STR/03/027/MT 1 Polishing of Fibre Optic Connectors a to z fibre optic services in uk london by the most perfectionist company azfo – This study reports the development of high efficiency polishing protocols of fibre connectors, by replacing the presents three step polishing with two step polishing protocols, and by using various abrasives as polishing mediums, to achieve necessary geometrical and optical qualities for physical contact transmission using a commercially available connector polisher. The study focuses on surface integrity of the fibre and ferrule as well as other geometry of the polished connector end faces covering the fibre height or undercut, the apex offset, and the radius of curvature. Optical performances of the polished connectors were evaluated by return and insertion losses. Finally, relationship between geometrical qualities and optical performances was established. This study shows that the efficiency for polishing fibre connectors can be raised by at least 30% by selection of suitable abrasives and grit sizes. Keywords: Fibre optic connectors, Polishing, Surface roughness, Fibre height/undercut, Apex offsets, Radius of curvature, Return loss, Insertion loss 1 INTRODUCTION In the applications of fibre optic technology, the fibre connectors have been playing a substantial role to bring together and hold the two cores of the glass fibre ends for communications. Among the many types of connectors, the physical contact (PC) type has emerged as the most popular connector, capturing the vast majority of instrumentation applications [1]. The physical contact connectors are designed to have a spherical end face with the fibre at the highest point. On connection, this allows the fibres to come into intimate optical contact and compress slightly for the optimum system performance over time, temperature, and vibration. The quality of the end face geometry determines the fibre-fibre interface where a perfect contact without air gap is expected over a long term or when intermated with another connector. Therefore, it is critical for finishing of a connector end face to obtain not only surface integrity for the fibre and ferrule and a reasonable radius of curvature of the end face, but also a minimum fibre height or undercut over the ferrule and a small apex offset between the fibre and ferrule. Ideally, a polished connector is expected to achieve a maximum light transmission with a minimum insertion loss. Practically, it is required to achieve of an insertion loss of < 0.3 dB and a return loss of < -45 dB for the physical contact connectors [1]. Currently, mechanical polishing is used to produce large volumes of connectors meeting or exceeding good optical and geometrical qualities as well as keeping high level of consistency from batch to batch. The polishing process often involves rough polishing of the cleaved and epoxy removed fibre connectors, intermediate polishing, and final polishing, taking at least 1.5 ~ 3 minutes for finishing such a process circle [2, 3]. This is expensive in consideration of the vast needs of connectors in fibre communications. Therefore, it is essential to develop the high efficiency polishing processes for connectors. Since fibre connector polishing involved abrasive machining of two brittle materials, the glass fibre and the ceramic ferrule simultaneously, it is likely to have different material responses to polishing in respects to surface roughness, material removal and form accuracy [4]. Furthermore, the lack of understanding of the relationship between the geometrical and optical qualities for connectors and the internationally standardized technical requirements for assessment of these qualities [5-6] has led to confusion in manufacturing fibre connectors. 2 OBJECTIVE The primary objective of this study is to develop high efficiency polishing protocols of fibre connectors to achieve necessary geometrical and optical qualities for physical contact transmission, using a commercially available connector polisher. Considering the abrasive machininginduced damage in machining brittle materials, the present study focuses on surface integrity of the fibre and ferrule as well as other geometry of the polished connector end faces covering the fibre height or undercut, the apex offset, and the radius of curvature. For optical performance, the polished connectors are investigated in terms of the return and insertion losses with an optic loss tester. Finally, we attempt to establish the relationship between the geometrical quality and the optical performance. Polishing of Fibre Optic Connectors 2 3 METHODOLOGY 3.1 Physical contact fibre ends The samples used in this investigation were the commercially available, unpolished single mode fibre connectors. The 125 µm diameter glass fibre was centered in the yettria partially stabilized zirconia ferrule and exited at its endface. The 2.5 mm diameter zirconia ferrule was keyed in a mechanical assembly to hold the cable rigid and aligned. The end faces were pre-convex, spherically curved with the radius of curvature of about 20 mm. The fibres were cleaved and the epoxy was removed. 3.2 Apparatus for fibre connector polishing The polishing experiments were conducted using a commercially available polishing machine (8671X-6100 Series, Molex), as shown in Fig. 1. The polishing machine had a timer ranging from 0 to 60 s, a constant pressure for polishing, and a universal connector holder for 12 connectors. An air cushion metal pad with the 12 circle orbits, and a flat metal pad and a flat rubber pad were applied for polishing. The machine provided an orbital polishing motion, which was a circular orbital oscillation. During polishing, each connector rotated independently along its circular orbit of a 17.5 mm diameter at a speed of 15 rpm; meanwhile it spun itself at a speed of 285 rpm. Disposable, self-adhesive polishing films of SiC with grit sizes of 3 µm and 5 µm, alumina with grits of 0.05 µm and 0.5 µm, and diamond with grits of 0.1 µm and 0.5 µm were selected. Typical sizes and distributions of the polishing films are shown in Fig. 2. The larger sizes of SiC grits are clearly identified in Fig. 2. The submicro alumina and diamond abrasives are shown in grit clusters in Fig. 2. Fig. 1. The polishing machine. 5 µm SiC 3 µm SiC 0.5 µm Alumina 0.5 µm Diamond 0.1 µm Diamond 0.05 µm Alumina Fig. 2. SEM micrographs of the polishing films. 3.3 Polishing procedures Usually, polishing of the single mode PC connectors involves three steps including rough polishing, intermediate polishing and final polishing for 30 s in each step, or 90 s in total for a polishing circle [2]. To reduce the polishing cycle time and to increase the polishing efficiency, in this investigation a two-step processes were proposed and conducted for a total circle time of 60 s, maintaining 30 s for each step. Before each polishing step, a pure grade of isopropyl alcohol was applied on the top of the polishing film as a lubricant. In each process, six randomly loaded connectors were polished to evaluate the repeatability of the results. The detailed polishing procedures for the industryapplied three-step processes and the two-step processes proposed are listed in Table 1. Before and after each step, the fibre connector end faces were carefully cleaned with alcohol and optical tissues for further polishing or analyses. 3.4 Characterization methods Before and after each polishing step, the connector end faces were evaluated using a fibre connector microscope and an optical interference profiler (WYKO 3300, Veeco) to investigate the polishing scratches, the surface roughness for the fibre and ferrule, and the relative fibre height or undercut. The fibre and ferrule Polishing of Fibre Optic Connectors 3 Table 1. The detailed description of the polishing processes Process Step 1 Step 2 Step 3 A 3 µm SiC film Metal air cushion pad Alcohol 30 second polishing 0.5 µm alumina film Metal flat pad Alcohol 30 second polishing 0.05 µm alumina film Rubber flat pad Alcohol 30 second polishing B 3 µm SiC film Metal air cushion pad Alcohol 30 second polishing 0.1 µm diamond film Metal flat pad Alcohol 30 second polishing 0.05 µm alumina film Rubber flat pad Alcohol 30 second polishing C 5 µm SiC film Metal air cushion pad Alcohol 30 second polishing 0.1 µm diamond film Metal flat pad Alcohol 30 second polishing D 5 µm SiC film Metal air cushion pad Alcohol 30 second polishing 0.1 µm diamond film Rubber flat pad Alcohol 30 second polishing E 3 µm SiC film Metal air cushion pad Alcohol 0.5 µm diamond film Metal flat pad Alcohol surface roughness in terms of the arithmetic average roughness (Ra) was measured using a vertical scanning interferometry (VSI) in a measured area of 185 × 243 µm2 covering the fibre region of the end face. The relative fibre height or undercut is evaluated using the vertical distance of the highest point of a fibre end face to the epoxy between the fibre and ferrule. The mean values and the standard deviations of these measurements were determined from the six connector end faces studied in each step polishing. The Wyko optical interferometer lacks the functions for fibre array and precision alignment and for multiple parameter measurements acquired in geometrical quality assessment. Therefore, the final geometry quality for the polished connector end faces was further and thoroughly evaluated using an automated non-contact interferometer system for array-type fibre optic connectors (AC-3005, Norland). The geometrical parameters included the fibre and ferrule surface roughness, the fibre height or undercut, the apex offset, and the radius curvature. The fibre height or undercut is defined as the distance between the fibre end face and the best fit of the spherical surface of the average ferrule end face [7]. When a fibre is recessed inside a ferrule, it is named a fibre undercut. When a fibre protrudes above the ferule it is called a fibre height. The radius of curvature is defined as the radius of the best fit of the spherical ferrule end face [7]. The apex offset is the linear distance from the center of the fibre to the apex or highest point on the best fit of the spherical end face [7]. This can also be expressed as angle between these two points, namely the angular offset [7]. The former is designated as the linear apex offset [7]. The optical performance in term of the return and insertion losses was measured using a loss test set (LTS-3900, EXFO), which combines a stable optical source, an optical power meter and an optical return loss meter. The return loss designates the total fractional power that is reflected from a test unit, which is defined as Return Loss = -10 log10 (Pin / Pback), where Pin is the input power and Pback the reflected power. The insertion loss is the amount of optical power lost at the interface of two connectors, which can be written as Insertion Loss = 10 log10 (Pi /P0), where Pi is the initial power and P0 the power after the connector is applied. Polishing of Fibre Optic Connectors 4 4 RESULTS 4.1 Effects of polishing processes on surface integrity of connector end faces The effects of polishing on the fibre surface roughness Ra are summarized in Fig. 3. After the first 30 s of polishing, the surface roughness values for the cleaved fibres of about 125 ~ 350 nm Ra dropped to about 50 ~ 150 nm Ra. It is noticed that Processes A, B and E using 3 µm SiC films generated the better fibre surfaces than those using 5 µm SiC in the first 30 s polishing. After the second 30 s polishing, the fibre surface roughness values were reduced to less than 50 nm Ra in all the processes. The coarser fibre surfaces were generated in Process A with the average value of 50 nm Ra; the finer were made in Processes B, C, D, and E with the values of 15 ~ 30 nm Ra. For the third 30 second polishing involved only in Processes A & B, the roughness did make a significant progress from 50 nm Ra to about 15 nm Ra in the former; however, it improved very slightly obtaining about 15 nm Ra in the latter. The results indicate that the two-step polishing in Processes C, D, and E is capable of achieving as a good fibre surface finish as that obtained in the three-step polishing processes A and B. It is therefore possible to use the two-step polishing to replace the threestep processes for improvement in the fibre surface roughness, shortening the polishing circle time from 90 s to 60 s. The effects of polishing on the ferrule surface roughness Ra values are plotted in Fig. 4. The initial ferule surface roughness obtained in the cleaving and epoxy removal process can be very rough with the maximum roughness value of about 1900 nm Ra. After the first 30 s polishing, the ferrule surface roughness values were tremendously diminished to < 75 nm Ra for all the processes. After the second 30 s polishing, the ferrule roughness values dropped to < 20 µm Ra in Processes C, D and E, and to about 30 nm Ra in Processes A and B. After the third 30 s polishing in Processes A and B, it is observed that the ferrule surface roughness slimly improved to < 25 nm Ra. It is thus concluded that the two-step polishing of Processes C, D and E is also more effective than the three-step polishing of Processes A and B for finishing the ferrule surfaces. The effects of polishing on the relative fibre heights or undercuts is shown in Fig. 5. The initial relative fibre heights after cleaving and epoxy removing could be as high as about 12 µm. After the first 30 s polishing, the relative fibre heights were reduced below 2 µm in all the processes. In the second step polishing, the relative fibre heights in Processes C, D and E dropped rapidly to about 0.05 µm; however, in Processes A and B, they just reduced to 0.2 µm and 0.15 µm respectively. Finally, after 90 s polishing in Processes A and B, the relative fibre heights diminished to about 0.05 µm. The threestep polishing processes A and B appear to have disadvantages in reducing the fibre heights. Fig. 6 shows the optical views of the fibre end face taken before and after each polishing step in Process C. Fig. 6(a) is a typical image for a cleaved and epoxy-removed fibre connector end face. Fig. 6(b) is the image taken after 30 s polishing, in which many polishing marks and scratches can be observed. Fig. 6(c) is image taken after the second-step polishing, where no visible scratches and damage can be seen. Fig. 7 shows the two-dimensional and threedimensional optical interference images corresponding to the optical images in Figs. 6(a), 6(b) and 6(c), in Process C. Fig. 7(a) shows the estimation of the relative fibre height above the epoxy of 5.8 µm for the cleaved and epoxyremoved connector end face. Fig. 7(b) exhibits the estimation of the relative fibre height of about 2 µm and the polishing scratches after 30 s polishing. Fig. 7(c) shows the estimation of the relative fibre height of about 50 nm and the damage free end face after the second-step polishing. Polishing Time (s) -30 0 30 60 90 120 Fiber Roughness in Ra (nm) 0 50 100 150 200 250 300 350 400 Process A Process B Process C Process D Process E Polishing Time (s) 30 60 90 120 Fiber Roughness in Ra (nm) 0 25 50 75 100 125 150 175 200 Process A Process B Process C Process D Process E Fig. 3. Influence of polishing on fibre surface roughness. Polishing of Fibre Optic Connectors 5 Polishing Time (s) -30 0 30 60 90 120 Ferrule Roughness in Ra (nm) 0 250 500 750 1000 1250 1500 1750 2000 Process A Process B Process C Process D Process E Polishing Time (s) 30 60 90 120 Ferrule Roughness in Ra (nm) 0 25 50 75 100 Process A Process B Process C Process D Process E Fig. 4. Influence of polishing on ferrule surface roughness. Polishing Time (s) -30 0 30 60 90 120 Relative Fiber Height (µm) 0 2 4 6 8 10 12 14 Process A Process B Process C Process D Process E Polishing Time (s) 60 90 120 Relative Fiber Height (µm) 0.00 0.05 0.10 0.15 0.20 0.25 Process A Process B Process C Process D Process E Fig. 5. Influence of polishing on the relative fibre heights. 50 µm 50 µm 50 µm (a) (b) (c) Fig. 6. Optical images of a connector end face polished in Process C. (a) Before polishing, (b) After the first-step polishing, and (c) After the second-step polishing. Polishing of Fibre Optic Connectors 6 Fig. 7. Corresponding optical interference images of the connector end face shown in Fig. 6, demonstrating the two- and three-dimensional views, and the relative fibre height assessment. (a) Before polishing. Fig. 7(b). The first-step polished connector end face. Fig. 7(c). The second-step polished connector end face. 4.2 Geometrical quality assessment for polished fibre connectors The fibre and ferrule surface roughness Ra values of the polished fibre connectors were also assessed using an automated non-contact interferometer system for array-type fibre optic connectors. The fibre surface roughness values were smaller than 20 nm Ra in Processes A, B, C, and E and smaller than 30 nm in Process D. The ferrule surface roughness values are smaller than 20 nm Ra in either the two-step or three-step polishing processes. It indicates that suitable selection of polishing films and grit sizes enables the achievement of a good surface finish for both fibres and ferrules and the reduction of polishing time. The fibre heights/undercuts for the polished fibre connectors measured using the automated noncontact interferometer system are summarized in Fig. 8. In Processes A, B, and C, the fibres are protruded above the ferrules with the fibre heights of smaller than 75 nm. In Processes D and E, the fibres are either recessed below the ferrules or protruded above the ferules within a range of ±25 nm for either case. In terms of the fibre heights or undercuts, it is also comparable for the two-step polishing in Processes C, D and Polishing of Fibre Optic Connectors 7 E with the three-step polishing in Processes A and B. Fiber Height/Undercut (nm) -100 -50 0 50 100 A B C D E Process Fig. 8. Fibre heights/undercuts of the polished connectors measured with an automatic non-contact interferometer system. The radius of curvature of the polished fibre connectors measured using the automated noncontact interferometer system is shown in Fig. 9. The ranges of the radius of curvature generated in Processes A, B, and C are about 20 ~ 75 mm. In Process D, the radius of curvature is about 20 mm with a very small variation. However, in Process E, the radii of curvature oddly ranged from about 50 ~ 250 mm, which were about 4 times larger than those obtained in the threestep polishing in Processes A and B. These indicate that the two-step polishing in Processes C and D, excepting Process