Demineralized dentin 3D porosity and pore size distribution using mercury porosimetry

Demineralized dentin 3D porosity and pore size distribution using mercury porosimetry

d e n t a l m a t e r i a l s 2 5 ( 2 0 0 9 ) 729–735 available at www.sciencedirect.com journal homepage: www.intl.elsevierhealth.com/journals/dema...

1MB Sizes 1 Downloads 51 Views

d e n t a l m a t e r i a l s 2 5 ( 2 0 0 9 ) 729–735

available at www.sciencedirect.com

journal homepage: www.intl.elsevierhealth.com/journals/dema

Demineralized dentin 3D porosity and pore size distribution using mercury porosimetry Elsa Vennat a,∗ , Christine Bogicevic b , Jean-Marie Fleureau a , Michel Degrange c a

Laboratoire Mécanique des Sols, Structures et Matériaux, Ecole Centrale Paris, Grande Voie des Vignes, 92295 Chatenay Malabry Cedex, France b Laboratoire Structures Propriétés et Modélisation des Solides, Ecole Centrale Paris, Grande Voie des Vignes, 92295 Chatenay Malabry Cedex, France c Unité de Recherche Biomatériaux et Interfaces, Faculté de Chirurgie Dentaire, University Paris-Descartes, 1 rue Maurice Arnoux, 92120 Montrouge, France

a r t i c l e

i n f o

a b s t r a c t

Article history:

Objectives. The objectives of this study were to assess demineralized dentin porosity and

Received 22 April 2008

quantify the different porous features distribution within the material using mercury intru-

Received in revised form

sion porosimetry (MIP) technique. We compared hexamethyldisilazane (HMDS) drying and

15 December 2008

lyophilization (LYO) (freeze-drying) in sample preparation.

Accepted 17 December 2008

Methods. Fifty-six dentin discs were assigned into three groups. The control (CTR) group discs were superficially acid-etched (15 s 37% H3 PO4 ) to remove the smear layer and then freeze-dried whereas LYO and HMDS groups samples were first totally demineralized using

Keywords:

EDTA 0.5 M and then freeze-dried and HMDS-dried respectively. MIP was used to determine

Dentin

open porosity and pore size distribution of each pair of samples. Field emission scanning

Demineralization

electron microscopy (FESEM) was used to illustrate the results.

Porosity

Results. The results showed two types of pores corresponding either to tubules and micro-

Pore size distribution

branches or to inter-fibrillar spaces created by demineralization. Global porosity varied from

Mercury porosimetry

59% (HMDS-dried samples) to 70% (freeze-dried samples). Lyophilization drying technique

FESEM

seems to lead to less shrinkage than HMDS drying. FESEM revealed that collagen fibers

Drying techniques

of demineralized lyophilized samples are less melted together than in the HMDS-dried samples. Significance. Demineralized dentin porosity is a key parameter in dentin bonding that will influence the hybrid layer quality. Its characterization could be helpful to improve the monomers infiltration. © 2008 Academy of Dental Materials. Published by Elsevier Ltd. All rights reserved.

1.

Introduction

A micromechanical interlocking principle is currently proposed as the prime mechanism of dentin bonding [1]. Since intact mineralized dentin does not permit much monomer diffusion in a clinically relevant time, dentin is acid-etched prior



to resin infiltration. A collagen fiber network is revealed and the challenge is to maintain the spaces between fibrils after hydroxyapatite crystals have been removed to obtain a durable seal. Characterization of the open porosity of this network is crucial to predict bonding efficiency, which is not always achieved by practitioners.

Corresponding author. Tel.: +33 1 41 13 17 06. E-mail address: [email protected] (E. Vennat). 0109-5641/$ – see front matter © 2008 Academy of Dental Materials. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.dental.2008.12.002

