1.

Introduction

The color of teeth is of particular concern to a large number of people seeking dental treatment. It has been reported that 28% of adults in the United Kingdom and 34% in the United States are dissatisfied with the appearance or color of their teeth.¹

Tooth bleaching is a well-accepted method of treating discolored tooth. It refers to the clinical application of a chemical solution to a tooth surface in order to achieve a lightening effect. The method is based upon hydrogen peroxide (HP) as the active agent.² HP can be applied directly or produced in a chemical reaction from carbamide peroxide. It acts as a strong oxidizing agent through the formation of free radicals, reactive oxygen molecules, and hydrogen peroxide anions.² These reactive molecules attack the long-chained, dark-colored chromophore molecules and split them into smaller, less colored, and more diffusible molecules.²

Although there is little question about the efficacy of tooth bleaching, the safety of this technique has been questioned.² A primary concern is that the structure of sound enamel may be weakened by the bleaching agent. Numerous studies have evaluated the effects of the peroxide-containing bleaching agents on the surface morphology of enamel. Some reported no or little topographic alterations of bleached enamel, ^{3, 4, 5, 6} but others have found increased porosity, pitting, erosion, and demineralization of enamel prism peripheries.^{7, 8, 9} Additionally, surface hardness and wear resistance of enamel have also been investigated, and the results have varied from no effect to significant decrease in hardness and fracture resistance of enamel. ^{10, 11, 12, 13} Several studies have evaluated the effects of bleaching agents on the chemical composition of enamel using energy dispersive spectroscopy,¹⁴ x-ray diffraction, and infrared spectroscopy.^{15, 16} The limitation of the above-mentioned analytical techniques is that they are invasive to the enamel sample. Moreover, these techniques provided very limited information about the organic matter that comprises only a minor part of enamel, and they fail to reach general agreement.

This paper is an application of laser Raman spectroscopy to the study of the structure of human tooth enamel. This technique was selected because it is a very powerful, noninvasive analytical method and involves only minimal sample preparation. It has been proven to be of great value for studying the molecular structure of enamel mineral. ^{17, 18, 19, 20, 21} The vibration modes of ${\mathrm{PO}}_4^{3-}$ , ${\mathrm{CO}}_3^{2-}$ , and ${\mathrm{OH}}^{-}$ groups can be studied by Raman spectroscopy.²¹ The frequency and the band shape of the most prominent ${\nu }_1$ ${\mathrm{PO}}_4^{3-}$ depend on the local mineral environment and therefore change with ionic incorporation and crystallinity.^{21, 22} Meanwhile, the intensity of ${\nu }_1$ ${\mathrm{PO}}_4^{3-}$ is linearly proportional to phosphate group concentration within the hydroxyapatite molecule, and it thus can be used to analyze changes in the phosphate group concentration.^{17, 23}

Laser-induced fluorescence, which is usually regarded as the interference of Raman scattering, can also be measured by a Raman spectrometer.^{24, 25, 26} It is well known that laser-induced fluorescence is one of the most striking features of enamel.^{27, 28} Dental caries can be measured as a change in fluorescence of the tooth substance when exposed to laser light.^{27, 28} Based on recent studies,^{26, 29} it is highly possible that the laser-induced fluorescence of sound enamel originates from the organic matter in enamel. Thus, the laser-induced fluorescence during Raman measurement may provide important information about the organic matter.

The aim of the study was to explore the effects of 30% HP on sound tooth enamel by applying Raman scattering and laser-induced fluorescence. In addition, microhardness testing and scanning electron microscopy (SEM) were also used.

2.

Materials and Methods

The study protocol was reviewed and approved by the Ethics Committee of the School and Hospital of Stomatology, Wuhan University. Patients from whom teeth were being extracted were asked to read and sign a consent form prior to the extraction.

2.3.

Raman Spectroscopic Detection

Raman/fluorescence spectra were recorded with a confocal micro-Raman spectrometer (LabRam HR800, Jobin Yvon Horiba, France) equipped with a 20-mW $\mathrm{He}\text{‐}\mathrm{Ne}$ laser (632.8-nm wavelength). A dry Olympus $100\times$ objective $(\mathrm{NA}=0.9)$ in combination with an adjustable confocal hole was used. Raman scattering was detected by using an air cooled $1024\times 256$ pixels CCD camera. Peak frequencies and rapid checking of instrumental performance were calibrated with the silicon phonon line at $520{\mathrm{cm}}^{-1}$ .

Ten specimens in each group were used for Raman spectroscopy. The specimens were marked and washed in running DW for $30\mathrm{s}$ before the measurements. The Raman/fluorescence spectra were acquired at six points near the marker. Three points were measured using a confocal hole of $100\mu \mathrm{m}$ . For each spectrum, an average of three acquisitions of $30\mathrm{s}$ was used. The other three were measured using a confocal hole of $1000\mu \mathrm{m}$ . An average of three acquisitions of $10\mathrm{s}$ was used. The specimens were measured repeatedly at the same location at 0, 15, 30, and $60\min$ after treatment, and the white light images were recorded. The Raman spectrum of 30% HP solution was also acquired.

Spectral data were visualized on a computer and processed using LabSpec 5 spectroscopic software. Profile data were then imported into Origin 7.0 software (Origin-Lab Corporation, Northampton, Massachusetts). Raman absolute intensity (RAI), Raman relative intensity (RRI), and fluorescence intensity (FI) at $960{\mathrm{cm}}^{-1}$ of the Raman/fluorescence spectra were calculated (Fig. 1 ). The RAI is the intensity of the Raman peak at $960{\mathrm{cm}}^{-1}$ before the spectrum is baselined. The RRI is the intensity of the same peak after the spectrum is baselined between 990 and $930{\mathrm{cm}}^{-1}$ . The FI is equal to RAI minus RRI. All values were transformed to percentage values, where the values of the baseline were 100% and the changed values were calculated as a percentage of the baseline.

Fig. 1

Raman absolute intensity (RAI), Raman relative intensity (RRI), and fluorescence intensity (FI) at $960{\mathrm{cm}}^{-1}$ of the representative Raman/fluorescence spectra. The spectra were recorded with a $100\times$ objective, and the confocal hole was set to $1000\mu \mathrm{m}$ . An average of three acquisitions of $10\mathrm{s}$ was used.

3.

Results

3.1.

Raman Scattering and Laser-Induced Fluorescence of Enamel

The main features of the Raman scattering of enamel were well consistent with previous results^{17, 18} (Fig. 1). The strongest band at $960{\mathrm{cm}}^{-1}$ arose from ${\nu }_1$ ${\mathrm{PO}}_4^{3-}$ . The bands at 1045 and $1024{\mathrm{cm}}^{-1}$ were assigned to ${\nu }_3$ ${\mathrm{PO}}_4^{3-}$ , 610 and $580{\mathrm{cm}}^{-1}$ to ${\nu }_4$ ${\mathrm{PO}}_4^{3-}$ , and $430{\mathrm{cm}}^{-1}$ to ${\nu }_2$ ${\mathrm{PO}}_4^{3-}$ . The bands at 1068 and $3573{\mathrm{cm}}^{-1}$ were attributed to ${\nu }_3$ ${\mathrm{CO}}_3^{2-}$ and ${\mathrm{OH}}^{-}$ stretch, respectively. The laser-induced fluorescence of enamel appeared as a featureless background in the Raman/fluorescence spectra (Fig. 1).

Both RRI and FI of enamel increased greatly with the opening of the confocal hole from 100 to $500\mu \mathrm{m}$ , but they showed little variations with the opening of the confocal hole from 500 to $1000\mu \mathrm{m}$ [Fig. 2a ]. When the enamel was illuminated by the laser of the Raman spectrometer, the RRI remained almost unchanged with increasing time, while the FI decreased considerably [Fig. 2b].

Fig. 2

Effects of confocal hole (a) and laser illumination (b) on the RRI and FI. The spectra were recorded with a $100\times$ objective. An average of three acquisitions of $1\mathrm{s}$ was used.

3.2.

Effects of Treatments on the Raman Scattering of Enamel

In all groups, the frequency of Raman bands, Raman bandwidth at $960{\mathrm{cm}}^{-1}$ , and carbonate phosphate intensity ratio showed little variation after treatment. However, significant changes in RRI were found in groups HP and $\mathrm{HCl}$ (Figs. 3 and 4 ).

Fig. 3

Raman/fluorescence spectra of enamel at each time point in different groups. The spectra were recorded with a $100\times$ objective. An average of three acquisitions of $30\mathrm{s}$ (confocal hole= $100\mu \mathrm{m}$ ) or $10\mathrm{s}$ (confocal hole= $1000\mu \mathrm{m}$ ) was used.

Fig. 4

Percentage RRI and FI changes of different groups using a confocal hole of either $100\mu \mathrm{m}$ [(a) and (c)] or $1000\mu \mathrm{m}$ [(b) and (d)].

Using a $100\text{‐}\mu \mathrm{m}$ hole, the one-way RM ANOVAs (or Friedman RM ANOVAs on ranks) revealed that the percentage RRI remained unchanged in group DW $(\mathrm{P}=0.730)$ , but changed over time in groups HP and $\mathrm{HCl}$ ( $\mathrm{P}=0.016$ and $\mathrm{P}=0.011$ ). Unlike the continual decrease in group $\mathrm{HCl}$ , the percentage RRI of group HP increased significantly in the first $30\min$ and then decreased at $60\min$ . The one-way ANOVAs (or one-way ANOVAs on ranks) showed that there was no significant difference among three groups at $15\min$ after treatment $(\mathrm{P}=0.680)$ but that there were significant differences at $30\min$ and $60\min$ after treatment ( $\mathrm{P}=0.013$ and $\mathrm{P}=0.003$ ). The percentage RRI of group $\mathrm{HCl}$ was significantly lower than that of group HP at $30\min$ , and it was significantly lower than those of groups HP and DW at $60\min$ after treatment.

Using a $1000\text{‐}\mu \mathrm{m}$ hole, the results suggested that the percentage RRI remained unchanged in groups DW and HP ( $\mathrm{P}=0.694$ and $\mathrm{P}=0.410$ ) but changed over time in group $\mathrm{HCl}$ $(\mathrm{P}<0.001)$ . The percentage RRI of group HCl increased significantly at $15\min$ and then decreased continually at $30\min$ and $60\min$ . There was no significant difference among the three groups at $15\min$ and $30\min$ after treatment ( $\mathrm{P}=0.052$ and $\mathrm{P}=0.705$ ), but there were significant differences at $60\min$ after treatment $(\mathrm{P}=0.002)$ . The percentage RRI of group $\mathrm{HCl}$ was significantly lower than those of groups HP and DW at $60\min$ after treatment.

In group HP, a small band at $876{\mathrm{cm}}^{-1}$ became pronounced over time (Fig. 3). It was identified as $\mathrm{O}–\mathrm{O}$ stretching of HP when compared with the Raman spectrum of 30% HP solution (data not shown). This weak band disappeared quickly when the specimens were removed from the HP solution (data not shown).

3.4.

Microhardness Testing

The one-way RM ANOVAs revealed that the percentage microhardness remained unchanged in group DW $(\mathrm{P}=0.761)$ but changed over time in groups HP and $\mathrm{HCl}$ (both $\mathrm{P}<0.001$ ; Fig. 5 ). There was no significant difference among the three groups at $15\min$ after treatment $(\mathrm{P}=0.127)$ , but there were significant differences at $30\min$ and $60\min$ after treatment ( $\mathrm{P}=0.029$ and $\mathrm{P}=0.002$ ). Both the percentage microhardness of group $\mathrm{HCl}$ and that of group HP were significantly lower than that of group DW at $30\min$ and $60\min$ after treatment. And the percentage microhardness of group $\mathrm{HCl}$ was significantly lower than that of group $\mathrm{HCl}$ at $30\min$ and $60\min$ after treatment.

Fig. 5

Percentage microhardness changes of enamel in different groups.

3.5.

Morphological Observations

White light images of the enamel surface in different groups presented different variations (Fig. 6 ). The enamel surface of group DW remained almost unchanged during the treatment period. Those of group $\mathrm{HCl}$ and HP showed obvious morphological changes after treatment. However, the change of the enamel surface in group HP was milder than that in group $\mathrm{HCl}$ at each time point. In both groups, the morphological changes became more pronounced as the treatment time increased.

Fig. 6

White light images of enamel surface at each time point in different groups.

The SEM images revealed that the enamel surface in group DW was smooth, flat, and polished (Fig. 7 ). The surfaces in groups $\mathrm{HCl}$ and HP showed distinct structures of enamel, which included enamel rods and narrow interrods. Under higher magnification, the nanocrystals in the rods and interrods became distinguishable from each other.

Fig. 7

SEM images of enamel surface after treatment in different groups. R: enamel rods; IR: enamel interrods.

4.

Discussion

In the present study, we monitored the changes of enamel subjected to different treatments in Raman scattering, laser-induced fluorescence, microhardness testing, and surface morphology. Although the results of morphological observations and microhardness testing suggested that both HP and $\mathrm{HCl}$ caused changes in the enamel structure, the chemical information obtained from these tests was indirect and limited. The results of Raman scattering and laser-induced fluorescence provided new insights into the chemical changes in the enamel structure. Based on these results, we propose models for the changes in the enamel structure caused by HP and $\mathrm{HCl}$ (Fig. 8 ).

Fig. 8

Proposed models for the changes in the enamel structure caused by HP and $\mathrm{HCl}$ .

An important finding in the study is that significant changes in RRI were found using a $100\text{‐}\mu \mathrm{m}$ hole in both group $\mathrm{HCl}$ and group HP. This indicated that the phosphate group concentration within the enamel surface had changed. It is very interesting that the change of RRI in group $\mathrm{HCl}$ showed similar trends to that of microhardness. The results are not unexpected, because phosphate group concentration within the enamel is a good indicator of the degree of mineralization,^{23, 30} while microhardness determination can provide indirect evidence of mineral loss or gain in the dental hard tissues.³¹

The different trends of RRI changes between groups $\mathrm{HCl}$ and HP using a $100\text{‐}\mu \mathrm{m}$ hole indicated that $\mathrm{HCl}$ resulted in stronger demineralization of enamel than HP did, although they had the same pH value. It is somewhat unexpected to observe that the RRI of enamel in group HP increased significantly from 0 to $30\min$ . A recent study has shown that the polishing procedure could produce a smear layer about 280-nm thick on the enamel surface.⁹ The increase of RRI could be explained by assuming that the phosphate group concentration in the smear layer is lower than that of the underlying enamel. Since the white light images suggested that the smear layer was slowly removed during the first $30\min$ , the exposure of the underlying enamel results in the increase in RRI at $30\min$ . After then, the RRI of the enamel in group HP decreased significantly, which indicated that the freshly exposed underlying enamel might be demineralized markedly from $30\min$ to $60\min$ . Unlike HP, the $\mathrm{HCl}$ might have removed the smear layer completely and demineralized the exposed underlying enamel greatly during the first $15\min$ . Thus, the increase of RRI was not observed in group $\mathrm{HCl}$ .

The RRI changes using 1000- and $100\text{‐}\mu \mathrm{m}$ holes showed different trends in both group $\mathrm{HCl}$ and group HP. This is mainly because the confocal Raman spectrometer has different depth resolutions with different confocal holes. When spectra are acquired with the $100\times$ objective, the depth resolutions determined with a silicon wafer were about $2\mu \mathrm{m}$ using a confocal hole of $100\mu \mathrm{m}$ and $8\mu \mathrm{m}$ using a $1000\text{‐}\mu \mathrm{m}$ hole (as reported by the vendor). Because the optical property of dental enamel is different from silicon wafer, the resolutions of the confocal Raman spectrometer should be changed when the enamel was measured. Although the precise depth resolutions could not be determined in present study, it is quite certain that a Raman spectrometer with a $1000\text{‐}\mu \mathrm{m}$ hole has a much larger analyzed depth than that with a $100\text{‐}\mu \mathrm{m}$ hole. This means that the analyzed volume with the $100\text{‐}\mu \mathrm{m}$ hole is more representative of the surface than that with a $1000\text{‐}\mu \mathrm{m}$ hole.

The significant increase in RRI at $15\min$ in group HP using a $1000\text{‐}\mu \mathrm{m}$ hole also indicated that deeper enamel might have higher phosphate group concentration. And the RRI of the exposed enamel should be even higher without the demineralization of the superficial enamel. Unlike that of group $\mathrm{HCl}$ , the RRI of group HP using a $1000\text{‐}\mu \mathrm{m}$ hole remained unchanged. This suggested that the demineralization caused by HP was likely limited to the enamel surface.

The results of Raman scattering also provided direct evidence that HP can penetrate the enamel. A weak band at $876{\mathrm{cm}}^{-1}$ , which became pronounced after treatment in the Raman/fluorescence spectra of group HP, is due to $\mathrm{O}–\mathrm{O}$ stretching of HP. This band was also observed in dentin subjected to HP but with much stronger intensity (data not shown). The difference in band intensity between enamel and dentin is probably because enamel is more compact than dentin. After the specimens were removed from the HP, the $876{\mathrm{cm}}^{-1}$ band in both enamel and dentin disappeared quickly.

Another important finding in this study is that the FI decreased dramatically in group HP using both 100- and $1000\text{‐}\mu \mathrm{m}$ holes, compared with those in groups DW and $\mathrm{HCl}$ . This indicated that the matter that causes the laser-induced fluorescence might be changed greatly during the treatment. It has been suggested that fluorescence excited by different wavelength light may have different originations.²⁷ Likely fluorophore candidates range from dityrosine and cross-linked chains of structure proteins to protoporphyrins produced by bacteria.²⁷ Hibst and Paulus suggested that fluorescence of sound teeth induced by red laser light $\left(655\mathrm{nm}\right)$ might be the result of combining the inorganic matrix with a low concentration of absorbing organic molecules.²⁹ Recently, a study by Fattibene suggested that the organic materials responsible for laser-induced fluorescence were likely to be electron paramagnetic resonance (EPR) active, while the paramagnetic centers associated with the EPR native signal were located in the protein–mineral interface.²⁶

To understand the origin of laser-induced fluorescence better, the enamel structure should be explained. It is well known that dental enamel is the most highly mineralized and hardest biological tissue. It is comprised of approximately 96% mineral, 3% water, and 1% organic matter (noncollagenous protein) by weight.³² Although the protein comprises only a minor part of enamel, it is contained in the spaces between mineral crystals, where it serves as a “glue” between crystallites.³³ According to the model proposed for enamel by Brik, ³⁴ the organic component of enamel can be described with two subsystems, conventionally called “organic-1” and “organic-2.” The first of these subsystems disintegrates under heating in the temperature range 160 to $250°\mathrm{C}$ , and the temperature range of disintegration of the second is 300 to $550°\mathrm{C}$ . “Organic-1” fills the space between the enamel rods and is 10- to 20-nm thick, while “organic-2”covers the hydroxyapatite nanocrystals and is 2- to 5-nm thick. Fattibene modified Brik’s model and pointed out that the protein and mineral components of calcified tissues should not be considered as separate phases, but rather as an ensemble, where proteins and mineral crystals are chemically linked.²⁶ They proposed that dentin and enamel could be modeled as made up of three phases: a bulk mineral, a bulk organic, and a protein–mineral interface. The latter phase was identified as the “organic-2” component, which is responsible for laser-induced fluorescence.

Weak but significant FI reduction was observed in group DW using both 100- and $1000\text{‐}\mu \mathrm{m}$ holes. This is probably due to the photobleaching because the specimens were repeatedly measured at the same place during the treatment period. The FI reduction in group $\mathrm{HCl}$ is likely due to the combination of photobleaching and demineralization. Since $\mathrm{HCl}$ removed the outer enamel layer, which was most affected by the laser, the FI reduction in group $\mathrm{HCl}$ should be less than that in group DW if there are no other factors. However, the FI reduction of group $\mathrm{HCl}$ did not show significant difference from that of group DW at each time point. This is probably because $\mathrm{HCl}$ can remove the organic matter in the outer enamel surface by demineralization.

The greatest and significant FI reduction in group HP should be mainly attributed to the strong oxidizing ability of HP. HP is a strong oxidant, and it is acidic. Demineralization of enamel by HP may result in some FI reduction, but the effect should be very weak because the demineralization of enamel in group HP was milder than that of group $\mathrm{HCl}$ . We suggest that the great reduction in FI results from the fact that HP can penetrate the enamel through the boundaries between nanocrystals, and it may attack the organic matter not only in the outer enamel but also in the inner enamel during the penetration.

The microhardness loss of enamel in group HP also provided indirect evidence for the destruction of organic matter. The enamel in group HP showed higher microhardness loss at $60\min$ compared with that in group $\mathrm{HCl}$ at $15\min$ . However, both the morphological observation and Raman spectroscopy suggested that enamel surface in group HP was much less demineralized. The possible explanation is that the microhardness loss is due to the combined effects of demineralization and destruction of organic matter by HP. Several other studies have demonstrated significant decrease in hardness and fracture resistance of bleached enamel.^{12, 13} The authors also hypothesized that HP or carbamide peroxide might cause changes to the organic component of enamel and dentin by altering of the organic matrix or protein oxidation.

This study is related to the use of HP in tooth bleaching. In a clinical situation, both HP and carbamide peroxide can be used as bleaching agents. It should be pointed out that some commercially available bleaching products have a pH value close to neutral. The demineralization effect of these agents should be minimal; however, the oxidative effect is still there. As the free radical reaction is not specific, the destruction of the organic matter in enamel is likely unavoidable. Since the little organic matter plays an important role in the integrity of enamel, the clinical application of HP-containing agents should be performed with caution. The effects of destruction of organic matter on the structure and function of the enamel need to be further investigated.

5.

Conclusions

In this study, Raman scattering and laser-induced fluorescence provided new insights into the changes in the enamel structure induced by HP. The 30% HP caused significant change in the RRI in surface enamel and reduced the FI of enamel greatly and significantly. Therefore, the 30% HP may have adverse effects on both the mineral and the organic matter of human tooth enamel, and it should be used with caution.