applied sciences

Article Development of Auto-Seeding System Using Image Processing Technology in the Growth Process via the Kyropoulos Method

Churl Min Kim 1, Sung Ryul Kim 1,* and Jung Hwan Ahn 2

1 Precision Manufacturing & Control R&D Group, Korea Institute of Industrial Technology, 1,30, Gwahaksandan 1-ro 60 beon-gil, Gangseo-gu, Busan 46742, Korea; [email protected] 2 School of Mechanical Engineering, Pusan National University, 2, Busan daehak-ro 63 beon-gil, Geumjeong-gu, Busan 46241, Korea; [email protected] * Correspondence: [email protected]; Tel.: +82-051-974-9259

Academic Editor: Giorgio Biasiol Received: 2 February 2017; Accepted: 5 April 2017; Published: 7 April 2017

Abstract: The Kyropoulos (Ky) and Czochralski (Cz) methods of are used for large-diameter single . The seeding process in these methods must induce initial by initiating contact between the seed crystals and the surface of the melted material. In the Ky and Cz methods, the seeding process lays the foundation for ingot growth during the entire growth process. When any defect occurs in this process, it is likely to spread to the entire ingot. In this paper, a vision system was constructed for auto seeding and for observing the surface of the melt in the Ky method. An algorithm was developed to detect the time when the internal convection of the melt is stabilized by observing the shape of the spoke pattern on the melt material surface. Then, the vision system and algorithm were applied to the growth furnace, and the possibility of process automation was examined for sapphire growth. To confirm that the convection of the melt was stabilized, the position of the island (i.e., the center of a spoke pattern) was detected using the vision system and image processing. When the observed coordinates for the center of the island were compared with the coordinates detected from the image processing algorithm, there was an average error of 1.87 mm (based on an image with 1024 × 768 pixels).

Keywords: sapphire; growth; Kyropoulos method; auto-seeding; image processing; spoke pattern

1. Introduction Single-crystal sapphire is a material that has high hardness, excellent chemical stability, and optical transparency in a wide range of wavelengths. Due to these advantages, it is widely used in various industries including engineering, military supply, aviation, optics, and healthcare. At present, large quantities of sapphire ingots are being produced due to the prevalence of light-emitting diode back light units (LED BLUs), which use sapphire substrates, and the demand for this material is increasing rapidly [1]. Accordingly, many studies are being conducted on the production of large-diameter single crystals because they provide greater benefits than small-diameter single crystals in terms of productivity and market price. For the growth of large-diameter single-crystal sapphire, the Kyropoulos (Ky) and Czochralski (Cz) methods are predominately used. The Cz method is a wide spread single crystal growth technique in which alumina in melted in a crucible and the seed is pulled up; the seed is simultaneously rotated after it contacts the surface of the molten metal [2]. The Ky method is mainly used for the single crystal growth of large-diameter sapphire. Though its basic growth furnace is similar to that of the

Appl. Sci. 2017, 7, 371; doi:10.3390/app7040371 www.mdpi.com/journal/applsci Appl. Sci. 2017, 7, 371 2 of 14

Cz method, the seed is not rotated after it contacts the molten alumina, but the heater temperature is slowly lowered so that the single crystal grows downward from the seed [3]. Many studies have been conducted on the Ky and Cz sapphire growth methods. Timofeev et al. [4,5] conducted a study on 3D simulations of the sapphire growth method and the effect of heating conditions on melt convection in a Ky furnace. Chen [6] and Demina [3] researched the effects of temperature conditions on ingot shape during growth in the Ky method. Nouri [7] investigated the interaction between the heat zone and the internal melt in the seeding process of the Ky method. These researchers focused on analyzing the growth shape of single crystal sapphire by numerical analysis, paying attention to heater arrangement and power. The following studies are also related to growth process monitoring. Kozik and Nezhevenko [8] investigated methods for measuring the diameter of ingots during the single crystal growth of sapphire using vision system. Xiang et al. [9] proposed measuring the melt level in Cz crystal pullers using an image-based laser-triangulation measurement system. In addition, Winkler [10,11] investigated the automation of the growth furnace in the Cz method, but this research deals with partial automation after the seeding process. As in previous papers, it is difficult to find an entirely automated process for single crystal growth. This is because the optimization of the process conditions for seeding has not been established. Most studies that used Ky and Cz methods stated that seeding was performed when the temperature field of the melt material in the growth furnace was stabilized [12–14]. Until now, the seeding process in ingot growth, which has great influence on the quality of single crystal sapphire, has depended on the experience and technique of skilled workers. The seeding process lays the foundation for ingot growth in both Ky and Cz methods, and when any defect occurs in this process, it is likely to spread to the entire ingot. So, the growth process must be standardized and stabilized to grow a single crystal sapphire of good quality. The aim of this research was to establish and optimize the conditions of the seeding process in the Ky method. The vision system and algorithm were developed to detect the exact time when the internal convection of the melt is stabilized by observing the shape of the spoke pattern on the melt surface. The possibility of sapphire process automation was also examined and applied to the growth furnace.