730

d e n t a l m a t e r i a l s 2 5 ( 2 0 0 9 ) 729–735

Monomers do not always fully penetrate the porous substrate [1–3] and a phase separation was reported by Spencer and Wang [4] and Wang and Spencer [5], the collagen network playing the role of filter between resin phases. These defects of the hybrid layer lead to further degradation, which affects bonding durability [6,7]. The collagen network impregnation is crucial [8]. Its porosity and pore size distribution characterization is essential to aim at improving infiltration. Mercury porosimetry is an experimental method that leads to these features assessment and has never been used on demineralized dentin (to the authors’ knowledge). Mercury porosimetry characterizes a material’s porosity by applying various levels of pressure to a sample immersed in mercury. The pressure required to intrude mercury into the sample’s pores is inversely proportional to the pores size. Previously, dentin structure and demineralized dentin porosity have been explored using microscopic techniques as transmission electron microscopy [9,10], scanning electron microscopy [11], and atomic force microscopy [12]. These techniques give section characteristics but no information on the core of the substrate. In a recent study, Figueiredo de Magalhães et al. [13] have quantified dentin 3D open porosity using a technique of immersion and penetration with xylol but pore size distribution was not obtained. Demineralized dentin presents two types of porosity: one due to the tubules, the other due to the collagen meshwork. Tubule density and diameter have been widely observed. The porosity due to the tubules has been reported to be 12–32% [14] and 21% [15] with tubule mean diameter of 3–3.5 ␮m [16] for demineralized dentin, whereas in undemineralized dentin it has been reported to be 1–22% [14] and 10% [15] with tubule mean diameter of 1.2–2.5 ␮m [15] and 0.8–2.5 ␮m [17]. Note that the tubules are enlarged by demineralization. Observed under the transmission electron microscope (TEM), spaces between collagen fibrils have been reported to be around 20 nm by Tay et al. and Van Meerbeeck et al. [9,10] but little appears to have been done in terms of global 3D porosity or pore distribution of demineralized dentin. Nevertheless, studies of mineral fraction dentin [18] indicate 50 vol% of mineral and 20 vol% of fluid. Thus, the potential porosity of an uncollapsed, totally demineralized and dried dentin is 70%, that is to say a porosity of roughly 55% for the collagen network alone. Mercury intrusion porosimetry (MIP) enables the 3D measurement of open porosity, and has already been used on paper [19], hydrogels [20], dental or bone substitute cements [21–23], and investments [24]. Our work focuses on the applicability of MIP to quantification of porosity and assessment of pore size distribution of demineralized dentin discs. MIP requires the specimens to be dry and we assessed its ability to monitor the influence of the drying technique. The process used to dry biological specimens is crucial for scanning electron microscope (SEM) sample preparation, as discussed by Carvalho et al. [25] and Perdigao et al. [26], and for MIP sample preparation too. We used MIP to compare hexamethyldisilazane (HMDS) drying and lyophilization (LYO) (or freeze-drying). HMDS-drying generates few dimensional changes in demineralized dentin [25,26], whereas the more common lyophilization drying technique is used before mercury porosimetry tests [20].

2.

Materials and methods

2.1.

Sample preparation

Fifty-six extracted non-carious human third molars were used for this study. The teeth were gathered following informed consent according to the protocols approved by the review board of the Dental Faculty of Paris-Descartes University. They were stored for a maximum of 3 months in a 1% chloramine-T solution at 4 ◦ C. Dentin discs were cut from crown segments parallel to the occlusal surface at the top of the pulp chamber using a water-cooled, low-speed diamond saw (Isomet, Buehler, Evanston, IL, USA). After enamel removing by grinding, the specimens were reduced to various thicknesses (from 0.8 to 1.5 mm) by wet grinding with SiC paper decreasing in grain size to #2400 to create a smooth, uniform surface on both sides. Any specimens in which pulp horn or enamel was detected were eliminated. The samples were randomly assigned to three groups. In the control (CTR) group, both sides of the 10 discs were acid etched with 37% phosphoric acid for 15 s and then rinsed to remove the smear layer formed during preparation procedures. Etching also demineralized a few microns of the dentin surface. The discs were immersed in liquid nitrogen for half an hour to freeze the water instantaneously and then transferred as quickly as possible into the chamber of an Alpha Christ 2-4 L.S.C. freeze-dryer (Bioblock Scientific, Illkirch, France), under vacuum. Freeze-drying was carried out for 24 h under 7.5 × 106 Torr (1 Pa), with a condenser temperature fixed at −43 ◦ C. With sublimation of the frozen water, the sample temperature increased progressively to +10 ◦ C. The temperature was then gradually increased to +60 ◦ C over 3 h while maintaining the vacuum. In both the HMDS and LYO groups, respectively, 11 and 15 discs were demineralized by immersing them for 3 weeks in 0.5 M EDTA pH 7 at 4 ◦ C. The effectiveness of demineralization was checked by EDS X-ray microanalysis (IMIX-PTS, PGT, Princeton, England) of randomly selected broken samples. The absence of Ca and P peaks indicated complete demineralization. After fixation in 2.5% glutaraldhehyde in 0.1 M sodium cacodylate buffer for 10 h, the 21 discs of the HMDS group were rinsed with 0.2 M sodium cacodylate buffer pH 7.4, dehydrated in ascending grades of ethanol (25% for 20 min, 50% for 20 min, 75% for 20 min, 95% for 20 min, and 100% for 60 min), immersed in HMDS for 10 min, placed on a filter paper inside a covered glass vial, and air-dried at room temperature. Porosimetry tests were performed on 20 discs and the remaining disc was used for SEM observations. After demineralization, the 25 discs of the LYO group were freeze-dried. Porosity was assessed on 20 discs, one disc was used for SEM observations, as the four last LYO discs were used to validate the method. Sample treatments are summarized in Table 1.

2.2.

Mercury intrusion porosimetry

Mercury porosimetry [27], which has been used to evaluate open porosity and pore size distribution of various materials,

d e n t a l m a t e r i a l s 2 5 ( 2 0 0 9 ) 729–735

Table 1 – Treatments of the different groups. Group Treatments

CTR

LYO

15 s 37% H3 PO4 Lyophilization

3 weeks in EDTA Lyophilization

3 weeks in EDTA HMDS-drying

4 cos  P

(1)

A mercury porosimeter (Micromeritics Autopore IV, Micromeritics, USA) was used with a small sample holder or penetrometer (ref. 13-0776). After weighing the sample (Sartorius TE214S, Labandco, France) with a precision of 0.1 mg, application of a low pressure, evacuated gas and filled the sample holder with mercury and porosimetry was then performed at about 0.003–0.20 MPa. After weighing the assembly including penetrometer, mercury and sample (to obtain the sample volume using the porosimeter as a pycnometer), pressures of between 0.20 and 200 MPa were applied. The contact angle and surface tension assumed for all tests were 130◦ and 485 dyn/cm, respectively [28]. Measurements of intruded volume of mercury versus applied pressure were obtained and the pressures were converted into pore sizes using Eq. (1). The higher intrusion pressure was 200 MPa corresponding to a pore diameter around 0.01 ␮m (Eq. (1)). Two samples per group were combined for each test to have sufficient material in the penetrometer. However, the measured mercury intrusion volume was between 3% and 10% of the maximum intrusion volume, which is below the manufacturer’s recommendation of between 20% and 80%. This is why four supplementary LYO discs were used for a test within the recommended range to validate the tests with two discs.

2.3.

Results

3.1.

Mercury intrusion porosimetry

HMDS

is based on the fact that mercury does not wet most substances and, therefore, will not penetrate pores by capillary action unless it is forced to do so. Entry into pore spaces requires application of pressure to balance the capillary pressure drop. By assuming that the pores are cylindrical, and this is a major drawback of the method, the applied pressure P can be related to the pore diameter D via  and , the surface tension and contact angle of mercury (Eq. (1)): D=−

3.

Porosimetry data are presented in Figs. 1–4 and Table 2. A typical plot of the cumulative (A) and differential (B) pore size distributions was obtained for each test of the CTR group (Fig. 1). The curves obtained for CTR dentin discs had the same shape and a peak with a mode around 0.6 ␮m. The LYO samples exhibited a bimodal pore size distribution with two narrow peaks (Fig. 2). The largest pores were around 1 ␮m and the smallest around 50 nm. Data on demineralized HMDS-dried specimens were more scattered (Fig. 3). For this type of sample, two peaks corresponding to a bimodal distribution of pore sizes were seen, but the second one was more variable and broader than for lyophilized samples. Table 1 summarizes the data on total porosity and pore size distribution. The mean porosity of CTR dentin discs was 4% (Table 2). After demineralization and drying, mean porosity increased significantly. It was 59% for HMDS-dried samples and 70% for

Statistics

Porosity data were analyzed using the Kruskal–Wallis test and the Games–Howell test, which is useful for the 2 by 2 comparison of batches with different variances. The significance level was fixed at p = 0.05.

2.4.

731

Field emission scanning electron microscopy

Field emission scanning electron microscopy (FESEM) was used to qualitatively compare the pore structure of the freeze-dried (LYO group) and HMDS-dried (HMDS group) samples using an FEG LEO 1530 microscope (LEO Elektronenmikroskopie GmbH, Oberkochen, Germany) operated at low voltage (1–5 kV). Observations were made on sample sections and on whole samples after nitrogen fracturation. All specimens were coated with gold using coater Bio Rad SC 500 (Microscience division, Elescience).

Fig. 1 – Cumulative (A) and differential (B) pore size distributions obtained for CTR dentin specimens.

732

d e n t a l m a t e r i a l s 2 5 ( 2 0 0 9 ) 729–735

Fig. 2 – Cumulative (A) and differential (B) pore size distributions obtained for demineralized lyophilized dentin specimens (LYO).

Fig. 3 – Cumulative (A) and differential (B) pore size distributions obtained for demineralized HMDS-dried dentin specimens (HMDS).

Fig. 4 – Histogram of the pore size distribution of the three groups.

d e n t a l m a t e r i a l s 2 5 ( 2 0 0 9 ) 729–735

Table 2 – Mean porosity (SD), part due to the pore diameter D > 0.2 ␮m and part due to the pore diameter D < 0.2 ␮m.

Mean porosity (%) Mean porosity corresponding to D > 0.2 ␮m (%) Mean porosity corresponding to D < 0.2 ␮m (%)

CTR

LYO

HMDS

4 (0.8) 4 (0.8)

70 (8) 33 (4)

59 (16) 34 (7)



37 (6)

25 (15)

lyophilized samples, with a higher standard deviation for the former (Table 2). A histogram of pore size distribution (Fig. 4) compares the porous structures of the samples. The first peak was significantly higher for demineralized samples, which alone had a second peak corresponding to smaller pores. Compared with the LYO samples, the two peaks of the HMDS-dried samples were shifted to smaller pore diameters. The test performed with four LYO discs gave values for porosity (64%) and the ratio between nano-scale (53% of total porosity) and micro-scale porosity (47% of total porosity) similar to those recorded in the tests with two discs.

3.2.

733

was observed between LYO and HMDS porosities. Analysis of nano-porosity, corresponding to the inter-tubular and peritubular zones, showed that the porosity of the LYO samples was significantly higher than that of the HMDS samples. HMDS dehydration of the collagen meshwork led to more shrinkage than freeze-drying.

3.3.

Field emission scanning electron microscopy

The CTR group was not studied except for the EDS X-ray analysis because the smear layer removal resulted in opening of the tubules and the images would not have reflected the real porosity of intact dentin. This highlights the value of mercury porosimetry data recorded using penetration inside the sample, as opposed to 2D data. FESEM imaging (Fig. 5 and Fig. 6) showed a collapsed superficial collagen layer. Nevertheless, inter-fibrillar porosities were observed in depth in fractured samples and were between 50 and 200 nm wide (Fig. 5A). At this magnification, the collagen bands can be observed. In the HMDS group (Fig. 6), tubules and micro-branches diameters ranged from 0.5 to 3 ␮m and inter-fibrillar spaces from less than 50–150 nm. The collagen bands cannot be seen.

Statistics

4. Total porosity differed significantly between the groups (p < 0.0001). It was statistically higher in the two groups of demineralized samples than in the control group, but no difference

Discussion

Our data confirm basic knowledge about undemineralized and demineralized dentin pore structure. The CTR samples (only

Fig. 5 – LYO specimens showing only one type of porosity (tubules) (A) but when fractured (B), inter-fibrillar spaces of the inter-tubular zone are revealed.

Fig. 6 – HMDS specimen showing two types of porosity: (A) tubules (and micro-branches) and collagen fiber network. (B) A higher magnification of porous inter-tubular dentin.

734

d e n t a l m a t e r i a l s 2 5 ( 2 0 0 9 ) 729–735

superficially demineralized) had only one pore size of around 0.6 ␮m corresponding to tubules and micro-branches (see ref. [29] for more information about the density and sizes of these features). The average porosity was 4% (depending on the location of the extracted dentin), which is in good agreement with a theoretical tubular porosity between 1% and 22% depending on the location within the material [14]. This theoretical porosity does not take into account odontoblastic processes and their lamina limitans, which explains why the measured porosity is in the lowest part of the theoretical range. In a recent study, Figueiredo de Magalhães et al. [13], using xylol absorption method, have found on intact dentin, open porosity values varying from 1.11 to 3.08%. These lower average values could be explained by the desiccation process of their samples before testing. Lyophilized demineralized dentin clearly showed two pore sizes (Fig. 2). The first one is around 1 ␮m and illustrates the opened tubules and micro-branchings. This first pore size mean diameter and the corresponding porosity (33%) are higher than in the CTR group because highly mineralized peritubular dentin has been removed by demineralization. This porosity value is in good agreement with data from Garberloglio modified by Nakabayashi and Pashley [14], who report 13.4–33.8% of surface area occupied by tubules in acid-etched dentin. Our value is in the highest part of this range because this pore size is not only due to the tubules: micro-branchings are obviously part of the first peak of porosity. This pore diameter is smaller than the diameters reported for opened tubules (from 0.8 to 3 ␮m) because of the presence of microbranchings. As regards with the second porosity class, mercury porosimetry indicated pore diameters of 1–100 nm whatever the drying mode (see Fig. 4), which is confirmed by FESEM imaging (Fig. 5B). The inter-fibrillar spaces reported to be around 20 nm on TEM images by Tay et al. [10] and Van Meerbeeck et al. [9] are in between the pore size range assessed by MIP. Kinney et al. [30] described hydroxyapatite crystallites whose description is helpful to validate our results: the spaces occupied before demineralization by crystallites will constitute the interfibrillar spaces after demineralization. The small-angle X-ray scattering study by Kinney et al. [30] indicated that crystallites are needle-shaped near the pulp and progress to a more platelike shape with increasing proximity to the dentin–enamel junction. The thickness of crystallites is reported to be around 5 nm and constant across dentin. The plate-like crystallites had a gyration radius of roughly 50 nm. These results confirm the range of inter-fibrillar spaces found using MIP. Although showing two different types of pores, the mercury porosimetry data on HMDS-dried demineralized dentin samples are more scattered (Fig. 3). The discrepancy of these results may stem from the collapse of the collagen network. Moreover, comparing the pore sizes with lyophilized samples, each type of pore seems to be shifted to smaller diameters (Fig. 4), indicating greater shrinkage than that shown by LYO samples. However, FESEM images (Fig. 6) clearly show two types of porosity that are not always detected by mercury porosimetry: the opened tubules and micro-branchings that are clearly identifiable on FESEM images (Fig. 6A). The microbranch diameter is around 0.7 ␮m and the tubule diameter is around 2.5 ␮m. And Fig. 6B shows inter-fibrillar spaces. We

may wonder whether this drying technique is reliable when it is to be used with mercury porosimetry, which did not corroborate FESEM imaging. FESEM imaging clearly indicates important differences between the collagen network obtained by lyophilization and the one obtained by HMDS drying. The gap-overlap structure of collagen fibers is only visible in the lyophilized collagen network. Although a porous collagen network was revealed by fracturing LYO samples, confirming the porosimetry results, no explanation has been found for the denatured and collapsed inter-tubular surface of these samples. And HMDS drying seems to link the fibers together, corroborating the assumption by Nation [31] that HMDS crosslinks proteins and stiffens the tissue, thus preventing its collapse. Mercury porosimetry clearly reveals the core porosity of a sample. Lyophilized demineralized dentin shows a collapsed section (Fig. 5A) and FESEM imaging does not assess directly and quantify its inter-fibrillar porosity. To obtain Fig. 5B, the sample had to be fractured to enter the material core. Mercury intrusion porosimetry investigates the core of the sample and highlighted and quantified the two types of demineralized dentin porosity. Moreover, mercury porosimetry shows that the core porosity in lyophilized demineralized dentin is greater than in HMDS-dried samples (Table 2). Caution is required in interpretation of mercury porosimetry data. As we already pointed out, the theory behind the conversion of applied pressures into pore diameter assumes that the pores are cylindrical, which is obviously an approximation in the present case, especially concerning the inter-fibrillar spaces. Another shortcoming of the method is that small pores are overestimated since the diameter considered is the entry diameter. Mercury porosimetry can only be used on dried samples (like SEM and TEM), so it is crucial to know to what extent the material changes during drying. The lyophilized samples gave less variable results (see the SD data in Table 2) concerning global porosity than the HMDS-dried samples. HMDS drying obviously causes more shrinkage and freeze-drying seems more reliable for mercury porosimetry of demineralized dentin. Nonetheless, it should be noticed that lyophilization could tend to widen pore volume since water expands while freezing. The HMDS drying method has been considered to generate few artefacts [25], [26], but our MIP investigation seemed to induce in samples global (3D) deformations not detected by FESEM (2D) imaging. Finally, it should be pointed out that our study was conducted on smaller porous volumes than usual because of the small dentin samples. This leads to an error of roughly 10% in the demineralized dentin porosity. But the tests are repeatable (see Figs. 1 and 3), especially for the lyophilized samples (CTR and LYO), even though dentin is a natural material which varies a lot (with age, location, person, etc). This repeatability, together with the consistency of the results with current knowledge, corroborates the legitimacy of this method. The test done on four LYO samples (with 3% precision in total porosity) confirmed that the use of two samples per test is sufficient.

5.

Conclusion

We have demonstrated that mercury porosimetry characterizes the porous structure of dried demineralized dentin. This

d e n t a l m a t e r i a l s 2 5 ( 2 0 0 9 ) 729–735

technique not only provides information about total 3D pore volume that can be used to quantify porosity, but also indicates pore size distribution. Despite some limitations due to the technique itself, the material features observed by FESEM have been confirmed and additional data such as dried demineralized porosity and pore distribution have been collected. Furthermore, our findings indicate that freeze-drying seems to be the most reliable of the two drying techniques used in this study. It would be interesting to extend the mercury porosimetry assessment on specimens treated with other postfixation drying techniques as the critical-point drying technique. In restorative dentistry, to obtain an efficient infiltration of superficially demineralized dentin, a key parameter is the substrate porosity. In order to have the highest porosity, pretreatment of dentin can be used and mercury porosimetry might be useful for investigating the effect of different pretreatments on demineralized dentin.

Acknowledgements The authors would like to thank Mrs. F. Garnier, Mrs. F. Karolak, Mrs. M. Paleczny and Mr. S. Legoff for technical assistance in SEM observations, freeze-drying, mercury porosimetry and dentin sample preparation, respectively.

references

[1] Nakabayashi N, Kojima K, Masuhara E. The promotion of adhesion by infiltration of monomers into tooth substrates. J Biomed Mater Res 1982;16:265–73. [2] Sano H, Shono T, Takatsu T, Hosoda H. Microporous dentin zone beneath resin-impregnated layer. Oper Dent 1994;19:59–64. [3] Spencer P, Wang Y, Walker MP, Wieliczka DM, Swafford JR. Interfacial chemistry of the dentin/adhesive bond. J Dent Res 2000;79:1458–63. [4] Spencer P, Wang Y. Adhesive phase separation at the dentin interface under wet bonding conditions. J Biomed Mater Res 2002;62:447–56. [5] Wang Y, Spencer P. Hybridization efficiency of the adhesive/dentin interface with wet bonding. J Dent Res 2003;82:141–5. [6] Hashimoto M, Tay FR, Ohno H, Sano H, Kaga M, Yiu C, et al. SEM and TEM analysis of water degradation of human dentinal collagen. J Biomed Mater Res B Appl Biomater 2003;66:287–98. [7] Yang B, Adelung R, Ludwig K, Bö␤mann K, Pashley DH, Kern M. Effect of structural change of collagen fibrils on the durability of dentin bonding. Biomaterials 2005;26: 5021–31. [8] Breschi L, Mazzoni A, Ruggeri A, Cadenaro M, Di Lenarda R, De Stefano Dorigo E. Dental adhesion review: aging and stability of the bonded interface. Dent Mater 2008;24: 90–101. [9] Van Meerbeek B, Dhem A, Goret-Nicaise M, Braem M, Lambrechts P, VanHerle G. Comparative SEM and TEM examination of the ultrastructure of the resin-dentin interdiffusion zone. J Dent Res 1993;72:495–501.

735

[10] Tay FR, Moulding KM, Pashley DH. Distribution of nanofillers from a simplified-step adhesive in acid-conditioned dentin. J Adhes Dent 1999;2:103–17. [11] Carvalho RM, Ciucchi B, Sano H, Yoshiyama M, Pashley DH. Resin diffusion through demineralized dentin matrix. Rev Odontol Univ São Paulo 1999;13:417–24. [12] Habelitz S, Balooch M, Marshall SJ, Balooch G, Marshall Jr GW. In situ atomic force microscopy of partially demineralized dentin collagen fibrils. J Struct Biol 2002;138:227–36. [13] Figueiredo de Magalhães M, Neto Ferreira RA, Grossi PA, de Andrade RM. Measurement of thermophysical properties of human dentin: effect of open porosity. J Dent 2008;36:588–94. [14] Nakabayashi N, Pashley DH. Hybridization of dental hard tissues. Tokyo: Quintessence Publishing Co. Ltd.; 1998. pp. 14–15. [15] Garberoglio R, Brännström M. Scanning electron microscopic investigation of human dentinal tubules. Arch Oral Biol 1976;21:355–62. [16] Fosse G, Saele PK, Eide R. Numerical density and distributional pattern of dentin tubules. Acta Odontol Scand 1992;50:201–10. [17] Pashley DH. Dentin: a dynamic substrate – a review. Scanning Microsc 1989;3:161–74. [18] Driessens FCM, Verbeeck RMH. Biominerals. Boca Raton: CRC Press; 1990. pp. 163–178. [19] Moura MJ, Ferreira PJ, Figueiredo MM. Mercury intrusion porosimetry in pulp and paper technology. Powder Technol 1995;160:61–6. [20] Kim SH, Chu CC. Pore structure analysis of swollen dextran-methacrylate hydrogels by SEM and mercury intrusion porosimetry. J Biomed Mater Res 2000;53:258–66. [21] Milutinovic-Nikolic AD, Medic VB, Vukovic ZM. Porosity of different dental luting cements. Dent Mater 2007;23:674–8. [22] Watson KE, Tenhuisen KS, Brown PW. The formation of hydroxyapatite-calcium polyacrylate composites. J Mater Sci Mater Med 1999;10:205–13. [23] Hesaraki S, Sharifi D. Investigation of an effervescent additive as porogenic agent for bone cement macroporosity. Biomed Mater Eng 2007;17:29–38. [24] Mueller HJ. Particle and pore size distributions of investments. J Oral Rehabil 1986;13:383–93. [25] Carvalho RM, Yoshiyama M, Brewer PD, Pashley DH. Dimensional changes of demineralized human dentin during preparation for scanning electron microscopy. Arch Oral Biol 1996;41:379–86. [26] Perdigao J, Lambrechts P, Van Meerbeek B, Vanherle G, Lopes ALB. Field emission SEM comparison of four postfixation drying techniques for human dentin. J Biomed Mater Res 1995;29:1111–20. [27] Webb PA, Orr C. Analytical methods in fine particle technology. Norcross: Micromeritics Instrument Corporation; 1997. [28] Mikijelj B, Varela JA, Whittemore OJ. Equivalence of surface areas determined by nitrogen adsorption and by mercury porosimetry. Ceram Bull 1991;70:829–31. [29] Mjör IA, Nordahl I. The density and branching of dentinal tubules in human teeth. Arch Oral Biol 1996;41:401–12. [30] Kinney JH, Pople JA, Marshall GW, Marshall SJ. Collagen orientation and crystallite size in human dentin: a small angle X-ray scattering study. Calcif Tissue Int 2001;69:31–7. [31] Nation JL. A new method using hexamethyldisilazane for preparation of soft insect tissues for scanning electron microscopy. Stain Technol 1983;58:347–51.