Nanoscale Mobility of Aqueous Polyacrylic Acid in Dental Restorative

Subscriber access provided by UNIV OF DURHAM

Biological and Medical Applications of Materials and Interfaces

Nanoscale Mobility of Aqueous Polyacrylic Acid in Dental Restorative Cements Marcella Cabrera Berg, Ana Raquel Benetti, Mark T. F. Telling, Tilo Seydel, Dehong Yu, Luke L. Daemen, and Heloisa N. Bordallo ACS Appl. Mater. Interfaces, Just Accepted Manuscript • DOI: 10.1021/acsami.7b15735 • Publication Date (Web): 05 Mar 2018 Downloaded from http://pubs.acs.org on March 5, 2018

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

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

Page 1 of 33

ACS Applied Materials & Interfaces

1 2 3 4 5 6 7 8 9 10 Marcella C. Berg1,2,*, Ana R. Benetti3, Mark T. F. Telling4,5, Tilo Seydel6, Dehong Yu7, Luke L. Daemen8, Heloisa N. 11 12 1,2 13 Bordallo 14 15 1 16 The Niels Bohr Institute, University of Copenhagen, DK-2100, Copenhagen, Denmark 17 18 19 2 European Spallation Source ESS ERIC, PO Box 176, SE-221 00 Lund, Sweden 20 21 22 3 Department of Odontology, Faculty of Health and Medical Sciences, University of Copenhagen, DK-2200, 23 24 25 Copenhagen, Denmark 26 27 28 4 ISIS Facility, Rutherford Appleton Laboratory, Chilton, Oxon, UK OX11 0QX 29 30 31 5 Department of Materials, University of Oxford, Parks Road, Oxford, UK 32 33 34 6 35 Institut Max von Laue - Paul Langevin, CS 20156, F-38042 Grenoble, France 36 37 7 38 Australian Nuclear Science and Technology Organisation, New Illawarra Road, Lucas Heights, NSW 39 40 41 8 Oak Ridge National Laboratory, P.O. Box 2008, Oak Ridge, TN 37831 42 43 44 45 Keywords: conventional Glass Ionomer Cements, confinement, nanoscale mobility, quasi-elastic neutron scattering, 46 47 48 inelastic incoherent neutron scattering, aqueous solution of polyacrylic acid. 49 50 51 52 53 54 55 56 57 58 59 ACS Paragon Plus Environment 60

Nanoscale Mobility of Aqueous Polyacrylic Acid in Dental Restorative Cements

1


Page 2 of 33

1 2 3 4 ABSTRACT: Hydrogen dynamics in a time range from hundreds of femtoseconds (fs) to nanoseconds (ns) can be 5 6 directly analyzed using neutron spectroscopy, where information on the inelastic and quasi-elastic scattering, hereafter 7 8 9 INS and QENS, can be obtained. In this study, we applied these techniques to understand how the nanoscale mobility 10 11 of the aqueous solution of polyacrylic acid (PAA) used in conventional Glass Ionomer Cements (GIC) changes under 12 13 confinement. Combining the spectroscopic analysis with calorimetric results we were able to separate distinct motions 14 15 16 within the liquid and in the GIC’s. The QENS analysis revealed that the self diffusion translational motion identified 17 18 in the liquid is also visible in the GIC. However, as a result of the formation of the cement matrix and its setting, both 19 20 translational diffusion and residence time differed from the PAA solution. When comparing the local diffusion 21 22 obtained for the selected GIC, the only noticeable difference was observed for the slow dynamics associated to the 23 24 25 polymer chain. Additionally, over short-term ageing, progressive water binding to the polymer chain occurred in one 26 27 of the investigated GIC. Finally, a considerable change in the density of the GIC without progressive water binding 28 29 indicates an increased polymer crosslinking. Taken together, our results suggest that accurate and deep understanding 30 31 32 of polymer-water binding, polymer crosslinking as well as material density changes occurring during the maturation 33 34 process of GIC are necessary for development of advanced dental restorative materials. 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 ACS Paragon Plus Environment 60

2

Page 3 of 33


INTRODUCTION 1 2 3 4 5 Oral health is an integrated part of the general public health, and does not only affect the quality of life, but also the 6 7 8 social healthcare systems through related economic costs. Despite great progress globally in oral health related issues, 9 10 dental caries is still one of the major problems, affecting 60-90% of schoolchildren and the vast majority of adults. 11 12 Therefore, dental restoration takes one of the most significant shares in the health care cost in industrialized 13 14 1,2 15 countries . Consequently, developing and improving dental restorative materials is an important matter. 16 17 The dental restorative materials known as glass ionomer cements (GIC) are acid-based and widely recognized as 18 19 being biologically active 3–5. The base for GIC is a fluoroaluminosilicate glass powder, formed by different cations, 20 21 such as calcium, strontium and aluminium, and the aqueous solution contains a polyacrylic acid (PAA). PAAs are 22 23 24 weak polyelectrolytes with many carboxyl groups that partially dissociate in water, leaving the PAA main-chain with 25 26 negative charges and ionized hydrogen ions in the bulk solution. PAA is an essential industrial polymer also used as 27 28 emulsifying agents, in pharmaceuticals, in cosmetics and in paints and have a particularly important role in 29 30 6 31 dispersants . Previous studies on liquid PAA solutions used in preparation of GIC similar to the ones studied here 32 33 showed two distinct water populations, one unbound and in its native bulk water state (hereafter called ‘bulk water’) 34 35 and the other hindered by the PAA chain with glass behavior7 (hereafter called ‘glassy water’). It is well known that 36 37 3 38 these distinct populations influence the properties of the GIC . During the first steps of the GIC setting process, water 39 40 will act as the reaction medium where the leached cations cross-link with the PAA, thus forming the polyacrylate 41 42 matrix. The conformation and flexibility of the polyelectrolyte chain will in turn govern the degree of crosslinking in 43 44 the polyacrylate matrix thus influencing the cement properties. Furthermore, the water also hydrates the siliceous 45 46 8 47 hydrogel and the metal polyacrylate salt, which both become an essential part of the cement structure . The 48 49 conformation of this structural water is evidenced by the vibrational motions of the water molecules moving around 50 51 their equilibrium position9. If water is lost from the cement by desiccation upon setting, the cement-forming reaction 52 53 54 will either stop or be retarded. On the contrary, in the first 3 to 6 minutes of the cement’s hydration, initial contact 55 56 with water can be damaging, and therefore the surface of the restoration is kept protected3,10. During maturation, bulk 57 58 water can remain unbound, become bound or part of the cement structure. Loosely water associated with the hydration 59 ACS Paragon Plus Environment 60 3

ACS Applied Materials & Interfaces 9,11

Page 4 of 33

12

or confined in the cement pores with a bulk-like behaviour with slight 1 shell of the cation-polyacrylate 2 3 confinement and will hereafter be referred to as ‘confined bulk-like’. Water tightly bound in hydration regions around 4 5 the polymer chain will be referred to as ‘confined glassy water’. The loosely bound water is liable to change which 6 7 13,14 , unless the cement is in an 80% relative humidity9. In addition, with ageing, the 8 can lead to cracks in the material 9 10 pore structure changes12, and as such it is believed that the ratio of confined water to loosely bound water 11 12 increases11,14–16. 13 14 15 3,17 and neutron scattering4,12 have indeed confirmed 16 Previous work using Fourier Transform Infrared Spectroscopy 17 18 the presence of distinct water populations in GIC. As in other systems, these populations are expected to demonstrate 19 20 distinct dynamic behaviours due to the complex GIC pore structure.18,19 However, to date, information on how the 21 22 water dynamics change under confinement and throughout GIC maturation is scarce20–24. For example, information is 23 24 25 still lacking regarding how much liquid remains in its original bulk-state and how the pore structure evolution changes 26 27 the dynamics of the aqueous solution, in turn influencing the cement properties. To deepen this understanding, we 28 29 used neutron spectroscopy to probe the dynamics of bulk and confined PAA solution in two selected GIC from the 30 31 32 same manufacturer at different ages of maturation. One cement, herein referred to as Poly, was mixed mechanically 33 34 using an aqueous PAA solution21, and the other, where the glass powder contains vacuum dried PAA (referred to as 35 36 Aqua cement), was hand-mixed with distilled water23. These particular GICs were selected for this study since each 37 38 39 has a distinctive mechanical property. Indeed based on flexural strength measurements the Aqua cement shows lower 40 12 41 strength that is however constant over the whole maturation period . Furthermore, based on previous knowledge, the 42 43 protons bind differently to the microstructure12; which is to a certain extent controlled by the nature of the aqueous 44 45 PAA solution7. Finally, one must also taken into account that the structure and the coiled nature of the aqueous PAA, 46 47 25–27 for the solution under investigation, while the porosity of 48 which depends on its pH- value, is mostly semi-recoiled 49 50 the GIC studied here ranges from micropores (smaller then 20 Å) to macro pores (bigger then 500 Å)12. 51 52 53 By understanding the different types of confinement that both the water and the pure PAA undergo over time, 54 55 56 unique insight into the distribution of the aqueous populations in GIC cements can be obtained. To this end, neutron 57 58 spectroscopy was desirable, since the neutron is not only able to penetrate deep into matter without altering 59 ACS Paragon Plus Environment 60 4