2. Principles

2.1. Detection of Auto Seeding Point in Ingot Growth In single crystal sapphire growth methods that use a cylindrical crucible, such as Ky and Cz, a stream line generated from the edge to the center of the crucible on the surface of the melt when convection occurs (after the internal alumina is fully molten). This melt material flows because of the temperature gradient caused by the heater of the outer crucible. Generally, the flow rises on the surface of the crucible wall and moves to the center of the crucible before descending. As convection progresses, multiple stream lines enter through the center of the crucible, forming a spoke pattern [4,15]. The center of the spoke pattern where the melt flow descends is called the island. This pattern on the surface of the melt is also measured in sapphire growth methods that apply heat to a cylindrical crucible in axial symmetry [16–18]. Furthermore, this spoke pattern is measured in natural convection phenomena [19]. If the temperature field of the melt in the growth furnace is stabilized, then convection in the growth furnace is stabilized, and the island is located in the center of the crucible during growth as shown in Figure1. The seeding process in the Ky method succeeds only when the top-down melting movement is dominant at the seeding point on the melt surface. Upward melt flow at the seeding point cause the seed to melt [20]. Therefore, if the position of the island on the surface of the melt can be measured automatically, the time when the internal convection of the melt is stabilized and when the seeding process is initiated can be detected Appl. Sci. 2017, 7, 371 3 of 14 Appl. Sci. 2017, 7, 371 3 of 14

Figure 1. Spoke patterns on the surface of the melt according to temperature distribution in Figure 1. Spoke patterns on the surface of the melt according to temperature distribution in the crucible. the crucible. 2.2. Image Processing Algorithm 2.2. Image Processing Algorithm The sapphire growth process of the Ky method consists of the following stages: the heating stage, inThe which sapphire alumina growth is put into process the crucible of the Ky and method heated; cons the meltingists of the stage, following in which stages: the alumina the heating is molten; stage, in thewhich seed alumina preparation is put stage, into in the which crucible the temperatures and heated; ofthe melted melting alumina, stage, stream in which lines, the and alumina island is molten;position the are seed observed preparation to prepare stage, for in seeding;which the and te themperatures seed descending of melted stage. alum Seedina, cleaningstream lines, occurs and islandthrough position the first are contact observed between to prepare the seed for and seed theing; melt, and and the the seed second descending contact between stage. theSeed seed cleaning and occursthe meltthrough for sapphire the first growth. contact between the seed and the melt, and the second contact between the seed andThe the process melt for observed sapphire in growth. this study is the seed preparation stage. In this process, the worker generallyThe process monitors observed the surface in this of thestudy melt is in the the seed growth pr furnaceeparation through stage. a In view this point process, with thethe naked worker generallyeye. Convection monitors becomesthe surface activated of the frommelt in the the heater growth around furnace the crucible through after a vi theew internalpoint with alumina the naked is eye.fully Convection molten. At becomes this time, activated stream lines from begin the toheater form around in areas the where crucible the temperature after the internal is lower alumina than the is fullysurrounding molten. At areas this ontime, the stream surface oflines the begin melt, andto form a spoke in areas pattern where can be the observed temperature as a result. is lower When than thethe surrounding melt material areas temperature on the surface inside theof the growth melt, furnace and a isspoke stabilized, pattern the can convection be observed of the as alumina a result. Whenmelt the is alsomelt stabilized, material andtemperature the island, inside is located the grow at theth center furnace of the is growthstabilized, furnace the convection crucible. Then, of the aluminawhen themelt island is also size stabilized, grows to and 10–20 the mm, island, seeding is located begins. at Tothe detect center the of seedingthe growth point furnace accurately, crucible. a Then,technique when isthe required island for size automatic grows detection to 10–20 of mm, island seeding size and thebegins. time whenTo detect the center the ofseeding the island point coincides with the center of the crucible. accurately, a technique is required for automatic detection of island size and the time when the center To measure the melt surface inside the growth furnace using an image, a charge coupled device of the island coincides with the center of the crucible. (CCD) camera is installed at a view point on the top of the growth furnace. However, due to the To measure the melt surface inside the growth furnace using an image, a charge coupled device location and size of the view point, only a portion of the actual melt material surface can be measured (CCD)as shown camera in Figureis installed2. at a view point on the top of the growth furnace. However, due to the locationTherefore, and size of an the algorithm view point, was developed only a portion to measure of the theactual shape melt of material the spoke surface pattern can from be a measured limited as shownfield of vision,in Figure to detect2. stream lines using various data obtained from the surface of the melt, and to measureTherefore, the island, an algorithm or the intersection was developed point to of measure the stream the lines shape of the spoke pattern from a limited field of Thevision, image to detect processing stream procedure lines using is divided variousinto data the obtained image preprocessingfrom the surface stage, of the in whichmelt, and the to measurecollected the images island, are or binarized,the intersection and the point image of processing the stream stage lines in which the center point of the island isThe traced image through processing binarized procedure data, as shown is divided in Figure into3. the image preprocessing stage, in which the collected images are binarized, and the image processing stage in which the center point of the island is traced through binarized data, as shown in Figure 3. Image preprocessing is not only important for acquiring accurate stream-line data, but is also related to the accuracy of the measured data. The luminance plane is extracted to obtain information related to the brightness of the melt surface, which is obtained from the CCD camera images. The extracted image goes through a binarization process that distinguishes particles (stream lines) from non-particles (background). Global binarization, which is one of the image binarization methods that uses threshold values, provides the advantage of minimal computation. However, both the surface temperature distribution and brightness of the extracted image are irregular. It is difficult to obtain accurate stream-line data through global binarization alone. Hence, the Niblack binarization Appl. Sci. 2017, 7, 371 4 of 14

Appl.algorithm Sci. 2017, 7, is371 used to overcome this condition. Niblack binarization takes threshold values by4 of region 14 and then performs binarization. The Niblack binarization equation is shown below: algorithm is used to overcome this condition. Niblack binarization takes threshold values by region and then performs binarization. The Niblac, k binarization, equation ∙, is shown below: (1)

