Study of the optical quality and strength of the glass edge after grinding and polishing

Glass edges are obtained from cut glass sheets with subsequent additional finishing. Mechanical disturbances of the brittle glass material lead to defects and cracks on the edge surfaces. These defects affect the strength of the entire glass. As part of a research project at the Faculty of Architecture of the Technical University of Dresden, the characteristic visible effect of grinding and polishing processes on glass edges and their strength was investigated. Therefore, special attention in this research project was given to the effect of various cup polishing wheels on the bevelled surfaces of annealed glass. This article presents some of the main steps involved in glass edge surface treatment, presents the grinding and polishing cup wheels in question, and describes the experimental tests that have been carried out.
Microscopic analysis allows characterization of typical surface defects. In addition, four-point bending tests were carried out to determine the tensile bending stress at failure. The combination of these two methods allows the analysis of defects leading to failure before failure, as well as the correlation between the optical quality of the surface and the bending tensile stress. In addition, the microscope can be used to set up the grinder and check for reproducible edge quality. Evaluations have shown that a specially designed chamfering polishing wheel improves surface quality and hence edge strength.
With modern architectural demands and stringent requirements for building physics, mechanical and thermal loads on window and facade structures are increasing. As the glass thickness increases and the glass structure becomes more complex, the imposition of mechanical loads, especially thermal ones, can lead to critical stresses, especially in the edge regions. According to current design regulations in various European countries, the estimated cost of annealed glass must be reduced to the size of the glass edge (Feldmann and Kasper, 2014, p. 55). In the case of German and Austrian standards (DIN 18008-1 2019 and ÖNORM B 3716-1 2016), the characteristic bending strength of the annealed glass must be reduced by 80%. This reduction takes into account the high dispersion of edge strength due to processing and represents the lowest level of glass edge strength.
The reduced strength resistance of the annealed glass edge and the lack of adaptability of the glass edge to load scenarios lead to insecurity. To ensure a certain degree of structural safety of the glass structure and not to damage the glass, further in-depth inspections of the glass edges must be carried out. Nowadays, planners often use tempered glass, which has higher edge strength. However, higher edge strength comes with the risk of spontaneous fracture due to nickel sulfide inclusions and reduced optical quality due to apparent anisotropy. In addition, tempered glass is very expensive to use. There is a need for more efficient use of annealed glass. For this reason, safe design methods and the production of annealed glasses with acceptable edge strength and low scattering are needed.
The glass edge is obtained by cutting and additional processing of the glass sheet. The cutting process determines the geometry and dimensions of the glass sheet. During the subsequent grinding and polishing process, the material is eroded at the edge surfaces and at the edges to ensure dimensional accuracy and improve optical quality. The so-called finished edges reduce the risk of injury and allow the glass to be further tempered.
On fig. 1 shows a typical geometry of a cut edge at the back and a further polished edge and two diagonal chamfers at the front. Five areas were defined, divided into surfaces and lines, to account for the milling process and to match crack locations to defined areas. The edge surface (e) describes the area perpendicular to the surface (p) of the glass sheet. Chamfer (c) defines the surface of the cut edge, typically at a 45° angle, extending from either side of the glass sheet surface to the edge surface. In addition, the transitions between the edge and the bevel (tc) and between the bevel and the surface of the glass plate (tp) are defined.
In the construction industry, different types of glass edge are classified according to DIN 1249-11 (2017). Cut edges are described as having sharp edges, clean cut areas, and irregular breaks in scratched areas. In addition to cut edges, there are four types of additionally processed edges with increased processing steps and optical quality. To create angularity, sharp edges and irregular breaks of the cut edges are cut off. The edge surfaces are not necessarily machined. Grinding edges, smooth grinding edges and polished edges depend on further grinding and polishing steps. When sanding, a rough abrasive is always done first, resulting in a rough appearance. The surface is then polished with finer, better quality tools, resulting in a flat, transparent surface. However, during the grinding and polishing process, new defects and cracks appear on the edge surface, which affect the tensile flexural stress at break and may need to be taken into account when calculating the edge strength of the entire glass.
