Seventy healthy premolars extracted for orthodontic reasons from patients under 25 years-old with complete apical formation and mild radicular curvature (0° to 15°)  were used. Teeth with calcified canals and apical or coronal fractures were excluded. Teeth were decoronated 2 millimeters at the cement-enamel junction and kept in a 10% formalin saline solution.
Root Canal Procedures
Canal patency using a #10 K-type file (Maillefer, Ballaigues, Switzerland) was established in 60 teeth, with the working length -0.5mm of major foramen. Single file, Wave One Gold® Primary (0.25mm/.07 Dentsply/Sirona Ballaigues, Switzerland) was used for cleaning and shaping using a X-smart plus® motor (Dentsply/Sirona Ballaigues, Switzerland); done by thirds (cervical, middle and apical), using 15ml of 5.25% sodium hypochlorite (NaOCl) as irrigant (Enzohip-5 Eufar®, Medellin, Colombia). One ml of ethylenediaminetetraacetic acid solution (EDTA) 17% (Eufar®, Medellin, Colombia) was used as final irrigant for one minute and deactivated with 2 ml of saline solution.
All canals were filled using lateral compaction; #35 standard Gutta-percha as master cone (Maillefer, Ballaigues;Switzerland), resin-epoxy based sealer (Topseal® (Dentsply/Sirona, Ballaigues, Switzerland) and standard #20 and #15 accessory cones (Maillefer, Ballaigues;Switzerland) with an A25 spreader (Dentsply/Sirona, Ballaigues, Switzerland). The coronal gutta-percha was removed 2 mm below the cement-enamel junction, followed with vertical compaction using a B60 plugger (Dentsply/Sirona, Ballaigues, Switzerland). Coronal access was sealed with Vitrebond® (3M ESPE). The quality of obturation phase was checked using periapical radiographs. Subsequently, all teeth were stored for two months in sterile jars under 95% humidity and 37°C conditions.
Two control groups were used: i) Control group 1 (negative control): 5 randomly chosen teeth without any treatment; and ii) Control group 2 (positive control): 5 teeth with the pulp mechanically removed and subsequently irrigated with NaOCl 5.25%, 17% EDTA and saline solution.
The coronal seal was removed using a #2 diamond round sterile bur and a #1 Peeso reamer (Maillefer, Ballaigues, Switzerland) was used to remove root coronal filling material. The specimens were divided into 3 groups (20 teeth each) for testing the substances, as follows:
- Group 1: Mechanical retreatment without any solvent.
- Group 2: 1ml of Xylol (Enzohip-5 Eufar®, Medellin, Colombia) for 1 minute followed by mechanical retreatment
- Group 3: 1ml of a 10% citric acid for 1 minute followed by mechanical retreatment.
Due to the chelating action of citric acid and concentration, one minute was used for a complete elimination of the smear layer and to prevent erosion of the calcium-rich peritubular dentine . Additionally, all samples were irrigated with 15ml of 5% of sodium hypochlorite during the procedure (Enzohip-5, Eufar®, Medellin, Colombia) to rinse debris and 1ml saline solution to neutralize citric acid .
Procedure for sample analysis
All samples, including controls, were split along the longitudinal axis to the root. Buccal and lingual sections were used for Raman spectroscopy analysis and micro-hardness test (Vickers), respectively.
Raman measurements were performed by an Ocean Optics IDR-Micro 785 Spectrometer with 15 seconds dwell time (10 scans). A Laser line of 785 nm wavelength with optical power of 10.0 mW and 4.0 mW was applied to measure wall dentine and canal dentine, respectively. Raman spectra for the different radicular dentine thirds (cervical, middle and apical) were taken and their intensity was normalized to unity. No significant differences were found between the spectra when compared with previous reports  (Figure1). Therefore, dentine spectra for every tooth were recorded as reference.
Additional reference Raman spectra were taken for gutta-percha and AH-Plus® sealer (Dentsply/Sirona, Ballaigues, Switzerland) (Figure 2a). The uppermost spectrum shown in Figure 2b, was measured on the middle third section of a radicular canal. Basically, this spectrum can be considered as a superposition of the previously taken reference Raman spectra of dentine (*) plus sealer (+), indicating a canal with sealer. Additional features attributed to surface roughness (**) due to canal concavity are also visible. Clearly, Raman spectroscopy was useful to demonstrate an incomplete removal of filling materials however the goal of the retreatment procedure was not accomplished.
The intrinsic hardness of dentinal structure depends on size, density, orientation and degree of mineral and hydroxyapatite content of microtubules . Therefore, demineralization effects due to chemical application could be evaluated through microhardness testing, using an indenter probe under a specific load that penetrates a surface during a defined dwell time. The imprinted size or depth left by the indenter, is inversely proportional to surface hardness. The hardness Vickers (HV number) is given by the ratio between the force applied for the diamond indenter in kilograms-force and the surface area of the resulting indentation in square millimeters. For a diamond pyramid with 136° angle, HV is given by the relationship (ASTM E384: Standard Test Method for Knoop and Vickers Hardness of Materials):
Here, is the average of the two diagonal lengths of the imprinted diamond pyramid.
To achieve an accurate measurement using the indenter tip, flat surfaces are recommended; extremely rough surfaces reduces data accuracy. Therefore, selected teeth were immersed into acrylic and subsequently polished with #320 and #600 sandpaper until a flat surface was obtained. Indentation was not performed at the canal itself, but at 0.5mm from the canal edge on a flat dentinal surface.
A Laizhou Lyric HVS-1000 microhardness tester with a load of 50 g-f (0.49 N) applied to a standard 136° diamond during 20s was used. Microhardness is inversely proportional to tubular density, which increases from apical to cervical dentine , different HV values were expected over different tooth regions. Therefore, indentation of cervical, middle and apical thirds was performed for all 70 teeth. This information combined with data from Raman spectroscopy was used to evaluate morphological changes in dentine surface.
Analysis of the information
The normality of the data was determined by the Kolmogorov Smirnov and Shapiro Wilk tests, and the equality of variances using the Levene´s test. The one-way ANOVA test was performed to analyze the treatments and controls groups independently of where the analysis of the root third was performed, and two-way ANOVA to consider the root third- dependent results. For the Post-Hoc analysis, T-Student and Tukey test were used. In all cases p<0.05 was considered significant. Chi-Square and Fisher's exact tests were used to compare the percentage of material remnant.