In this equation, m(x, y) and, s (x, y) denote, the average ∙, and standard deviation in a regional(1) window, and k denotes the user defined variable for binarization. Niblack binarization determines In this equation, m(x, y) and s(x, y) denote the average and standard deviation in a regional the critical value using the average and standard deviation of pixel values inside the regional window, window, and k denotes the user defined variable for binarization. Niblack binarization determines and the parts above the threshold value are extracted by comparing the critical value with the original the critical value using the average and standard deviation of pixel values inside the regional window, image; these parts are then binarized. and the parts above the threshold value are extracted by comparing the critical value with the original In this study, the deviation value of 0.2 was used in a window with 32 × 32 pixels. Because the image; these parts are then binarized. parts detected on the image are stream lines with low temperatures, dark objects were extracted. In this study, the deviation value of 0.2 was used in a window with 32 × 32 pixels. Because the After binarization, a mask method is used to delete data associated with the rod part, in which parts detected on the image are stream lines with low temperatures, dark objects were extracted. the seed is mounted, and the outer parts, which are not necessary for the measurement of stream After binarization, a mask method is used to delete data associated with the rod part, in which lines on the surface of the melt. The region of interest is extracted from the original image using the the seed is mounted, and the outer parts, which are not necessary for the measurement of stream Appl. Sci. 2017mask, 7 ,method 371 after binarization because the boundary line in the image may appear as data4 ofif 14 lines on the surface of the melt. The region of interest is extracted from the original image using the binarization is performed after extraction. mask method after binarization because the boundary line in the image may appear as data if binarization is performed after extraction.

Figure 2.FigureComparison 2. Comparison of the of spoke the spoke pattern pattern and and observed observed pattern pattern on on the thesurface surface of the the melt. melt. Figure 2. Comparison of the spoke pattern and observed pattern on the surface of the melt.

Figure 3. Image processing procedure to find the center point of the spoke pattern. FigureFigure 3. Image 3. Image processing processing procedure procedure toto findfind the the center center point point of ofthe the spoke spoke pattern. pattern.

Image preprocessing is not only important for acquiring accurate stream-line data, but is also related to the accuracy of the measured data. The luminance plane is extracted to obtain information related to the brightness of the melt surface, which is obtained from the CCD camera images. The extracted image goes through a binarization process that distinguishes particles (stream lines) from non-particles (background). Global binarization, which is one of the image binarization methods that uses threshold values, provides the advantage of minimal computation. However, both the surface temperature distribution and brightness of the extracted image are irregular. It is difficult to obtain accurate stream-line data through global binarization alone. Hence, the Niblack binarization algorithm is used to overcome this condition. Niblack binarization takes threshold values by region and then performs binarization. The Niblack binarization equation is shown below:

(x, y) = m(x, y) + k ·s(x, y) (1)

In this equation, m(x, y) and s(x, y) denote the average and standard deviation in a regional window, and k denotes the user defined variable for binarization. Niblack binarization determines the critical value using the average and standard deviation of pixel values inside the regional window, and the parts above the threshold value are extracted by comparing the critical value with the original image; these parts are then binarized. Appl. Sci. 2017, 7, 371 5 of 14

In this study, the deviation value of 0.2 was used in a window with 32 × 32 pixels. Because the parts detected on the image are stream lines with low temperatures, dark objects were extracted. After binarization, a mask method is used to delete data associated with the rod part, in which the seed is mounted, and the outer parts, which are not necessary for the measurement of stream lines on the surface of the melt. The region of interest is extracted from the original image using the mask methodAppl. Sci. after 2017 binarization, 7, 371 because the boundary line in the image may appear as data if binarization5 of 14 is performed after extraction. AfterAfter the the region region of of interest interest isis extracted,extracted, noisenoise elements, elements, excluding excluding the the stream stream lines lines on onthe the meltmelt surface, surface, are are removed removed using using aa low-passlow-pass filterfilter and pixel size. size. Figure Figure 4 4shows shows the the image image preprocessingpreprocessing process. process.

(a) (b)

(c) (d)

(e) (f)

Figure 4. Image preprocessing procedure: (a) acquiring image information; (b) extracting brightness Figure 4. Image preprocessing procedure: (a) acquiring image information; (b) extracting brightness information; (c) the binarization process; (d) extracting the region of interest; (e) removal of noise information;elements; (f ()c )result the binarizationof image preprocessing. process; (d ) extracting the region of interest; (e) removal of noise elements; (f) result of image preprocessing. The stream lines on the surface of the melt are distinguished through image preprocessing procedures.The stream In linesthe next on image the surface processing of thestage, melt the areisland distinguished position is traced through using image stream preprocessing lines. procedures.First, In through the next the image Hough processing transform stage,of the theacquir islanded pixel position data, isa straight traced using line representative stream lines. of each stream line is extracted. The Hough transform is illustrated in Figure 5. One point on the x-y coordinate system appears as one curve on the ρ-θ coordinate system, and one straight line on the x-y coordinate system is expressed as one point on the ρ-θ coordinate system. The region with the highest number of intersection points on the ρ-θ coordinate system is judged as a straight line in the x-y coordinate system, and a straight line in the image is detected in this way. Appl. Sci. 2017, 7, 371 6 of 14

First, through the Hough transform of the acquired pixel data, a straight line representative of each stream line is extracted. The Hough transform is illustrated in Figure5. One point on the x-y coordinate system appears as one curve on the $-θ coordinate system, and one straight line on the x-y coordinate system is expressed as one point on the $-θ coordinate system. The region with the highest number of intersection points on the $-θ coordinate system is judged as a straight line in the x-y coordinateAppl. Sci. 2017 system,, 7, 371 and a straight line in the image is detected in this way. 6 of 14

