Implant‐cemented Co–Cr frameworks produced by hard and soft milling: The impact of porcelain firing cycles on microstructure and fit accuracy

Abstract

This study examines how porcelain firing cycles and fabrication methods impact the microstructure and the marginal and internal fit of implant‐cemented Co–Cr frameworks. Two fabrication processes for Co–Cr specimens were used: hard milling or soft milling. Each type of framework was then allocated to use either as‐fabricated or as porcelain‐fired. The microstructural analysis employed x‐ray diffraction (XRD) and scanning electron microscopy with energy‐dispersive x‐ray spectroscopy (SEM/EDX). A total of ten Co–Cr frameworks were fabricated for each method, and discrepancies in margin and internal fit were evaluated using X‐ray microtomography (micro‐CT) before and after the porcelain firing. Statistical analysis was performed using two‐way ANOVA and Tukey's multiple comparison tests (α = 0.05). The findings indicated a combination of γ‐fcc and ε‐hcp cobalt‐based phases, with porcelain firing notably enhancing the ε‐Co content, particularly among soft milled specimens. Soft milled specimens exhibited a diffuse presence of pores, while hard milled specimens revealed a random distribution of small pores. The soft milled specimens exhibited a significant reduction in discrepancies after the application of firing cycles. Although not statistically significant, the hard milled specimens also showed a trend toward reduced discrepancies post‐firing. All subgroups maintained clinically acceptable marginal adaptation, with soft milled frameworks outperforming their hard milled counterparts in overall fit.

INTRODUCTION

The goal of modern prosthetic dentistry is to deliver restorations that emphasize both esthetics and functionality [1, 2]. When it comes to implant‐supported prosthetics, the selection of materials for restorations—whether all‐ceramic or metal–ceramic—can depend on the preferences of either the clinician or the patient, as the clinical results for both types are quite comparable [3]. Nonetheless, metal–ceramic prostheses remain a widely accepted treatment option, primarily due to their reduced production costs [4]. Regarding implant‐cemented restorations, the durability and functionality is significantly influenced by the marginal and internal fit [5]. Microscopic cracks or leaks at the marginal gaps can lead to the dissolution of cement [6, 7, 8], which allows the adherence of bacteria that can cause inflammation of peri‐implant tissues and subsequent peri‐implant disease [9]. Furthermore, the internal discrepancy along the axial walls of the abutment plays a crucial role in the retention of the prosthesis and its occlusal relationship [10]. Recognizing the importance of these two factors, many researchers have investigated the accuracy of fit between metal frameworks and support (natural tooth/implant abutment). Their findings indicate that: (i) the choice of metal alloys, (ii) the fabrication methods employed [11, 12, 13], and (iii) the number of porcelain firing cycles can significantly impact this fit accuracy [14].

Cobalt–chromium alloys (Co–Cr) are commonly selected for fabricating frameworks metal–ceramic restorations due to their excellent strength, effective thermal resistance, low density, and biocompatibility [15, 16, 17]. These properties are influenced by the alloy's microstructure (e.g., porosity, defects) and/or phase composition (i.e., γ‐fcc/ε‐hcp ratio [face‐centered cubic and hexagonal close‐packed, respectively]), the presence of precipitates, as well as heat treatment processes such as annealing [17, 18]. Although the conventional casting method is straightforward, it can be time‐intensive for dental laboratories [19] and is susceptible to problems such as metal shrinkage and the formation of porosities or other defects during the solidification process [18]. Recent developments in computer‐aided design and manufacturing (CAD/CAM) have enhanced the fabrication of Co–Cr restorations through additive techniques such as selective laser melting (SLM) and subtractive methods including hard milling and soft milling [2, 18, 20, 21, 22]. Hard and soft milling involve the milling of hard and soft metal blocks, respectively, using a milling machine [2, 23, 24, 25, 26]. Compared with hard milling, soft milling minimizes stress on the milling equipment, significantly decreases the milling time, and prolongs the lifespan of the milling bur, thereby streamlining the process of creating metal frameworks for metal–ceramic restorations [27, 28].

Multiple ceramic firing cycles, at temperatures ranging from 850°C to 1000°C, are often necessary in laboratory practice to attain the desired morphology and occlusal relationship for metal–ceramic restorations [27]. According to: (i) the Co–Cr equilibrium phase diagram generated by Oikawa et al. [29] and Ishida and Nishizawa [30], (ii) the Co–Mo phase diagram [31], and (iii) the Co–W phase diagram [32], the temperature of the porcelain firing cycles occurs approximately at the γ‐fcc/ε‐hcp phase boundary (Figure 1), so the fcc‐hcp ratio may change significantly due to the repetitive firing cycles/heat‐treatments within the thermodynamically stable hcp region. Consequently, these structure changes may affect both the marginal and internal accuracy [27]. Despite researching the accuracy of Co–Cr frameworks before and after undergoing porcelain firing cycles [2, 5, 27, 33, 34, 35, 36, 37, 38, 39, 40] these in vitro studies have overlooked the impact of structural characteristics (i.e., crystalline phase ratios and microstructure) on framework precision. So, this remains a significant gap in comprehensive research on this topic.

