X-ray photoelectron spectroscopy study of the ... - Dental Materials

dental materials Dental Materials 15 (1999) 229–237 www.elsevier.com/locate/dental

X-ray photoelectron spectroscopy study of the dentin–glass ionomer cement interface H.E. Sennou a,*, A.A. Lebugle b, G.L. Gre´goire a a

Laboratoire de Biomate´riaux, Faculte´ d’Odontologie, 3 chemin des Maraıˆchers, 31062 Toulouse cedex, France b Laboratoire des Mate´riaux, Equipe sur les Phosphates ENSCT-UPRESA CNRS 5071, Toulouse, France Received 30 October 1997; received in revised form 24 August 1998; accepted 13 April 1999

Abstract Objectives: The work was carried out with a view to identifying the elements composing the glass ionomer under study, and then to characterising the interactions occurring between this particular glass ionomer and the dentin substrate on which it was placed and with which it interacted. Methods: The samples studied were sections of healthy human dentin on which a very thin film of auto-polymerisable cement, composed of a powder and a liquid, was deposited under para-clinical conditions. After separation, the interfaces on the dentin side and on the glass ionomer side were studied using X-ray Photoelectron Spectroscopy (XPS). Results: This study showed that the dentin and glass ionomer cement exchanged mineral and organic elements. The acid contained in the liquid showed a certain degree of aggressivity, despite the presence of the glass ionomer. The dentin protein was, in fact, rapidly denuded from the very first minute. Migration of the mineral elements from one substrate to the other led to the formation of an intermediate layer on the surface of the materials. Significance: This layer, which forms an interphase, enables the material to adhere to the dentin. q 1999 Academy of Dental Materials. Published by Elsevier Science Ltd. All rights reserved. Keywords: Dentin–cement interface; Glass ionomer cement; X-ray Photoelectron Spectroscopy; Electron Spectroscopy for Chemical Analysis

1. Introduction Glass ionomer cement has remarkable properties of adhesion to dentinal tissue [1]. Numerous studies have thus been devoted to these properties, but these have generally been directed to mechanical characteristics: compressive and tensile strength [2,3]. However, very little research has been devoted to chemical characterisation of the dentin–glass ionomer cement interface. Lin et al. [4] have studied adhesion from various viewpoints, mechanical, physical and chemical, using different analytical methods, including XPS (X-ray Photoelectron Spectroscopy); or ESCA (Electron Spectroscopy for Chemical Analysis). However, the work of these authors concerned interactions between bovine dentin, ground and polished, and different types of glass ionomer (essentially light-polymerisable) and, for the purpose of comparison, one auto-polymerisable type. They studied the surface * Corresponding author. Tel.: 1 33-5-62-17-2929; fax: 1 33-5-61-254719. E-mail address: [email protected] (H.E. Sennou)

morphology of the samples using SEM and confocal microscopy. Chemical composition was determined by means of SIMS (Secondary Ion Mass Spectrometry) and XPS; however, the two techniques last mentioned were applied to dentin samples that had been coated with both glass ionomer and an adhesive system. We thought of interest to carry on this study on interaction between glass ionomer cement and dentin in depth, but focused solely on XPS chemical analysis, to arrive at a clearer understanding of the mechanisms conditioning adhesion to the dentin. This technique is perfectly suitable for such a study, thanks to its two main characteristics. The first is its depth of analysis, ˚, which is very small, being in the order of a few tens A which makes it possible to study solely those phenomena occurring at the interface. The second characteristic is the variation in the binding energy of emitted photoelectrons, depending on the chemical environment in which the atoms are present, this being, in our case, on either side of the interface. In order to avoid the problem of the smear layer encountered in the above-mentioned studies, due to preparation of the samples by polishing, we carried out a pilot study on dentin that was fractured, and thus free from any surface

0109-5641/99/$20.00 + 0.00 q 1999 Academy of Dental Materials. Published by Elsevier Science Ltd. All rights reserved. PII: S0109-564 1(99)00036-6

230

H.E. Sennou et al. / Dental Materials 15 (1999) 229–237

Fig. 1. XPS general spectrum of dentin.

contamination that affects numerous studies. After preparation, we deposited an auto-polymerisable glass ionomer cement on this dentin. The study was initially concerned with the product itself, then with the dentin and, finally, with the interfaces produced after separation of the dentin from the glass ionomer cement. Thus, we were able to identify their interactions and to highlight the exchanges that had taken place.