Figure 5. Principle of Hough transform. Figure 5. Principle of Hough transform. Figure 6 shows the image processing procedure for detecting the position of stream–line Figureintersection.6 shows Figure the 6a image shows processingan image that procedure underwent for the detecting preprocessing the positionprocedure of on stream–line the ρ-θ intersection.coordinate Figure system.6a Figure shows 6b an shows image a histogram that underwent that represents the preprocessing the number of intersections procedure onof each the $-θ coordinatecurve. system.The region Figure with6 ba showslarge number a histogram of intersections that represents of curves the on number the ρ-θ ofcoordinate intersections system of is each actually expressed as a straight line on the x-y coordinate system. Thus, representative peak values curve. The region with a large number of intersections of curves on the $-θ coordinate system is actually are selected from each intersection region and moved to the x-y coordinate system region to generate expressed as a straight line on the x-y coordinate system. Thus, representative peak values are selected straight lines as shown in Figure 6c. In particular, because too many straight lines are likely to appear, fromonly each points intersection that have region 100 or and more moved intersections to the x are-y coordinate transformed system into straight region lines, to generate as shown straight in lines asFigure shown 6c. Furthermore, in Figure6c. the In particular,intersection becausecount of curves too many on the straight ρ-θ coordinate lines are system likely is to assigned appear, to only pointseach that straight have 100 line oras morea weight intersections value. are transformed into straight lines, as shown in Figure6c. Furthermore,Figure the 6d intersection shows the stage count in of which curves the on coordinates the $-θ coordinate of the island system center is are assigned detected to using each straightthe line asstraight a weight line value. data from the x-y coordinate system acquired in the previous stage. To scan the entire Figureimage,6 ad circular shows region the stage of 20 in× 20 which pixels theis created coordinates on the x of-y coordinate. the island At center this time, are detectedthe intersections using the straightin the line scanned data from region the arex -detectedy coordinate and the system sum of acquired the weight in values the previous of the straight stage. lines To involved scan the in entire the creation of each detected intersection is calculated. However, if the same straight line is included image, a circular region of 20 × 20 pixels is created on the x-y coordinate. At this time, the intersections twice when summing the weights of straight lines, it is calculated only once. The region that has the in the scanned region are detected and the sum of the weight values of the straight lines involved in the largest sum of intersection weights in the inspected image is detected and the center of the creationintersections of each detected in that region intersection is calculated; is calculated. this is re However,presented as if thethe samecoordinate straight of the line island is included center in twice whenthe summing growth furnace the weights melt material. of straight Table lines, 1 shows it is calculatedthe parameter only of once.image Theprocessing region to that find has the thecenter largest sum ofpoint intersection of the spoke weights pattern. in the inspected image is detected and the center of the intersections in that region is calculated; this is represented as the coordinate of the island center in the growth furnace melt material. Table1 shows the parameterTable of 1. image Parameter processing of image toprocessing. find the center point of the spoke pattern. Image Porcessing Parameters Value TableColor 1. PlaneParameter Extraction of image processing. HSL (Luminance Plane) Local threshold : Kernel size 32 × 32 Image Porcessing(Niblack) Parameters: Deviation factor 0.20 Value Image : Filter size 3 × 3 Low-passColor filter Plane Extraction HSL (Luminance Plane) binarization : tolerance 50% : Kernel size 32 × 32 Local threshold (Niblack) : Iterations 1 : Deviation factor 0.20 Removal small objects : Pixel frame shape Square frame (3 pixel × 3 pixel) Image binarization : Filter size 3 × 3 Low-pass filter : Connectivity 4 (Horizon or vertically adjacent) Threshold : tolerance100 50% Hough transform : Iterations 20 × 20 (Circular) 1 Removal small objects : Pixel frame shape Square frame (3 pixel × 3 pixel) : Connectivity 4 (Horizon or vertically adjacent) Threshold 100 Hough transform 20 × 20 (Circular) Appl. Sci. 2017, 7, 371 7 of 14

Appl. Sci. 2017, 7, 371 7 of 14

(a) (b)

(c) (d)

(e)

Figure 6. Image processing algorithm for detecting the island center point: (a) Hough transform; Figure 6. Image processing algorithm for detecting the island center point: (a) Hough transform; (b) intersection histogram; (c) straight line generation on x-y; (d) detection of central point; (e) detection (b) intersection histogram; (c) straight line generation on x-y;(d) detection of central point; (e) detection of island center point from the original image. of island center point from the original image. 3. Experiment Method and System Composition 3. Experiment Method and System Composition In the experiment, 32 kg grade sapphire growth equipment (Insight 200, ASTEK, Jeonnam, Korea) wasIn used, the experiment, and the crucible 32 kg diameter grade sapphire is 200 mm. growth Figu equipmentre 7 shows the (Insight single 200, crystal ASTEK, growth Jeonnam, of sapphire Korea) wasusing used, the and Ky themethod, crucible as well diameter as a sapphire is 200 mm. ingot. Figure In this7 showsstudy, the Ky single method crystal was growth used to of measure sapphire usingthe surface the Ky of method, the melt as in well the as seed a sapphire preparation ingot. stage In this of sapphire study, the growth. Ky method The experimental was used to system measure thewas surface set up of as the shown melt in in Figure the seed 8. This preparation system consists stage of of sapphirethe following: growth. a vision The experimentalmodule unit, which system wasconsists set up of as a shownCCD camera in Figure to take8. This photographs system consists of the ofmelt the material following: surface a vision and modulea frame unit,grabber which to consistsconvert of the a CCD camera camera images to take into photographs processable of signals; the melt the materialshutter module surface unit, and which a frame has grabbera shutter to convertand a pneumatic the CCD camera actuator images for damage into processable prevention; signals; and the the motion shutter module module unit, unit, which which measures has a shutter the location of the seed and moves the seed up and down for seeding. Appl. Sci. 2017, 7, 371 8 of 14 Appl. Sci. 2017, 7, 371 8 of 14 The CCD camera used in this experiment (model UI-6230SE, IDS, Obersulm, Germany) has a The CCD camera used in this experiment (model UI-6230SE, IDS, Obersulm, Germany) has a resolution of 1024 × 768 pixels and an image capture speed of 40.0 fps. The amount of light that enters resolution of 1024 × 768 pixels and an image capture speed of 40.0 fps. The amount of light that enters the camera was reduced using a neutral density filter (ND filter) because intense light is generated the camera was reduced using a neutral density filter (ND filter) because intense light is generated inside the growth furnace at 2000 °C or higher after alumina melting and before temperature inside the growth furnace at 2000 °C or higher after alumina melting and before temperature stabilization. The collected image signals are processed through an algorithm in a controlled PC and Appl.stabilization. Sci. 2017, 7, 371The collected image signals are processed through an algorithm in a controlled PC8 of and 14 the results are analyzed to detect the central point. Shutter open/close control signals and rod the results are analyzed to detect the central point. Shutter open/close control signals and rod up-and-down movement commands are issued through the PC. up-and-down movement commands are issued through the PC. and aThe pneumatic experiment actuator progressed for damage as shown prevention; in Figure and 9. the The motion camera module shutter unit, open/close which measures cycle is 5 the s, The experiment progressed as shown in Figure 9. The camera shutter open/close cycle is 5 s, andlocation the total of the cycle seed takes and just moves one the minute, seed upduring and downwhich forthe seeding.surface images are collected and analyzed. and the total cycle takes just one minute, during which the surface images are collected and analyzed.