FIGURE 1.

(A) State of the art: Co–Cr equilibrium phase diagram generated by Oikawa et al. [29] and Ishida and Nishizawa [30]. The vertical red dashed line indicates the chromium content (wt.%) in the materials being examined (Co–Cr alloy for: SM 28.0 wt.%, and HM 24.8 wt.%). Additionally, the (1) red shaded area denotes the firing temperature range for porcelain, and the (2) purple line marks the temperature at which oxidation occurs. (B) Flowchart depicting the experimental procedure. XRD: x‐ray diffraction; SEM/EDX: scanning electron microscopy with energy‐dispersive x‐ray spectroscopy; HM: hard milling; SM: soft milling; SG: subgroup.

The objective of this in vitro study was to evaluate and compare the effects of repeated porcelain firing cycles and fabrication techniques on the microstructure, as well as the marginal and internal fit of metal–ceramic implant‐cemented Co–Cr frameworks produced by soft milling and hard milling methods. The null hypotheses formulated for this research posited that there would be no significant differences in the marginal and internal fit between Co–Cr framework and abutment: (i) among the specimens measured before and after the porcelain firing within the same group, or (ii) across the two production techniques.

MATERIAL AND METHODS

The experimental protocol was split into two main sections. The first section was dedicated to characterizing the Co–Cr alloys, while the second section focused on evaluating the marginal and internal fit between the Co–Cr framework of a cemented metal–ceramic prosthesis and the implant abutment.

Fabrication and characterization of Co–Cr specimens

Two groups of specimens were fabricated using common commercial Co–Cr alloys (the chemical compositions of these commercially available Co–Cr alloys were chosen to be as similar as possible, given that there is no specific Co–Cr alloy used for the production of frameworks both via hard milling and soft milling methods). The first group consisted of specimens created using the hard milling technique (N = 6). These disc‐shaped specimens, measuring 10 mm in diameter and 7 mm in thickness, were produced from Co–Cr dental alloy (Mediloy M‐Co, Bego; chemical composition: Co 63.8, Cr 24.8, Mo 5.1, W 5.3, Si 1.0 (in wt.%)) utilizing CAD software (Siemens NX 12) and an Ultrasonic 10 milling machine (DMG Mori; the available specifications from the manufacturer are: a wet 5‐axis milling process, travel in the X/Y/Z axis: 120/120/200 mm, A axis: −10 ° to 120° and C axis: 360°, max milling speed range: 40,000 rpm, feed speed range: up to 10,000 mm/min, and compressed air: 6 bar) using milling burs ∅ 2.5, 1.0, 0.6, and 0.3 mm. The second group comprised specimens fabricated by the soft milling technique (N = 6). Similar in shape and dimensions (10 mm in diameter and 7 mm in thickness), these specimens were made from Co–Cr dental alloy (Ceramill Sintron, Amann Girrbach; of a nominal composition of Co 66, Cr 28, Mo 5, Si < 1, Fe < 1, Mn < 1 (in wt.%)) utilizing CAD software (Siemens NX 12) and a Ceramill Motion 2 milling machine (Amann Girrbach; the available specifications from the manufacturer are: a dry 5‐axis milling process, motor speed 60,000 rpm, and compressed air: 6 bar) using milling burs ∅ 2.5, 1.0, 0.6, and 0.3 mm. After the soft milling procedure, the green (non‐presintered) Co–Cr specimens were subjected to sintering at 1280°C in a high‐temperature furnace filled with Ar gas for 5 h. Finally, the Co–Cr specimens underwent ultrasonic cleaning.

The Co–Cr specimens produced by each method were equally divided into two subgroups. The first subgroups (SG hard milled‐1 and SG soft milled‐1) comprised the as‐fabricated disc‐shaped specimens. The second subgroups (SG hard milled‐2 and SG soft milled‐2) comprised the porcelain‐fired specimens that underwent seven simulated porcelain firing cycles (Table 1) in a dental porcelain furnace (P510 Programat; Ivoclar Vivadent), following the manufacturer's guidelines (Initial MC, GC), without any ceramic material applied to their surfaces.

TABLE 1.

Conditions of the porcelain firing procedure according to the manufacturer's instructions (Initial MC, GC), applied in the present in vitro study.