2. Materials and methods The studies were carried out on samples of human dentin on which an glass ionomer cement was deposited under the conditions described below. The dentine samples studied were obtained from third molar teeth of young adults (aged between 20 and 40). These teeth were used immediately after extraction. They were fractured in such a way as to avoid any risk of surface contamination due to the use of rotary instruments. The teeth were fractured after making a circular cut in the amelary region, inserting a gouge chisel into the groove and tapping smartly with a surgical hammer. The section was taken at mid-coronal level so as to obtain average mineralisation values, the degree of mineralisation being known to depend on the coronal level examined. The glass ionomer cement (GIC) chosen was auto-polymerisable GC Fuji Type II (Lot 350 1806 GC Corp. ITABASHI-KU Tokyo, Japan). The manufacturer’s handling recommendations were observed (one drop of liquid, 1 g, for one spoonful of powder—2.7 g—and mixing with the plastic spatula for 20 s). A uniform thickness of glass ionomer, in the order of 1.5 mm, was deposited on the dentin. To approximate to normal clinical conditions, the samples were then placed in a steam saturated bath (100% Rh) under thermostatic control (at 348C) for 7 min. Under these conditions, initial setting was completed; in addition, any risk of the glass ionomer cement being denatured by excess water was avoided. After which, the samples were dehydrated in a

special recipient for 15 h, in a primary vacuum, with a view to subsequent XPS analysis. After desiccation, the interfaces between the dentin and the glass ionomer were easily obtained by splitting, started by means of a wedge inserted between the two parts. The latter will be referred to below as the dentin side interface (DI) and the glass ionomer side interface (GII). The different samples (dentin, glass ionomer cement, interfaces: DI and GII) underwent XPS analysis. A number of surface characterisations were also carried out using SEM (JMR 64000) with pictures being obtained with secondary or back scattered electrons. XPS (X-ray Photoelectron Spectroscopy) permits qualitative and quantitative analysis of the surface of various materials (organic, mineral or metallic), whether conductive ˚ . Unlike other or insulating, to a depth in the order of 40 A techniques, it detects all the elements in the periodic table, except hydrogen and helium, which have no core electrons. It does not require optical polishing and can even be applied to rough surfaces. In addition, its sensitivity to the chemical environment, as already pointed out, makes it possible to demonstrate the presence of atoms that have partial ionic charges owing to their bond with atoms of different electronegativity. The proportion of aliphatic, hydroxylated or carboxylic carbon atoms can thus be determined. The samples were studied after being fixed in place with double-sided adhesive tape, and were analysed using an ESCALAB Mk II (V.G.) spectrometer equipped with an Al Ka source (photon energy equal to 1486.6 eV) and a 3 Channeltrons detector in a vacuum in the order of 10 28 Torr. The surface analysed was in the order of 6 mm 2, making it possible to obtain a global response from the surface and to avoid any error due to local concentrations or artefacts. To prevent any degradation from irradiation, the power of the X-rays was limited to 100 W (10 mA, 10 kV) and pass energy was fixed at 50 eV. Energy calibration was effected with reference to aliphatic carbon C1s (Eb 285 eV). A general spectrum was made for each sample in order to


231

Table 1 Comparison of the organic and mineral ratios for the different parts studied Ratio

Dentin

Dentin interface

Glass ionomer interface

Powder glass ionomer cement

Al/Si Sr/Si Na/Si F/Si Ca/Si Ca/N Ca/P C/N Ca/C

– – – – – 1 1.59 4.25 0.20

0.94 0.31 0.14 0.76 0.46 0.62 0.9 19 0.03

0.62 0.24 0.08 0.59 0.08 – 0.43 – 0.01

0.81 0.46 0.11 1.07 – – – – –

identify the elements and regional spectra were made around selected elements to permit their quantitative analysis and study their chemical environment. On the basis of the integrated intensity of the photoelectronic peaks of the chemical elements, and using the efficient sections of Scofield photoionisation [5] it proved possible to undertake a quantitative analysis of the composition of the samples. Accuracy was in the order of 5% in the presence of standards.

3. Results 3.1. The dentin The general XPS spectrum (0–1000 eV) of the fractured dentin is shown in Fig. 1. It is comparable with those reported by Ruse [6], Lin [7] and Elfersi [8]. It presents the peaks of the major elements composing the dentin, originating either from the mineral part (carbonated hydroxyapatite), or from the organic part (essentially type I collagen): Ca, P, C, N and O. The peaks of lesser intensity located at 320 and 500 eV are due to the Auger transitions of magnesium and sodium. Indeed, the spectrum obtained between 1000 and 1350 eV

shows the XPS peaks for magnesium Mg1s at 1305 eV and sodium Na1s at 1072 eV. Observations for these elements concorded with the chemical composition of the dentin proposed by Legeros [8], in particular in the case of the mineral elements: Ca: 35%, P: 16.9%, and then Mg: 1.23% and Na: 0.6%. The other elements indicated by this author, in far smaller proportions (K: 0.05%, F: 0.06%, Cl: 0.01%) could not be detected owing to the sensitivity limitations of the method. The carbon and nitrogen observed were essentially due to the collagen. One member of our team demonstrated, in fact, that the carbon from the carbonate ions in the apatite played very little part in the intensity of the signal [9]. The atomic ratios of the main constituents of the dentin were determined from the regional spectra Ca 2p, P 2p, C 1s and N 1s. The ratios Ca/P, Ca/N and C/N are given in Table 1. These ratios will be used later in Section 4, but it should be pointed out here that the Ca/N ratio reflects the proportion of mineral substances (apatite) to organic substances (collagen). 3.2. The glass ionomer cement The powder of the glass ionomer cement was studied first. The general XPS spectrum is shown in Fig. 2. It permits

Fig. 2. XPS general spectrum of glass ionomer powder.

232


Fig. 3. XPS regional spectrum (C1s) for glass ionomer cement.

Fig. 4. (a) XPS general septrum (0–400 eV) of the dentin interface, (b) XPS general spectrum (400–800 eV) of the dentin interface.


233

and 33%, respectively. The peak at 285 eV corresponds to atoms of aliphatic carbon, while the other at 289 eV, corresponds to the atoms of carboxylic carbon –COO. The relative proportions of these two types of carbon show that two out of three carbon atoms are aliphatic and one carboxylic. These proportions correspond precisely to what is determined in the case of polyacrylic acid[–CH2CH (CO2H)–]. These findings are in agreement with the composition given for the liquid in this GIC, which is essentially polyacrylic acid based. 3.3. The interfaces

Fig. 5. Dentin interface viewed using SEM in secondary electrons.

identification of the basic constituents of glass: aluminium, silicon and fluorine. However, it also reveals the presence of very large quantities of strontium, while calcium is absent, at the sensitivity limit of the method used (1%). The atomic ratios between these elements were determined on the basis of the regional spectra and are given in Table 1. There was a large amount of strontium. The powder was also analysed using ICP-MS (Inductive Coupling Plasma Mass Spectometry) and chemical analysis. These techniques confirmed the above results, namely the abundance of strontium and the very small proportion of calcium. It would thus seem that, in this glass ionomer, the calcium element, which ensures setting, was replaced almost entirely by strontium. This substitution will be discussed later. The glass ionomer cement, obtained after mixing the powder with the liquid, was also studied using XPS. The general spectrum enabled the previously indicated elements to be identified, but the presence of a large quantity of carbon (derived from polyalkenoic acid) was also noted. The C1s spectrum of the carbon (Fig. 3) appears clearly dissymmetrical. It was decomposed into peaks, at 285 and 288.6 eV, the relative intensities of which are equal to 66

The dentin side and glass ionomer side interfaces were obtained under the conditions described earlier, and then studied. The general spectrum for the DI (Fig. 4(a) and (b)) shows the elements in the dentin, but also those of the GIC. SEM analysis carried out with secondary electrons showed that the interface was continuous and that there were no residual patches of cement (Fig. 5). It was very thin as it did not show up in the picture obtained with back scattered electrons in which dentin tubuli were observed (Fig. 6). These electrons ˚ . The results come from a depth of several thousand A obtained using XPS were thus clearly representative of the composition at the interfaces. It was noted that peak N1s was located at 400 eV. This peak corresponded to the nitrogen atom of the peptide bond in the collagen. The general spectrum of the GII (Fig. 7(a) and (b)) shows not only the elements contained in the GIC, but also calcium from the dentin. The absence of the N1s peak from the spectrum showed that there was no longer a collagen structure in the GII. The atomic ratios between the elements observed in the two interfaces were determined and set out in Table 1 together with the ratios observed in the glass ionomer cement and the dentin. Their variations will be discussed in Section 4. 4. Discussion

Fig. 6. Dentin interface as viewed using SEM with back scattered electrons.

A considerable amount of research has been devoted to studying auto-polymerisable and light-polymerisable glass ionomer cements [10–12]. However, these studies have generally been related either to the setting mechanism of these cements or to the selective release of its constituent elements in contact with pure water or water acidified by lactic acid. Thus, Forss [13] demonstrated that samples of light-polymerisable or auto-polymerisable glass ionomer placed in solution for periods ranging from 1 to 122 days continued to release their mineral components: Al, Si, Na, F, Sr, Ca. However, little work has been carried out on the interactions that take place when the cement is brought into contact with the dentin. Ideas in this connection are often confused, and there is a need for clarification. It is essential

234


Fig. 7. (a) XPS general spectrum (0–400 eV) of the glass ionomer interface, (b) XPS general spectrum (400–800 eV) of the glass ionomer interface.

to differentiate between short-term and long-term interactions. Short-term interactions are those that occur when the freshly prepared glass ionomer cement is brought into contact with the dentin. They correspond to the rapid interdiffusions that occur between the elements in the dentin and those of the glass ionomer cement when the cement is not completely set. These interdiffusions enable the GIC to adhere to the dentin. They cease once the cement has set completely. Long-term interactions are those that occur over a long period of time. They correspond to the slow diffusion of some elements of the glass ionomer cement through the dentin. They can be caused by the presence of water in the buccal environment. Studies devoted to glass ionomer cements, whether they relate to setting mechanisms or to their interaction with the dentin, have been conducted using electron microscopy, generally in association with EDS or WDS microanalysis. However, these techniques analyse to a depth in the order a micron, and the information on adhesion, which relates only

to the first atomic layers, is thus diluted. SEM has, in fact, shown that a very thin continuous film forms at the interface. It is so thin that it cannot be studied by microanalysis. XPS, on the contrary, has the advantage of examining only the adhesion layer, the depth of analysis being only a few ˚ . In addition, electronic microanalysis is localised (in tens A the order of a square micron) and can be distorted by surface irregularities. However, the area examined using XPS, which is in the order of several tenths of a mm 2, is far more representative of the sample. XPS does not require pre-treatment of the sample, such as metallisation, whereas this is essential for electronic microanalysis. Furthermore, XPS spectroscopy permits qualitative and semi-quantitative analysis of all the elements in the periodic table (except for hydrogen and helium, which have no core electron), a characteristic which is not frequent in microanalysis. EDS, in fact, only analyses elements above Na, whereas WDS, which is not often available on the apparatus used, involves very long acquisition times. Finally, XPS has the advantage of clearly showing the chemical form in which the elements occur, whereas this is not possible with EDS and WDS


235

Fig. 8. Solubility isotherms of calcium phosphates.

microanalysis. We were thus able to show that the carbon in the GIC was basically due to polyacrylic acid. To sum up, XPS is the best technique for studying phenomena that occur on the interface between the dentin and the glass ionomer cement. The study undertaken made it possible to determine the atomic ratios between the elements observed in the interfaces, as well as those observed in the GIC and the dentin (Table 1). This table makes very interesting reading and calls for the following comments. Considering the characteristic elements of the dentin (Ca, P, N), we note that the Ca /N and Ca/P ratios in the DI are smaller than in the dentin, and that calcium is present in the cement interface, whereas the glass ionomer powder does not contain any. There can be no doubt that calcium

migrates from the dentin to the glass ionomer through diffusion. Moreover, this increase cannot be caused by the retention of dentin fragments at the glass ionomer interface as its XPS spectrum does not reveal the presence of peptidic nitrogen due to the collagen in the dentin. In addition, such retention would not have any effect on the Ca/N ratio. Decalcification of the dentin thus takes place in contact with the GIC. However, although the decalcifying action of polyacrylic acid is well known [14], it is not known whether it retains this property when mixed with other constituents of the glass ionomer. In order to check this, an experiment was conducted by exposing the dentin to the action of freshly prepared GIC for 1 min. The cement, which had not yet completely set, was then removed by flushing with water. XPS analysis clearly showed the

236


Table 2 Characteristics of calcium and strontium

Ionic radius Q (F2) Q (O2) Solubility

Calcium

Strontium

˚ 0.99 A 2.90 Kcal 1.60 1.6 mg/100 cc

˚ 1.13 A 2.90 Kcal 1.20 1.2 mg/100 cc

disappearance of the calcium and the phosphorus, while there remained only the carbon and the nitrogen of the protein part of the dentin. It would thus appear that a part of the polyacrylic acid in the cement, although mixed with the glass ionomer powder, is still capable of having a decalcifying effect on the dentin and of laying bare the collagen fibers. We can thus understand why there is not any need to pre-treat the dentin with a decalcifying agent before a GIC is applied. Moreover, if water is not added subsequently, reprecipitation of a calcium phosphate probably occurs, with a different Ca/P ratio. The solubility isotherms produced by software developed by Heughebaert [15] show that, at experimental temperatures, a calcium phosphate exposed to a pH below that of the singular point A becomes covered by a surface layer of a more acidic phosphate (Fig. 8). Under these pH conditions, a calcium phosphate is formed that has a lower Ca/P ratio than that of the initial dentin, as observed, moreover, with XPS. If we consider the characteristic elements of glass ionomer cement (Al, Si, Sr, F, Na), we note that the ratio of the “diffusible” elements Al, F, Sr, Na, to Si on the silicon increase from the GII towards the dentin-interface: Al, Sr, F, and Na depletion occurs in the cement interface, these elements migrating preferentially towards the dentin. Here again, the variations in atomic ratio between the cement and the different interfaces clearly indicate that ion diffusion takes place. The presence of glass ionomer fragments at the dentin interface thus appears to have no influence on the ratios in the two interfaces.

Fig. 9. Chart showing chemical exchanges between dentin and glass ionomer cement.

If we compare the characteristic elements of the dentin and the glass ionomer with one another, we see that the decrease in the Ca/C and the increase in the C/N ratios noted in the dentin interface, with regard to the dentin, reflects the migration of polyacrylic acid towards the dentin. The Ca/Si and Ca/C ratios show that the dentin interface is formed of the elements both of the dentin and of the glass ionomer cement, while the glass ionomer interface essentially contains the elements of the glass ionomer cement and a little calcium, but no longer contains any peptidic nitrogen. A layer, composed of a mixture of calcium phosphate that has re-precipitated, polyacrylate and salts from the elements released by glass ionomer attack thus appears on the interface between the dentin and the GIC. It has long been the practice of certain manufacturers to introduce strontium into various filling materials for the purpose of increasing their radio-opacity, generally in the form of sulphate on account of the high insolubility of this compound. Other manufacturers have introduced rare earth oxides for the same purpose. All these chemical elements have in common a high atomic number “Z”, thus favouring the adsorption of X-rays, directly as a function of the number of electrons of the atom in question. Recently, strontium has been partially substituted for calcium in certain types of glass ionomer cement, as pointed out by Forss et al. [13] and Schmalz et al. [16], with a view to increase the radio-opacity of the cements. However, this substitution leads to a release of strontium, as time goes by, when the cement is placed in water or water acidified by lactic acid. The substitution of strontium for calcium is complete in the case of GIC Dyract p (De Trey Dentsply) or Base Line p (Kerr), but these are light-polymerisable cements in which setting is partially due to the opening of unsaturated double bonds of TCB (resin obtained by reaction of tetracarboxylic butan with hydroxy-ethylmethacrylate) in the case of Dyract p. The results that we obtained show that, in the glass ionomer under study, practically all of the calcium was replaced by strontium. We know that calcium plays a decisive role in the setting of chemo-polymerisable cements. The crosslinking of the polymer chains of polyacrylic acid derives partly from the chelation of the calcium [17]. There is room for doubt as to whether, in the event of total substitution, this property can be fully preserved in an auto-polymerisable cement. Comparative study of the physico-chemical properties of calcium and strontium showed that such substitution is possible as these elements have similar characteristics. Their ionic radii are, in fact, approximately the same, and the heat values for reaction with fluorine and oxygen are comparable (Table 2) [18]. In the same way, the solubility of their fluorides is similar. This property is important as the fluorides take the form of microprecipitations formed as the glasses ionomer cool down. These micro-precipitations have to dissolve on the surface so that the cement can set.


5. Conclusion This study shows that an interphase is formed by reciprocal diffusion of the different elements forming the glass ionomer cement and the dentin, with the exception of collagen. The fracture occurs in the interphase when there are no longer any collagen fibers to contribute to reinforcing strength according to a mechanism schematically represented in Fig. 9. It would be of interest to study the variations undergone by this interphase when the GIC is applied, not to young healthy dentin, but to a dentin that is sclerotic or carious by reason of its partial demineralisation and the deterioration to the collagen. Acknowledgements We thank Prof Michel Delamar (ITODYS-PARIS 7) for his support. References [1] K. Shigeru, I. Tatsuya, F. Benji, in: S. Katsuyama (Ed.), Glass Ionomer Dental Cement. The Materials and their Clinical Use, Ishiyaku Euro-America, 1990, pp. 168. [2] S.B. Mitra, B.L Kedrowski, Long-term mechanical properties of glass ionomers, Dent. Mater. 10 (1994) 78–82. [3] R.J. Heys, M. Fitzgerald, Microleakage of three cements bases, J. Dent. Res. 70 (1) (1991) 55–58. [4] A. Lin, N.S. Mc Intyre, R.D. Davidson, Studies on adhesion of glass ionomer cements to dentin, J. Dent. Res. 71 (11) (1992) 1836–1841. [5] J.H. Scofield, Hartree-slater subshell photoionization cross-sections at 1254 and 1487 eV, J. Electron Spectrosc. Relat. Phenom. 8 (1976) 129–137.

237

[6] N.D. Ruse, D.C. Smith, Adhesion to bovine dentin-surfaces characterisation, J. Dent. Res. 70 (6) (1991) 1002–1008. [7] S. Elfersi, A.A. Lebugle, G.L. Gregoire, Apport de la technique ESCA dans la caracte´risation des surfaces dentine-adhe´sifs-mate´riaux de reconstitutions coronaires esthe´tiques, J. de Biomat. Dent. (7) (1992) 141–153. [8] R.Z. Legeros,Calcium Phosphates in Oral Biology and Medicine, New York University College of dentistry, New York, 1991. [9] A.A. Lebugle, G.L. Gregoire, H.E. Sennou, S. Elfersi, submitted for publication in Dentin Analysis by X-ray Photoelectron Spectroscopy. [10] B.E. Causton, N. Johnson, The role of diffusible ions species in the bonding to polycarboxylate cements to dentin: an in vitro study, J. Dent. Res. 58 (4) (1979) 1383–1393. [11] A.D. Wilson, D. Groffman, A.T. Kuhn, The release of fluoride and other chemical species from a glass ionomer cement, Biomaterials 6 (1985). [12] P.V. Hatton, I.M. Brook, Characterisation of the ultrastructure of glass ionomer (polyalkenoate) cements, Br. Dent. J. 173 (1992) 275–277. [13] H. Forss, Release of fluoride and other elements from light-cured glass ionomer cements in neutral and acidic condition, J. Dent. Res. 72 (8) (1993) 1257–1262. [14] G.C. Eliades, International symposium: dentine bonding systems. Reaction paper, in: G. Vanherle, M. Degrange, G. Willems (Eds.), State of the Art on Direct Posterior Filling Materials and Dentinebonding, , 1993, pp. 49. [15] J.C. Heughebaert, S.J. Zawacki, G.M. Nancollas, The growth of octacalcium phosphate on beta tricalcium phosphate, J. Cryst. Growth 63 (1983) 83–90. [16] G. Schmalz, B. Thonemann, M. Riedel, R.J. Elderton, Biological and clinical investigations of a glass ionomer base material, Dent. Mater. 10 (1994) 304–313. [17] A.D. Wilson, H.J. Prosser, D.M.C. Powis, Mechanism of adhesion of polyacrilyte cements to hydroxyapatite, J. Dent. Res. 62 (5) (1983) 590–592. [18] P. Pascal, Nouveau traite´ de chimie mine´rale, Editeur Masson Tome IV:1–6, 1961.