Figure 7. The single crystal growth of sapphire using the Ky method. FigureFigure 7. 7.The Thesingle singlecrystal crystalgrowth growth of of sapphire sapphire using using the the Ky Ky method. method.

Figure 8. Block diagram of the auto seeding system. Figure 8. Block diagram of the auto seeding system. Figure 8. Block diagram of the auto seeding system. The CCD camera used in this experiment (model UI-6230SE, IDS, Obersulm, Germany) has a resolution of 1024 × 768 pixels and an image capture speed of 40.0 fps. The amount of light that enters the camera was reduced using a neutral density filter (ND filter) because intense light is generated inside the growth furnace at 2000 ◦C or higher after alumina melting and before temperature stabilization. The collected image signals are processed through an algorithm in a controlled PC and the results are analyzed to detect the central point. Shutter open/close control signals and rod up-and-down movement commands are issued through the PC. The experiment progressed as shown in Figure9. The camera shutter open/close cycle is 5 s, and the total cycle takes just one minute, during which the surface images are collected and analyzed. Appl. Sci. 2017, 7, 371 9 of 14 Appl. Sci. 2017, 7, 371 9 of 14

Figure 9. Flowchart of the auto-seeding process. Figure 9. Flowchart of the auto-seeding process. 4. Results 4. Results During sapphire growth in the Ky method, images were acquired from the melt surface of the aluminaDuring material sapphire using growth the propos in theed Ky system, method, and images the coordinates were acquired of the from stream-line the melt center surface were of the aluminaprocessed material through using an image the proposed processing system, algorithm. and After the coordinates the melt material of the caused stream-line convection center and were processedbefore it through was stabilized, an image 10 processing images were algorithm. acquired After at 5 the min melt intervals. material Figure caused 10 convection shows four and beforerepresentative it was stabilized, images10 among images the were 10 images acquired that at re 5 minsulted intervals. from the Figure calculation 10 shows using four the representative algorithm. imagesTo determine among thethe 10reliability images of that the resultedimage processing from the results, calculation the is usingland positions the algorithm. (shown To by determine a green thetriangle) reliability were of thecompared image processingwith the central results, point the islandpositions positions of stream (shown lines by detected a green through triangle) the were compareddeveloped with algorithm the central (shown point by positionsa red circle) of streamin the figures. lines detected through the developed algorithm (shownThe by ablack red circle)regions in obtained the figures. from image preprocessing represent the data of stream-line areas, whichThe had black low regions temperatures obtained in fromthe original image image. preprocessing Noise elements represent are thegenerated data of in stream-line certain regions areas, which(see hadFigure low 11) temperatures from the stream-line, in the original depicted image. as black Noise dots. elements This is are one generated problem inassociated certain regionswith (seeimage Figure binarization. 11) from These the stream-line, noise elements depicted are generate as blackd from dots. the Thisresults is of one binarization problem using associated regional with threshold values. Most noise can be removed by a low-pass filter and by removing pixels under image binarization. These noise elements are generated from the results of binarization using regional a certain pixel value. However, large black regions are not removed, as shown in the Figure 11. When threshold values. Most noise can be removed by a low-pass filter and by removing pixels under a these noise elements are calculated together with other stream-line elements in the Hough transform, certain pixel value. However, large black regions are not removed, as shown in the Figure 11. When unintended straight lines may be generated, causing a problem in the calculation of intersection these noise elements are calculated together with other stream-line elements in the Hough transform, points. Yet such unremoved shades are not a serious concern when detecting the center point of the unintendedisland because straight their lines intersection may be generated, count is smaller causing than a problem that of the in theactual calculation stream-line of intersection elements. Also, points. Yetthey such do unremoved not appear as shades main arestraight not alines serious on the concern ρ-θ coordinate when detecting system. the center point of the island becauseWhen their the intersection observed countand detected is smaller islands than were that comp of theared, actual in stream-linemost cases, the elements. coordinates Also, of they two do notpoints appear were as mainon the straight main stream lines onlines, the but$-θ coordinacoordinatete errors system. occurred according to the shape of the stream-lineWhen the intersection observed andpositions. detected When islands stream were lines compared, intersected in in most the shape cases, of the a straight coordinates line, there of two pointswas wereabout ona 7-pixel the main (approximately stream lines, 0.7 but mm) coordinate error for errors the island occurred coordinates. according However, to the shapewhen ofthe the stream-linestream lines intersection intersected positions. in a vortex When shape, stream as sh linesown in intersected Figure 12, in a themaximu shapem oferror a straight of 13.93 line, pixels there was(approximately about a 7-pixel 1.4 (approximatelymm) occurred in 0.7the mm)distance error between for the the island actually coordinates. observed coordinates However, and when the the streamcoordinates lines intersected obtained from in a the vortex algorithm. shape, as shown in Figure 12, a maximum error of 13.93 pixels (approximately 1.4 mm) occurred in the distance between the actually observed coordinates and the coordinates obtained from the algorithm. Appl. Sci. 2017, 7, 371 10 of 14 Appl. Sci. 2017, 7, 371 10 of 14