Porcelain firing cycles	Oxidation	Bonder	1st Paste opaque firing	2nd Paste opaque firing	1st Dentin firing	2nd Dentin firing	Glaze firing (With glaze powder)
Initial temperature	–	550°C	550°C	550°C	580°C	580°C	480°C
Holding time	–	6 min	6 min	6 min	6 min	6 min	2 min
Final temperature, heating rate, vacuum	Hard milled: 900°C 80°C/min ✓	980°C 80°C/min ✓	940°C 80°C/min ✓	930°C 80°C/min ✓	890°C 55°C/min ✓	880°C 55°C/min ✓	860°C 45°C/min −
	Soft milled: 980°C 80°C/min ✓
Holding time	Hard milled: 5 min	1 min	1 min	1 min	1 min	1 min	1 min
	Soft milled: 1 min

The Co–Cr specimens were assessed using the following techniques: x‐ray diffraction (XRD) was used to analyze the crystalline phases in the Co–Cr specimens of each of the four subgroups. The analysis was performed using a D8 Advance system (Bruker AXS, Billerica) with Cu Kα radiation (λ = 1.5406 Å), generated at 30 kV and 25 mA. Diffraction angles (2θ) were recorded between 20° and 100°, with a scanning step size of 0.004°/s. The resulting x‐ray diffractograms were compared with reference data from the International Center for Diffraction Data (https://www.icdd.com). To calculate the weight percentage of the ε‐Co phase in the alloys, the intensities (I) of the XRD peaks corresponding to the (101) ε and (200) γ planes were used, applying the formula originally proposed by Sage and Gillaud [41]:

$$\mathrm{hcp}\left(\mathrm{wt}.\%\right)=\left(\frac{I_{\left(101\right)\mathrm{hcp}}}{I_{\left(101\right)\mathrm{hcp}}+1.5I_{\left(200\right)\mathrm{fcc}}}\right)\times 100$$ (1)

where subscript hcp stands for hexagonal close‐packed lattice and fcc stands for face‐centered cubic lattice.

To analyze the microstructure of the specimens, several preparation steps were carried out. The specimens were first embedded in phenolic resin (Presi, Eybens). Then, the specimens were ground using SiC papers to a 2400 grit finish and polished to 1 µm with diamond sprays (Diamant Mecaprex Spray; Presi) using an Ecomet 6 polishing system (Buehler). After polishing, the specimens were cleaned in a distilled water bath with ultrasonic agitation and then dried. The grain structure was revealed by electrochemical etching of the polished surfaces using a solution of HCl/HNO₃/FeCl₃ (40 mL/1 mL/13 g) at 6 V for 15 s at room temperature [22]. Following etching, the Co–Cr specimens were rinsed with water, cleaned ultrasonically, and blow‐dried. Microstructural analysis was carried out using scanning electron microscopy (SEM, JSM‐6510 LV; JEOL), equipped with energy‐dispersive x‐ray spectroscopy (EDX, X‐Act; Oxford Instruments), and operated in secondary electron mode.

Evaluation of marginal and internal fit

The second section of the experimental procedure aimed to evaluate the marginal and internal fit between the Co–Cr framework of a cemented metal–ceramic prostheses and the implant abutment.

To create a simulation of a maxilla with an implant located in the 26 region (termed the modified model), modifications were applied to a maxillary model using a milling unit/parallelizer (S1 Basic, Schick Dental) to establish the drilling site. An implant analog (Friadent Implant Analog D5.5, Dentsply Sirona Implants) was subsequently placed in the 26 region. Additionally, to enhance the depiction of the adjacent anatomical structures, a wax simulation of the peri‐implant tissues was incorporated using Tenatex wax red (Kemdent).

A closed tray method was employed using a transfer coping/cap (Friadent, Transfer Coping/cap D5.5, Dentsply Sirona Implants) along with silicone duplicating material (Elite Double 16 Fast, Zhermack Dental) to form a mold. The process included the use of an implant analog (Friadent Implant Analog D5.5, Dentsply Sirona Implants), silicone material that simulates gingiva (Multisil Mask, Bredent), and Type IV dental stone (Fujirock EP Premium Pastel Yellow, GC) to produce the working model. Finally, an implant abutment (Friadent Esthetic Base Abutment straight D5.5/GH1/A0, Dentsply Sirona Implants) was securely attached to the implant analog within the working model.

A laboratory‐calibrated scanner, the Medit T710 (Medit), noted for its scanning precision of 4 µm (according to ISO 12836; as stated by the manufacturer), was employed in the digitalization process. Before the scanning occurred, the implant abutment received a coating of scan spray (HELLING 3D Laser Scanning Anti‐Reflection Spray MATT, Helling). The design of an anatomical implant‐cemented framework for tooth #26 (intended for a metal–ceramic restoration) featuring a minimal Co–Cr framework thickness of 0.6 mm and a cement gap thickness of 50 µm (the cement gap thickness begins 1 mm above the margin line of the implant abutment, in order to ensure a precise/tightest fit of the Co–Cr framework at the margin line), was accomplished using dental CAD software (exocad DentalCAD, exocad). The finalized digital framework design was saved in standard tessellation language (STL) format, which was subsequently utilized by CAM equipment to produce Co–Cr frameworks via two fabrication methods: hard milling and soft milling. The fabrication processes, equipment, and materials used for the implant‐cemented Co–Cr frameworks were consistent with those previously detailed for the Co–Cr specimens in both the hard milled (N = 10) and soft milled (N = 10) groups.