Page 5 of 33


1 composition, but also the high incoherent scattering cross section of the hydrogen atom makes the technique ideal for 2 3 extracting proton self correlations (dynamics) in hydrogenous materials. See Table S1 in the supplementary 4 5 information. Consequently, insight into the hydrogen bond network (i.e. the hydrogen bond between the base glass6 7 8 silicate and the aqueous PAA) and distinct states of water can be gleaned. In particular, incoherent neutron scattering 9 10 provides geometrical information on the molecular confinement of diffusive motions. Deuteration was however not 11 12 considered as means of separating different dynamics since it might influence the hydration process, the cement 13 14 4,28 15 setting reaction and consequently liquid diffusion . Furthermore, by combining thermal analysis and neutron 16 17 spectroscopy, a better understanding of the hydrogen-bond network formation was achieved, in addition to how 18 19 density changes influence this unique type of cement. 20 21 22 23 24 EXPERIMENTAL METHODS 25 26 27 Cement sample preparation: The two GICs described in the introduction were prepared following the manufacturer 28 29 specification. The Poly cement (Ionofil Molar AC, Voco GmbH, Germany) in an encapsulated version was prepared 30 31 32 by mixing the GIC powder with its respective liquid in a powder-to-liquid ratio of 3.7:1. Using a mechanical RotoMix 33 34 Capsule Mixing Device (3M ESPE AG, Germany), the mixture was agitated for 10 seconds followed by 3 seconds 35 36 centrifugation. The Aqua cement (Aqua Ionofil Plus, Voco GmbH, Germany) was hand-mixed using a powder-to37 38 39 liquid weight ratio of 5.6:1. Preparation was carried out at room temperature. 40 41 42 For the neutron experiments, 1 g of material was used. After mixing, the samples were uniformly distributed in 43 44 individual aluminium foil envelopes, and mounted in a flat aluminium sample holder. The sample holders were sealed 45 46 47 securely using Indium wire and aged for either 7 or 28 days at body temperature (37 °C). For the thermal analysis, 48 49 triplicates of each sample were prepared following the procedure described above. On the day of the thermal analysis 50 51 measurement, the sample holders were opened; one of the aged triplicates was gently ground using a pestle and 52 53 54 mortar, with the 2 other samples resealed. This procedure was repeated before each thermogravimetric analysis and 55 56 Fourier transform infrared spectroscopy (TGA-FTIR) experiment, the samples weights being approx. 40 mg. 57 58 59 ACS Paragon Plus Environment 60

5


Page 6 of 33

1 Thermogravimetric Analysis and Fourier Transform Infrared Spectroscopy (TGA-FTIR): TGA is a method by 2 3 which changes in the mass of a material is followed as a function of temperature. When coupled to the FTIR device, a 4 5 technique used to obtain an infrared spectrum of the sample absorption, any decomposing components in the material 6 7 8 can be identified. For this work, TGA-FTIR was used to characterize different proton states, i.e. bulk-like vs. bound 9 10 water, and thus indirectly obtain an estimate of the amount of structural hydrogen; an information crucial for the 11 12 analysis of the neutron data. 13 14 15 16 All samples were measured using a TG 209 F1 Libra PERSEUS from Netzsch coupled to a Fourier Transform 17 18 Infrared Spectrometer from Bruker Optics. The experimental conditions were: N2-atmosphere (20 mL/min), heating 19 20 rate of 10 K/min. The measurements were performed on the ground cement samples (described above) between 20 °C 21 22 23 (293 K) and 900 °C (1173 K) in a standard Al2O3 crucible. An empty crucible was also measured for instrument 24 25 correction. The data was processed using the software provided by Netzsch. A FTIR spectrum of the developed gases 26 27 was recorded for every 3 °C during the entire measurement. From the collected FTIR data, a selection of spectra at 28 29 30 temperatures of interest, based on the TGA data, was further analyzed. All 7 day-old materials were measured on the 31 32 same day. Due to the reproducibility of this triplicate data, only one 28 day-old sample was measured. The powder 33 34 samples were studied as provided by the manufacturer. 35 36 37 38 Neutron spectroscopy: Neutron spectroscopy is the ideal tool to study the structure and dynamics of water in 39 40 confinement, since the translational frequencies will change with the local water structure. Furthermore, since the 41 42 interatomic forces determine a particular phonon mode, the strength of the H-bond interaction can be obtained. 43 44 45 In neutron spectroscopy the scattering function, (, ), is the observable quantity of interest, where Q is the 46 47 magnitude of the scattering wave vector and ω is the energy transfer. (, ) can be divided into three spectroscopic 48 49 components: the elastic scattering function, (, = 0), which is related to the thermal fluctuations of the atoms 50 51 52 around their equilibrium position, the quasi-elastic signal, (, ≈ 0), that describes molecular translation and 53 54 rotational motions, and the inelastic signal, (, || >> 0), that probes phonon and molecular vibrations. In

55 56 57 hierarchical porous structures, such as in the cements studied here, proton nanoscale mobility occurs over different 58 59 timescales. Consequently, investigations using different neutron spectrometers, which probe different temporal ranges, ACS Paragon Plus Environment 60

6

Page 7 of 33


1 are necessary. As a result, this study was carried out using, four neutron instruments: VISION (giving information on 2 3 molecular vibrations occurring over a broad energy range from 1 to 1000 meV corresponding to the femtosecond, fs, 4 5 domain), PELICAN (covering the high picosecond range, ~ 60 ps, and allowing for observation of fast localized 6 7 8 motions of the confined or chemically bound hydrogen atoms), IRIS (covering the low picosecond, ~ 200 ps, time 9 10 scale thus allowing localized dynamics to be followed) and IN16B (covering slower dynamics in the nanosecond 11 12 regime; i.e ~ 5 ns ). Each instrument is described in detail below. 13 14 15 The vibrational spectrometer VISION, located at Spallation Neutron Source, Oak Ridge, USA follows those neutrons 16 17 that scatter inelastically from a sample with an exchange of energy, (, > 0), that is equal to vibrational states. 18 19 The scattered neutrons lose energy by exciting the vibrational modes of the scatterer, consequently the spectrum is 20 21 22 obtained in the neutron energy loss side (Stokes side). As the intensity of an INS band is proportional to the Debye– 23 24 Waller factor, whose magnitude is in part determined by the thermal motion of the molecule, this factor can be 25 26 reduced by collecting data at 10 K. Furthermore, the INS spectrum is directly proportional to the phonon density of 27 28 states and scattering lengths of the associated atoms. Because selection rules are not involved, INS measures all the 29 30 31 modes simultaneously. Thus, by comparing the INS spectrum obtained for each sample at different maturation 32 33 periods, it is possible to probe changes in the vibrational response as a function of cement age. Such information gives 34 35 insight into the molecular properties and development of the hydrogen bond network. 36 37 38 In addition to the INS data, confined and chemically bound hydrogen dynamics in the dental cements studied in this 39 40 work were probed via a comprehensive investigation of the QENS signal, (, ≈ 0). To obtain such information, 41 42 QENS measurements were performed over a broad temporal range by combining data obtained using PELICAN, IRIS 43 44 45 and IN16B. 46 47 48 The time-of-flight cold neutron spectrometer, PELICAN, located at Australian Nuclear Science and Technology 49 50 Organisation (ANSTO) in Australia, provides an elastic energy resolution, ∆E, corresponding to an upper 51 52 53 experimental observation time of ~ 60 ps and covers an angular scattering range of 22° < 2θ < 119°. In this study an 54 55 incident wavelength of λ = 6.0 Å was chosen, corresponding to ∆E = 63 µeV at full width at half maximum (FWHM), 56 57 in order to probe the fast dynamics mostly due to the loosely bound water either in the PAA or confined in the cement 58 59 ACS Paragon Plus Environment 60 structure.