The Fachverband Konstruktiver Glasbau eV (FKG) Edge Strength Working Group has carried out extensive experimental studies on edge strength with cut, angular and smooth edges. Of these, over 1,000 glass beams from six different manufacturers were tested in 33 test series for four-point bending (Ensslen 2013). In addition, selected samples from each test series were subjected to microscopic examination on the edge surfaces before destruction. Klederlein et al. (2014) demonstrate in Fig. 2 differences in the types of edge surfaces from different manufacturers. For each type of glass frame, there are images from three different manufacturers. The quality of the edges is especially different. Recordings of processing parameters showed that the edges were produced on different grinding machines (cross belts or grinding wheels).
The test procedure for determining the strength of glass edges refers to DIN 1288-3 (2000), which defines a four-point bending test for sheet glass, in which a load is applied to the weak axis. However, for a special edge strength test, the setup was modified to perform in-plane tests. This creates a uniform tensile stress on the edge, which increases the likelihood of tearing directly from the edge and allows the strength of the edge to be determined (Ensslen 2013). The test series were statistically evaluated using a two-parameter Weibull distribution to determine the 5% quantile at a 95% confidence level, which corresponds to the inherent bending strength of the glass.
Out of 33 test series, only two showed a quantile 5% below the recommended standard edge strength of 36 N/mm² for annealed glass (fk = 45 N/mm² for 80% according to DIN 18008-1 (2019)). The maximum quantile was 64.84 N/mm². The common belief that ground and polished edges are always of higher quality than cut edges can be refuted. A comparison of the finish of the same edge shows that the influence of the manufacturer is much greater than that of the finish itself. Thus, a well-cut edge may have higher strength than a poorly-ground edge with many imperfections. (Klederlein et al. 2014)
In addition to the work of the Working Group on Edge Strength, Lindqvist (2013) and Vandebrook (2014) also investigated the edge strength of annealed glass. Lindqvist (2013) performed a microscopic analysis to predict edge strength based on the identification of critical cracks. It turns out that critical defects are difficult to detect. Vandebroek (2014) investigated edge strength with bending tests to determine the effects of load history, stress corrosion, size and stress distribution. The check focuses on different types of edges, different glass manufacturers and sizes. The results confirmed the wide scatter of strength results from different manufacturers and indicated the need to explore different edge finishing methods.
Lohr (2019) investigated the grinding of thermally toughened glass. In some microscopic examination, it can be seen that the transition region between the tc and tp surfaces (see Fig. 1) shows many defects, and various breaks come from chamfers and transition regions. The possibility of transferring the results of thermally tempered glass to annealed glass should be investigated.
After extensive research on different types of glass edges, the Edge Strength working group focused on cutting edges. Based on the results of the first study and a deeper study of the cutting process, it is possible to create a “good” set of cutting parameters for 8 mm thick glass, resulting in a reproducible edge strength of at least 45 N./mm245 N/mm2 (Ensslen and Müller-Brown, 2017). In addition, optical properties related to edge strength can be determined, such as the depth of transverse and intermediate cracks. Therefore, the assessment of cut edge strength can be based on optical methods (Müller-Braun et al., 2020).
Previous studies by the FKG Working Group on Edge Strength (Ensslen 2013; Kleuderlein et al. 2014; Ensslen and Müller-Braun 2017) and Vandebroek (2014) have shown that grinding and polishing processes have a significant impact on the quality and strength of glass edges. . Despite standard designs, the quality of the optical edges varies considerably, as shown in Fig. 2. There are a large number of process parameters for all types of edges, and they are widely used among manufacturers (Kleuderlein et al., 2014).
