Full-Text Paper (PDF): Ion-Macromolecule Interactions Studied with Model Polyurethanes. ... Article (PDF Available) in J...
Ion-Macromolecule Interactions Studied with Model Polyurethanes Borja Fernández–d’Arlas,*[a],[b] Miguel Ángel Huertos[c],[d] and Alejandro J. Müller[b],[d] [a]
INAMAT (Institute for Advanced Materials) y Departamento de Física, Universidad Pública de Navarra (UPNA), Centro Jerónimo de Ayanz, Campus Arrosadía, 31006-Pamplona, Spain. *e-mail: [email protected]
[b] POLYMAT and Polymer Science and Technology Department, Faculty of Chemistry, University of the Basque Country UPV/EHU, Paseo Manuel de Lardizábal 3, 20018 Donostia-San Sebastián, Spain [c] Grupo de Investigación Química Inorgánica, Departamento de Química Aplicada, Universidad del País Vasco UPV/EHU, Avda. Manuel de Lardizábal 3, 20018, San Sebastian, Spain [d] IKERBASQUE, Basque Foundation for Science, Bilbao, Spain
Abstract Hypothesis The solubility and self-assembly of macromolecules in solution can be tuned by the presence of different salts. Natural proteins have been long manipulated with the aid of salts, and natural silk is processed in the gland tip across a gradient of different salts which modifies its solubility. Hence, the comprehensive understanding of the role of ion-macromolecule interactions should pave the way towards a biomimetic processing of macromolecules. Experiments +
A model polyurethane catiomer (PU ) with high density of hydrogen donors and acceptors (similar to proteins) has been designed and synthesized in order to study ion-macromolecule interactions by means of dynamic light scattering (DLS), infrared spectroscopy (FTIR) and 13 nuclear magnetic resonance ( C-NMR). Findings +
The PU solubility in the presence of different salts exhibited a reversed anion Hofmeister series + (i.e., the anion ability to precipitate the PU was F < Cl < Br < NO3 < CH3COO < H2PO4 < H2CO3 < I < ClO4 < SCN ). The ordering of this series was found to be predicted, for the first time, by the Born-Landé-Ephraim-Fajans-Bjerrum model used here to estimate the degree of macromolecule-ion pairing in water solution. This work also helps understanding the role of cations and anions nature on their interaction with macromolecules backbone.
Keywords: Hofmeister series • polyurethane • ion-macromolecule interactions • selfassembly • aqueous macromolecules
1. Introduction Ion-macromolecule interactions are involved in a large number of physico-chemical[1-9] and biological processes.[9-12] Pioneering works reported in the XIX century by Denis, Bernard and Matthieu and Urbain led to a deep understanding of the impact of certain salts over some protein solutions. With the aim of understanding the nature of the salt-protein interaction, Lewith and Hofmeister carried out a number of experiments to elucidate the role of the salt nature on either stabilizing or coagulating egg white aqueous solutions. These experiments are considered the first systematic studies of ions specificity on the ion-macromolecule interactions. According to the results of Hofmeister et al., the effectiveness of anions of sodium salts in precipitating egg white solutions followed the series: SO42- ≈ HPO42- > CH3COO- > HCO3- > CrO42- > NO3- > ClO3Although not as conclusive as with the anion series, comparing different chloride salts with different cations, Hofmeister et al. also came to the conclusion that the effectiveness of cations on coagulating egg white solutions followed the order: Li+ > Na+ > NH4+ > Mg2+ Extension of ion-macromolecule specific effects with other macromolecules and salts has led to the generalization of the anionic series in terms of its precipitating effectiveness as follows: CO32- > SO42- > H2PO4- > F- > CH3COO- > Cl- > Br- > NO3- > I- > ClO4- > SCNDifferent experiments carried out with different proteins have revealed that when the protons concentration lead to a net positively charged protein (pH < pI) the general tendency is opposite to the Hofmeister series.[18-20] Experiments performed with positively charged latex have shown that inverse series, in terms of precipitating efficiency, are obtained when different salts were added to their solutions.[4,21] At this point it is necessary to recall that when Hofmeister published their works, the concept of pH was still not introduced. The mayor component of egg white is ovoalbumin (pI = 4.5-5.1) and therefore it is expected that in Hofmeister experiments it was fulfilled that pH > pI. The Hofmeister series were first attempted to be explained by the effect of the ions on the three dimensional order of water, classifying them into chaotropes or kosmotropes, depending on their ability to disrupt or to reorganize water molecules, respectively, [22,23]
what would modulate water hydrating capacity. The stability of macromolecules in solutions would be therefore linked to measurable properties of ion-water interactions in solution, such as the viscosity coefficient, B. Nevertheless, different experiments suggest that the presence of ions has a minimal effect on the bulk structure of water far beyond its first shells of hydration.[24,25] In addition, different experiments have demonstrated the relevance of ion-macromolecule interactions for understanding the Hofmeister phenomena, minimizing the importance of water structure and emphasizing the importance of the nature of the solutes and the ions.[26,27,28] The complicated nature of proteins, composed of multiple amino acids, makes difficult the assignment of the nature of the interactions between the ions and the macromolecules that determines their behavior in solution. It is for this reason that studies with synthetic molecules, with a defined structure, are important in order to discriminate the nature of the ion-macromolecule predominating interactions. Studies of López-León et al.[4,21] with latexes positively functionalized with amidinium groups, R1C(=NHR)NR2+, or negatively with sulphate (RSO4-) charges, have pointed out the importance of macromolecules charge on determining the ordering of ions on influencing their stability in solution. Stability of positively charged latexes followed an inverse Hofmeister series, while negatively charged latexes followed a direct Hofmeister series in terms of latexes stability in solution. In addition, López-León et al. have also shown the relevance of the solute hydrophobicity on determining ion-specific effects.[29,30] Zhang et al. have studied the ion specific effects onto the solubility of poly(N-isopropylacrilamide), PNIPAN, chosen as a non-charged polymer model to study the specificity of amide-ion interactions, presumably present in proteins. They have found that the solubility of PNIPAN with the addition of salts follows a direct Hofmeister series (i.e., ClO4- > SCN- > Br- > NO3- > Cl-) and interpreted their observation in terms of different types of interactions between the anions and the PNIPAN amide (-NH-CO-) groups. Similar results have been obtained more recently by López-León et al. with thermally sensitive charged PNIPAM microgels.[31b] Naturally occurring proteins are complex polymers composed of different amino acids which can be characterized by having different types of side groups. Some amino acids like Lys, Hys, Arg, gain positive charges, while some others as Asp and Glu, get a negative net charge after ionization. Therefore, in proteins both electrostatic and dipole-dipole type of interactions with ions might be responsible of their stability in solutions. Fox et al. have rationalized that two mechanism might be responsible of the uncertainty concerning the interaction between proteins and ions in aqueous solution: (i) what attributes of ions and proteins determine the binding sites and the
degree of interactions; and (ii) how do ions alter the structure of solvating water molecules. Here, with the aim of elucidating ions binding sites on a defined synthetic macromolecular structure, a synthetic catiomeric polyurethane (PU+) has been design, synthetized and studied as a model macromolecule. The polyurethane has been designed to have a high density of urethane groups (-NH-CO-O-) in parallel to amide groups in proteins. In addition, the urethane groups are intercalated with one alkylammonium group, which has a positive charge in solution, as certain proteins do at pH values below their isoelectric point. The nature of the interactions between the PU and different salts has been studied by Fourier transformed infrared spectroscopy (FTIR) and nuclear magnetic resonance (13C-NMR).