(a)

(b)

(c)

(d)

Figure 10. Image processing results of the melt surface. (Red: island position, Green: position of image Figure 10. Image processing results of the melt surface. (Red: island position, Green: position of image processing result). processing result). Appl. Sci. 2017, 7, 371 11 of 14 Appl. Sci. 2017, 7, 371 11 of 14 Appl. Sci. 2017, 7, 371 11 of 14

Figure 11. Noise elements after image preprocessing. Figure 11. Noise elements after image preprocessing.

Figure 12. Error detection at the curved parts of stream lines. (Red: island position, green: position of Figureimage processing12. Error detection result). at the curved parts of stream lines. (Red: island position, green: position of Figure 12. Error detection at the curved parts of stream lines. (Red: island position, green: position of image processing result). image processing result). Ten images were acquired from the time the surface began to melt inside the growth furnace. The Tenpositions images for were the acquiredcenter point from of the the time island the assurf observedace began and to meltas detected inside thethrough growth the furnace. image Ten images were acquired from the time the surface began to melt inside the growth furnace. Theprocessing positions algorithm for the are center shown point in Figure of the 13. island Table as 2 showsobserved the andcoordinates as detected of the through island as the observed image The positions for the center point of the island as observed and as detected through the image processingand as detected algorithm from are the shown algorithm, in Figure as 13.well Table as the 2 shows distance the coordinatesbetween the of coordinatesthe island as inobserved pixels. processing algorithm are shown in Figure 13. Table2 shows the coordinates of the island as observed andThe differenceas detected ranged from fromthe algorithm, 6.32 pixels as (at well the minimum)as the distance to 34 pixelsbetween (maximum), the coordinates and the in average pixels. and as detected from the algorithm, as well as the distance between the coordinates in pixels. Thedifference difference was ranged18.76 pixels, from 6.32whic pixelsh is about (at the 1.9 minimum) mm in actual to 34 dipixelsstance. (maximum), One explanation and the for average these The difference ranged from 6.32 pixels (at the minimum) to 34 pixels (maximum), and the average differencedifferences was between 18.76 the pixels, two coordinatewhich is about values 1.9 is mmbecause in actual the stream-line distance. informationOne explanation at the forcenter these of difference was 18.76 pixels, which is about 1.9 mm in actual distance. One explanation for these differencesthe image was between lost: about the two 25% coordinate of the region values visible is because through the the stream-line view point information is occupied at bythe the center shape of differences between the two coordinate values is because the stream-line information at the center of theof the image rod andwas seed,lost: about as shown 25% inof Figure the region 14, resulting visible through in lower the accuracy. view point is occupied by the shape the image was lost: about 25% of the region visible through the view point is occupied by the shape of of theIn rod general, and seed, the asinternal shown convection in Figure 14,of resultingthe melt materialin lower accuracy.was stabilized after about 20 h in the the rod and seed, as shown in Figure 14, resulting in lower accuracy. heatingIn general,process, theand internal the center convection position of of the the melt island material in the wasmelt stabilizedmaterial isafter within about 10 20mm h inof the heatingcenter of process, the crucible and thein thecenter growth position furnace. of the Af islater ndthat, in whenthe melt the materialisland size is withingrows 10to 10–20mm of mm, the centerthe seed of is the lowered. crucible Figure in the 15 growth displays furnace. the distance After betweenthat, when the the center island of the size island grows observed to 10–20 by mm, the thenaked seed eye is lowered.and the centerFigure of15 thedisplays island the detected distance from between the algorithmthe center takenof the islandfrom images observed acquired by the nakedduring eyesapphire and the growth center at theof thecenter island position detected (750, from 600) ofthe the algorithm growth furnace taken fromcrucible. images All the acquired island duringcenters sapphirewere within growth 10 mm at the of thecenter center position of the (750, growth 600) furnace of the growth crucible. furnace They crucible.approached All the islandcenter centersof the crucible were within over time10 mm and of continuously the center of movedthe growth at distances furnace ofcrucible. 2–3 mm. They approached the center of the crucible over time and continuously moved at distances of 2–3 mm. Appl. Sci. 2017, 7, 371 12 of 14 Appl. Sci. 2017, 7, 371 12 of 14

Appl. Sci. 2017, 7, 371 12 of 14

(a) (b) (a) (b) Figure 13. Changes of island positions at the center of the growth furnace crucible: (a) observed Figure 13. Changes of island positions at the center of the growth furnace crucible: (a) observed Figurepositions; 13. Changes (b) the positions of island resulting positions from at image the center processing. of the growth furnace crucible: (a) observed positions;positions; (b) the (b positions) the positions resulting resulting from from image image processing.processing. Table 2. Comparison of coordinates between observed and measured positions (i.e., the results of Table 2. Comparison of coordinates between observed and measured positions (i.e., the results of Tableimage 2.imageComparison processing). processing). of coordinates between observed and measured positions (i.e., the results of image processing). DistanceDistance between between the the ObservedObserved MeasuredMeasured Positions No.No. Coordinates/PositionsCoordinates/Positions Positions (ImageMeasured Processing Positions Results) Distance between the Positions (Image Processing Results) (in Pixels) No. Observed Positions (Image Processing Results) Coordinates/Positions(in Pixels) (in Pixels) 1 1 (647, (647, 555) 555) (653, 553) 6.32 6.32 1 (647, 555) (653, 553) 6.32 22 2 (663, (663, (663, 576) 576) 576) (679, (679, 606) 606) 34.00 34.00 34.00 33 3 (680, (680, (680, 585) 585) 585) (701, (701, 606) 606) 29.70 29.70 29.70 44 4 (694, (694, (694, 569) 569) 569) (702, (702, 576) 576) 10.63 10.63 10.63 55 5 (713, (713, (713, 583) 583) 583) (698, (698, 556) 556) 30.89 30.89 30.89 6 (717, 563) (717, 556) 7.00 6 6 (717, (717, 563) 563) (717, 556) 7.00 7.00 77 (733, (733, 571) 571) (728, (728, 558) 558) 25.94 25.94 87 (738, (733, 573) 571) (731, (728, 573) 558) 25.94 7.00 8 (738, 573) (731, 573) 7.00 98 (742, (738, 591) 573) (745, (731, 613) 573) 7.00 22.20 9 (742, 591) (745, 613) 22.20 109 (728, (742, 611) 591) (741, (745, 616) 613) 22.20 13.93 10 (728, 611) (741, 616) 13.93 10 (728, 611) (741, 616) 13.93

Figure 14. Loss of stream-line data caused by the outline of the rod and seed. Figure 14. Loss of stream-line data caused by the outline of the rod and seed. Figure 14. Loss of stream-line data caused by the outline of the rod and seed.

In general, the internal convection of the melt material was stabilized after about 20 h in the heating process, and the center position of the island in the melt material is within 10 mm of the center of the crucible in the growth furnace. After that, when the island size grows to 10–20 mm, the seed is lowered. Figure 15 displays the distance between the center of the island observed by the naked eye and the center of the island detected from the algorithm taken from images acquired during sapphire growth at the center position (750, 600) of the growth furnace crucible. All the island centers were within 10 mm of the center of the growth furnace crucible. They approached the center of the crucible over time and continuously moved at distances of 2–3 mm. Appl. Sci. 2017, 7, 371 13 of 14 Appl. Sci. 2017, 7, 371 13 of 14

Figure 15. Distance from the center of the crucible in the growth furnace to the island center point. Figure 15. Distance from the center of the crucible in the growth furnace to the island center point. 5. Conclusions 5. ConclusionsIn this study, an alumina melt material surface measuring system was constructed that utilized Ina this vision study, system an aluminafor auto seeding melt material in a single surface crystal measuring sapphire ingot system growth was furnace constructed using the that Ky utilized method. Furthermore, the island position of the melt material was detected using the developed a visionimage system processing for auto algorithm, seeding and in its aperformance single crystal was evaluated. sapphire ingot growth furnace using the Ky method. Furthermore,To periodically the acquire island images position of the of alumina the melt melt material material was surface detected during using sapphire the growth, developed an image processingexperimental algorithm, system and was its performanceconstructed with was a visi evaluated.on module unit that contained a CCD camera, a Toshutter periodically module unit acquire to protect images the camera, of the a alumina motion module meltmaterial unit for the surface seeding duringprocess, sapphireand the PC growth, an experimentalcontrol unit system for total wassystem constructed control. The withimages a visionwere stably module acquired unit through that contained this system. a CCD camera, a The stream lines and background regions on the alumina melt material surface were detected shutter module unit to protect the camera, a motion module unit for the seeding process, and the PC separately using an image-processing algorithm, and the island center coordinates were calculated controlusing unit forthe intersections total system of control. the stream The lines. images were stably acquired through this system. The streamWhen linesthe observed and background coordinates regionsof the island on the center alumina were meltcompared material with surfacethe coordinates were detected separatelydetected using from an the image-processing image processing algorithm, algorithm, ther ande was the an islandaverage center error rate coordinates of 1.87 mm (based were calculatedon using the1024 intersections × 768 pixels). ofThe the reason stream for this lines. was that the accuracy level dropped for observed coordinates Whencompared the observedto algorithmically coordinates detected of coordinates the island when center some werestream compared lines intersected with in the a vortex coordinates shape and when the detection of stream lines was limited due to the outline of rod and seed shapes detected from the image processing algorithm, there was an average error rate of 1.87 mm (based on in the visual field. However, since the maximum error was about 3 mm when converted to actual 1024 × distance,768 pixels this). error The did reason not cause for this a significant was that proble the accuracym in judging level the droppedstable condition for observed of the internal coordinates comparedconvection to algorithmically of the melt material detected during coordinates the actual growth when furnace some process. stream lines intersected in a vortex shape and whenConsidering the detection the above ofresults, stream the linesvision was system limited and algorithm due to the could outline be applied of rod to andthe heater seed shapes in the visualtype growth field. method However, of axial since symme thetry maximum in a circular error cylindrical was crucible about 3 using mm the when Cz method converted as well to actual as the Ky method. When applied to an actual system, these results can provide objective indicators distance, this error did not cause a significant problem in judging the stable condition of the internal for the seeding preparation process during sapphire growth. If the problems discovered in this study convectionare addressed, of the melt such material as the limited during field the of actualview an growthd the error furnace occurring process. in the curved parts of stream Consideringlines, automated the abovetechniques results, for the the seeding vision process system as and wellalgorithm as the entire could process be of applied single crystal to the heater type growthsapphire method growth of will axial be possible. symmetry in a circular cylindrical crucible using the Cz method as well as the KyAuthor method. Contributions: When applied C.M.K. and to anS.R.K. actual conceived system, and thesedesigned results the experiments; can provide C.M.K. objective performed indicators the for the seedingexperiments; preparation C.M.K., S.R.K. process and duringJ.H.A. analyzed sapphire the data; growth. C.M.K. Ifwrote the the problems paper. discovered in this study are addressed,Conflicts such of as Interest: the limited The authors field declare of view no conflict and theof interest. error occurring in the curved parts of stream lines, automated techniques for the seeding process as well as the entire process of single crystal sapphire growth will be possible.