These facts highlight the need for a deeper understanding of further finishing processes, the definition and specification of process parameters, and the production of comparable edge quality. The production parameters of grinding and polishing have not been scientifically verified. Only reproducible edge quality, based on defined process parameters, minimizes the variation in edge strength regardless of different manufacturers. This is necessary to achieve consistent glass edge quality and intrinsic flexural strength, annealed during processing, without the need to further shorten the design process.
The purpose of this study is to better understand the effect of finishing processes on materials and provide guidance on how to set up the process to achieve reproducible edge quality supported by optical techniques. Therefore, the process of grinding and polishing was studied and the adjustable parameters were determined. First, consider the manufacturer’s edge processing process. Particular attention in this control is given to reducing the number of defects and cracks in the chamfers (c) and transition zones (tc) to improve and create a reproducible optical quality and higher glass edge strength. Therefore, three different polishing tools and chamfering sizes are different.
To assess the influence of parameters, a method of experimental tests was developed using microanalytical methods. Microscopic analysis helps to understand the effect of finishing on the material, characterizing the resulting surface and evaluating the resulting defects. The control procedure included microscopy of the edge surface before destructive testing and localization of defects leading to failure after failure in a four-point bend test. Finally, a defect leading to failure is associated with a certain flexural tensile stress at failure.
In practice, various grinding machines are used for further processing of glass edges. For example, cross belt machines consist of moving belts coated with abrasives. For processing, the glass must be manually pressed against the conveyor belt. Some manufacturers use cross band machines to make the edges. The other is a CNC grinder that grinds complex shapes under computer control. However, the most common method for obtaining additional processing on the edges of glass, especially smooth ground and polished edges, is the edger.
It can be divided into vertical, single side, horizontal and double side edger. The samples in this study were processed with a Neptun Rock 11 single-sided vertical edger as shown in Figure 3. Therefore, the following description of the grinding process is based on this method. The machine consists of eleven grinding and polishing stations. At each station, multi-grain tools, so-called cup wheels, finish the edges. Only the cup wheel motors are visible from the outside. Pictograms under the pictures in fig. 3 should help in assigning processing steps and individual jobs.
The glass sheets are stacked on vertical and horizontal guide belts on the right and then transported to the left by the edger. The edge to be processed is facing down. The horizontal guide belt transports the glass to the first station. The glass is then guided to other stations using two side guide belts. They hold the glass in a straight line and regulate the speed. At each station, a rotating cup wheel or polisher works the edge from below while cooling water is continuously added. The cooling water prevents the cup wheels from overheating. Cup wheels for stations 1 to 3 and stations 4 and 6 are rigid. The cup polishing wheels of Station 5 and Stations 7-11 are hydraulically pressed against the edge of the glass. The contact pressure of the mobile station can be adjusted. Sanding depth describes the amount of material removed at the end of the process, controlled by the position of the horizontal guide belt.
The first station in the grinding process ensures material removal and dimensional accuracy along the glass edge surface. Then, at stations 4 to 7, chamfers are created with a grinding wheel at an angle of 45°. On fig. 4 shows the alignment of the grinding wheel during the processing of edges and chamfers.
The edger shown has two stations for each bevel: a coarse grinding wheel with diamond grains as the abrasive and a fine grinding wheel for further polishing. The cups become thinner towards the end of the process at stations 8 to 11. They reduce surface roughness and provide high transparency. Depending on the type of edge, the required stations can be opened individually during the process. The polished edges run through all stations and are therefore of the highest quality. A consistent grinding wheel configuration and other process parameters are critical to the production of high quality glass edges.
Grinding with grinding and polishing cup wheels removes microscopically small and geometrically indefinite chips from the glass edge. They differ in the composition and grain size of the combined system. On fig. 5 shows examples of various cup grinding and polishing wheels and their microscopic surfaces.
