In vitro synchrotron-based radiography of micro-gap formation at the implant–abutment interface of two-piece dental implants

Micro-radiography using hard X-ray synchrotron radiation is the first potential tool to allow an evaluation of the mechanical behavior of the dental implant–abutment complex during force application, thus enabling the enhancement of the design of dental implants which has been based on theoretical analysis to date.


Introduction
Two main designs of the dental implant have emerged within the last century, i.e. the two-piece and one-piece implant (Binon, 2000). To date, extensive research has been performed on the mechanisms of osseointegration of these implants and a high predictability of success for dental implants has been demonstrated (Albrektsson et al., 2008). This status of knowledge obviously influenced a recent recognizable shift of research objectives to the composition of the implant components and their mating zone (Tsuge et al., 2008;Semper et al., 2009Semper et al., , 2010. Two-piece implants consist of two separate components: the endosteal implant and the abutment carrying the prosthetic restoration connected by a screw joint (Binon, 2000). Unlike one-piece implants, two-piece implants are commonly used because they can be individually loaded with different types of abutments. Two-piece implants feature a mating zone in which the implant-abutment connection is ensured.
The mating zone utilized in all two-piece implants can be differentiated into two principles: a butt-joint connection or one based on conical surfaces (cf. Fig. 1). The implant-abutment interface in butt-joint connections reveals a micro-gap (Jansen et al., 1997;Coelho et al., 2007). Adjacent to the microgap, when placed into the bone or gingiva, an inflammatory reaction has been described (Broggini et al., 2006). The stimulus for this inflammatory reaction has been discussed to originate from the micro-gap (Broggini et al., 2006;Hermann et al., 2001). The micro-gap allows microbial colonization of the internal cavity of the implant-abutment complex as well as penetration of bacterial endotoxins into the surrounding tissue initiating a pathophysiological process that can result in bone Two principles of mating zone utilized in all two-piece implants: based on (a) conical surfaces and (b) butt-joint connection. loss and eventually implant loss (Jansen et al., 1997;Broggini et al., 2006;Steinebrunner et al., 2005).
Direct observations of this micro-gap at the implant-abutment interface using X-rays are challenging owing to the limited resolution and contrast of the available laboratorybased methods: in vivo radiography and computed tomography are only applied to assess success and stability or failure of dental implants (see, for example, Yip et al., 2004;Brä gger, 1998). In vitro studies reported in the literature are very scarce and commonly limited to butt-joint connections where the micro-gap is visualized or its size estimated indirectly, e.g. via reference points (cf. Tsuge et al., 2008;Coelho et al., 2007). One approach to visualizing the micro-gap in both types of implant-abutment connection designs is by using micro-focus X-ray tubes for in vitro micro-radiography, where again only the gap in butt-joint connections was accessible (Zipprich et al., 2007). Other methods where the micro-gap is commonly inspected after cyclic loading of butt-joint connections include scanning electron microscopy, optical microscopy, scanning laser microscopy or theoretical approaches via finite-element modeling (see, for example, Tsuge et al., 2008;Coelho et al., 2007;Hecker et al., 2006). An in vitro observation of a micro-gap at the implant-abutment interface with conical-shaped connections has not been reported yet, hence even its non-existence was concluded due to this (Zipprich et al., 2007). Recent leaking tests showed only the lack of sealing capability of this type of connection (Coelho et al., 2008;Harder et al., 2010).
In order to overcome limitations of the imaging technique we apply hard X-ray synchrotron radiation. The advances towards X-ray imaging using laboratory sources are the several orders of magnitude higher photon flux density available and the almost parallel beam propagation. This allows for extending the sample-to-source distance to up to more than 100 m and therefore to suppress the influence of the finite source size on the spatial resolution. The high monochromatic photon flux density increases the contrast while reducing artifacts. Synchrotron micro-imaging was established during the 1990s, and nowadays is available with spatial resolutions up to the sub-micrometer and time resolutions up to the microsecond range (Koch et al., 1998;Rack et al., 2009a). Besides the improved resolution, imaging using synchrotron light sources also gives access to more sophisticated contrast modes like inline phase contrast or holotomography (Cloetens et al., 1999); for further details see, for example, the book by Banhart (2008). Besides the numerous applications in materials science, archaeology or cultural heritage [cf. Baruchel et al. (2002Baruchel et al. ( , 2006 or Stock (2008)], the development is also approaching fast medical and even clinical applications (Keyrilainen et al., 2008;Baruchel et al., 2008;Stiller et al., 2009;Issever et al., 2008;Weitkamp et al., 2008;Zabler et al., 2006).
The purpose of this study is the in vitro visualization of a micro-gap formation at the implant-abutment interface with conical-shaped connection. The images taken show dimensions and the development of the micro-gap under different mechanical loads, knowledge which is important for under-standing the functionality of implants with conical-shaped connection as well as to optimize and develop further their clinical applications.

Dental implant and test stand
A virgin dental implant with a conical connection and a diameter of 4.1 mm (bone level implant, L = 14 mm, ref: 021.4114, lot: G6582; Straumann AG, Basel, Switzerland) and the corresponding rotation-safe abutment (NC-Mesosekundaerteil -Titan, ref: 022.2202, lot: F6601, Straumann AG, Basel, Switzerland) were assembled and screw-tightened with a torque of 0.25 N m using the system-specific screw driver and ratchet. An individually fabricated steel-ball was glued to the abutment (Superglue X60; HBM Germany, Darmstadt) according to EN ISO Norm 14801:2003.
The implant-abutment assembly was embedded in an individually fabricated brass cylinder (Fraunhofer Institut Werkstoffmechanik, Freiburg, Germany) using Superglue X60 (HBM Germany, Darmstadt, Germany) according to EN ISO Norm 14801:2003; a crestal bone level 3 mm below the implant shoulder was simulated. The brass cylinder carrying the implant-abutment assembly was screw-fastened to an individually fabricated test stand made from stainless steel (V4A, Klaus Ellinger CNC Zerspannung GmbH, Berlin, Germany).
A static force (nominal 0 N, 30 N, 60 N, 100 N) was applied at a 90 angle to the implant axis onto the ball. The force application was monitored using a digital force gauge, model SH-500 [PCE-group OHG (serial No. 5808062790)].

Synchrotron-based micro-imaging
Measurements were carried out at the BAMline of the third-generation synchrotron light source BESSY-II (Helmholtz Zentrum Berlin fü r Materialien und Energie, Germany) (Gö rner et al., 2001;Rack et al., 2008). Numerous successful studies have already proven that this experimental station is excellently suited for synchrotron-based micro-imaging (see, for example, Kamenz & Weidemann, 2009;Rack et al., 2009b;Zabler et al., 2007;Manke et al., 2007). The white radiation from the wavelength-shifter insertion device of the BAMline was filtered with 0.2 mm Cu and 0.2 mm Be before passing through a double-multilayer monochromator which selected X-ray photons with an energy of 50 keV for imaging. The resulting photon flux density is of the order of 10 10 photons s À1 mm À2 with an energy bandwidth of 1.7% (Rack et al., 2008). Radiographic projection images were acquired using an indirect detector, based on the concept as introduced by Hartmann et al. (1975) as well as Bonse & Busch (1996); the luminescence image of a scintillator screen is optically coupled to a camera via diffraction-limited visible-light optics. A principle sketch of the detector design and the experimental set-up is displayed in Fig. 2. For this experiment, a 50 mm thin CdWO 4 (CWO) single crystal glued on top of a 500 mm-thick undoped Y 3 Al 5 O 12 (YAG) substrate was chosen as scintillator screen (Nagornaya et al., 2005). The luminescence image of research papers 290 A. Rack et al. Two-piece dental implants the crystal is read via a visible-light microscope, designed and manufactured by the company Optique Peter (Lyon, France): an Olympus objective Uplsapo (10Â/0.4 NA) in combination with 2Â eye-piece projects the image with an effective 20Â magnification onto a CCD camera (0.43 mm effective pixel size; this value allows one to convert the size of the features in the images from pixels into meters). The diffraction-limited resolution of the objective is not reached owing to the thickness of the scintillator which exceeds the depth of focus of the objective. Hence, based on the thickness of the scintillator crystal we can estimate the spatial resolution of our detector system to be approximately 4 mm (250 line-pairs mm À1 ) (Koch et al., 1998). The thickness of scintillator and substrate were required in order to protect the visible-light optics from radiation damage owing to the intense high-energy X-ray beam applied. As camera a pco.4000 (PCO AG, Germany) was used. The camera is based on a Kodak KAI-11000 interline transfer CCD chip with 4008 Â 2672 pixels (each 9 mm in size); a dynamic range of 5000 : 1 was measured with exposure times between 0.1 s and 10 s, one signal unit (ADU) corresponds to a charge of 3 electrons in the corresponding potential well of the CCD chip, peak quantum efficiency above 50% at 500 nm. The field of view of the complete detector is approximately 1.7 mm Â 1.1 mm. As the experiment is located 35 m away from the X-ray source (Rack et al., 2008), the finite source size has no influence on the resolution (e.g. by penumbral blurring). The radiographic projection images shown in Fig. 3 were acquired with exposure times of 10 s. The distance between sample and detector was roughly 4 cm. Despite this relatively large propagation distance, common phase-contrast edge enhancements are only slightly present. This is due to the X-ray photon energy and spatial resolution of the employed indirect detector set-up, as well as the coherence properties of the BAMline (approximately 165 mm horizontal source size).

Results
The set of images taken during the in vitro measurement is shown in Fig. 3. The contrast is given by the X-ray absorption of the specimen which is determined roughly by its material and density as well as the effective thickness of the specimen along the X-ray beam path. The main components of the implant are marked and the position of the detector's field of view with respect to the sample is sketched in Fig. 2. Furthermore, the orientation of the applied force F is also shown. Stripe modulations are present in all images which originate from the double-multilayer monochromator used; they are not a feature of the investigated sample. Owing to the limited field of view of a high-resolution indirect X-ray pixel detector, only a small part of the specimen is investigated. As the length of the gap is roughly 0.7 mm and the height of the synchrotron beam is limited, two images were acquired and later merged into one in order to illustrate the complete gap.
In all stages with different values for the applied force F (nominal 0 N, 30 N, 60 N, 100 N), a micro-gap between the abutment and the implant can be detected. The specimen as Sketch of the experimental set-up: the hard X-rays coming from an insertion device of the light source BESSY-II (right;Gö rner et al., 2001;Rack et al., 2008) are transmitted to the sample under load (photograph with zoom inset and sketch); the attenuated beam is converted into visible light by a scintillator screen. This luminescence image is captured via visible-light optics and a digital camera (Hartmann et al., 1975). Only a small part of the specimen can be imaged owing to the detector's limited field of view (cf. Fig. 3).
imaged without mechanical load (0 N) shows a micro-gap which is slightly below the resolution limit of our detector as it is only visible owing to a local reduction of the attenuation but not sampled by several pixels (cf. with the image at nominal 60 N mechanical load). Hence, we estimate its size to be in the range >1 mm and <<4 mm. For a nominal value of the applied force F = 30 N the micro-gap becomes clearly visible, sampled by between eight and ten pixels, so its size is in the range of 4 mm. When moving towards higher mechanical load, the gap at the implant-abutment interface opens further. At a nominal force F = 60 N, the size is around 11 mm (26 pixels) with the surfaces of the implant and the abutment running almost parallel. Finally, at a mechanical load of nominal F = 100 N, the micro-gap shows a non-parallel shape: at the upper end of the gap its size is roughly 22 mm (50 pixels) while at the lower end the size is around 15 mm (35 pixels).

Discussion
Animal studies have shown that the design of the implantabutment connection has proven to be of high relevance for the stability of the soft and hard tissue surrounding the implant (Weng et al., 2008). The exact mechanisms responsible for the biologic reaction of the bone in correlation to the micro-gap are still unclear. To date the illustration of the micro-gap in conical implant-abutment connections has not been feasible, as the conventional laboratory radiographic methods utilized did not allow distinct identification of clinically relevant gaps in the micrometer range (Zipprich et al., 2007).
Implant-abutment assemblies are screwed joints that are exposed to dynamic loading owing to the masticatory process with axial and extra-axial forces (Binon, 2000). These forces can be up to 450 N, varying with the angle of application (Morneburg & Proeschel, 2002;Mericske-Stern et al., 1992). Forces of up to 110 N applied at a 90 angle to the implant axis have been described to occur on the abutment carrying the implant-retained restorations (Mericske-Stern et al., 1992). It has been shown with radiography using micro-focus X-ray tubes that butt-joint connections present an increase in microgap when extra-axial force is applied (Zipprich et al., 2007).
Radiography using laboratory sources has been used to evaluate the micro-gaps of various systems but this method does not allow the detection of a micro-gap in internal conical implant-abutment connections. This is due to the limitations in resolution and contrast, given by the limited photon flux density, non-parallel beam propagation and influence of the finite source size in comparison with a synchrotron light source. The use of monochromatic hard X-ray synchrotron radiation to perform micro-radiography allowed for the first High-resolution radiographic images of the micro-gap formation at the implant-abutment interface for different mechanical load (the stripe modulations within the images originate from the X-ray monochromator used and are not a feature of the specimen). The relative position of this field of view with respect to the complete implant can be found in Fig. 2. time the visualization of a micro-gap in internal conical implant-abutment joints.
The size of the micro-gap visualized varied, depending on the mechanical load, between approximately 1 mm and 22 mm, clearly ranging above the size for oral pathogens found responsible for a periimplantitis. The smallest size of oral bacteria found in the oral biofilm is 0.1 mm, whereas bacterial endotoxins with a size clearly smaller than 0.1 mm seem to be of importance. Endotoxins are lipopolysaccharide molecules (size of 10 kDa) found as part of the cell wall of gram-negative bacteria, and are released primarily upon cell lysis. These small-sized pathogenic molecules induce an inflammatory process within their vicinity (Broggini et al., 2006). The size of the micro-gap has not been proven to correlate with the degree of inflammation; it is rather important that its existence seems to influence the periimplant environment (Hermann et al., 2001).
As in all screwed joints the two mating components do not form a leak-proof tightness and have proven to show bacterial leakage in in vitro studies even without load application (Jansen et al., 1997;Coelho et al., 2007;Steinebrunner et al., 2005;Harder et al., 2010). Optical microscopy of implant-abutment connections based on a butt-joint principle have shown that there is a micro-gap of up to 10 mm with only punctual contact/fulcra of the mating zones which has been proposed to occur owing to the machining of the mating parts. To date there are no data available about the condition of the surface of the mating zone in conical dental implant joints (Coelho et al., 2007). A continuous micro-gap without punctual contact of the mating components was seen within the unloaded specimen evaluated in this study. This location/site examined represents one cross section of the implant and shows that there is a missing surrounding surface contact. This incongruence allows for the correspondence between the external and internal environment of the screwed implant assembly even in unloaded conditions, as proposed in in vitro experiments (Jansen et al., 1997). A possible explanation for this phenomenon is the imprecision of the machining of the parts which has been correlated to a repositioning instability of the abutment and to technical complications encountered in dental implant-retained restorations (Semper et al., 2009(Semper et al., , 2010Jansen et al., 1997;Coelho et al., 2007). For generalization, this assumption needs further investigation with numerous samples, which has become accomplishable now by applying monochromatic hard X-ray synchrotron radiation.
Besides allowing for a precise illustration of the joint gap in conical dental implant-abutment connections by using monochromatic hard X-ray synchrotron radiation, a thorough investigation of the mechanical behavior in various loading situations of the components has become possible. Elucidation of the mode of the mechanical behavior of the implantabutment joint under various loading scenarios regardless of their design will provide information to enhance the design and function of the joints and minimize the technical complications encountered to date. All materials were purchased by the hospital and are therefore free of any commercial interest. The research was funded by the HU Internal Research Funding.