2. Materials and Methods 2.1.Materials The following chemicals were used in this investigation: Hexamethylene diisocyanate, HDI (Sigma-Aldrich), N-methyldiethanolamine, MDEA (Riedel de Haën), Ethylene glycol, EG (Panreac) LiBr (Panreac), LiCl (Panreac), CH3COONa (Fluka), H2CO3Na (Probus), NaH2PO4 (Probus), NaF (Panreac), NaCl (Panreac), NaBr (Panreac), NaI (Probus), NaSCN (Pantreac), NaClO4 (Panreac), NaNO3 (Probus). The HDI was dried at 110ºC under vacuum for 14 h before being used. MDEA and EG were dried (before use) into a rotary vapor for 2h at 80ºC, with reduced pressure and constant air flux. Salts were used as received, except LiCl and LiBr, which were dried under vacuum at 100 ºC, before being used for the preparation of solutions.
2.2. Polyurethane synthesis. Here two polyurethanes were synthesized. One consisted in a homopolymer (PU0) synthesized by the stoichometrical reaction between 1,6-hexamethylene diisocyanate (HDI) and N-methyl diethanol amine (MDEA). For this 3.516 g of HDI were introduced into a vial and 2.463 g of MDEA were added dropwise while stirring vigorously at room temperature. The second polyurethane (PU3/2) was a copolymer based on HDI, MDEA and ethylene glycol (EG). The synthesis was performed mixing 1.364 g of MDEA with 0.472 g of EG. Afterwards 3.235 g of HDI were added dropwise while stirring vigorously. In both cases, when a sudden rise of temperature occurred, the reaction vial was immersed into an ice bath in order to avoid lateral reactions. After solidification, the vials were further annealed at 90 ºC for 2 h. Data related to polyurethanes structure and thermal properties are gathered in Table 1 of the main manuscript.
2.3. Dynamic Light Scattering (DLS). Dynamic light scattering of PU2/3 in the presence of different ions was performed at a Malvern Correlator, measuring the intensity scattered at 90º from the incident beam. Data was collected from 5 acquisitions of 10 s of duration. The PU2/3-salts solutions were prepared by mixing 0.2 mL of 2 wt% PU3/2 in 0.1 N HCl, with 0.1 mL of each of the salt solution in distilled water, and 0.7 mL of pure distilled water. The total volume of the solutions was of 1 mL and final PU3/2 and salts concentrations were of 0.4 wt% and 0.02 N, respectively. Under these conditions the ratio salt/alkyl ammonium groups in the PU3/2 was of 2.6. Data acquisitions were performed immediately after mixing the PU3/2 and salt solutions.
2.4. FTIR spectroscopy. Analysis of PU-ions interactions by FTIR was carried out with PU0, for having a unique type of carbonyl group. The data was acquired by attenuated total reflectance using an ATR-FTIR (Nicolet Magna 6700) spectrometer averaging 20 scans at a resolution of 2 cm
taken at the
surface of films casted from solutions. The films for the study of the impact of the nature of cations and anions were prepared by mixing 3 mL of PU0 at 5 wt% in HCl 0.1 N with 1 mL of different salts solutions with a concentration of 1.50 ± 0.01 N. In the case of NaSCN (with lower +
solubility) 1.3 mL of a 1.21 N solution was used instead. It was estimated that with this PU /salt ratio every alkyl ammonium or urethane group (summing 0.01 mol/gPU) would have available a pair of incoming ions. Water was evaporated in a vacuum oven with a cycle of 48 h at 60 ºC followed by 2 h at 90 ºC. The experiments carried out to elucidate the impact of counter ion nature and salts (NaBr, LiBr, LiCl) concentration were performed with films prepared by mixing 1 mL of the PU0 solution at 5 wt% in HCl with either 0.05, 0.1, 0.2 and 0.4 salt solutions at 1.5 N. In this case water was evaporated in a vacuum oven with a cycle of 48h at 80 ºC followed by 2h at 90 ºC and additional 22h at 110 ºC
2.5. C-NMR spectroscopy. 13
Nuclear Magnetic Resonance ( C-NMR) spectra were recorded on Bruker AVD 500 spectrometer at room temperature. Data was acquired from solutions prepared by mixing 400 L of PU0 solution at 8 wt% in HCl with 200 L of salt solution in H2O at 1.5 N and 100 L of D2O as an internal reference. The structure of the PU0 along with its solution is gathered in Figure S1 (below).