Large particles in combination with hard, metal or resin bonds are used in the grinding stage. The left image of Figure 5 shows a metal bonded diamond grinding wheel that achieves high material removal rates by using extremely hard diamond particles. They create roughness on the treated surfaces. The removed material accumulates in the segments and is washed out by the cooling water. The image in the center shows a synthetic resin bonded diamond wheel, which is similar to a metal bond but is softer due to the use of a synthetic resin bonding system. This ensures smooth material removal.
On the right side of Fig. 5 shows a polished cup with fine-grained corundum. Polishing wheels are made by combining abrasive grains such as silicon carbide, corundum or cerium oxide with elastically bonded backing materials such as polyurethane, rubber or modified synthetic resins with varying degrees of elasticity. The chemical structure of the respective binder system, in combination with the abrasive grit, influences the final performance of the tool. Smaller particles of a softer and more elastic glue are used to create smooth and transparent surfaces. Cerium oxide is a suitable grit for fine polishing because it can be removed by mechanical and chemical grinding. The chemical reaction dissolves the atoms in the glass, which then protrude above the surface. In this way, smoothing of small irregularities of the edges is achieved.
Cup wheels should always have a few open grains on their surface to finish the edges of the glass. So it can work effectively. To obtain the desired abrasive effect, the abrasive particles must not wear faster than the bonded system, destroy the bonded system, or expose individual particles (Figure 6).
The final edge quality is the result of the combination and number of cup wheels, several process parameters and their correct interaction. Currently, there are no well-known optimized process parameters for grinding and polishing processes to achieve a certain quality or edge strength. In fact, the optimization is based on the visual quality of the glass edges. Producers and shredders have a lot of influence and are not well understood. However, the influence of individual process parameters can be deduced from the principle of operation of grinding. The high rotational speed of the cup wheel combined with the low speed results in more frequent contact of the wheel with the glass surface. This should improve the physical material removal and polishing process. In addition, the quantity and purity of the cooling water can have a big impact on quality.
Beveled surfaces and transition regions are of particular interest in this observation. Therefore, three different cup wheels for chamfering and polishing were investigated in terms of optics and edge strength.
To assess the quality of the edges and surfaces with a chamfer and conclusions about defects that could lead to fracture, a three-stage experimental test methodology was developed. First, edges and bevelled surfaces are documented and checked under a microscope. Then a four-point bending test was carried out. At the final stage, the source of destruction is localized and the defect leading to destruction is identified by comparison with the recorded image of the surface before destruction.
The fabricated specimens were beams measuring 125 mm x 1100 x 10 mm with polished edges (Fig. 7). The dimensions are based on the four-point bend test stand design. Polished glass edges were chosen because obtaining high optical quality edges was one of the objectives of the research project. In order to examine a sufficiently large surface and thus obtain more information about the grinding and polishing process, a thickness of 10 mm was chosen for the sample. According to figure 7, the definition of chamfers as chamfer 1 and chamfer 2 is necessary for clear comparability. The classification is determined by the grinding method. Since the grinding machine required a height of at least 250 mm to process the edges, specimens were made oversized and then trimmed.
Microscopic analysis of the surface is carried out using a Zeiss digital light microscope with a magnification from 34x to 1100x. Mark and record the edges and beveled surfaces of all specimens. This makes it possible to characterize the quality of the edges and additionally localize defects leading to fractures. The registered surface covers the middle of the sample about 200 mm long (this region is marked in Fig. 7). This special area corresponds to the loaded area in the following four-point bend test. Therefore, the occurrence of the fracture likely occurred in this region. Edge surfaces were recorded at 70x magnification and beveled surfaces were recorded at 100x magnification. On fig. Figure 8a shows a tilt-mounted microscope used to record images of a beveled surface, and an example of microscopic recording of a beveled surface approximately 20 mm in size (Fig. 8b).