Author Contributions: C.M.K. and S.R.K. conceived and designed the experiments; C.M.K. performed the experiments; C.M.K., S.R.K. and J.H.A. analyzed the data; C.M.K. wrote the paper. Conflicts of Interest: The authors declare no conflict of interest. Appl. Sci. 2017, 7, 371 14 of 14

References

1. Tang, H.; Li, H.; Xu, J. Growth and Development of Sapphire Crystal for LED Applications; InTech: Rijeka, Croatia, 2013. 2. Golubovic, A.; Nikolic, S.; Djuric, S.; Valcic, A. The growth of sapphire single crystals. J. Serbian Chem. Soc. 2001, 66, 411–418. 3. Demina, S.E.; Bystrova, E.N.; Lukanina, M.A.; Mamedov, V.M.; Yuferev, V.S.; Eskov, E.V.; Nikolenko, M.V.; Postolov, V.S.; Kalaev, V.V. Numerical analysis of sapphire crystal growth by the Kyropoulos technique. Opt. Mater. 2007, 30, 62–65. [CrossRef] 4. Timofeev, V.V.; Kalaev, V.V.; Ivanov, V.G. 3D melt convection in sapphire crystal growth Evaluation of physical properties. Int. J. Heat Mass Transf. 2015, 87, 42–48. [CrossRef] 5. Timofeev, V.V.; Kalaev, V.V.; Ivanov, V.G. Effect of heating conditions on flow patterns during the seeding stage of Kyropoulos sapphire crystal growth. J. Cryst. Growth 2016, 445, 47–52. [CrossRef] 6. Chen, C.H.; Chen, J.C.; Lu, C.W.; Liu, C.M. Effect of power arrangement on the crystal shape during the Kyropoulos sapphire crystal growth process. J. Cryst. Growth 2012, 352, 9–15. [CrossRef] 7. Nouri, A.; Delannoy, Y.; Chichignoud, G.; Lhomond, L.; Helifa, B.; Lefkeir, I.K.; Zaidat, K. Numerical investigation of an experimental Kyropoulos process to grow silicon ingots for photovoltaic application. J. Cryst. Growth 2017, 460, 48–58. [CrossRef] 8. Kozik, V.I.; Nezhevenko, E.S. Measuring silicon monocrystal diameters during melt growth. Optoelectron. Instrum. Data Process. 2008, 44, 392–401. [CrossRef] 9. Xiang, S.; Pan, F.; Xiang, K.; Wang, X. Melt level measurement for the CZ crystal using an improved laser triangulation system. J. Meas. 2017, 103, 27–35. [CrossRef] 10. Winkler, J.; Robenack, K. Observer based determination of the crystal diameter in Czochralski crystal growth utilizing algorithmic differentiation. In Proceedings of the Systems, Signals and Devices (SSD), 2012 9th International Multi-Conference, Chemnitz, Germany, 20–23 March 2012; pp. 1–6. 11. Winkler, J.; Neubert, M.; Rudolph, J. A review of the automation of the Czochralski crystal growth process. Acta Phys. Pol. 2013, 124, 181–192. [CrossRef] 12. Xu, C.H.; Zhang, M.F.; Meng, S.H.; Han, J.C.; Wang, G.G.; Zuo, H.B. Temperature field design, process analysis and control of SAPMAC method for the growth of large size sapphire crystals. Cryst. Res. Technol. 2007, 42, 751–757. [CrossRef] 13. Wang, G.; Zuo, H.; Zhang, H.; Wu, Q.; Zhang, M.; He, X.; Hu, Z.; Zhu, L. Preparation, quality characterization, service performance evaluation and its modification of sapphire crystal for optical window and dome application. Mater. Des. 2010, 31, 706–711. [CrossRef] 14. Chenghai, X.; Songhe, M.; Mingfu, Z.; Hongbo, Z.; Guigen, W. Thermal Stresses and Cracks During the Growth of Large-sized Sapphire with SAPMAC Method. Chin. J. Aeronaut. 2007, 20, 475–480. [CrossRef] 15. Demina, S.E.; Kalaev, V.V. 3D unsteady computer modeling of industrial scale Ky and Cz sapphire crystal growth. J. Cryst. Growth 2011, 320, 23–27. [CrossRef] 16. Jing, C.J.; Imaishi, N.; Yasuhiro, S.; Miyazawa, Y. Three-dimensional numerical simulation of oxide melt flow in Czochralski configuration. J. Cryst. Growth 2016, 216, 372–388. [CrossRef] 17. Jing, C.J.; Imaishi, N.; Yasuhiro, S.; Miyazawa, Y. Three-dimensional numerical simulation of spoke pattern in oxide melt. J. Cryst. Growth 1999, 200, 204–212. [CrossRef] 18. Kobayashi, M.; Tsukada, T.; Hozawa, M. Effect of internal radiative heat transfer on the convection in Cz oxide melt. J. Cryst. Growth 1997, 180, 157–166. [CrossRef] 19. Miller, D.C.; Pernell, T.L. Fluid flow patterns in a simulated garnet melt. J. Cryst. Growth 1982, 58, 253–260. [CrossRef] 20. Numerical Analysis and Optimization of Sapphire Crystal Growth by the Kyropoulos Technique. Available online: http://str-soft.com/products/CGSim/Kyropoulos_Sapphire/index.htm (accessed on 16 March 2017).

© 2017 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).