3. Results and discussion 3.1. Properties of the model polyurethanes.
C-NMR spectra in
The model polyurethanes (PU) conceived for this research consisted on a high urethane density copolymers consisting on quaternizable N-methyl diethanol amine (MDEA) and crystallomorphic ethylene glycol (EG) monomers linked with 1,6hexamethylene diisocyanate (HDI). The ratio of EG/MDEA allowed to control the hydrophobic/hydrophobic balance and PU crystallinity. Table 1 gathers the composition and thermal properties of the synthesized PUs. Both the glass transition and melting temperatures increased with EG content. Upon addition of HCl these PU quaternized through protonation of MDEA and become soluble in aqueous media, as long as the MDEA/EG ratio was maintained larger than 2/3.
3.2. Solubility of PU Catiomers in the Presence of Ions. The relevance of the influence of anion nature on the character of the aqueous dispersions of the PU catiomers (PU+) in diluted HCl, can be readily visualized in the photographs shown in Figure 1a. It is clearly seen that the PU tendency to precipitate from solution (i.e., appearance of turbidity) follows an inverse Hofmeister series (F- < Cl- < Br- < NO3- < CH3COO- < H2PO4- < H2CO3- < I- < ClO4- < SCN-). This trend is similar to that found by López-León et al. for positively charged latexes, functionalized with amidine groups, R1C(=NHR)NR2+, in terms of their stability in solution.[4,21] A similar trend has also been found by us for the TPU homopolymer (HDIMDEA), PU0. Solutions from Figure 1a were analyzed by DLS in order to estimate the impact of the nature of the anions on the hydrodynamic radii of the PU copolymers in solution. Figure 1b represents light intensity dispersed by PU3/2 dispersions, I, effective hydrodynamic diameter, Deffec, and number average hydrodynamic diameter, Dnum, as functions of the anion of the sodium salt employed, and ordered proportionally to the anion charge density, Z2/r. Relevant anion parameters are also listed in Table 2. As can be seen in Figure 1b, the PU3/2 macromolecules in solution gain a maximum hydrodynamic diameter when they are combined with anions whose charge density approximates the density of the positively charged ternary alkyl ammonium, -NR3H+, (see Table 2 data). This phenomenon is especially relevant with anions such as SCN-, I-, ClO4- and HCO3-. Anions with charge density values far away from the alkyl
Table 1. Polyurethanes composition and thermal properties. [c]
Tg [b] [ºC]
Glass transition temperature.
Melting point. Melting enthalpy.
Figure 1. a) Vials containing PU3/2 in the presence of sodium salts of the indicated anions. b) DLS data of the solutions in (a), where the scattered light intensity, I, and effective (Deffc.) and number averaged (Deffc.) diameters are represented for each anion, ordered proportionally to their charge density.
ammonium values do not appear to promote any relevant distortion of the PU in solution, in terms of their solution behavior as studied by DLS. 3.3. Counter Ion Pairing in Solution. According to the Born-Landé-Ephraim-Fajans-Bjerrum model the enthalpy of solution,
HS0, of an ionic compound is related to the radius and charge of the ions according to:
𝑒 2 𝑍+2 𝑒 2 𝑍−2 1 𝑁𝐴𝑒 2 𝑍+ 𝑍− 1 𝛥𝐻𝑆0 = −𝑁( + )(1 − ) + (1 − ) 2𝑟+ 2𝑟− 𝐷 𝑟+ + 𝑟− 𝑛
where N is Avogadro' s number, e the electron charge, Z+ and Z- the charges of the cation and anion, respectively, r+ and r- the radii of the cation and the anion, respectively, D the medium electric constant, A the structural constant of Madelung, and n the Born coefficient. Eq. (1) results from the summation[49b,c] of the solvation energy of the ions in a dielectric medium (first term) and the ionic network reticular energy, according to Born-Landé[49a] (second term). Using the Z2/r data of Table 2 it is possible to estimate the relative enthalpy of dissolution of different ion pairs. Comparing the Z2/r values for NH4+and (CH3CH2)4N+, it can be considered that the charged chemical groups of the TPUs, which might be simplified to (CH3CH2)2CH3NH+, could range between Z+2/r = 6-4 nm-1. Taking the weight values contributions of protons and alkyl CH3CH2-) groups to the Z2/r values, it
Table 2. Selected properties of relevant ions. B -1 [a] (dm ∙mol )
Z /r -1 (nm )
B: Viscosity coefficient. In square brackets the reference [b] from where the data was obtained. Calculated from data [c] of ref. 37, by subtracting -0.032 of bromide. Based on [d] dimensions of hydrated anions. According to ref. 46.
can be estimated that the TPU ionizable groups might have a Z+2/r 4 nm-1. Using this value, normalizing Eq. (1) by Ne2, taking (1-1/D) (1(1/n) ≈ 1, and considering A 2 for all the ion pairs, the relationship between HS0 (nm-1) and the ratio r+/r- has been represented in Figure 2a. It can be observed that the highest HS0 is obtained when r+/r- 1, that is when the cation and anion have similar radii. Data for the considered (anions have been highlighted by symbols. Anion ordering predicted by Eq. (1) is well followed by the empirical data of laser scattered intensity, in Figure 2b. Exceptions might arise from the fact that Eq. (1) consider ions as spheres and do not take into account molecular geometry and other physical phenomena like polarization. In any
Figure 2. Counter ion pairing. a) Enthalpy of solution (relative units) of the different anions with 2 -1 the TPU charged motif, with Z+ /r 4 nm , as predicted by Eq. (1) using data in Table 2. b) DLS data of the scattered light intensity by PU3/2 solutions, I, as function of the ration between the cation and anion radii. c) "Volcano plot" illustrating the degree of interaction between the + quaternary ammonium of the PU with different counter anions, as a function of the differences in their charge densities.