The four-point bending test refers to DIN EN 1288-3 (2000) and the FKG edge strength test, carried out by bending a loaded glass sample around a strong axis. The controlled edge faces down, where the bending tensile stress occurs. On fig. 9 shows a test setup and corresponding pictograms at the Dresden University of Technology. Ball bearings and side supports at the ends of the beams are combined into a ball bearing support system. The specimen was supported vertically on an 80 mm wide POM block with a span of 1000 mm. The load is applied pointwise over a span of 200 mm using small POM blocks placed on the upper edge of the glass. The fracture test was carried out at a constant load rate of 2 N/(mm2s2s) until failure. After failure, the breaking load was measured.
After the break test, determine the xglobglob element’s global break position to decide if the test should be evaluated (Figure 10). Only the centers of destruction in the loaded area are evaluated. Careful observation of the fracture mirror can determine the source of the original crack (Quinn 2016, 7–10). Further differentiation is made from edge surfaces (e), chamfered surfaces (c), transition areas (tc), or glass sheet surfaces (p) to localize fracture-causing defects in the recorded images of undamaged surfaces. On fig. 10, the center of the broken mirror indicates the initial crack on the second bevelled surface.
All samples were ground at a speed of 2 m/min and a grinding depth of 1 mm. Grinding depth describes the amount of material removed from the glass edge during the grinding process. Before the grinding process, the measured length of the visible intermediate crack formed during the cutting process (Müller-Braun et al., 2020) is about 300 µm. With a selected grinding depth of 1 mm, intermediate cracks visible during cutting are completely removed.
Therefore, it is assumed that the influence of the cutting process is excluded. However, this assumption should be verified in the next series of tests by grinding cutting edges of different qualities at constant grinding process parameters. The bond system and grit type of the cup wheels used are listed in Table 1. The adjustable contact pressure for each cup wheel is set and recorded based on the manufacturer’s experience and macro-optical results during manufacture. Since this depends on the wear of the cup wheels used, it may vary depending on the test series.
Table 2 provides an overview of the test series. A total of three different polishing cup wheels from Artifex were tested for chamfering at stations 5 and 7 (according to Fig. 3). For test series A, a resin-bonded cup wheel with very fine corundum particles was used. This polished cup is solid. Samples of experimental series B were made using a polished polyurethane cup wheel as a binder and medium-grained corundum.
Therefore, a softer polyurethane foam is obtained, and the polishing cup wheel is softer. As part of a research project for chamfering, a special polishing wheel was developed from a polyurethane binder system and internal granules. They foam into a fine-meshed but rigid polyurethane foam. The exact composition is subject to a non-disclosure agreement with Artifex Corporation. Samples of the experimental series C were made with specially designed cup polishing wheels.
In addition, three series of tests with different chamfer sizes were investigated to test the effect of different chamfer removal rates. Therefore, the burrs used for chamfering are manually adjusted. A typical adjustment removes about 1mm of 45° edge area on each side, resulting in a 1.4mm wide beveled surface. For small chamfers (KS test series) the adjustment should remove approximately 0.5 mm of edge width on each side, giving a chamfer width of 0.7 mm.
Large chamfers approximately 1.7 mm wide (test series GS) were obtained by adjusting the disc grinding wheel to remove 1.5 mm on each side. In addition, a test series with polished edges but no chamfers was made (test series O). The goal is to avoid defects caused by chamfering. For test series O, sites 4 to 7 were closed. For each series of tests, eight to twelve test specimens are made.
Microscopic analysis helps to obtain general information about machined edges and bevelled surfaces. Some typical surfaces and emerging defects are shown in Fig. 11. The impact of the cup wheel on the glass leaves an abrasive pattern on the surface, which depends on the direction of rotation of the cup wheel. An abrasive pattern can be thought of as a permanent pattern of grooves on a surface. In addition, typical defects were identified, such as individual scratches in the direction of the grinding pattern, defects in the transition zones, further described as chipping and conchoidal appearance.