The implant‐cemented Co–Cr frameworks (N = 10 for each fabrication method) were initially assessed for marginal and internal gaps, forming the first subgroups, SG hard milled‐1 and SG soft milled‐1. To prevent any confusion, the frameworks were numbered. Subsequently, the same frameworks underwent a series of dental porcelain layering and firing cycles in a dental porcelain furnace (P510 Programat; Ivoclar Vivadent), in accordance with the manufacturer's guidelines provided by Initial MC (GC, Austria; chemical composition: SiO₂ 55–65, Al₂O₃ 10–15, K₂O 10–15, Na₂O 5–8, Li₂O < 1, CaO 1–2, BaO 0–2, B₂O₃ 0.5–2, TiO₂ < 1, ZrO₂ < 1, P₂O₅ < 1, and pigments 0.1–3.0 wt.%), as detailed in Table 1. This process resulted in the formation of second subgroups, SG hard milled‐2 and SG soft milled‐2. Following the completion of the porcelain layering and firing cycles, the frameworks underwent further measurements of the marginal and internal gaps between the Co–Cr frameworks and the implant abutment. These measurements were conducted prior to and after the porcelain application process. X‐ray microtomography (micro‐CT) was employed to record the aforementioned measurements at both stages (before and after dental porcelain layering and firing cycles).

Micro‐CT scans were conducted utilizing a SkyScan 1172 system (Bruker, Kontich), which features a tungsten‐based x‐ray source and an 11 PM CCD camera (4000 × 2672 pixels) capable of achieving a high resolution of less than 0.8 µm per pixel. The scanning protocol involved setting the voltage to 100 kV and the current to 100 µA, with the addition of a combined aluminum and copper filter. All frameworks before and after porcelain firing cycles were scanned at a voltage of 100 kV and a current of 100 µA using a combined aluminum and copper filter. Images were acquired at a pixel size of 13.79 µm with a camera binning of 2 × 2. The exposure time was 2480 ms and scans were performed for a full rotation of 360° and a rotation step of 0.45°. This scanning protocol ensured the acquisition of high‐resolution images for accurate analysis.

The projection images were processed into cross‐sectional views utilizing SkyScan's NRecon software (Bruker), which incorporates a modified version of Feldkamp's back‐projection algorithm. The resulting cross‐sectional images were then imported into the ct analyser software (version 1.14.4.1, CTAn, Bruker) for the assessment of marginal and internal fit across two perpendicular axes. Additionally, volume renderings of the frameworks were generated with the help of SkyScan's ctvox software.

For each framework before and after porcelain firing cycles two cross‐section images for marginal and internal discrepancy measurements were used, that is, the middle cross‐section images of implant abutment at the coronal and at the sagittal planes. One marginal discrepancy measurement and three internal discrepancy measurements were made at each region (mesial, distal) of the mesiodistal cross‐section image of the coronal plane and at each region (buccal, palatal) of the buccopalatal cross‐section image of the sagittal plane. Therefore, 4 marginal discrepancy measurements and 12 internal discrepancy measurements for each framework in total were conducted and the marginal and internal discrepancy was calculated as the average of these individual measurements, respectively.

Assessments were conducted to verify the normality of all collected data using the Shapiro–Wilk test, as well as to test for homoscedasticity (equal variances) through an equality of variances test. The mean values and corresponding standard deviations were then calculated for the marginal and internal fit measurements of each group (N = 10). These measurements were derived from 4 and 12 different indentations, as outlined previously. The data were subject to further statistical analysis using two‐way analysis of variance (ANOVA), followed by the application of Tukey's multiple comparison post hoc test considering the manufacturing process and porcelain firing cycles as independent/discriminating variables (at 95% level of significance, α = 0.05). Sample size calculation and power analysis rationale were incorporated into the statistical assessment. Statistical computations were performed utilizing the ORIGINPRO SOFTWARE (OriginLab).

RESULTS

Figure 2 displays the x‐ray diffractograms for the Co–Cr specimens, including both the as‐fabricated (hard milled‐1 and soft milled‐1) specimens and those subjected to simulated porcelain firing cycles (hard milled‐2 and soft milled‐2). In these diagrams, each diffraction peak is indexed with the corresponding plane of the identified cobalt phase. From the peak analysis, it can be concluded that both the as‐fabricated (hard milled‐1 and soft milled‐1) and porcelain‐fired (hard milled‐2 and soft milled‐2) specimens demonstrate the presence of the γ cobalt phase, which is noted for its face‐centered cubic (fcc) structure (ICDD card number 00‐015‐0806), and the ε cobalt phase, which exhibits a hexagonal close‐packed (hcp) structure (ICDD card number 00‐005‐0727). Additionally, several less prominent peaks are detected, linked to M₂₃C₆ (a chromium‐rich carbide, ICDD card number 00‐035‐0783) in specimens made using both hard milling and soft milling techniques, with peak intensity noted to increase following the porcelain firing cycles. Finally, in the Co–Cr specimens of soft milled‐2 the XRD pattern also revealed the presence of SiO₂ (ICDD card number 12–708).

FIGURE 2.

X‐ray diffractograms of Co–Cr specimens fabricated by hard milling (A) and soft milling (B) processes. The identification of the peaks is also presented. On the right‐hand side (insets), the magnified 2θ regions within 40°–55° for each diffractogram show the peaks necessary to calculate the ε‐Co phase. HM: hard milling; SM: soft milling; SG: subgroup.

The determination of the weight percentage of the ε‐hcp phase was conducted based on the intensities (I) of the XRD peaks associated with the (101) ε‐hcp and (200) γ‐fcc reflections, in accordance with the formula proposed by Sage and Guillaud [41]. Figure 2 (insets) presents an amplified view of the 2θ range from 40° to 55° for each diffractogram, emphasizing the peaks necessary for the calculation of the ε‐hcp phase. As a result, the weight percentages of the ε‐hcp phase relative to the γ‐fcc phase were determined for the Co–Cr specimens made using hard milling and soft milling techniques, yielding values of 65.6% (hard milled‐1) and 67.1% (hard milled‐2) for the hard milled specimens, and 28.0% (soft milled‐1) and 38.8% (soft milled‐2) for the soft milled specimens.

The SEI images and area mapping results from hard milled and soft milled specimens are presented in Figures 3 and 4, respectively. The specimens produced using the hard milling method (Figure 3) revealed a random distribution of small pores or defects, both prior to (hard milling‐1) and following (hard milling‐2) porcelain firing cycles. The magnified figures (taken at 500x and 2000x magnification) depict a dispersed second phase apart from the matrix phase. EDX analysis revealed that the carbon‐containing precipitates, represented as small dots in the carbon mapping image, are uniformly distributed and correlate with the carbides identified in the XRD analysis (Figure 2). In the specimens fabricated through the soft milling method (Figure 4), the grain boundaries became distinctly apparent after electrochemical etching. Both subgroups (soft milled‐1 and soft milled‐2) exhibited similar microstructural characteristics; the images displayed in Figure 4 indicate a diffuse presence of pores with diameters ranging from approximately 1.5–3.0 µm. EDX analysis revealed that the uniformly distributed carbon‐containing precipitates, represented as small dots in the carbon mapping image, correspond with the carbides identified in the XRD analysis (Figure 2). Additionally, the presence of Si‐oxide, visible as small dots, further corroborates the findings from the XRD analysis (soft milled‐2).

FIGURE 3.

Microstructure and elemental analysis of the as‐fabricated (A) and porcelain‐fired (B) Co–Cr specimens fabricated by hard milling. HM: hard milling; SG: subgroup.

FIGURE 4.

Microstructure along with elemental analysis of Co–Cr specimens produced by soft milling, in their as‐fabricated state (A) and after porcelain firing (B). SM: soft milling; SG: subgroup.

Representative micro‐CT scan image of Co–Cr framework prior to the porcelain layering and firing process along with the measurement points indicating internal discrepancy and marginal discrepancy are illustrated in Figure 5. Table 2 summarizes the mean values along with their standard deviations for the marginal and internal gaps measured between the Co–Cr frameworks and the implant abutment. It is worth noting that the sample sizes used in this investigation exceeded the minimum recommended power level of 0.8 [42], with an actual power value of 1, ensuring sufficient statistical reliability. In the hard milled specimens, both marginal and internal gap remained relatively stable after the porcelain firing cycles, with marginal gaps changing from 39.6 ± 0.3 to 39.2 ± 0.3 µm and internal gaps shifting from 98.2 ± 0.2 to 97.9 ± 0.2 µm, representing only a minimal reduction. These differences were not statistically significant (p > 0.05). In the soft milled specimens, both marginal and internal gaps exhibited a reduction following the porcelain firing cycles, with the marginal gap decreasing from 35.5 ± 1.5 to 31.2 ± 1.3 and the internal gap shifting from 94.2 ± 0.3 to 89.4 ± 0.2. These differences were statistically significant (p < 0.05). When comparing the manufacturing processes, the Co–Cr frameworks created through the soft milling method showed enhanced precision (lower gap measurements) in both marginal and internal discrepancies, both before and after multiple firing cycles. Their measurements were significantly lower than those of the Co–Cr frameworks produced using the hard milling method (p < 0.05).

FIGURE 5.

Measurement points (A) and micro‐CT scan images of internal and marginal discrepancy of Co–Cr frameworks fabricated by hard milling (B) and soft milling (C) methods. ID: internal discrepancy; MD: marginal discrepancy; micro‐CT: x‐ray microtomography; HM: hard milling; SM: soft milling; SG: subgroup.

TABLE 2.

Mean values (± standard deviation) of marginal and internal discrepancy of Co–Cr frameworks before and after the porcelain firing procedure.

	Marginal gap (µm)		Internal gap (µm)
Fabrication method	Before	After	Before	After
Hard milled	39.6 ± 0.31A	39.2 ± 0.31A	98.2 ± 0.21A	97.9 ± 0.2 1A
Soft milled	35.5 ± 1.51B	31.2 ± 1.32B	94.2 ± 0.31B	89.4 ± 0.22B

Superscripts indicate statistical comparisons: numbers for row comparisons and capital letters for column comparisons. Different superscripts denote mean valueswith statistically significant differences (p < 0.05), while same superscripts denote mean values without statistically significant differences (p > 0.05).

DISCUSSION

The null hypothesis, which stated that there would be no notable differences in the marginal and internal fit of metal–ceramic implant‐cemented Co–Cr frameworks (i) within the same group before and after the porcelain firing process, or (ii) across different manufacturing methods, was partially rejected, as statistically significant differences were identified among the materials tested (Table 1).

The introduction of CAD/CAM methods has transformed the dental laboratory domain, resulting in significant advancements in restoration manufacturing methods [36]. One critical aspect influencing the durability of metal–ceramic implant‐supported prostheses is the marginal and internal adaptation [43, 44]. Recognizing the importance of these factors, several studies have investigated the fit accuracy between metal frameworks and abutments. Among the key findings, the number of porcelain firing cycles stands out as a critical element that influences the precision of the fit [14]. Several studies [2, 27, 33, 35, 36, 37, 38, 39, 40, 45] have examined the fit accuracy of Co–Cr frameworks, assessing them both before and after the application of repeated porcelain firing cycles. The discrepancies noted in the marginal gap are generally linked to the differences in the coefficients of thermal expansion (CTEs) between the veneering ceramic and the framework alloy, which can result in deformation of the framework [2, 27, 33, 35, 36, 37, 38, 39, 40, 45]. Given the minimal difference in the CTEs between Co–Cr alloy and porcelain (less than 0.5 to 1.0 × 10⁻⁶ K⁻¹), with the ceramic material having a lower CTE, it would be anticipated that the marginal gap of Co–Cr frameworks would decrease or remain unchanged (given the small CTE difference) after the porcelain firing cycles; however, the findings of the researchers remain controversial, with conflicting findings across studies [2, 27, 33, 35, 36, 37, 38, 39, 40, 45]. Additionally, the repeated firing of ceramics at temperatures ranging from 850°C to 1000°C, along with variations in heating and cooling rates, can modify the metallurgical structure and the ratio of the crystalline phases [46] of the alloy, potentially affecting both marginal and internal precision [27], a consideration that has been largely neglected in the majority of existing research.

The XRD analysis of the Co–Cr specimens fabricated by hard milling and soft milling methods (Figure 2) indicates the existence of both allotropes of cobalt: the high‐temperature stable form, γ‐Co (γ‐fcc), and the low‐temperature stable form, ε‐Co (ε‐hcp); results which are consistent with previous research [22, 26]. In pure cobalt, the phase transition (allotropic transformation) from the high‐temperature γ‐fcc phase to the low‐temperature ε‐hcp phase takes place at 417°C (690 K) via a displacive martensitic mechanism [16]. However, when alloying elements such as Cr, Mo, W, and Si are added, as in the alloy examined in this study, the transformation temperature increases to above 900°C, as these elements extend the stability range of the hcp phase [47].

It is widely recognized that the transformation from γ‐fcc to ε‐hcp (referred to as the ε‐martensitic transformation [22] in Co‐based alloys can occur through three main mechanisms: thermally (by rapid cooling from temperatures where the γ‐phase is stable), isothermally (through aging at temperatures between 650°C and 950°C), or via the application of plastic strain [22, 48]. For Co–Cr specimens produced using the hard milling technique, plastic strain appears to be the primary driver of the martensitic transformation. In Co–Cr specimens made using the soft milled method, the transformation is likely influenced by plastic strain induced during the soft milling process and/or by active recrystallization occurring during sintering.