case, the general matching of prediction and data suggests that the clouding (i.e., precipitation) of the TPU solutions with the corresponding counter-anions are related to the formation of more stable ion pairs. The behavior presented in Figures 1b, 2a and 2b, is in accordance with Collins model[3,22] of Matching Water Affinities, which aims to explain the stability of macromolecules in solution in the presence of ions, based on ions mutual affinity in relation to its affinity to hydration by water molecules. Collins observed that salts of monovalent ions releasing heat upon solution (i.e., with negative solubilization enthalpy) were composed on either chaotropic cations or kosmotropic anions (e.g., CsF) or, conversely by kosmotropic cations and chaotropic anions (e.g., LiI). Contrarily, the less soluble salts were composed of anions and cations with relatively similar ionic radii (i.e., the same kosmotropic or chaotropic behaviour). In a parallel way, in charged macromolecules such as proteins or ionomers such as PU+, counter ions would interact preferably with the charged motifs of similar charge density. Therefore, Collins model can explain our observations, as the anions with charge density similar to those of the charged alkyl ammonium groups are causing more shielding of the positively charged TPU and therefore promoting macromolecular agglomeration. This concept is conceptually depicted in Figure 2c by means of a “volcano-plot”, where the degree of counter ionic interaction, is represented, in a qualitative way, against the difference between PU alkyl ammonium and anions charge densities. As can be deduced from afore mentioned comments, the results obtained here for this particular quaternized TPU should not be considered an universal law applicable for all positively charged polymers, since depending on the ammonium functionality the charge density of the cation might vary so dramatically that its affinity for ion-paring can be substantially modified. For example, the trend observed here might not be comparable with that obtained with positively charged motifs such arginine, lysine or the N-terminus of triglicine,[51,52] with much higher charge density in comparison to the quaternized tri-alkyl ammonium of these TPUs. In addition, predictions based on Eq. (1) refer only to enthalpically driven processes, but in the course of ion pairing in aqueous media entropic phenomena related to ions hydration and de-hydration might be very relevant in certain cases. For this, the model of Matching Water Affinities (also known as Fajan's competition principle) preferably relies on the similarity of the free energy of hydration, Ghydr between the considered counter ions in order to predict their affinity: counter ions with similar values of Ghydr would trend to associate more
tightly. As can be observed by studying the Ghydr values listed in Table 2, this consideration also estimates the observed solubility of PU+ in the presence of different anions (i.e., more insoluble and higher scattering intensity when Ghydr(+)-Ghydr(-) ≈ 0) Also, it is necessary to highlight at this point that although the Matching Water Affinities model might partially explain our observations, the electrostatic phenomena might not be the only type of interaction between the ions and macromolecules which might be ascribed to certain particular specificity. Patel et al. and Thormann have studied uncharged, water soluble poly(propylene oxide) polymers (PPO) in aqueous solutions with anions, determining that ion specific effects play an important role on PPO stability and thermal behavior in solution, although this polymer is not electrically charged. Therefore these observations should be explained in terms of ion-dipole interaction between the ions and the macromolecules and, perhaps, to some extent, by certain contribution from water structuring by ions. In order to elucidate the weight of each functional group on the interactions responsible for the visual observations and results from DLS experiments, a set of spectroscopic (NMR and FTIR) analysis was performed and discussed below.
3.4. Polyurethane Backbone-Ion Interactions Studied by 13C-NMR in Solution. The NMR experiments in solution were aimed to elucidate the nature of the preferential ion-macromolecule interactions. In Figure S1 from the Supplementary material, the structure of the PU (PU0) used for this analysis, with the seven
differentiated signals, is depicted. The colloidal behavior of PU0 in the presence of ions is similar to PU3/2. The assignment of these signals in Figure S1 is in agreement with previous analysis of polyurethanes performed by
C-NMR. These signals shifted in
the presence of different ions. Shifts of 1H-NMR bands of peptides in the presence of different concentrations of salts were recently reported by Paterová et al. Figures 3b,c show the carbonyl (C=O) region of
C-NMR spectra of the PU0 in
aqueous solutions of different sodium and chloride salts, respectively. In Figure 3b it can be seen that this signal drifts towards lower chemical shifts as a function of the nature of the anion in the following order: F- > Cl- (≈ blank) > Br- > ClO4- > I-. In fact, in the presence of F-, the carbonyl signal appears downfield compared to the blank, or as compared to that of the PU0 in the presence of Cl-. This is consistent with a
Figure 3. C-NMR analysis of the carbonyl group. a) Molecular structure of PU 0 indicating the studied carbons, C1, C2 and C3. Carbonyl (C1) region for the PU0 in the presence of b) different sodium salts with different anions and c) different chloride salts with different cations. Aliphatic C2 and C3 regions for the PU0 in the presence of d) different sodium salts with different anions and e) different chloride salts with different cations.
higher electronegativity of the F- anion that can attract electrons from the carbonyl carbon. On the contrary, I-, is less electronegative and more polarizable and can provide electrons to the C=O group, moving the signal towards lowers chemical shifts. Studying anion-polymer interactions by
H-NMR with poly(N,N-diethylacrylamide)
(PDEA) as model polymer, Rembert at al. observed reductions of the chemical shifts of the alpha protons close to the carbonyl, in accordance with the shifts observed in the present work for the C=O carbon. On the contrary, the variation of the cations nature (Figure 3c) provokes none or very scarce chemical shifts in the case of LiCl (≈ -0.1 ppm).
The reason for this could be that cations may preferably interact with the C=O oxygen while anions could interact with the C=O carbon (electrophilic), more readily detected by 13C-NMR. Comparing the interactions of ions with C=O with respect to other carbon in the PU backbone, an opposite behavior occurs, as higher chemical shifts are promoted. This behavior can be observed for C2 and C3 carbons in Figures 3d,e. The impact of the anion nature on chemical shift increases follows a sequence of F- < Cl- < Br- < I- (Figure 5a). It should be noted that ClO4- did not promote an increase in chemical shift for C2, but it did in one peak of the C3 region which appears downfield the blank. Similarly, Ipromotes the duplication of the C2 peak, with one shoulder at higher chemical shift than the blank and the other at lower chemical shift. The reason for the peak duplication is not clear at this stage. The multiple peaks exhibited by the C3 carbon might be ascribed to the region-selectivity coming from the quaternary amine and large polarizable anions. As can be seen in Figure 3d, the impact of the cations on the upward chemical shift of the C2 band follows the trend K+ < Na+. The presence of Li+ appears not to promote a significant chemical shift with respect to the blank. The reason for the opposite shifts exhibited by C2, C3 as well as N-H adjacent carbons (see Supplementary information, Fig. S2) with respect to C1, could be explained by partial zwitterionic character of the urethane bond, in parallel to the explanation given by von Hippel for the amide group.[3,58] The urethane amine nitrogen would be partially charged with a positive charge due to electron delocalization along the carbonyl-amine system, and it would preferably interact with low density anions, which would stabilize this resonant structure. The nitrogen adjacent aliphatic carbons would be therefore more electronically unshielded and present higher chemical shifts.
3.5. Ion-Functional Group Interactions Analyzed by ATR-FTIR of PU/salt Films. Figure 4 show different regions of the FTIR spectra of the solid films composed of the PU0 and different sodium salts. As can be seen, the presence of different anions has a dramatic impact on the vibrational modes of the PU0 chemical groups. The upward arrows in Figure 4 indicate the peak of the bands associated to the groups under consideration. All the considered bands, except the N-H bending band in Figure 4c,
Figure 4. Influence of anions as studied by FTIR. Transmittance FTIR spectra of the PU0 in the presence of different sodium salts (10 mmol/gPU): a) N-H stretching region, b) ammonium region, d) N-H bending band, and d) -C-O- stretching region.
shifted towards higher frequencies as the anion density decreased (i.e., from F- to SCN-). As will be seen below, this band also shifts towards higher frequencies when the Na+ is replaced by Li+. The positions of these peaks as a function the parameter Z2/r, proportional to anion charge density,[33,41] are plotted below in Figure 5. It can be observed that, in general the anions with lower charge density shift the bands towards higher wavenumbers. Considering that absorption frequencies, , are proportional to √𝐾/𝜇, being K and , parameters related to the bond strength and the reduced mass, respectively, of the considered chemical group, it is hypothesized that the frequency changes could be related to variations in K (i.e., variations in the bond strength, dipole moment, etc.) rather to variations in reduced mass. This hypothesis is reinforced by the fact that the decrease in charge density is accompanied by an increase in , an effect that should display an opposite trend in comparison to that observed.
The trend of shifts in , follows approximately the same trend as the precipitation of an inverse lyotropic series, i.e., F- < Cl- < Br- < I- < SCN- < ClO4-. The band that displayed the highest shift in frequency with the addition of different anions is related to the quaternary ammonium (-NR3H+) absorption, which shifted 202 cm-1 in the presence of SCN-. The frequency shifts as function of Z2/r of the anions, also reveal the relative variations of different bands. It can be seen that bands related to N-H (i.e., ammonium and N-H stretching) develop about an order of magnitude higher shifts in the presence of low charge density anions than bands related to C-O stretching. This suggests that anion binding to the PU is preferably done via electrostatic binding to ammonium or via N-H dipole interactions. It should be highlighted, nevertheless, that anion binding might also
occur at polarized sites, even in the absence of N-H moieties, as has been determined by Rembert et al., and is also deduced from the
C-NMR analysis of the anion
interaction with C=O carbon, presented in this work. Interestingly, not all bands followed the same trend. The amide II band (N-H bending + C-N stretching), at 1530 cm-1, appears to be shifted towards lower wavenumbers as the anion density decreased, i.e., approximately in the opposite order as explained for the previous bands. As will be discussed below this band also shifted towards higher wave number when sodium or potassium were replaced by lithium (Supplementary Fig. S4). This tendency was also confirmed by comparing peaks shifts with different concentrations of LiBr and NaBr salts (below and Supplementary Fig. S3).
3.6. Assessment of the counter ion effect. The impact of the nature of the salt counter-ions on the degree of interaction of the individual ions with the different chemical groups of the PU was studied with a set of the following three salts: LiBr, LiCl and NaBr. By comparing LiBr with NaBr, the influence of the cation should be highlighted while by comparing LiBr with LiCl, the nature of the anion should be displayed. Sheth et al. studied the influence of LiCl on a non-charged polyurea/polyurethane, concluding that the interaction of Li+ with urea and urethane C=O oxygen, as well as Clinteraction with N-H hydrogen, prevents intermolecular association. Studying different
Figure 5. Peaks frequencies of the maxima of the PU0 spectra in Figure 5, represented as function of anions charge density. In b) the grey points correspond to the higher wavenumber shoulder
model amides Balasubramanian et al. yield a similar conclusion in which the Li+ interaction with the C=O oxygen could play an important role on determining their spectroscopic observations. Here, this analysis is expanded by studying a charged catiomeric PU in which anion-cation specific effects, as those described by Collins, could be manifested, and also because this situation should be more comparable to that of natural proteins below their isolectric point. Figures 6a-c show different FTIR spectra regions of PU0 mixed with different amounts of LiBr. Increasing the amount of LiBr increases the chemical shift of the considered bands. The same trend, although in a different degree, was observed for LiCl salt. Contrarily, in the case of NaBr some bands did not vary as much as in the case on LiBr or LiCl, or shifted towards lower wavenumbers (Supplementary Information, Fig. S4).
In Figures 6d-f the maxima of the considered bands are represented against the salt concentration in the PU, for the different salts. In these plots, the lines represent the fitting to the following equation:
𝜈 (𝑐𝑚−1 ) = 𝜈0 − 𝑐 ∙ [𝑆] +
𝐵 ∙ [𝑆] 𝐾 + [𝑆]
where andcorrespond to the peak frequency of the PU in the presence and absence of salt, respectively, [S], the salt concentration and c, B and K are constants. A similar relation has been used by Cremer et al. in order to model the shifts of 1HNMR frequencies[52,57,62] or variations of phase transition temperatures of elastine-like polypeptides, as a function of salt concentration. The third term of eq. 2 is similar to a Lagmuir isotherm. Fitting parameters obtained for different bands and different salts are gathered in Table 3. Another feature of Figures 6d-f is that shifts appear to be leveled in the range of urethane and ammonium saturation (both groups summing 10 mmol/gPU) suggesting that main PU+/ion interactions are mediated through this functional groups. As can be seen by analyzing Figures 6d-f and data in Table 3, the binding of LiBr (K = 15.7 mmol/gPU) to the ammonium band is much more stronger than NaBr (K = 0.1 mmol/gPU), in spite of the fact that they both share the Br - anion. Nevertheless, the nature of the counter cation is probably decisive. The strong interaction between the Li+ and the C=O oxygen[60,61] might disrupt intermolecular interactions between the C=O and the amines and quaternary ammonium, easing penetration and neutralization of this by the anion. In respect to the difference between LiBr and LiCl, this could arise from the higher affinity of Br- towards quaternary ammonium according to the law of Matching Water Affinities (see Table 2 and Figure 3). The evolution of the amide II (N-H bending + C-N stretching) and backbone C-C-O stretching bands is somewhat different. On one hand the shifts are smaller. On the other, the evolution of the bands with LiBr and LiCl was approximately similar (Figure 6e). Interestingly, the addition of NaBr leads to redshifts when the salt concentration is higher than 1.5 mmol/gPU.
Figure 6. FTIR analysis of the TPU (PU0) as a function of LiBr, LiCl and NaCl concentration (indicated in mmol/gPU). a) Ammonium region, b) Amide II region and c) -C-C-O stretching region as function of LiBr concentration (mmol/gPU). d-f) Frequencies of the maxima absorption of the different bands represented as a function of salt concentration. The dotted lines correspond to fittings to eq.2 (see Table 3).
The shift towards higher wavenumbers of amide II in the presence of Li+ is in agreement with results obtained by Balasubramian with amides and with more recent studies by Zhao et al. on peptides amide II band in the presence of different chloride salts. Accordingly, the blueshifts in the presence of LiBr and LiCl suggest a strong interaction between Li+ and C=O. In this work, a band appearing at 1636 cm-1 when salts containing Li+ were added to the PU (Supplementary Information, Fig. S3) could be associated to a partial induction of an imine bond (C=N) as consequence of the electron delocalization along the urethane bond, motivated by the favored Li+ interaction with C=O.
Table 3. Salt-PU0 binding constants obtained after fitting data to Eq.(2) for LiBr, LiCl and NaBr. Ammonium Amide II C-C-O stretching -1 -1 -1 (0 = 2623 cm ) (0 = 1530 cm ) (0 = 1139 cm ) Salt c B K c B K c B K LiBr
Units: c = cm ∙gPU/mmol; B = cm ; K = mmol/gPU.
On the other hand, the lower shifts in the amide II in the presence of LiBr and NaBr salts with respect to LiCl could be indicative of an increase of the reduced mass along the Br-:::H+-N- vibrational mode, in comparison to Cl-. In this direction, as has been presented in Figures 5c and 6c, a subtle but perceivable shift towards lower frequencies have been detected in the Amide II band in the presence of different sodium salts as a function of the anion in the following order I - < Br- < Cl- < F-. It should be therefore derived that Amide II band shifts in the presence of ions might be determined by the partial formation of imine bands (causing blue shifts) or by the increase in the reduced mass of the X-:::H+-N- vibrational mode (causing redshifts), being X- the anion interacting with the urethane bond. In the case of LiBr and LiCl the former effects would predominate due to the strong polarization of C=O (and subsequent pulling of the urethane to its zwitterion form containing an imide bond) due to strong interaction with Li+, while in the case of NaBr the increase in the reduced mass would be the predominating factor, causing shifts towards lower wavenumbers. Scheme 1 summarizes the most relevant ion-macromolecule interactions as conceptualized from results derived in this work. The addition of a salt with a less hydrated anion than chloride displaces the chloride anion to the bulk water while the low charge density anion (i.e., Br-, I-, SCN- or ClO4-) form a stronger ion pair with the polyurethane positively charged motif. The degree of the subsequent PU+ charge screening would be the predominating phenomena for determining the solubility of the PU+ (Scheme 1a and details in Scheme 1b-d). Both alkylamonium and urethane N-H groups interact preferably with low charge density anions, such as I- or SCN- (Schemes
Scheme 1. Ion exchange and ion-macromolecule pairing. a) Scheme of the substitution of a high density charge anion by the low density charge anion from an added salt (e.g. LiBr), and + precipitation caused by decrease in PU effective charge. b) Illustration of the repetitive unit + schematized (PUo). c) Detail of the PU in solution with the chloride counter anion, forming weak + ion-pairs. d) Detail of the PU with the presence of a salt with high charge density cation (i.e., + Li ) and low density anion (i.e., Br ). e) Scheme of the preferably interaction of the alkylamoniumm group with low density anions and f) Detail of the preferably interaction of the urethane group with low charge density anions (strongly) and high charge density cations (moderately).
9e, 9f). The urethane groups, in addition, also interact moderately with high density charge cations such as Li+.
4. Conclusions The behavior of the synthesized PU+ in solutions corresponds to an inverse lyotropic or Hofmeister series, which could be predicted by the Born-Landé-Ephraim-FajansBjerrum model for the enthalpy of mixing of an ionic pair. This correlation, despite omitting entropic effects, can be related to the ionic pairing between the different anions and the charged quaternary ammonium in the TPU. The ion interaction with the polyurethane backbone in solution has been analyzed by
C-NMR. The Carbonyl
(C=O) carbon band appears to be affected more strongly by anions than cations. Low
charge density anions displace the C=O carbon resonance bands towards lower shifts, indicating a partial electron shielding. This tendency is opposite for carbons adjacent to urethane amine (O-CO-NH-C). Both phenomena suggest the polarization of the urethane bond, which might act as a zwitterion (O-O--C=N+H-C), with the ammonium side stabilized by low charge density counter anions, and the oxygen side stabilized by high charge density cations. FTIR analysis of different polyurethane bands in the presence of different sodium salts revealed their shifts correspond, to a different degree, to the anion ordering observed in macroscopic experiments such as precipitation or dynamic light scattering. The large shifts observed in bands associated to quaternary ammonium suggests that ion pairing, according to the law of Matching Water Affinities, might play an important role in determining the evolution or the PU catiomer in aqueous solution. The analysis of the bands related to urethane N-H, suggests a higher interaction of the urethane bond with low charge density anions (such as I-) and high charge density cations (such as Li+), in parallel to what has been reported for amides in synthetic polyamides, peptides[26,65] and proteins. On the view of these results, the polyurethane structure might be considered as a biomimetic material which can be designed appropriately to mimic the aqueous behavior of some natural macromolecules such as certain proteins, and reproduce its solubility in the presence of different ionic species. The result presented here help filling the gap between the observed phenomenology of macromolecules in ionic solutions (i.e., salting-in or salting-out) and the role of ion-macromolecule interactions. This work shed light on understanding ion-macromolecule interactions and determining key macromolecule features that influence the aqueous behavior of synthetic and natural macromolecules in the presence of ions.
Acknowledgments BFD wants to acknowledge "Fundación Caja Navarra" and "Obra Social La Caixa" in the framewok of UPNA's program "Captación de Talento" for funding during the data analysis and writing of this manuscript.
P. S. Denis in Mémoire sur le sang (Ed. Acd. Sci. Baillière), Paris (France), 1858.
J. L Poiseuille, Ann. Chim. Phys. 1847, 21, 76–110.
K. D. Collins, Biophys. J. 1997, 2, 65-76.
(a) T. López-León in Efectos Hofmeister en sistemas coloidales. Ph. D Thesis, Universidad de Granada (Spain), 2004. (b) T. López-León, A. B. Jódar-Reyes, J. L. Ortega-Vinuesa, D. Bastos-González, J. Coll. Interf. Sci. 2005, 284, 139-148.
V. S. J. Craig, Curre. Op.Coll. & Interf. Sci. 2004, 9, 178–184.
C. C. Díaz-Ramírez, L. Villafuerte-Robbles, Rev. Mex. Cienc. Fram. 2011, 43, 31-45.
S. C. Flores, J. Kherb, P. S. Cremer, J. Phys. Chem. C 2012, 116, 14408-14413.
S. Murgia, M. Monduzzi, B. W. Ninham, Curr. Opin. Coll. Interf. Sci. 2004, 9, 102-106.
N. Singh, M. Kumar, J. F. Miravet, R. V. Ulijn, B. Escuder, Chem. Eur. J. 2017, 23, 981 – 993
S. P. L. Sørensen, Compt. rend. du Lab. de Carlsberg, 1909, 8, 1–168.
L. Lorenzo-Andreu, An. Estac. Exp. Aula Dei 1951, 2, 174-186.
J. J.García-Celma, L. Hatahet, W. Kunz, K. Fendler, Langmuir 2007, 23, 10074-10080
C. Bernard in Traité de chimie anatomique (Ed. C. Robin, F. Verdeil), París (France), 1853, pp. 299.
E. Mathieu, V. Urbain, Compt. Rend. L´Acad. Sci, 1874, 79, 665-698.
S. Bonilla-Mirat in Tratado Elemental de Química General (Ed. Impr. Hijos de Rodríguez), Valladolid (Spain), 1881, pp. 692-714.
(a) S. Lewith, Arch. Exptl. Pathol. Pharmakol. 1887, 24, 1-6; (b) F. Hofmeister, Arch. Exptl. Pathol. Pharmakol. 1887, 24, 247-260; (c) F. Hofmeister, Arch. Exptl. Pathol. Pharmakol. 1888, 25, 1-30; (d) F. Hofmeister, Arch. Exptl. Pathol. Pharmakol. 1891, 28, 210-238
Y. Marcus, Chem. Rev. 2009, 109, 1346-1370.
L. Medda, C. Carucci, D. F. Parsons, B. W. Ninham, M. Monduzzi, A. Salis, Langmuir 2013, 29, 15350-15358
L. Zhang, H. Tan, R. M. Fesinmeyer, C. Li, D. Catrone, D. Le, R. L. Remmele, J. J. Zhang, Pharm. Sci. 2012, 101, 965–977.
Y. Zhang, P. Cremer, PNAS 2009, 106, 15249-15253.
T. López-León, A. B. Jódar-Reyes, D. Bastos-González, J. L. Ortega-Vinuesa, J. Phys. Chem. B 2003, 107, 5696-5708.
K. D. Collins, G. W. Neilson, J. E. Enderby, Biophys. Chem. 2007, 128, 95-104.
B. Hribar, N. T. Southhall, V. Vlachy, K. A. J. Dill, J. Am. Chem. Soc. 2002, 16, 1230212311.
A. W. Omta, M. F. Kropman, S. Woutersen, H. J. Bakker, Science 2003, 5631, 347-9.
W. Wachter, W. Kunz, R. Buchner, G. Hefter, J. Phys. Chem. A 2005, 109, 8675-8683.
J. Peterova, K. B. Rembert, J. Heyda, Y. Kurra, H. I. Okur, W. R. Liu, C. Hilty, P. S. Cremer, P. Jungwirth, J. Phys. Chem. B 2013, 117, 8150-8158.
I. A. Katime, C. Abad, A. Campos, Afinidad 1983, 40, 257-261.
C. Calero, J. Faraudo, D. Bastos-González, J. Am. Chem. Soc. 2011, 133, 15025-15035.
T. López-León, J. L. Ortega-Vinuesa, D. Bastos-González, A. Elasïssari, J. Phys. Chem. B 2006, 110, 4629-4636.
T. López-León, M. J. Santander-Ortega, J. L. Ortega-Vinuesa, D. Bastos-González, J. Phys. Chem. C 2008, 112, 16060–16069.
(a) Y. Zhang, S. Furyk, D. E. Bergbreiter, P. S. Cremer, J. Am. Chem. Soc. 2005, 127, 14505-14510. (b) T. López-León, J. L. Ortega-Vinuesa, D. Bastos-González, A. Elaissari, J. Coll. Interf. Sci. 2014, 426, 300-307.
J. M. Fox, K. Kang, W. Sherman, A. Héroux, G. M. Sastry, M. Baghbazadeh, M. R. Locket, G. M. Whiteside, J. Am. Chem. Soc. 2015, 137, 3859-3866
B. Fernández-d’Arlas, A. J. Müller, Rev. Iber. Polím. 2017, 18, 9-20.
R. A. Robinson, R. H. Stokes in Electrolyte Solutions, (Eds. Butterworths Scientific Publications) London (England), 1959.
J. F. Shackelford, W. Alexander in CRC Materials Science and Engineering Handbook, 3rd ed. (Eds. CRC Press), Boca Ratón, Florida, 2001.
M. Y. Kiriukhin, K. D. Collins, Biophys. Chem. 2002, 99, 155–168.
C. Franco, J. M. Costa in Aplicación tratamiento del estado de transición a la viscosidad de disoluciones de Et4NBr en mezclas de acetonitrilo-agua. XVIII Reunión Real Sociedad Española de Física y Química, Comunication 17.3. Burgos, 29 Sep.-3 Oct., 1980.
Y. Marcus, Pure &AppI. Chem. 1987, 59, 1093-1101.
G. Rayner-Canham, T. Overton in Descriptive Inorganic Chemistry, 4 ed., (Eds. W.H. Freeman Publishing Co), New York (USA) 2006.
F. David, V. Vokhmin, G. J. Ionova, Mol. Liq. 2001, 90, 45–62.
B. Tansel, Separat. Purif. Technol. 2012, 86, 119-126.
D. C. Harris in Quantitative Chemical Analysis, 6th ed., (Eds. W.H. Freeman Publishing Co.), New York (USA), 2003.
J. B. Robinson, J. M. Strottmann, E. Stellwagen, Proc. Natl. Acad. Sci. USA, 1981, 78, 2287–2291.
P. E. Mason, G. W. Neilson, C. E. Dempsey, A. C. Barnes, J. M. Cruickshank, PNAS 2003, 100, 4557– 4561.
J. E. House in Inorganic Chemistry, (Eds. Academic Press), Burlington, Massachusets (USA), 2008.
Y. Marcus in Ion Properties (Eds. Marcel Dekker), New York (USA), 1997.
G. Georgievich in Electrolites: Supramolecular interacctions and non-equilibrium phenomena in concentrated solutions (Edt. CRS Press), Boca Ratón, Florida, 2014.
M. Y. Kiriukhin, K. D. Collins, Biophys. Chem. 2002, 99, 155-168.
a) M. Born, A. S. Landé, B. preuss. Akad. Wiss. 1918, 76, 1048-1068. b) N. Bjerrum, Berch. Deuts. Chem. Gell. 1929, 5, 1090-1103. c) E. Gutiérrez-Ríos in Química Inorgánica (Edit. Reverté), Barcelona, (Spain), 1994, pp. 131-140.
P. E. Mason, J. Heyda, H. E. Fischer, P. Jungwirth, J. Phys. Chem. B 2010, 114, 13853– 13860.
P. E. Heyda, T. Hrobárik, P. Jungwirth, J. Phys. Chem. A 2009, 113, 1969–1975
J. Paterová, K. B. Rembert, J. Heyda, Y. Kurra, H. I. Okur, W. R. Liu, C. Hilty, P. S. Cremer, P. Jungwirth. J. Phys. Chem. B 2013, 117, 8150−8158.
a) K. Fajans, Naturwiss. 1921, 9, 729-738; b) D. F. C. Morris, Struct. Bond 1968, 6, 157159.
K. Patel, B. Bharatiya, Y. Kadam, P. J. Bahadur, Surf. Deter. 2010, 13, 89-95.
E. Thormann, RSC Advances 2012, 2, 8297–8305.
a) W. Tang, R. J. Farris, W. J. MacKnight, C. D. Eisenbach, Macromolecules, 1994, 27, 2814-2819; b) S. Zhang, L. Cheng, J. Hu, J. Appl. Polym. Sci. 2003, 90, 257-260; c) B. S. Valdés, M. Guerra, G. Mieres, L. Alba, A. Brown, N. A. Rangel-Vázquez, M. Sosa, Y. de la Hoz, Rev. Iber. Polím. 2008, 9, 377-388.
K. B. Rembert, H. I. Okur, C. Hilty, P. S. Cremer, Langmuir 2015, 31, 3459-3464.
P. H. Von Hippel, V. Peticolas, L. Schack, L. Karlson, Biochemistry 1973, 12, 1256-1264.
H. I. Okur, J. Hladílkova, K. B. Rembert, Y. Cho, J. Heyda, J. Dzubiella, P. S. Cremer, P. Jungwirth, J. Phys. Chem. B 2017
S. P. Sheth, A. Aneja, G. L. Wilkes, Polymer 2004, 45, 5979-5984.
D. Balasubramanian, R. Shaikh, Biopolymers 1973, 12, 1639-1650.
K. B. Rembert, J. Paterova, J. Heyda, C. Hilty, P. Jungwirth, P. S. Cremer, J. Am. Chem. Soc. 2012, 134, 10039-10046.
Y. Cho, Y. Zhang, T. Christensen, L. B. Sagle, A. Chilkoti, P. S. Cremer, J. Phys. Chem. B 2008, 112, 13765-13771. I. Langmuir, J. Am. Chem. Soc. 1917, 39, 1848-1906.
J. Zhao, J. Wang, J. Phys. Chem. B 2016, 120, 9590–9598.
a) J. F. Reuther, J. D. DeSousa, B. M. Novak, Macromolecules 2012, 45, 7719-7728; b) J. E. B. García, J. L. I. Palafox, M. A. V. Rodríguez, R. A. V. García, J. V. G. Gómez, Rev. Inter. Contam. Ambie. 2015, 31, 415-426; c) G. H. Téllez, S. Bernès, A. Mendoza, F. J. R. Merino, G. E. Moreno, O. Portillo, R. Gutiérrez, Acta Cryst. 2016, 72, 350-354.