, M N., Naushad, M., Shirazian, S., Walker, G., & Mangwandi, C.

(2017) Activated lignin-chitosan extruded blends for efficient adsorption of

methy-lene blue Chemistry Engineering Journal, 307, 264–272

Bian, H., Wei, L., Lin, C., Ma, Q., Dai, H., & Zhu, J Y (2018) Lignin-containing cellulose

nanofibril-reinforced polyvinyl alcohol hydrogels ACS Sustainable Chemical

Engineering, 6, 4821–4828

Bouwstra, J A., Salomon-de Vries, M A., & van Miltenburg, J C (1995) The thermal

behaviour of water in hydrogels Thermochimica Acta, 248, 319–327

Bueno, V B., Bentini, R., Catalani, L H., & Petri, D F S (2013) Synthesis and swelling

behavior of xanthan-based hydrogels Carbohydrate Polymers, 92, 1091–1099

Burdock, G A (2007) Safety assessment of hydroxypropyl methylcellulose as a food

ingredient Food Chemistry and Toxicology, 45, 2341–2351

Câmara, A K F I., Vidal, V A S., Santos, M., Bernadinelli, O D., Sabadini, E., & Pollonio,

M A R (2020) Reducing phosphate in emulsified meat products by adding chia

(Salvia hispanica L.) mucilage in powder or gel format: A clean label technological

strategy Meat Science, 163, 1–11 https://doi.org/10.1016/j.meatsci.2020.108085

Carr, H Y., & Purcell, E M (1954) Effects of diffusion on free precession in nuclear

magnetic resonance experiments Physical Reviews, 94, 630

Çorman, M E., Bereli, N., Özkara, S., Uzun, L., & Denizli, A (2013) Hydrophobic

cryogels for DNA adsorption: Effect of embedding of monosize microbeads into

cryogel network on their adsorptive performances Biomedical Chromatography, 27,

1524–1531

da Silva Lacerda, V., López-Sotelo, J B., Correa-Guimarães, A., Hernández-Navarro, S.,

Sánchez-Báscones, M., & Navas-Gracia, L M (2015) Rhodamine B removal with

activated carbons obtained from lignocellulosic waste Journal of Environmental

Management, 155, 67–76

Doose, S., Neuweiler, H., & Sauer, M (2005) A close look at fluorescence quenching of

organic dyes by tryptophan ChemPhysChem, 6, 2277–2285

Flury, M., & Wai, N N (2003) Dyes as tracers for vadose zone hydrology Reviews of

Geophysics, 41, 1002–1039

Foo, K Y., & Hameed, B H (2010) Insights into the modeling of adsorption isotherm

systems Chemical Engineering Journal, 156, 2–10

Hatakeyama, T., & Hatakeyama, H (2017) Heat capacity and nuclear magnetic

relaxa-tion times of non-freezing water restrained by polysaccharides, revisited Journal of

Biomaterials Science Polymer Edition, 28, 1215–1230

Kaewprasit, C., Hequet, E., Abidi, N., & Gourlot, J P (1998) Application of methylene

blue adsorption to cotton fiber specific surface area measurement: Part I.

Methodology The Journal of Cotton Science, 2, 164–173

Kalaskar, D M., Gough, J E., Ulijn, R V., Sampson, W W., Scurr, D J., Rutten, F J., et al.

(2008) Controlling cell morphology on amino acid-modified cellulose Soft Matter, 4,

1059–1065

Kaur, G., Grewal, J., Jyoti, K., Jain, U K., Chandra, R., & Madan, J (2018) Oral

con-trolled and sustained drug delivery systems: Concepts, advances, preclinical, and

clinical status In A M Grumezescu (Ed.) Drug targeting and stimuli sensitive drug

delivery systems (pp 567–626) William Andrew

Kim, S J., Lee, C K., & Kim, S I (2004) Characterization of the water state of hyaluronic

acid and poly(vinyl alcohol) interpenetrating polymer networks Journal of Applied

Polymer Science, 92, 1467–1472

Knowles, A., & Gurnani, S (1972) A study of the methylene blue-sensitized oxidation of

amino acids Photochemistry and Photobiology, 16, 95–108 https://doi.org/10.1111/j.

1751-1097.1972.tb07341.x

Liu, Y., Shen, X., Zhou, H., Wang, Y., & Deng, L (2016) Chemical modification of

chit-osan film via surface grafting of citric acid molecular to promote the

biominer-alization Applied Surface Science, 370, 270–278

Lochhead, R Y (2017) The use of polymers in cosmetic products, in cosmetic science and

technology In K Sakamoto, R Y Lochhead, H I Maibach, & Y Yamashita (Eds.).

Cosmetic science and technology: Theoretical principles and applications (pp 171–221).

Elsevier

Maleki, H (2016) Recent advances in aerogels for environmental remediation

applica-tions: A review Chemical Engineering Journal, 300, 98–118

Mali, K K., Dhawale, S C., Dias, R J., Dhane, N S., & Ghorpade, V S (2018) Citric acid

crosslinked carboxymethyl cellulose-based composite hydrogel films for drug

de-livery Indian Journal of Pharmaceutical Sciences, 80, 657–667

Marani, P L., Bloisi, G D., & Petri, D F S (2015) Hydroxypropylmethyl cellulose films crosslinked with citric acid for control release of nicotine Cellulose, 22, 3907–3918

Martins, B F., Toledo, P V O., & Petri, D F S (2017) Hydroxypropyl methylcellulose based aerogels: Synthesis, characterization and application as adsorbents for waste-water pollutants Carbohydrate Polymers, 155, 173–181

Pei, Y., Chu, S., Chen, Y., Li, Z., Zhao, J., Liu, S., et al (2017) Tannin-immobilized cel-lulose hydrogel fabricated by a homogeneous reaction as a potential adsorbent for removing cationic organic dye from aqueous solution International Journal of Biological Macromolecules, 103, 254–260

Peng, H., Yang, C Q., & Wang, S (2012) Nonformaldehyde durable press finishing of cotton fabrics using the combination of maleic acid and sodium hypophosphite Carbohydrate Polymers, 87, 491–499

Qu, J., Zhang, Q., Xia, Y., Cong, Q., & Luo, C (2015) Synthesis of carbon nanospheres using fallen willow leaves and adsorption of Rhodamine B and heavy metals by them Environmental Science and Pollution Research, 22, 1408–1419

Ramette, R W., & Sandell, E B (1956) Rhodamine B equilibria Journal of the American Chemical Society, 78(19), 4872–4878

Reddy, N., & Yang, Q (2010) Citric acid cross-linking of starch films Food Chemistry,

118, 702–711

Richard, D M., Dawes, M A., Mathias, C W., Acheson, A., Hill-Kapturczak, N., & Dougherty, D M (2009) L-tryptophan: Basic metabolic functions, behavioral re-search and therapeutic indications Internal Journal Tryptophan Rere-search, 23, 45–60

Schmidt-Rohr, K., & Spiess, H W (1994) Multidimensional solid-state NMR and polymers (1st ed.) New Work: Academic Press

Scotti, K L., & Dunand, D C (2018) Freeze casting – A review of processing, micro-structure and properties via the open data repository, FreezeCasting.net Progress in Material Science, 94, 243–305

Silverstein, R M., Webster, F X., Kiemle, D J., & Bryce, D L (2014) Spectrometric identification of organic compounds (8th ed.) John Wiley & Sons

Smith, G J (1978) Photo-oxidation of tryptophan sensitized by methylene blue Journal

of the Chemical Society Faraday Transactions, 2(74), 1350–1354

Tchobanian, A., Van Oosterwyck, H., & Fardim, P (2019) Polysaccharides for tissue engineering: Current landscape and future prospects Carbohydrate Polymers, 205, 601–625

Toledo, P V O., Limeira, D P C., Siqueira, N C., & Petri, D F S (2019a).

Carboxymethyl cellulose/poly(acrylic acid) interpenetrating polymer network hy-drogels as multifunctional adsorbents Cellulose, 26, 597–615

Toledo, P V O., Marques, L R., & Petri, D F S (2019b) Recyclable xanthan/TiO 2

composite cryogels towards the photodegradation of Cr(VI) ions and methylene blue dye International Journal of Polymer Science, 1–13 https://doi.org/10.1155/2019/

8179842 8179842.

Toledo, P V O., Martins, B F., Pirich, C L., Sierakowski, M R., Teixeira-Neto, E., & Petri,

D F S (2019) Cellulose based cryogels as adsorbents for organic pollutants Macromolecular Symposia, 383, Article 1800013

Tsuruta, T (2010) On the role of water molecules in the interface between biological systems and polymers Journal of Biomaterials Science Polymer Edition, 21, 1831–1848

Türkmen, D., Bereli, N., Derazshamshir, A., Perçin¸, I., Shaikh, H., & Yılmaz, F (2015) Megaporous poly(hydroxy ethylmethacrylate) basedpoly(glycidylmethacrylate-N-methacryloly-(l)-tryptophan)embedded composite cryogel Colloids and Surfaces B: Biointerfaces, 130, 61–68

Ulker, Z., & Erkey, C (2014) An emerging platform for drug delivery: Aerogel based systems Journal of Controlled Release, 177, 51–63

Vidal, V A S., Bernardinelli, O D., Paglarini, C S., Sabadini, E., & Pollonio, M A R (2019) Understanding the effect of different chloride salts on the water behavior in the salted meat matrix along 180 days of shelf life Food Research International, 125, Article 108634

Wang, H., Cui, H., Wang, X., Lin, C., Xia, S., Hayat, K., et al (2020) Metal complexed-enzymatic hydrolyzed chitosan improves moisture retention of fiber papers by mi-grating immobilized water to bound state Carbohydrate Polymers, 235, Article

115967

Weber, J W., & Morris, J C (1963) Kinetics of adsorption of carbon from solution Journal of the Sanity Engineering Division, 89, 31–39

Wei, S., Tian, B Q., Jia, H F., Zhang, H Y., He, F., & Song, Z P (2018) Investigation on water distribution and state in tobacco leaves with stalks during curing by LF-NMR and MRI Drying Technology, 36, 1515–1522

Wimley, W C., & White, S H (1996) Experimentally determined hydrophobicity scale for proteins at membrane interfaces Nature Structural & Molecular Biology, 3, 842–848

Zhang, L., Hailong, L., Yu, J., Wang, H., McSporran, E., Khan, A., et al (2019) Preparation of high-strength sustainable lignocellulose gels and their applications for antiultraviolet weathering and dye removal ACS Sustainable Chemical Engineering, 7, 2998–3009

Zhang, S., Wang, Z., Zhang, Y., Pan, H., & Tao, L (2016) Adsorption of methylene blue on organosolv lignin from rice straw Procedia Environmental Sciences, 31(3), - 11

Zhao, G., Xu, X., Di, Y., Wang, H., Cheng, B., Shi, L., et al (2019) Amino acid clusters supported by cellulose nanofibers for proton exchange membranes Journal of Power Sources, 438, Article 227035