The only visible scratch in the direction of grinding (Fig. 11, left) can be caused by glass abrasive removal of small, exposed or broken abrasive particles on the cup wheel. Chipping and peeling can occur due to the removal of material in the edge and chamfer area. They occur naturally at the point of failure of the material (Fig. 11, center and right). Since defects are formed during grinding or polishing, they are classified as typical defects. The following analysis contains a comparison of these typical defects in various test series.
Microscopic images of the surface allow a preliminary assessment and characterization of the quality of the edges. On fig. 12 shows representative surface sections of bevel 1, edge and bevel 2 of the test series (test series A, B and C) with different cup wheels. The location of the images corresponds to the transition areas of adjacent surfaces. The chamfer surface is 4 times larger than the edge surface.
Comparison of the edge surfaces of individual test series showed no significant differences. On closer inspection, you can see the grinding with small parallel grooves. However, grinding patterns visible at the microscopic level are barely visible at the macroscopic level. Therefore, the edge surfaces are unambiguously classified as polished.
Observations of the bevelled surfaces of the test series revealed various qualities that can be traced back to the characteristics of the various beveled cup polishing wheels.
Thus, grinding cup C showed the best optical results with the fewest chips and nicks. Macroscopically, test series C samples also showed the best quality of the optical bevel and the fewest defects.
On fig. 13 shows microscopic images of a series of tests (test series GS, KS and O) with different chamfer sizes. Chamfering cup wheel C was used for chamfering because it showed the best results in the cup wheel polishing study. Garbage and shells of experimental series KS and GS were present in the range of 80 µm. Only one sample in the GS test series showed unusual clasts and shells in the 450 µm range. The surface quality of the chamfer is reproducible, except for this single sample. Different sizes are not recognized. It is assumed that the amount of material removed can be seen in the transition zone by the presence of chipping and peeling.
One of the bevelled surfaces of test series KS and GS showed the same smooth finish as test series C in Figure 12, with several visible directions. In test series KS, the grinding pattern is visible on chamfer 2, and in test series GS, grinding pattern 1 is shown on the surface of the chamfer. Due to the grinding pattern of the edge and surface of the chamfer, it can be confirmed that the chamfer specification is correct. Since the manual adjustment of the chamfering cup wheels at stations 4 and 6 (according to Fig. 3) is the only difference between the KS and GS test series, the interaction between grinding pressure and material removal has a noticeable effect. In addition, random measurements of the width of selective chamfers have shown that manual adjustment will result in different chamfer sizes.
For test series O, there is no chamfer. Therefore, in Fig. 13 only the edge surfaces are shown. The edge surface corresponds to the optical quality of other test series. In the region of transition from the edge surface to the surface of glass 2 of experimental series O, individual chips and shells with a length of about 300 μm were found. Material removal can be traced to the edge of the glass as no bevels are created.
A total of 62 specimens were tested to study bevel polishing cup wheels and bevel sizes. According to the Euler-Bernoulli beam theory, the flexural tensile stress of each specimen was determined from the measured fracture load. The analysis took into account the global location of the Xglob fracture, as well as the exact origin of the fracture. Cracks were found emanating from the edge surface (e), the chamfer surface (c), the transition region between the edge and the chamfer surface (tc) and the glass surface (p). 10 samples showed fracture centers on the glass surface, which were not taken into account in the evaluation. In addition, nine specimens failed outside the loaded region, where maximum tension is expected. These samples were also not included in the evaluation. Since only the loaded area was recorded under the microscope, these defects could not be characterized and additionally compared with microscopic images.
On fig. 14 shows the determined tensile bending stress for each series of tests as a rectangular chart. The thick line inside the box indicates the average value of the estimated value, and the number (n) in the box indicates the number of estimated samples. The gray chart contains all samples from the test series that have a discontinuity in the load area. Further subdivided by crack origin, initial cracks originate from edges (purple chart), transitions (blue chart), and chamfers (green chart).


Post time: Mar-22-2023

Send your message to us: