1.

Introduction

Since their introduction to commercial use in the 1960s, lasers have been investigated as an alternative to conventional mechanical methods in dental caries treatment. The initial research efforts carried out using continuous wave or pulsed ${\mathrm{CO}}_2$ lasers (wavelength ranging from 9.3 to $10.6\mu \mathrm{m}$)^1–4 and Nd:YAG lasers with wavelength 1064 nm^5,6 failed to satisfy the medical requirements. Continuous wave lasers do not allow controlling precisely the laser–material interaction time, often leading to extensive melting, carbonization, and thermomechanical cracking of the teeth and, in extreme cases, necrosis of the pulpal tissue.⁷ Pulsed ${\mathrm{CO}}_2$ and Nd:YAG lasers, with pulse duration in the ${10}^{-3}$ to ${10}^{-9}\mathrm{s}$ range, allow reducing thermal damage but, due to the relatively long interaction time, heat transfer from the interaction region to the tooth bulk is still important, leading to melting, thermomechanical cracking,^1,3,8 loss of collagen, and decomposition of hydroxyapatite.⁹ Moreover, the ablation rates achieved with those lasers are insufficient for the clinical applications.¹⁰

Several approaches have been tested to improve these results. Er,Cr:YSGG and Er:YAG lasers, which emit at 2.79 and $2.94\mu \mathrm{m}$, respectively, allow achieving relatively high ablation rates with minimal thermal degradation of the teeth and became the main type of lasers used in clinics.^11,12 This is due to the fact that their radiation wavelengths are within a strong absorption band of water.^13,14 As a result, radiation is efficiently absorbed by the water contained in the tissues, which warms up and boils explosively, leading to the mechanical disruption of the tissue matrix.^15–17 This ablation mechanism minimizes heat transferred to the tooth, but it has been reported that the shockwave created by the rapid water evaporation may lead to the formation of sub-superficial cracks.¹⁸ These thermomechanical effects can be minimized by a proper selection of the processing parameters and by using a cooling water spray.¹⁹ Another approach to reduce thermal effects is using UV lasers.^20–22 Eugenio et al.²² and Sivakumar et al.²³ showed that intertubular dentin can be ablated by a KrF excimer laser with a 248-nm radiation wavelength and 30-ns pulse duration with negligible thermal effects and no degradation of the remaining material. The ablation mechanism is predominantly photochemical for intertubular dentin and photothermal for peritubular dentin.^22,24 However, excimer lasers are not suitable for clinical application due to their low ablation rate and the potential health hazard caused by exposure to intense UV radiation. A more effective method to reduce thermal effects is using ultrafast lasers. Femtosecond lasers have been shown to ablate dentin^25–27 and bone²⁸ efficiently and with negligible thermal degradation of the underlying tissue. For such short pulse durations, the interaction time is shorter than the characteristic energy relaxation times such as the electron-lattice energy transfer time and heat conduction time.²⁸ Therefore, no energy is transferred to the lattice during the laser pulse and most of the absorbed radiation energy is spent on the ablation process and carried away by the ablated material. Moreover, since femtosecond laser radiation absorption by dielectric materials is highly nonlinear, the interaction zone is confined to a very small region, similar to the focal volume. With an ultrafast laser, the energy efficiency is very high and the ablation rates achieved are higher than those obtained with nanosecond pulse duration infrared²⁹ and UV lasers.³⁰ Surface analysis of the ultrafast laser ablation surfaces of dentin^26,27 and bone²⁸ revealed that no significant changes in the tissue composition and structure occurred.

The ablation of enamel by femtosecond lasers was studied by a number of authors^25,27,29,31 but no detailed characterization on the ablated surfaces was performed. In the present paper, the chemical and structural modifications induced by ultrafast laser ablation on enamel are studied and related with enamel ablation mechanisms.

2.

Materials and Methods

3.

Results

3.1.

Ablation Threshold

The ablation threshold, calculated using the D-scan method described in Sec. 2, decreases with increasing number of laser pulses [Fig. 1(a)], indicating the existence of an incubation effect.²⁸ This effect was explained by the formation of defects in the material for radiation intensities below the ablation threshold, which create electronic energy states within the energy gap that facilitate radiation absorption for subsequent laser pulses.³⁴ The single pulse ablation threshold and the incubation coefficient were calculated from multiple pulse experiments’ results using the equations proposed by Jee et al.³⁴

Eq. (4)

$$F_{\mathrm{th}}(N)=F_{\mathrm{th}}(1)·N^{S-1},$$

Eq. (5)

$$N·F_{\mathrm{th}}(N)=F_{\mathrm{th}}(1)·N^S,$$

where $F_{\mathrm{th}}(N)$ and $F_{\mathrm{th}}(1)$ are the ablation threshold values for $N$ pulse and a single pulse, respectively, and $S$ is the incubation factor, which characterizes incubation.

Fig. 1

(a) Variation of the ablation threshold with number of pulses and (b) plot of the accumulated fluence versus number of pulses.

The relationship between the accumulated fluence $N·F_{\mathrm{th}}(N)$ and number of pulses $N$ is depicted in Fig. 1(b). The slope of the line represents the incubation factor $S$. The calculated values of the single pulse ablation threshold and of the incubation coefficient are $3.2\pm 0.3\mathrm{J}/{\mathrm{cm}}^2$ and $0.68\pm 0.08$, respectively.

A plot of $D^2$ versus $Ln(F)$ for single pulse ablation experiments ($F$ and $D$ are fluence and diameter of the craters, respectively) is shown in Fig. 2. The single pulse ablation threshold, determined by extrapolating $D^2$ to zero, is $3.4\pm 0.2\mathrm{J}/{\mathrm{cm}}^2$, in good agreement with the value determined by the D-scan method.

Fig. 2

Semilogarithmic plot of $D^2$ as a function of fluence and linear fitting of the experimental data.

3.2.

Morphology of the Laser-Treated Specimens

The morphology of enamel surfaces treated with five laser pulses and fluences of 4, 9, and $14\mathrm{J}/{\mathrm{cm}}^2$ are shown in Figs. 3(a)–3(f). The surface is covered by a resolidified material layer, which presents round pores and traces of collapsed gas bubbles, showing that the tissue was melted and vapor was generated within the liquid, leading to the formation of gas bubbles that collapsed upon reaching the surface. The continuity and thickness of the resolidified layer and the size of collapsed gas bubbles increase with increasing fluence, as shown in Figs. 3(c)–3(f), indicating that surface temperature increases with laser fluence.

Fig. 3

Morphology of craters created on enamel surface using five laser pulses with radiation fluences of (a, b) $4\mathrm{J}/{\mathrm{cm}}^2$, (c, d) $9\mathrm{J}/{\mathrm{cm}}^2$, and (e, f) $14\mathrm{J}/{\mathrm{cm}}^2$.

The surface morphology changes drastically when the laser beam is scanned continuously [Figs. 4(a), 4(c), and 4(e)]. The laser-treated surfaces are covered with ablation debris, mostly aggregated in clusters, and resolidified droplets. The thickness of this layer and the size of the clusters increase with the fluence, and particles aggregate with $\sim 3$ to $4\mu \mathrm{m}$ in diameter are observed for $14\mathrm{J}/{\mathrm{cm}}^2$. The layer of ablation debris layer of the specimens treated with 4 and $9\mathrm{J}/{\mathrm{cm}}^2$ is poorly adhered to the surface and easily removed by ultrasonication, exposing a layer of molten material underneath [Figs. 4(b) and 4(d)]. The layer formed at $14\mathrm{J}/{\mathrm{cm}}^2$ adheres better to the surface and a significant amount of ablation debris remains after ultrasonication [Fig. 4(f)]. The observation of the cross-section of a specimen treated with $14\mathrm{J}/{\mathrm{cm}}^2$ (Fig. 5) reveals a layer of resolidified material, with a thickness varying in the range of 300 to 900 nm.

Fig. 4

Morphology of laser-treated enamel surfaces before and after ultrasonication. Fluences: (a, b) $4\mathrm{J}/{\mathrm{cm}}^2$, (c, d) $9\mathrm{J}/{\mathrm{cm}}^2$, and (e, f) $14\mathrm{J}/{\mathrm{cm}}^2$.

Fig. 5

Cross-section of a specimen treated with $14\mathrm{J}/{\mathrm{cm}}^2$.

3.3.

Surface Constitution

3.3.1.

XRD results

The x-ray diffractograms of polished and laser-treated enamel surfaces are shown in Fig. 6. The peaks of untreated enamel can be indexed as nonstoichiometric hydroxyapatite, according to JCPDS card number 09-0432.²⁸ The diffractograms of the specimens treated with 4 and $9\mathrm{J}/{\mathrm{cm}}^2$ are similar to the diffractogram of the polished specimen, but new peaks appear at 36.2 deg and 42.4 deg in the specimen treated with $14\mathrm{J}/{\mathrm{cm}}^2$, which indicate that new phases formed as a result of the laser treatment. These peaks could be indexed as the (131) and (303) diffraction peaks of $\beta$-tricalcium phosphate ${\mathrm{Ca}}_3{({\mathrm{PO}}_4)}_2$ (TCP, JCPDS card number 09-0169),³⁵ respectively, but the diffraction pattern of this compound presents two intense peaks at $\sim 31.5\mathrm{deg}$ and 34.5 deg, which are not seen in Fig. 6. On the other hand, these peaks cannot be attributed to the other two compounds that may form by the decomposition of hydroxyapatite at high temperature, tetracalcium phosphate ${\mathrm{Ca}}_4{({\mathrm{PO}}_4)}_2\mathrm{O}$ (TTCP, JCDPS 70-1379), and calcium pyrophosphate $\beta \text{-}{\mathrm{Ca}}_2{\mathrm{P}}_2{\mathrm{O}}_7$ (JCDPS 9-0346).³⁶ Therefore, we tentatively ascribe them to the (131) and (303) diffraction peaks of $\beta$-TCP.³⁵ In order to investigate the depth distribution of this phase, GIXRD analysis was performed with several angles of incidence between 0.2 deg and 1 deg with 0.2 deg step. The diffractograms normalized to the peak at 25.8 deg, corresponding to the diffraction of enamel apatite are shown in Fig. 7(a). The regions around the newly formed peaks are magnified and displayed in Fig. 7(b). The intensities of the peaks at 36 deg and 42 deg decrease for increasing incidence angle, showing that $\beta$-TCP exist mainly near the surface, probably in the ablation debris and in the resolidified layer.

Fig. 6

Diffractograms of polished and laser-treated enamel. Angle of incidence: 0.2 deg. The indexed peaks correspond to hydroxyapatite. The arrows indicate the peaks corresponding to $\beta$-TCP, present in the specimen treated with $14\mathrm{J}/{\mathrm{cm}}^2$.

Fig. 7

(a) Full range x-ray diffraction spectra obtained for different incoming angles of enamel specimen treated with $14\mathrm{J}/{\mathrm{cm}}^2$, and (b) the magnified view of the spectra.

3.3.2.

Raman results

The Raman spectra of the enamel specimens, normalized to the strongest peak, at $960{\mathrm{cm}}^{-1}$, which corresponds to the ${\nu }_1$ vibration of the phosphate group in hydroxyapatite,²⁸ are shown in Fig. 8. The peaks at 470, 530, and 1025 to $1085{\mathrm{cm}}^{-1}$ can be assigned to ${\nu }_2$, ${\nu }_4$, and ${\nu }_3$ vibrational modes of the phosphate group.³⁷ The peak at $1070{\mathrm{cm}}^{-1}$, which corresponds to the ${\nu }_1$ vibration of the carbonate group,^28,38 is also observed due to the partial replacement of hydroxyl and phosphate anions by the carbonate anion.

Fig. 8

Raman spectra of polished and laser-treated enamel.

The relative intensity of the ${\nu }_1$ peak of ${\mathrm{PO}}_4^{3–}$ decreases with increasing fluence, which indicates the formation of nanocrystalline and/or amorphous calcium phosphate (ACP) at the surface of the sample.^39,40 The presence of ACP is supported by the increase of relative intensity in the region around $950{\mathrm{cm}}^{-1}$ in the spectra of the laser-treated specimens, which corresponds to the ${\nu }_1$ Raman peak of that compound.^39–41 This peak is still present in the Raman spectrum of the specimen treated with $14\mathrm{J}/{\mathrm{cm}}^2$ after ultrasonication, but its relative intensity decreases (Fig. 9), showing that ACP exists in the ablation debris. Raman spectra of the specimens treated with 4 and $9\mathrm{J}/{\mathrm{cm}}^2$ also show traces of ACP even after all the ablation debris were removed. The results show that the resolidified material also contains ACP.

Fig. 9

Raman spectra of specimen treated with $14\mathrm{J}/{\mathrm{cm}}^2$ before and after ultrasonication.

4.

Discussion

The enamel single pulse ablation threshold of enamel determined in the present work by D-scan³³ and $D^2$³² methods were $3.2\pm 0.3$ and $3.4\pm 2\mathrm{J}/{\mathrm{cm}}^2$, respectively. This value cannot be directly compared with those previously reported in the literature because they were determined for multiple pulse ablation.^25,42 Krüger et al.⁴² found a threshold of $0.6\mathrm{J}/{\mathrm{cm}}^2$ for 100 pulses of 615 nm radiation wavelength with 350 fs duration, while Neev et al.²⁵ reported a value of $0.7\mathrm{J}/{\mathrm{cm}}^2$ for 100 pulses, 350-fs pulse duration, and 1050-nm radiation wavelength. By substituting the data reported by Neev et al.²⁵ and Krüger et al.⁴² in Eq. (5) and using the incubation coefficient value found in the present work ($0.68\pm 0.08$), single pulse ablation thresholds of 3.0 and $2.7\mathrm{J}/{\mathrm{cm}}^2$, respectively, were obtained, in good agreement with our result. On the contrary, the value obtained here is much higher than the single pulse ablation threshold proposed by Ji et al.³¹ ($0.58\mathrm{J}/{\mathrm{cm}}^2$) for 800-nm wavelength, and 85-fs pulse duration.³¹ This difference could be explained by the much shorter pulse duration and reduced wavelength of the laser used by Ji et al.,³¹ but Rode et al.²⁹ found a single pulse ablation threshold of $2.2\mathrm{J}/{\mathrm{cm}}^2$ using laser parameters similar to Ji et al. (805-nm laser wavelength and 92-fs pulse duration). The large scatter of the ablation threshold values found in the literature reflects sample to sample variations typical for dental tissues and other natural materials. The calculated value of the incubation coefficient $0.68\pm 0.08$ is in good agreement with the value reported by Cangueiro et al.²⁸ for bone.

Several mechanisms have been proposed for the ultrashort laser ablation of dielectric materials. At low radiation intensities, Coulomb explosion is predominant,^43,44 which is characterized by the laser-induced emission of photoelectrons from the surface of dielectric materials at the beginning of the laser pulse, resulting in the buildup of a high positive charge density.⁴³ Once the electrostatic repulsion forces between the positive ions exceed the binding forces, these ions are ejected.⁴³ In this ablation mechanism, material is removed mainly by the electrostatic process and thermal effects are neligible.⁴⁴ At higher radiation intensities, part of the absorbed energy is transferred to the lattice by electron–phonon interactions, leading to extremely fast heating of the material, which transforms into a highly overheated liquid.^44,45 Eventually, the liquid phase may undergo different ablation processes including liquid spallation and phase explosion.^44,46 In the present study, a thin layer of resolidified material is observed at the surface of the laser-treated specimens, independently of the processing parameters used. Melting of enamel indicates that a surface temperature above 1600°C was reached,^9,47 and that the ablation process was dominated by thermal processes. On the other hand, the pores and traces of burst bubbles observed in the resolidified layer (Fig. 3) show that the liquid decomposed, resulting in the release of gas. According to Holcomb and Young,⁴⁸ enamel releases gaseous water and ${\mathrm{CO}}_2$ in the solid state at temperatures of 100°C and 250°C, respectively, which are much lower than the melting point of the tissue.^9,48 These reactions cannot occur in the present case due to the extremely fast heating rate and decomposition of enamel only occurs after melting, leading to gas release within the liquid.

The ablation behavior of enamel observed in this study is essentially different from the ablation behavior of bone²⁸ and dentin²⁶ with the same laser system and fluences in the ranges of 0.55 to $2.18\mathrm{J}/{\mathrm{cm}}^2$ and 1 to $3\mathrm{J}/{\mathrm{cm}}^2$, respectively. The values of the single pulse ablation threshold of dentin and bone are significantly lower than enamel’s (0.6, 0.79, and $3.2\mathrm{J}/{\mathrm{cm}}^2$, respectively). On the other hand, no melting was observed in both cases. This difference in ablation behavior can be explained by the different constitution of these tissues. While enamel consists essentially of hydroxyapatite, bone and dentin contain significant amounts of organic compounds and water.^49,50 Moreover, the cohesion of dentin and bone is ensured by a network of collagen fibrils, reinforced by hydroxyapatite nanocrystals. Collagen presents an ablation threshold much lower than the ablation threshold of hydroxyapatite ($0.06\mathrm{J}/{\mathrm{cm}}^2$⁵¹ as compared to $3.3\mathrm{J}/{\mathrm{cm}}^2$⁵²). Therefore, the cohesive strength of bone and dentin can be reduced at lower fluences, facilitating the tissue’s removal.

The surface of the specimens treated with a scanning laser beam was covered by ablation debris, which could be removed by ultrasonication, revealing a layer of resolidified material underneath. The number of pulses per surface position in these experiments was 403, which corresponds to an ablation threshold value of $0.48\mathrm{J}/{\mathrm{cm}}^2$, calculated using Eq. (4). Since the fluences used were significantly higher than this threshold, intense ablation was expected, leading to the ejection of a large amount of ablation debris, which redeposited on the surface after the collapse of the ablation plume,⁴⁵ forming a continuous layer of material loosely adherent to the surface. Below this layer of ablation debris, a very thin layer of resolidified material exists, which for the specimen treated with $14\mathrm{J}/{\mathrm{cm}}^2$ presents a thickness varying within the range of 300 to 900 nm.

Since the ablation debris and the resolidified material were formed by rapid cooling of calcium phosphate from the liquid phase, products such as ACP, tricalcium phosphate, and tetracalcium phosphate can be expected.⁵³ The present results show that the ablation debris and the resolidified material contain ACP, formed by rapid solidification of hydroxyapatite. Traces of $\beta$-TCP were detected by x-ray diffraction in the specimen treated with $14\mathrm{J}/{\mathrm{cm}}^2$, but this identification needs to be confirmed. No other decomposition products of hydroxyapatite were detected in laser-treated surfaces.

5.

Conclusions

Enamel can be effectively ablated using a femtosecond laser with 560-fs pulse duration and 1030-nm wavelength. The single pulse of the ablation threshold is $3.3\mathrm{J}/{\mathrm{cm}}^2$. The ablation threshold of enamel decreases with increasing number of laser pulse indicating an incubation effect characterized by an incubation coefficient of 0.68. The ablation surfaces are smooth and present no evidence of carbonization or cracking. They are covered by a layer of resolidified material, whose thickness increases with increasing fluence. The morphology of the resolidified material indicates that hydroxyapatite gaseous decomposition products (probably ${\mathrm{CO}}_2$ and ${\mathrm{H}}_2\mathrm{O}$) were released within the liquid. In the specimens treated with a scanning laser beam, the resolidified layer is covered by a layer of ablation debris loosely attached to the surfaces, which can be partially removed by ultrasonication. The thickness of the resolidified material layer varied between 300 and 900 nm in the specimen treated with $14\mathrm{J}/{\mathrm{cm}}^2$. Below this thin altered layer, the structure and composition of enamel is preserved. ACP was detected in the ablation debris and in the resolidified material. On the other hand, a new phase, probably $\beta$-tricalcium phosphate ${\mathrm{Ca}}_3{({\mathrm{PO}}_4)}_2$, was observed at the surface of the specimen treated with $14\mathrm{J}/{\mathrm{cm}}^2$. The present results show that the ablation of enamel involves the melting of hydroxyapatite, which is the predominant phase in enamel, but the thickness of the altered layer is very small and thermal damage of the remaining material is negligible.