7


Page 8 of 33

The backscattering spectrometer IRIS, located at the ISIS Facility in the UK, provides an elastic energy resolution of 1 2 3 17.5 µeV at FWHM, corresponding to an upper experimental observation time of ~200 ps, achieved by neutrons 4 5 scattered with λ = 6.7 Å. This resolution is constant over the entire Q-range of 0.501 to 1.960 Å−1 (angular scattering 6 7 8 range of 25° < 2θ < 160°). IRIS probes the intermediate dynamics of the loosely and tightly bound water. 9 10 11 Finally, the high-energy-resolution analyser (∆E = 0.75 µeV at FWHM, λ = 6.3 Å) provided by the backscattering 12 13 spectrometer IN16B, located at ILL in France Grenoble, allows an upper experimental observation time of ~ 5 ns. This 14 15 16 time window corresponds mostly to the slower polymer-related motion. The instrument resolution is constant over the 17 18 entire Q-range of 0.190 to 1.895 Å−1 (angular scattering range of 11° < 2θ < 142°). IN16B was used here with its 19 20 standard Si(111) monochromator and analyzer crystal setup29. 21 22 23 24 For all experiments, the angle between the plane of the sample and the incident neutron beam was 135°. 25 26 Consequently, the last detectors were self-shielded by the edge of the sample holder and thus discarded from analysis. 27 28 The data was reduced using the software packages LAMP30 (PELICAN and IN16B) and MANTID31 (IRIS) and 29 30 32 31 analysed using the DAVE software . 32 33 34 Insight into the diffusive behaviour of confined liquid in Aqua and Poly cements: Based on the analysis of the 35 36 QENS response, the nanoscale diffusion of the water molecules was analyzed by fitting the quasi-elastic (QE) signal 37 38 using a sum of Lorentzian functions, L (Γ,ω). Since individual trajectories of all the distinct relaxation processes 39 40 33 41 cannot be distinguished, the scattering function, S(Q,ω), was simplified and described using Equation 1. The full 42 43 equation is given in the supplementary information as Equation S1. 44 45 (1) 46 S(Q,ω) = DWF (A0 (Q)δ(ω) +(1-A0 (Q)) ∑#&' L# (Γ,ω)) ⨂ R(Q,ω)( + B(Q) 47 48 where DWF represents the Debye Waller factor. A* (Q), defined as the averaged elastic incoherent structure factor 49 50 (EISF), describes the time averaged spatial distribution of all scattering centers. The term ∑#&' L# (Γ,ω) indicates the 51 52 53 sum of Lorentzians representing the broadened energy distribution that results from neutron-nucleus collisions, 54 55 corresponding to the population statistics of distinct relaxation processes. Considering that the GIC has a large 56 57 distribution of pore sizes the system was assumed to be dynamically heterogeneous and the different types of 58 59 34 . R(Q, ω) denotes the resolution function of the instrument that ACS Paragon Plus Environment 60 hydrogen motions were superimposed independently 8

Page 9 of 33


1 defines the observation time of the spectrometer. B(Q) represents a background term, including those dynamics that 2 3 are too fast for to be seen in the temporal range probed by the instrument. Finally, δ(ω), is a delta function describing 4 5 those particles seen as immobile. Additional information can be extracted by looking at the ratio between the elastic 6 7 8 intensity and the total intensity from both the elastic (E) and of the quasi-elastic (QE) signal, referred to as the 9 10 EISF35,36 11 12 13 (2) 14 EISF = - 15 16 17 In general terms, the EISF36, A* (Q), expresses the probability that a particle can be found in the same volume of space 18 19 after a time interval, t, (corresponding to the instrument resolution) has passed. It can therefore be used to determine 20 21 22 the geometry of the motion of a molecule or molecular specie in a molecule. Since the type of confinement influences 23 24 the local geometry of the molecule, the evolution of the EISF can shed light on the specific type of local restrictions on 25 26 the molecule35–37. In this work, however, since we modeled our data using two Lorentzian components, the EISF 27 28 29 measured is in fact a global, or effective, value which gives insight on the collective behaviour of all motions detected 30 37 31 within the observation time of the spectrometer in use . 32 33 34 The behavior of the extracted half width at half maximum (Γ, HWMH), of the QE signal, was modeled as a function 35 36 of Q using the expression for the random-jump-diffusion38: 37 38 / 39 Γ(Q) = . , (3) '-. / 01 40 41 42 where τ0 corresponds to the residence time and D is the translational diffusion constant given in terms of the mean 43 44 jump length < 3 > as: 45 45 46 = 47 6 = 789:; (4) ?@AB ?@ C'*D = 2.9 ∙ 10HI m= /s and τ* = 0.35 ps, listed in 31 32 33 Table 1. 34 35 36 In contrast, there is incomplete information about the dynamics of aqueous PAA solutions48,49. In the particular 37 38 case of the PAA studied here, it is known that gel-like polymeric networks are formed40, and that, from a combination 39 40 7 41 of neutron spectroscopy and calorimetric analysis , two water populations can be distinguished; one that resembles 42 43 bulk water and the other suggesting glassy behavior originating from regions of hydration around the polymer chain. 44 45 Therefore using two Lorentzians in Equation 1, one representing the dominating faster bulk water motion (hereafter 46 47 48 referred to by its HWHM, Γbulk) and the second representing a population with a more hindered diffusion 49 50 representative of glassy water behavior (hereafter referred to by its HWHM, Γglassy), the evolution of the QENS signal 51 52 53 of PAA collected at PELICAN at body temperature was analysed. All parameters were unconstrained during the 54 55 fitting procedure. The resulting HWHM, also analysed using Equation 3, are listed in Table 1. From this analysis we 56 HI = 57 observe that Γ bulk (D QRST = 2.9 ∙ 10 m /s and τ* = 0.37 ps) and Γwater are basically the same, indicating that the 58 59 onEnvironment the value of τ0) are not modified in the aqueous PAA ACS (reflected Paragon Plus 60 faster bulk water motion and its residence time 13


Page 14 of 33

1 solution. However, the diffusion coefficient for Γglassy (DVS?WWX = 1.7 ∙ 10HI m= /s and τ* = 9.9 ps) is ~2 times slower 2 3 than Γbulk in regards to the D coefficient and ~30 times in regards to the residence time τ* . Further confirmation of the 4 5 6 coexistence of bulk and constrained regions of polymer hydration in the PAA network is realized by looking at the 7 8 effective EISF. Here we clearly see that the water states probed by PELICAN, Fig. 3(a), fit very well with the 9 10 modified analytical EISF model, Equation 5, for confined water; namely diffusion of a point particle inside a sphere50: 11 12 13 ] ^_` ab ] cd^ ab = 14 H ab (ab)/ 15 Y* () = Z + (1 − Z) \ f (5) e 16 17 18 19 where d is the radius of the idealised spherical confinement, and p accounts for the immobile fraction of the protons. 20 21 22 To better map the more hindered water population that gives rise to Γglassy, as well as to verify the existence of 23 24 51 25 possible slower dynamics in the PAA , QENS measurements were performed using the backscattering spectrometers 26 27 IRIS and IN16B. By analysing the IRIS signal using the approach described above, and using a fixed value of Γbulk, 28 29 the two water populations detected during the PELICAN measurements were also observed. However, on IRIS, with 30 31 32 its narrower energy resolution, the slower Γglassy motion was seen to be the dominating spectral component; Γbulk 33 34 contributing to a slightly higher offset in the ‘background’. From the evolution of the effective EISF, Fig. 3 (b), also 35 36 37 analysed using Equation 5, we observe a lower immobile fraction of the protons, p (see Table 1). This corroborates the 38 39 idea that distinct water states have different mobility. Furthermore, by imposing the linewidth of Γglassy while fitting 40 41 Equation 1 to the IN16B data, a new, slower dynamic process, not discernible in the calorimetric analysis, was 42 43 H'* m= /s, which is ~ 10 times slower than DVS?WWX), 44 detected. Considering the diffusion value for this motion (0.6 ∙ 10 45 46 we attribute it to the PAA itself51, hereafter named by its HWHM as ΓPAA (see Table 1). Moreover, unlike IRIS and 47 48 49 PELICAN, the evolution of the effective EISF for ΓPAA cannot be modeled using the confined water model, Fig. 3(c), 50 51 indicating changes in the structuring of the hydrogen network52. This point will be further discussed later. 52 53 54 55 56 57 58 59 ACS Paragon Plus Environment 60