As mentioned in the Introduction section, according to: (i) the Co–Cr equilibrium phase diagram generated by Oikawa et al. [29] and Ishida and Nishizawa [30], (ii) the Co–Mo phase diagram [31], and (iii) the Co‐W phase diagram [32] the temperature of repeated porcelain firing cycles coincides with the γ‐fcc/ε‐hcp phase boundary (Figure 1). Consequently, the fcc‐hcp phase ratio may undergo significant changes due to the thermal cycles performed within the thermodynamically stable hcp region. The effect of repeated firing cycles on the phase composition of Co–Cr specimens fabricated using the hard milling and soft milling techniques has been evaluated. The results indicate that the proportion of the ε‐hcp phase relative to the γ‐fcc phase increased following the firing cycles, with values rising from 65.6% (hard milled‐1) to 67.1% (hard milled‐2) in the hard milled specimens, and from 28.0% (soft milled‐1) to 38.8% (soft milled‐2) in those made using the soft milling process. Notably, the lower percentage of the ε‐phase in the soft milled specimens than in the hard milled specimens can be attributed to the sintering process (1280°C for 5 h) undergone by the soft milling‐fabricated specimens, which falls within the γ‐fcc stability range (Figure 1) corresponding to their specific elemental compositions. Calculations based on Equation (1) suggest a higher percentage of the ε‐phase in the porcelain‐fired specimens compared with the as‐fabricated specimens in both groups, which is most likely a consequence of aging at the temperatures experienced during thermal porcelain cycling (>860°C, isothermal mechanism). However, it is also important to note that there was no significant increase in the percentage of the ε‐phase in the porcelain‐fired specimens. This can be attributed to the sluggish nature [49] of the martensitic transformation in Co–Cr‐Mo alloys, which justifies the predominance of fcc Co in both hard milled and soft milled porcelain‐fired Co–Cr alloys.

In the diffractograms, alongside the γ‐fcc and ε‐hcp peaks, several weak peaks were detected, which correspond to M₂₃C₆ (Cr₂₃C₆; in hard milled‐1, hard milled‐2 and soft milled‐2) carbides as well as SiO₂ in soft milled‐2 (which will be addressed in the next paragraph). These findings from the XRD analysis (Figure 2) are further validated by the elemental analysis results (Figures 3 and 4), suggesting that the majority of the precipitates are likely nanoscale carbides. The simulated porcelain firing cycles did not lead to the dissolution of these carbide precipitates. Notably, in soft milled specimens, the firing cycles resulted in an increased presence of carbides, as shown by the XRD pattern (Figure 2). This indicates that the strength, wear resistance, and corrosion resistance of the final material were preserved, as these properties largely depend on the presence of the precipitates [16, 18].

Co–Cr specimens produced by the hard milling method exhibit a random distribution of small pores, whereas the soft milling method results in a more diffuse pore distribution. As illustrated in Figures 3B and 4B, the simulated porcelain firing cycles do not cause any noticeable macroscopic changes to the microstructure. The porosity observed in Figure 4A and B has been reported in earlier studies of Co–Cr alloys created using the soft milling process [21, 50]. This porosity is typically attributed to the necking of metallic powder particles and the evaporation of the organic binder during the sintering process (at 1280°C for 5 h). Figure 4A and B demonstrate that most of the pores are filled with Si‐oxide, a finding consistent with the work of Lekatou et al. [22]. According to their study, the presence of silica within the pores arises from the diffusion of oxygen and silicon into the pores, a process driven by the high temperatures during milling, and their consequent inter‐reaction.

Numerous studies have concentrated on the marginal and internal fit of implant‐supported metal–ceramic restorations, as a poor marginal fit can result in heightened cement dissolution, microleakage, and plaque build‐up, potentially leading to periodontal and peri‐implant diseases [9]. To achieve the highest level of measurement accuracy for assessing marginal and internal fit, this in vitro research utilized micro‐CT technology, among other methods. Micro‐CT is a non‐destructive technique capable of reliably and repeatedly capturing measurements at the same or at various locations with micrometer‐level precision [34, 36].

According to American Dental Association (ADA) Specification No. 8, the marginal gap width range between 25 and 40 µm has been suggested as a clinical goal, but this goal is seldom achieved [40]; as a result, most researchers regard a marginal gap of less than 100–120 µm as acceptable for clinical purposes [36, 51]. Regarding the results of this in vitro study, the values observed across all subgroups, as shown in Table 2, (i) achieve the ADA Specification No. 8 of the clinical goal (25–40 µm [40]) and consequently (ii) fell within clinically acceptable ranges of 100–120 µm [36, 51] even though the application of porcelain firing cycles influenced the marginal discrepancies. While no definitive standard has been set for internal discrepancies, certain researchers propose that the internal gaps should not surpass 100 µm at the occlusal surface. Additionally, the effectiveness of bonding diminishes for misfits exceeding 450–500 µm beneath the occlusal area due to polymerization shrinkage stress [34, 52]. The internal discrepancy values for all subgroups were also below the 100 µm threshold.