E, are comparable in regards to the radii of curvature of the polished connector end faces. Process Radius of Curvature (mm) 0 50 100 150 200 250 A B C D E Fig. 9. Radii of curvature of the polished connectors measured with an automatic non-contact interferometer system The linear and angular apex offset values for the polished connectors measured using the automated non-contact interferometer system are plotted in Figs. 10(a) and 10(b), respectively. The linear apex offsets of smaller than 150 µm and the angular apex offsets of smaller 0.2 degree are generated in Processes A, B, and C. In Process D, the linear and angular offsets are smallest, 30 µm and 0.15 degree, respectively. In Process E, however, both the linear and angular apex offsets increased significantly to 600 µm and 0.45 degree, respectively. These indicate that the different two-step polishing processes could generate either better or worse apex offsets than those obtained in the threestep polishing processes. Process Linear Apex Offset (µm) 0 100 200 300 400 500 600 700 A B C D E Process Angular Apex Offset (degree) 0.0 0.1 0.2 0.3 0.4 0.5 A B C D E (a) (b) Fig. 10. (a) Linear and (b) angular apex offsets of the polished connectors assessed with an automatic noncontact interferometer system. 4.3 Optical quality assessment for polished connectors The corresponding return losses for the polished connectors are given in Fig. 11(a). For the connectors polished in Processes A, B, C and D, the return loss values are all smaller than the critical value of –45 dB. In Process E, however, the return loss values are larger than –45 dB and even the two are greater than -20 dB, indicating that a high apex offset could cause more return losses for the polished connectors. The corresponding insertion losses for the polished connectors are plotted in Fig. 11(b), suggesting Polishing of Fibre Optic Connectors 8 that the insertion losses achieved using in all the polishing processes are smaller than the required value of 0.3 dB. Process Insertion Loss (dB) 0.0 0.3 0.6 A B C D E (a) (b) Processes Backreflection (dB) -60 -45 -30 -15 0 A B C D E Fig. 11. Optical performance of the polished connectors. (a) Return loss and (b) Insertion loss. 5 CONCLUSIONS The following conclusions are drawn from the present investigation on the polishing processes and the assessment of geometrical quality and optical performance of the polished fibre optic connectors. • The efficiency for polishing the physical contact connectors having the return loss of < – 45 dB and the insertion loss < 0.3 dB can be raised by at least 30%, involving two-step polishing and consuming 1 minute circle time, by selection of suitable abrasives and grit sizes. • The high linear and angular apex offsets of the polished connector end faces could significantly deteriorate the optical quality, especially the return loss. • The relationship between the geometrical quality and optical quality for the polished connectors has been established. The fibre and ferrule surface roughness Ra values of < 30 nm, the fibre height or undercut of –25 ~ 75 nm, the radius of curvature of 10 – 75 mm, the linear apex offset of < 150 µm and angular apex offset of < 0.2 degree appear to be acceptable for the physical contact fibre connectors. 6 INDUSTRIAL SIGNIFICANCE The current research has developed a high efficiency polishing processes for single mode physical contact connectors. The polishing protocols can be applied in optic fibre industry for mass production of connectors. REFERENCES [1] D. Derickson, Fibre Optic Test and Measurement, Prentice-Hall, New Jersey, pp. 621-638, (1998). [2] Fibre Optic Polishing Machine 8671X-6100 Series, Molex, Illinois, (2001). [3] OFL-12 Series OFL-126001 & PFL0127001 Mass Production Polisher Instruction Manual, Seiko Instruments Inc., Chiba, (1995). [4] S. Jahanmir, H.K. Xu and L.K. Ives, “Mechanisms of materials removal in abrasive machining of ceramics”, in Machining of Ceramics and Composites, S. Jahanmir, M. Ramulu and P. Koshy (Ed.), Marcel Dekker, New York, pp. 11-84, (1999). [5] T. Kanda, M. Misuhashi, T. Ueda, A. Toyohara and K. Yamamoto, “New micro-finish surface technology for the fabrication of optical device endfaces”, in Proceedings of the International Conference on Optical Fabrication and Testing, T. Kasai (Ed.), Vol. 2576, pp. 84-91, (1995). [6] T. Karaki-Doy, T. Satoh, J. Watanabe and K. Matsunaga, “Development of a new automatic processing machine for optic-fibre connector ends”, Bulletin of Japan Society for Precision Engineering, Vol. 22(3), pp. 216-222, (1988). [7] TIA Standard, Telecommunications Industry Association, Arlington, VA, USA, (2002).