14

Page 15 of 33


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 2 40 Figure 3: Left side: Evolution of the HWHM for Γbulk, Γglassy and ΓPAA as a function of Q obtained for the PAA sample 41 42 and modelled using the theory developed to describe random jump diffusion. Right side: Effective EISF, obtained 43 44 from the analysis of the quasi-elastic spectra, plotted as a function of Q and fitting Equation 5 to the data, solid lines. 45 46 47 The horizontal dotted lines are representative of the amount of immobile protons (p) in the PAA as seen using 48 49 spectrometers with differing energy resolutions. The results were extracted from the data collected at 310 K using (a) 50 51 PELICAN, (b) IRIS and (c) IN16B. Note that the simple diffusion of a point particle inside a sphere does not fit the 52 53 54 IN16B data very well. 55 56 57 58 59 ACS Paragon Plus Environment 60

15


Page 16 of 33 2

1 Table 1: Parameters Dt and τ0 obtained from fitting the HWHM’s using Equation 1 as a function of Q and modelled 2 3 using the analytical function developed to describe a random jump diffusion. Different water populations are observed 4 5 6 in PAA; namely Γbulk, Γglassy and ΓPAA. The parameters d and p were obtained, for each spectrometer, by fitting the Q 7 8 dependence of the effective EISF by the diffusion of a point particle inside a sphere model50. The parameters were 9 10 extracted from data collected at 310K. The values obtained for bulk water measured at 310K using the IRIS 11 12 13 spectrometer also analysed using the random jump diffusion model are given for comparison. 14 15 Population Dt (10-9m2/s) p (%) d (Å) τ0 (ps) 16 Spectrometer 17 18 PELICAN 2.96±0.4 0.37±0.02 0.22±0.02 3.6±0.2 Γbulk 19 20 1.72±0.1 9.9±0.3 0.11±0.01 5.3±0.3 IRIS Γglassy 21 22 Not possible Not possible 23 to fit with a to fit with a IN16B 0.06±0.01 30±10 ΓPAA 24 simple model simple model 25 26 IRIS 27 2.96 ±0.04 0.35±0.02 Γwater 28 (bulk water) 29 30 31 32 33 34 Modification to the dynamical behavior of water in aqueous PAA when confined in the GIC as observed by neutron 35 36 spectroscopy 37 38 39 From the TGA analysis we know that a fraction of the bulk water, Γbulk, remains in both cements even after 28 40 41 42 days of aging. Therefore, to analyse the QENS signal observed in the cements when using PELICAN and IRIS, two 43 44 Lorentzians were used, but with the HWHM describing the bulk water, Γbulk, fixed. All other parameters were 45 46 unconstrained. The results obtained for the Poly and Aqua samples using PELICAN and IRIS are shown in Fig. 4 and 47 48 49 Fig. 5 and listed in Table 2. 50 51 In the timescale covered by PELICAN, and for all cement samples (Fig. 4, Table 2) independent of ageing, the 52 53 54 quasi-elastic signal comes from a water state for which the diffusion coefficient was found to be in the order of ~ 1.5 55 56 57 times higher than Γglassy, an order of ~ 3 times slower than Γbulk and with a residence time ~ 7 times higher than Γglassy. 58 59 Based on this observation, plus the fact that a fraction of the liquid used in preparing the GIC either remains in its bulk ACS Paragon Plus Environment 60

16

Page 17 of 33

ACS Applied Materials & Interfaces 12

53,54

, we conclude that the slower motion 1 state, is confined in the small cement pores or is assigned to cationic water 2 3 with a bulk-like behaviour is ascribed to confinement (hereafter referred to after its HWHM, Γbulk-confined). In contrast, 4 5 6 the broader component describes unreacted liquid, Γbulk. Surprisingly, even though the strength and porosity of the 7 8 cement samples differ12, in the ps timescale the dynamical behavior of the water in aqueous PAA when confined in the 9 10 GIC, Γbulk-confined, does not. Indeed, independent of the cement’s age, the diffusion coefficient is approximately the 11 12 H'* m= /s and τ*HQRSTHhijkljAm = 0.5 ps. However, the amount of 13 same for all cements; D QRSTHhijkljAm = 1.2 ∙ 10 14 15 hydrogen in the form of unbound water does vary, as depicted by the values of p, which accounts for the immobile 16 17 fraction of the protons, listed in Table 2. Once again, the evolution of the effective EISF as a function of Q is well 18 19 36 20 described using the modified analytical form for confined proton moblity , with the p-values (pAA=0.22, pPoly7=0.82, 21 22 pAqua7=0.78, pPoly28=0.86 and pAqua28=0.82), being consistent with the TGA measurements, compare Table 2 with 23 24 Table S2 given in the supplementary information. This finding confirms the highest water content to be in the 7 days 25 26 27 old Aqua material while the lowest is seen to be in the 28 days old Poly sample. Comparison of the extracted d values, 28 29 which represents the radius of the idealised spherical confinement, shows dPAA= 3.6 Å ± 0.2 Å, dPoly7=3.4 Å ± 0.1Å 30 31 and daqua7=3.3 Å ± 0.1 Å; the average radius for the confined water molecule not differing within error. This result 32 33 34 implies that the type of confinement probed on the ps timescale is comparable. If we, however, compare the d values 35 36 extract for the 28 day-old samples (dPoly28=3.0 Å ± 0.1 and dAqua28=3.3Å ± 0.1Å) with the 7 day-old samples, we 37 38 observe a slight decrease in the radius of confinement for the Poly cement, see Fig 4. This change, not observed in the 39 40 41 Aqua cement, indicates a variation of local geometry in the Poly material between 7 and 28 days. Previous work on 42 12 43 these specific cements showed a change in mechanical strength over the same time period for the Poly cement only . 44 45 46 47 48 49 Table 2: Parameters D and τ obtained from fitting the evolution of the HWHM’s as a function of Q2 using the t 0 50 51 52 analytical function developed to describe a random jump diffusion. Different water populations in the GIC samples as 53 54 a function of ageing; namely Γbulk, Γbulk-confined and Γglassy-confined. The parameters d and p were obtained, for each of the 55 56 selected spectrometer, from the fit of the Q dependence of the effective EISF as a function of Q using the diffusion of 57 58 50 59 a point particle inside a sphere model . The parameters were extracted from the data collected using the spectrometers ACS Paragon Plus Environment 60 17

ACS Applied Materials & Interfaces 1 PELICAN, IRIS and IN16B at 310K. Bulk water is known to exist in these samples 2 3 considered during the fitting procedure. 4 5 6 Dt (10-9m2/s) τ0 (ps) Sample Spectrometer Populations 7 8 2.96 (fixed) 0.37 (fixed) Γbulk 9 10 PELICAN 11 1.12±0.12 1.7±0.4 Γbulk-confined 12 Poly 13 14 15 0.97±0.09 15.9±1.1 Γglassy-confined IRIS 16 17 18 19 2.96 (fixed) 0.37 (fixed) Γbulk 20 21 PELICAN 22 1.09±0.17 2.1±0.1 Γbulk-confined 23 Aqua 24 25 26 0.12±0.01 64.8±6.8 Γ glassy-confined IRIS 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 ACS Paragon Plus Environment 60

Page 18 of 33

and therefore its contribution was

p (%)

d (Å)