As noted by Li et al. [46] and Kocaağaoğlu et al. [27], repeated firings of dental porcelain can alter the structural characteristics (i.e., crystalline phase ratios and microstructure) of an alloy which in turn influence both the marginal and internal fit. In the soft milled specimens, the percentage of the ε‐hcp phase increased from 28.0% to 38.8% after undergoing the simulated porcelain firing cycles. This change led to a statistically significant difference in both marginal and internal discrepancy measurements (p < 0.05, Table 2). In contrast, the hard milling group exhibited no significant changes in either marginal or internal discrepancy measurements (p > 0.05, Table 2), likely due to the minimal increase in the ε‐phase, which rose from 65.6% to 67.1%.

A detailed examination and correlation of the XRD and micro‐CT results among specimens, before and after the porcelain firing cycles, produced using the same fabrication method indicate that: (i) an increase in ε‐phase content is observed following the porcelain firing cycles, and (ii) both marginal and internal gap values decrease after porcelain firing cycles (Table 2). These findings suggest that an increase in the ε‐hcp phase is associated with a reduction in the gap between the framework and the abutment. Consequently, when comparing manufacturing techniques, it is anticipated that frameworks produced via hard milling will exhibit smaller gaps. This expectation is based on the higher ε‐hcp phase content observed in hard‐milled frameworks compared with those fabricated using the soft milling method. However, exactly the opposite was observed. This may be due to the fact that after the soft milling process, the sintering process at 1280°C for 5 h follows, after which the final specimen is estimated to have a shrinkage of ∼ 11% (according to the manufacturer). However, this is not binding, that is, the shrinkage may be larger. Thus, although one would expect the soft milled frameworks to have a larger gap, perhaps the ∼11% shrinkage compensates for the expected increase in the gap. The findings of this study align with those from other research groups [35, 38, 53] who observed a narrower marginal and/or internal gap between the framework and abutment in specimens produced using the soft milling method, as opposed to those fabricated with the hard milling method.

This research has a number of limitations. To begin with, while the Co–Cr alloys examined had comparable elemental compositions, they were not entirely the same. Additionally, the study focused exclusively on one variety of dental porcelain, restricting the analysis to particular firing cycles. Notably, with respect to the specimens created through the soft milling technique, the pre‐set sintering temperature and holding time specified by the manufacturer were utilized. Future studies may thus benefit from investigating how alterations in these parameters would impact the microstructure, as well as the extent of marginal and internal discrepancies.

The findings of this study hold significant clinical implication for both dental clinicians and technologists. In general, all Co–Cr frameworks, regardless of their fabrication method or whether they underwent porcelain firing, exhibited marginal adaptation according to the clinical goal of American Dental Association (ADA) Specification No. 8 standard (25–40 µm) [40] and consequently fell within clinically acceptable ranges of 100–120 µm ([36, 51]), suggesting a reduced likelihood of cement dissolution. It is important to note that the Co–Cr frameworks cemented to implants and produced through the soft milling process showed statistically significantly better marginal and internal fit with the implant abutment, particularly in comparison with those constructed using the hard milling process.

AUTHOR CONTRIBUTIONS

Conceptualization: Michael Nikolopoulos, Konstantinos Dimitriadis. Methodology: Michael Nikolopoulos, Konstantinos Dimitriadis. Investigation: Michael Nikolopoulos, Konstantinos Dimitriadis. Formal analysis: Michael Nikolopoulos, Konstantinos Dimitriadis, Pepie Tsolka. Data curation: Michael Nikolopoulos, Konstantinos Dimitriadis. Visualization: Michael Nikolopoulos, Konstantinos Dimitriadis, Stavros Yannikakis, Athanasios K. Sfikas, Aristidis Galiatsatos, Pepie Tsolka. Writing—original draft: Michael Nikolopoulos, Konstantinos Dimitriadis. Writing—review & editing: Michael Nikolopoulos, Konstantinos Dimitriadis, Stavros Yannikakis, Athanasios K. Sfikas, Aristidis Galiatsatos, Pepie Tsolka.

CONFLICT OF INTEREST STATEMENT

The authors declare no conflicts of interest.

FUNDING INFORMATION

The author(s) received no financial support for the research and authorship of this article. Open Access funding was obtained from Heal‐Link Greece.

Nikolopoulos M, Dimitriadis K, Yannikakis S, Sfikas AK, Galiatsatos A, Tsolka P. Implant‐cemented Co–Cr frameworks produced by hard and soft milling: The impact of porcelain firing cycles on microstructure and fit accuracy. Eur J Oral Sci. 2025;133:e70031. 10.1111/eos.70031