-

-

0.82±0.01 at 7 d

3.53±0.057 at 7 d

0.86±0.01 at 28 d

3.0±0.1 at 28 d

0.73±0.01 at 7 d

3.37±0.18 at 7 d

0.75±0.01 at 28 d

3.39±0.16 at 28 d

-

-

0.78±0.01 at 7 d

3.3±0.03 at 7 d

0.83±0.01 at 28 d

3.3±0.07 at 28 d

0.72±0.01 at 7 d

3.16±0.16 at 7 d

0.79±0.01 at 28 d

3.58±0.18 at 28 d

18

Page 19 of 33


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 2 37 Figure 4: Left side: Evolution of the HWHM for Γbulk and Γbulk-confined as a function of Q , for GIC samples Aqua (a) 38 39 and Poly (b), extracted from data collected on PELICAN and analysed using the random jump diffusion model. Note 40 41 that the HWHM does not change as a function of ageing. Right side: The effective EISF as a function of Q obtained 42 43 44 from the analysis of the quasi-elastic spectra for Aqua (a) and Poly (b) samples aged at 7 and 28 days. The solid and 45 46 dashed lines were obtained using the model described by Equation 5. The horizontal dotted line is a guide to the eye to 47 48 allow the fraction of immobile protons (p) in the cements to be ascertained as a function of ageing. 49 50 51 To further understand how the polymeric-water network is changing during the reaction with the silicate glass, 52 53 54 as observed using the intermediate timescale covered by IRIS (~200 ps), we followed the same procedure described 55 56 above. The results are summarized in Fig. 5 and Table 2. By comparing the diffusion coefficient values obtained for 57 HI = 58 the Poly cement at different ages (D VS?WWXH{|}#~eH|8 = 0.99 ∙ 10 m /s) to the value for glassy water in the PAA 59 ACS Paragon Plus Environment 60

19


Page 20 of 33

1 solution (DVS?WWX = 0.7 ∙ 10 m /s), we observe that they are of the same order. We can therefore, conclude that the 2 3 water molecules that occupy coordination sites around the polymers in the PAA solution do not change their dynamic 4 5 6 response when confined. Furthermore, the dramatic increase in the residence time of τ*_VS?WWXH{|}#~eH|8 = 27 ps 7 8 compared to τ*_VS?WWX = 3.9 ps clearly indicates a strong deviation from the ideal water structure. On the other hand, 9 10 11 for the Aqua cement at different stages of maturation, the obtained diffusion coefficient (DVS?WWXH{|}#~eH4 = 12 13 0.12 ∙ 10HI m= /s with τ*HVS?WWXH{|}#~eH4 = 65 ps) is ~ 10 times slower than Dglassy and ~ 6 times 14 15 16 slower DVS?WWXH{|}#~eH|8 , with a much longer residence time, τ0. We therefore conclude that the polymeric-water 17 18 network is more hindered in the Aqua sample. Previous observation12 showed that the hydrogen protons bind faster to 19 20 the microstructure of the Aqua cement on the intermediate timescale of ~200 ps, which was earlier connected to 21 22 23 different proton (population) mobility in the Aqua Cement. 24 25 50 26 Additionally, the diffusion of a point particle inside a sphere model cannot fully describe the evolution of the 27 28 effective EISF collected at IRIS over the full Q-range for the water confined in the cement matrix . The effective EISF 29 30 deviates at around Q= 0.8 Å-1 and decreases again at Q=1.4 Å-1, indicating a restricted rotational diffusion18, see (*) 31 32 55 33 and (**) in Fig. 5 , that in fact cannot be fully separated from translational diffusion at room temperature . 34 35 Furthermore we note that changes in the immobility fraction p of the effective EISF are more evident for the Aqua 36 37 cement upon ageing, which is in full agreement with the TGA measurements and the QENS data analysis obtained 38 39 40 from the PELICAN data. Finally we observe that on the timescale covered by IRIS there is a slight change in the 41 42 radius of the confinement, reflected by the effective EISF, for the differently aged Aqua cements. In contrast, for the 43 44 Poly cement this happens on the timescale covered by PELICAN. Knowing that water gets trapped in the PAA, we 45 46 47 interpret our IRIS results as probing exactly this phenomenon, showing more confinement on the Aqua cement. 48 49 50 51 52 53 54 55 56 57 58 59 ACS Paragon Plus Environment 60 HI

=

20

Page 21 of 33


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 2 37 Figure 5: Left side: Evolution of the HWHM for Γglassy-confined as a function of Q obtained for the GIC samples, Aqua 38 39 (a) and Poly (b), extracted from data collected on IRIS and analysed using a random jump diffusion model. Note that 40 41 the HWHM does not change as a function of ageing. Right side: The effective EISF as a function of Q obtained from 42 43 44 the analysis of the quasi-elastic spectra for GIC samples Aqua (a) and Poly (b) aged at 7 and 28 days modelled using 45 46 Equation 5, solid and dashed lines. The horizontal dotted lines are a guide to the eye to enable the percentages of 47 48 immobile protons (p) in the cements to be ascertained as a function of ageing. Here (*) and (**) show where the data 49 50 deviates from the modified confined water model for, Poly cement and Aqua cement, respectively. 51 52 53 54 Modification to the dynamical behavior of the polymer in PAA when confined in the GIC as analyzed by neutron 55 56 spectroscopy 57 58 59 ACS Paragon Plus Environment 60

21


Page 22 of 33

1 When analysing dynamics on the ns timescale (IN16B) the motions from Γbulk, Γbulk-confined and Γglassy-confined contribute 2 3 to a ‘background’ intensity and therefore only one Lorentzian was needed to model the data. The variation in the 4 5 6 HWHM as a function of Q for the GIC cannot be satisfactorily described using the random jump diffusion model 7 56 8 (Equation 5), but it is instead in good agreement with the Chudley-Elliott model (Equation 6), specially in the high Q 9 10 limit: 11 12 13 ' Wlj(8 Wlj ) (6) 14 Γ(Q) = 01 1 − 8 Wlj 15 16 17 where is the residence time and 3 the jump length. * 18 19 20 Although, for small Q values the evolution of HWHM as function of Q tends to the random jump model, Equation 21 22 23 3, the Chudley-Elliott model gives the best fit when Q > 0.5 Å-1. Independent of the modelling, the extracted HWHM 24 25 is very similar between both cements types, with the mobile population diffusing at a rate similar to as ΓPAA, see Table 26 27 28 1. As a result, we can conclude that the PAA motion is hindered by a geometrical confinement from changes in the 29 30 chemical bonding, i.e. polysalt formation in the cement. This population, hereafter named ΓPAA-confined, is illustrated in 31 32 Fig. 6 and Table 3. Such changes in confinement can be interpreted as follows; taking the formation of the aluminum 33 34 35 poly salts as an example, by mixing PAA, aluminum silicate glass and water, some CH-COO groups from the PAA 36 +3 37 will form CH-COO -Al and others will remain free, these can interchange, thus creating a possible hydrogen 38 39 diffusion in a diffusive landscape52. These different hydrogen networks will then have distinct hydrogen bond 40 41 2 -2 -2 42 lifetimes, originating the periodic bumps at Q = 0.7 Å and 2.2 Å seen at the HWHM evolution in the GIC’s at the 43 44 longer time scale probed by the high resolution backscattering instrument IN16B. 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 ACS Paragon Plus Environment 60

22

Page 23 of 33


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 2 37 Figure 6: Evolution of the HWHM for ΓPAA-confined as a function of Q described by the Chudley-Elliott model. The 38 39 results were extracted from data collected using IN16B for 7 days old Aqua (a, in blue) and 7 days old Poly (b, in red). 40 41 42 Table 3: Residence time (* ) and 3 the jump length describing the dynamical behavior of the polymer in PAA 43 44 confined, ΓPAA-confined, in Poly and Aqua cements aged for 7 days obtained from the data collected at 310K on IN16B 45 46 47 analysed using the Chudley-Elliott model. 48 49 50 l(Å) τ0 (ps) Sample Spectrometer Population 51 52 12 ±1 275±9 ΓPAA-confined Poly IN16B 53 54 12 ±1 275±9 ΓPAA-confined 55 Aqua IN16B 56 57 58 59 ACS Paragon Plus Environment 60

23


Page 24 of 33

1 Modification of the hydrogen-bond interaction and phonon density in the PAA when confined in the GIC analyzed by 2 3 neutron spectroscopy at 10K: 4 5 6 Similar for all cement samples the librational modes at 10 meV assigned to confined water (ice-like) is observed in the 7 8 9 samples shown in Fig. 7. Modes 20 and 40 meV can be assigned to the water in the polymer network. The distortion 10 11 from the bulk water (ice) modes in the spectrum is more evident in the 28 day-old samples, as the ratio of bulk-water 12 13 to glassy-water decreases. The first peak that can be fully assigned to the silicate matrix is visible at 130-190 meV and 14 15 matches well the signal detected by the FTIR (~1000-1500 cm-1 or ~125-188 meV) illustrated in Fig. S1 (see 16 17 18 supplementary information), which corresponds to the Si-O-Si stretching vibration from the powder. The small 19 20 vibration at 210 meV most likely corresponds to the carboxylic acid C=O stretching also detected in the FTIR (~1700 21 22 cm-1 or ~212 meV). Furthermore, at 370 meV (~2900 cm-1) we observe a C-H stretching coming from the PAA in all 23 24 25 samples. 26 27 28 First we notice that the signal to noise ratio is the same in all samples. Then, by comparing the intensity obtained from 29 30 VISION spectrometer from samples aged during 7 and 28 days, we observe that the largest change in signal occurs in 31 32 the Poly cement (see Fig. 7). The intensity of the signal is directly proportional to the phonon density, which we can 33 34 35 interpret as the variation in density over time. 36 37 38 In putting all findings together, we first consider the intensities from the QENS measurements, and notice that the 39 40 signal from the 28 day-old Aqua is higher than the 7 day-old, thus confirming the TGA findings, which show that the 41 42 amount of unbound water decreases while the amount of bond water increases. This is however not the case for the 43 44 45 Poly cement, that in contrast reveals a bigger change in density over time, based on the intensities from the VISION 46 47 measurements. This suggests that in the Aqua cement the degree of crosslinking is lower than in the Poly cement, and 48 49 instead more H-bonds are being formed. In the Poly on the other hand, the crosslinking continues even after 28 days 50 51 52 with limited H-bond formation. 53 54 55 56 57 58 59 ACS Paragon Plus Environment 60

24

Page 25 of 33 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


ACS Paragon Plus Environment

25


Page 26 of 33

1 Figure 7: Vibrational transition energies obtained at 10 K using the vibrational spectrometer VISION. (a) Aqua 2 3 cement samples (blue) aged for 7 and 28 days and (b) Poly cement samples (red) aged for 7 and 28 days. The intensity 4 5 of the signal is directly proportional to the phonon density. The peaks are assigned to the librational and vibrational 6 7 8 modes of the samples. Similar data for bulk water is given in the supplementary information for comparison. 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 Figure 8: Schematic overview of the hydrogen populations present in the aqueous PAA solution (left side) and in the 31 32 33 GIC (right side). Three hydrogen populations are seen in the PAA solution, namely bulk water, glassy water 34 35 originating from the water incorporated in the PAA polymeric chain (referred to as PAA) - and the hydrogen in the 36 37 polymeric chain itself. When incorporated in the GIC new populations are derived: unbound water to the cement 38 39 matrix (unreacted) that remains in its bulk state, loosely bond water that behaves similar to bulk, tightly bound water 40 41 42 that has a glassy behaviour and the confined hydrogen in the PAA chain. 43 44 45 46 47 CONCLUSION 48 49 50 To date relatively few studies have been carried out on the dynamics of GIC, this one being the first that maps the full 51 52 hydrogen dynamics in the aqueous PAA solution used in conventional GIC to its behaviour when confined in the 53 54 55 cement matrix. In this study, combining neutron spectroscopy and calorimetric analysis we were able to separate 56 57 distinct hierarchically superimposed molecular motions within the liquid and in the cements. Consequently a better 58 59 understanding of the hydrogen-bond network formation and density changes of this unique type of GIC was achieved. ACS Paragon Plus Environment 60

26

Page 27 of 33


1 A key component that influences sample properties is water. Water is necessary for initial setting but also for the 2 3 maturation, since it hydrates the cement matrix and further allows it to react. While it is known that water binds to the 4 5 structure during the first steps of the setting8, there are still some ambiguities regarding water binding during 6 7 8 maturation. In this study, we have added to this understanding as follows. From the thermogravimetric method 9 10 different amounts of bulk water in the samples after maturation were identified. This analysis showed that for these 11 12 samples, bulk water binds to the polymer chain, while an increase in the amount of structural water is not necessarily 13 14 15 observed. However, progressive water binding due to ageing occurred in only one of the GIC, namely the weaker 16 17 Aqua sample. As the mechanical strength of the Aqua cement does not change with aging in the maturation timeframe 18 19 studied here12, we can hypothesize that the progressive water binding does not influence the strength of the material. 20 21 Clear changes in the material density were only observed in the Poly cement as detected by the INS data. Thus we 22 23 24 interpret that the continuous crosslinking on the Poly cement is the main cause for its increase in strength during 25 26 maturation. 27 28 29 QENS analysis revealed three distinguishable hydrogen populations in the PAA aqueous solution, which were also 30 31 32 visible in both GICs. This result is summarized in Fig. 8. By varying the observation time, it was not possible to detect 33 34 changes in the self-diffusing motions between samples aged 7 or 28 days; only a change in the intensity of the elastic 35 36 signal was observed. The residence time in the cement itself differed from that of the PAA solution due to changes in 37 38 the local environment: the most noticeable difference between the GIC samples being in the dynamics associated to 39 40 41 the confined glassy water; the water eventually just binding to the polymer. The QENS analysis revealed that the 42 43 hydrogen binding varied between the two GIC, thus the shape and structure of the polyelectrolyte chain will influence 44 45 the degree of crosslinking and hydrogen bonding in the PAA matrix and consequently influencing the cement 46 47 48 properties. The dynamics of the hydrogen on the PAA itself showed that for both GIC localised diffusivity dominates 49 50 parts of the spatial range. 51 52 53 To conclude, these results indicate that the advantages of this material can be improved not only by modifying the 54 55 glass-silicate, particle size and chemical formula, so that its disintegration and binding to tissue is improved. But also 56 57 58 that the understanding of polymer-water binding and polymer crosslinking is important in material development as it 59 ACS Paragon Plus Environment 60 can shade light on material strength and longevity.

27


Page 28 of 33

1 2 3 4 AUTHOR INFORMATION 5 6 7 Corresponding Author 8 9 * Marcella C. Berg: [email protected] to whom correspondence should be addressed. 10 11 12 13 Author Contributions 14 15 MCB, ARB and HNB conceived the project. MCB and ARB prepared the samples and MCB characterized the 16 17 18 samples using the TGA/FTIR. MCB, ARB and HNB designed the neutron scattering experiments, and details were 19 20 discussed a priori with MTFT, DY and TS. All authors carried out the neutron experiments. MCB analyzed all data 21 22 with input from MTFT, TS, LSD, DY, ARB and HNB. MCB, ARB and HNB wrote the paper with input from the co23 24 25 authors. The manuscript was approved by all coauthors in its final version 26 27 28 29 30 Funding Sources 31 32 33 ACKNOWLEDGMENT 34 35 36 The work was funded by the European Spallation Neutron Source, the Niels Bohr Institute and the Danish Agency 37 38 for Science, Technology and Innovation through DANSCATT. HNB acknowledges support from the CoNext project. 39 40 41 MCB and HNB gratefully acknowledge the financial support from MAX4ESSFUN of the European Regional 42 43 Development Fund Interreg Öresund-Kattegat-Skagerrak (project KU-025). MCB and ARB acknowledge the financial 44 45 supported offered by the ILL to cover their travel expenses. 46 47 48 Supporting Information. Tables with coherent and incoherent cross-sections, mass loss in percentage obtained from 49 50 2 51 the thermal analysis experiments as well as the evolution of the HWHM vs Q obtained form the QENS data analysis 52 53 for bulk water and for samples aged at 28 days are supplied as supporting information. Supplementary TGA/FTIR 54 55 data, selected fits of the QENS spectra and the equation describing the dynamic structure factor as a sum of distinct 56 57 58 exponential (Debye type) relaxation processes are also provided. 59 ACS Paragon Plus Environment 60

28

Page 29 of 33


1 2 3 REFERENCES 4 5 (1) WHO | What Is the Burden of Oral Disease? WHO 2010. 6 7 (2) Patel, R. The State of Oral Health in Europe Report Commissioned by the Platform for Better Oral Health in 8 Europe. 2012. 9 10 Sidhu, S.; Nicholson, J. A Review of Glass-Ionomer Cements for Clinical Dentistry. J. Funct. Biomater. 2016, 11 (3) 7 (3), 16. 12 13 Tian, K. V; Yang, B.; Yue, Y.; Bowron, D. T.; Mayers, J.; Donnan, R. S.; Dobó-Nagy, C.; Nicholson, J. W.; 14 (4) Fang, D.-C.; Greer, A. L.; et al. Atomic and Vibrational Origins of Mechanical Toughness in Bioactive Cement 15 16 during Setting. Nat. Commun. 2015, 6, 8631. 17 18 (5) Ngo, H. Glass-Ionomer Cements as Restorative and Preventive Materials. Dent. Clin. North Am. 2010, 54, 551– 19 563. 20 21 (6) Swift, T.; Swanson, L.; Geoghegan, M.; Rimmer, S. The pH-Responsive Behaviour of Poly(acrylic Acid) in 22 Aqueous Solution Is Dependent on Molar Mass. Soft Matter, 2016 vol: 12 (9) pp: 2542-2549 23 24 (7) Berg, M. C.; Jacobsen, J.; Momsen, N. C. R.; Benetti, A. R.; Telling, M. T. F.; Seydel, T.; Bordallo, H. N. 25 Water Dynamics in Glass Ionomer Cements. Eur. Phys. J. Spec. Top. 2016, 225 (4), 773–777. 26 27 (8) Barry, T. I.; Clinton, D. J.; Wilson, A. D. The Structure of a Glass-Ionomer Cement and Its Relationship to the 28 Setting Process. J. Dent. Res. 1979, 58 (3), 1072–1079. 29 30 (9) Wilson, A. D.; Nicholson, J. W. Acid-Base Cements: Their Biomedical and Industrial Applications; Cambridge, 31 UK: Cambridge University Press, 1993. 32 33 34 (10) Earl, M. S. A.; Mount, G. J.; Humet, W. R. The Effect of Varnishes and Other Surface Treatments on Water Movement across the Glass Ionomer Cement Surface. II. Aust. Dent. J. 1989, 34 (4), 326–329. 35 36 37 (11) Lohbauer, U. Dental Glass Ionomer Cements as Permanent Filling Materials? – Properties, Limitations and Future Trends. Materials (Basel). 2009, 3 (1), 76–96. 38 39 40 (12) Benetti, A. R.; Jacobsen, J.; Lehnhoff, B.; Momsen, N. C. R.; Okhrimenko, D. V.; Telling, M. T. F.; Kardjilov, 41 N.; Strobl, M.; Seydel, T.; Manke, I.; et al. How Mobile Are Protons in the Structure of Dental Glass Ionomer 42 Cements? Sci. Rep. 2015, 5, 8972. 43 44 (13) Wasson, E. A.; Nicholson, J. W. New Aspects of the Setting of Glass-Ionomer Cements. J. Dent. Res. 1993, 72 45 (2), 481–483. 46 47 (14) Nicholson, J. W.; Wilson, A. D. The Effect of Storage in Aqueous Solutions on Glass-Ionomer and Zinc 48 Polycarboxylate Dental Cements. J. Mater. Sci. Mater. Med. 2000, 11 (6), 357–360. 49 50 (15) Roberts, H.; Berzins, D. Thermal Analysis of Contemporary Glass-Ionomer Restorative Materials. J. Therm. 51 Anal. Calorim. 2013, 115 (3), 2099–2106. 52 53 54 (16) Characterization of Glass-Ionomer Cements 5. The Effect of the Tartaric Acid Concentration in the Liquid Component. J. Dent. 1979, 7 (4), 304–312. 55 56 57 (17) Tadjiev, D. R.; Hand, R. J. Surface Hydration and Nanoindentation of Silicate Glasses. J. Non. Cryst. Solids 2010, 356 (2), 102–108. 58 59 ACS Paragon Plus Environment A. Water Dynamics in Hardened Ordinary Portland Cement Paste or 60 (18) Bordallo, H. N.; Aldridge, L. P.; Desmedt,

29

ACS Applied Materials & Interfaces 1 2 3 (19) 4 5 6 7 (20) 8 9 10 (21) 11 12 13 (22) 14 15 (23) 16 17 18 (24) 19 20 21 22 (25) 23 24 25 26 (26) 27 28 29 30 (27) 31 32 33 (28) 34 35 36 (29) 37 38 39 40 (30) 41 42 (31) 43 44 45 (32) 46 47 48 (33) 49 50 51 (34) 52 53 54 (35) 55 56 57 58 (36) 59 60

Page 30 of 33

Concrete: From Quasielastic Neutron Scattering. J. Phys. Chem. B 2006, 110 (36), 17966–17976. Le, P.; Fratini, E.; Ito, K.; Wang, Z.; Mamontov, E.; Baglioni, P.; Chen, S. H. Dynamical Behaviors of Structural, Constrained and Free Water in Calcium- and Magnesium-Silicate-Hydrate Gels. J. Colloid Interface Sci. 2016, 469, 157–163. Kent, B. E.; Lewis, B. G.; Wilson, A. D. Glass Ionomer Cement Formulations: I. The Preparation of Novel Fluoroaluminosilicate Glasses High in Fluorine. J. Dent. Res. 1979, 58 (6), 1607–1619. Wilson, A. D.; Kent, B. E. A New Translucent Cement for Dentistry. The Glass Ionomer Cement. Br. Dent. J. 1972, 132 (4), 133–135. Wilson, A. D. Resin-Modified Glass-Ionomer Cements. Int. J. Prosthodont. 1990, 3 (5), 425–429. McLean, J. W.; Wilson, A. D.; Prosser, H. J.; Mondell, J. Development and Use of Water-Hardening GlassIonomer Luting Cements. J. Prosthet. Dent. 1984, 52 (2), 175–181. Wilson, A. D. A Hard Decade’s Work: Steps in the Invention of the Glass-Ionomer Cement. J. Dent. Res. 1996, 75 (10), 1723–1727. Sulatha, M. S.; Natarajan, U. Origin of the Difference in Structural Behavior of Poly(acrylic Acid) and Poly(methacrylic Acid) in Aqueous Solution Discerned by Explicit-Solvent Explicit-Ion MD Simulations. Ind. Eng. Chem. Res. 2011, 50 (21), 11785–11796. Sulatha, M. S.; Natarajan, U. Molecular Dynamics Simulations of PAA–PMA Polyelectrolyte Copolymers in Dilute Aqueous Solution: Chain Conformations and Hydration Properties. Ind. Eng. Chem. Res. 2012, 51 (33), 10833–10839. Sulatha, M. S.; Natarajan, U. Molecular Dynamics Simulations of Adsorption of Poly(acrylic Acid) and Poly(methacrylic Acid) on Dodecyltrimethylammonium Chloride Micelle in Water: Effect of Charge Density. Mikel, S. E.; Biernacki, J. J.; Gnäupel-Herold, T. A Neutron Diffraction-Based Technique for Determining Phase-Resolved Strains in Portland Cement. ACI Mater. J. 2009, 106 (5), 455–460. Frick, B.; Mamontov, E.; van Eijck, L.; Seydel, T. Recent Backscattering Instrument Developments at the ILL and SNS. Zeitschrift für Phys. Chemie 2010, 224 (1–2), 33–60. LAMP, the Large Array Manipulation Program. http://www.ill.eu/data_treat/ lamp/the-lamp-book/. Arnold, O. et al. Mantid - Data Analysis and Visualization Package for Neutron Scattering and µSR Experiments. Nucl. Instr. Meth. Phys. Res A 2014, 764, 156–166. Azuah, R. T. et al. DAVE: A Comprehensive Software Suite for the Reduction, Visualization, and Analysis of Low Energy Neutron Spectroscopic Data. J Res Natl Inst Stand Technol. 2009, 114, 341. Bée, M. (Marc). Quasielastic Neutron Scattering : Principles and Applications in Solid State Chemistry, Biology, and Materials Science; Adam Hilger, 1988. Bicout, D. J. ILL Millennium Symposium 2001. In Incoherent Neutron Scattering Functions for Combined Dynamics; 2001; p 60. Jacobsen, J.; Rodrigues, M. S.; Telling, M. T. F.; Beraldo, A. L.; Santos, S. F.; Aldridge, L. P.; Bordallo, H. N.; Dobbs, R.; Neville, A. M.; Isaia, G. C.; et al. Nano-Scale Hydrogen-Bond Network Improves the Durability of Greener Cements. Sci. Rep. 2013, 3, 69–76. Dianoux, A.; Volino, F.; Hervet, H. Incoherent Scattering Law for Neutron Quasi-Elastic Scattering in Liquid Crystals. Mol. Phys. 1975, 30 (4), 1181–1194. ACS Paragon Plus Environment 30

Page 31 of 33 1 (37) 2 3 4 5 (38) 6 7 8 (39) 9 10 11 (40) 12 13 14 (41) 15 16 17 18 (42) 19 20 21 (43) 22 23 24 (44) 25 26 27 (45) 28 29 30 31 32 (46) 33 34 35 36 (47) 37 38 39 40 (48) 41 42 43 (49) 44 45 46 (50) 47 48 49 (51) 50 51 52 (52) 53 54 55 56 57 (53) 58 59 60


Zhang, C.; Arrighi, V.; Gagliardi, S.; McEwen, I. J.; Tanchawanich, J.; Telling, M. T. F.; Zanotti, J.-M. Quasielastic Neutron Scattering Measurements of Fast Process and Methyl Group Dynamics in Glassy Poly(vinyl Acetate). Chem. Phys. 2006, 328 (1–3), 53–63. Singwi, K. S.; Sjölander, A. Diffusive Motions in Water and Cold Neutron Scattering. Phys. Rev. 1960, 119 (3), 863–871. Berzins, D. W.; Abey, S.; Costache, M. C.; Wilkie, C. A.; Roberts, H. W. Resin-Modified Glass-Ionomer Setting Reaction Competition. J. Dent. Res. 2010, 89 (1), 82–86. Roberts, H. W.; Berzins, D. W. Early Reaction Kinetics of Contemporary Glass-Ionomer Restorative Materials. J. Adhes. Dent. 2015, 17 (1), 67–75. Khan, A. S.; Khalid, H.; Sarfraz, Z.; Khan, M.; Iqbal, J.; Muhammad, N.; Fareed, M. A.; Rehman, I. U. Vibrational Spectroscopy of Selective Dental Restorative Materials. Appl. Spectrosc. Rev. 2016, 1–34. Cárdenas, G.; Muñoz, C.; Carbacho, H. Thermal Properties and TGA–FTIR Studies of Polyacrylic and Polymethacrylic Acid Doped with Metal Clusters. Eur. Polym. J. 2000, 36 (6), 1091–1099. Max, J.-J.; Chapados, C. Infrared Spectroscopy of Aqueous Carboxylic Acids: Comparison between Different Acids and Their Salts. Krishnamurthy, S.; Bansil, R.; Wiafe-Akenten, J. Low-Frequency Raman Spectrum of Supercooled Water. J. Chem. Phys. 1983, 79 (12), 5863. Cho, M.; Fleming, G. R.; Saito, S.; Ohmine, I.; Stratt, R. M. Instantaneous Normal Mode Analysis of Liquid Water Instantaneous Normal Mode Analysis of Orientational Motions in Liquid Water: Local Structural Effects Instantaneous Normal Mode Analysis of Liquid Water. J. Chem. Phys. J. Chem. Phys. J. Chem. Phys. J. Chem. Phys. Na J. Chem. Phys 1994, 100 (105), 10825–19281. Bellissent-Funel, M.-C.; Teixeira, J. Dynamics of Water Studied by Inelastic Neutron Scattering. J. Mol. Struct. Elsevier Sci. Publ. B.V 1991, 250, 213–230. Berg, M. C.; Dalby, K. N.; Tsapatsaris, N.; Okhrimenko, D. V.; Sørensen, H. O.; Jha, D.; Embs, J. P.; Stipp, S. L. S.; Bordallo, H. N. Water Mobility in Chalk: A Quasielastic Neutron Scattering Study. J. Phys. Chem. C 2017, 121 (26), 14088–14095. Li, J.; Zhao, K. Effect of Side-Chain on Conformation of Poly(acrylic Acid) and Its Dielectric Behaviors in Aqueous Solution: Hydrophobic and Hydrogen-Bonding Interactions and Mechanism of Relaxations. Yamazaki, S.; Noda, I.; Tsutsumi, A. 13C NMR Relaxation of Poly(acrylic Acid) in Aqueous Solution. Effects of Charge Density on Local Chain Dynamics. Polym. J. 2000, 32 (1), 87–89. Volino, F.; Dianoux, A. J. Neutron Incoherent Scattering Law for Diffusion in a Potential of Spherical Symmetry: General Formalism and Application to Diffusion inside a Sphere. Mol. Phys. 1980, 41 (2), 271–279. Sierra-Martin, B.; Retama, J. R.; Laurenti, M.; Fernández Barbero, A.; López Cabarcos, E. Structure and Polymer Dynamics within PNIPAM-Based Microgel Particles. Adv. Colloid Interface Sci. 2014, 205, 113–123. Roosen-Runge, F.; Bicout, D. J.; Barrat, J.-L. Analytical Correlation Functions for Motion through Diffusivity Landscapes. J. Chem. Phys. J. Chem. Phys. J. Chem. Phys. Chaos J. Chem. Phys. 2041, 144 (144), 124105– 143129. Gates, W. P.; Bordallo, H. N.; Aldridge, L. P.; Seydel, T.; Jacobsen, H.; Marry, V.; Churchman, G. J. Neutron Time-of-Flight Quantification of Water Desorption Isotherms of Montmorillonite. J. Phys. Chem. C 2012, 116 (9), 5558–5570. ACS Paragon Plus Environment

31


Page 32 of 33

1 (54) González Sánchez, F.; Jurányi, F.; Gimmi, T.; Loon, L. Van; Unruh, T.; Diamond, L. W.; Maier-Leibnitz, F. H. 2 Translational Diffusion of Water and Its Dependence on Temperature in Charged and Uncharged Clays: A 3 Neutron Scattering Study. J. Chem. Phys. 2008, 129. 4 5 (55) Teixeira, J.; Bellissent-Funel, M.-C.; Chen, S. H.; Dianoux, A. J. Experimental Determination of the Nature of 6 Diffusive Motions of Water Molecules at Low Temperatures. 7 8 (56) Chudley, C. T.; Elliott, R. J. Neutron Scattering from a Liquid on a Jump Diffusion Model. Proc. Phys. Soc. 9 1961, 77 (2), 353–361. 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 Table Of Contents (TOC) 54 55 56 57 58 59 ACS Paragon Plus Environment 60

32

Page 33 of 33

0.5

a) Poly 7 days old at Q= 0.6 Å

PELICAN IRIS IN16B

Normalized Int ensit y

1

0.6

0.4

Bulk Fixed Fast Wat er

0.4

Cationic water Fast Wat er

Fixed Confined

0.3 0.2 0.1 0.0 0.0

0.1

b)

0.4

0.01

0.001

-0.3

0.2

-0.2

-0.1

0.0

0.1

0.2

1.0

2.0 2

0.5

HWHM (meV)

0.8

-1

HWHM (meV)

1.0

0.3

3.0

-2

Q (Å ) Fixed Water confined the PAA Slow Wat erinConfined Bulk Fixed Fast Wat er

0.3 0.2 0.1 0.0

E (meV)

0.0

1.0

2.0 2

c) 0.004

3.0

-2

Q er (Å Confined ) Protons in the PAA Polymer Wat

HWHM (meV)

Normalized Int ensit y

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


0.003

0.0 -0.3

-0.2

-0.1

0.0

0.1

Energy Transfer (meV)

0.2

0.3

0.002 0.001 0.000 0.0

1.0

2.0 -2

3.0

-2

Q (Å )

ACS Paragon Plus Environment

33

Nanoscale Mobility of Aqueous Polyacrylic Acid in Dental Restorative

Recommend Documents