Pii - 0924-0136 (94) 01314-Q

You might also like

Download as pdf
Download as pdf
You are on page 1of 30
ELSEVIER Journal of Materials Processing Technology 49 (1995) 85-114 Non-destructive in-situ replication metallography S. Jana School of Mechanical and Production Engineering, Nanyang Technological University, ‘Singapore 2263, Singapore Received 3 September 1993 Industrial Summary By using plastic films and other materials and making replicas of microstructures, the scope and usefulness of conventional optical metallography have been extended to field studies. However, replication does not always produce the same high-quality micrographs as are obtained with normal metallography. In this investigation some of the techniques of replication, the factors affecting such techniques and several alternative replicating materials were further studied, analysed and assessed. Replication cannot be adapted easily to underwater structures: the paper also describes a simple laboratory-made fixture which can be used successfully to etch. and replicate such structures, 1, Introduction Replication is essentially a “non-destructive” recording technique by which the microstructures of surfaces of interest can be copied for later indirect examination using the light microscope. Field or in-situ metallography is often not feasible due to inaccessibility of the area of interest to microscopes and in some cases it may not even be advisable to cut up the specimen for laboratory examination. Replicas are thus very useful in such situations. However, replication for the purpose of field metallography is still very much an art and often does not produce the same high-quality micrographs as are obtained with normal metallography [1,2]. Replica techniques have been of considerable use for examining surface structures by means of scanning electron microscopy (SEM). However, such techniques are restricted primarily to observation of fractures or other surfaces of high topographical features. In in-situ (field) microscopy, observation of polished (using portable hand grinders) and etched microstructures assumes critical importance. Overheating of boiler tubes, stress corrosion /hydrogen damage to pro- cess pipe-lines, decarburization of heat-treated components, intergranular precipitation (0924-0136)95/S09.50 (© 1995 Elsevier Science S.A. All rights reserved SSDI 0924-0136(94)01314-Q 86 S. JanaJournal of Materials Processing Technology 49 (1995) 85-114 in stainless steel weldments, etc, are some of the practical examples where field metallography is often required. The effectiveness of replication techniques with regard to such metallographic structures which involve low topographical features was felt to be in need of study and documentation for the benefit of non-destructive test personnel. In the work reported in this paper, some of the techniques of metallo- graphic replication were further studied, tested and evaluated. Experiments were performed to investigate the technique parameters involved in the replication proced- ure and through statistical analysis of the results the optimum combination of technique parameters and factors has been obtained. Several readily-available mater- ials were also tested as replicating materials and their suitability assessed. Another problem with replication is that it is not directly suited for submerged structures. While polishing can be performed on a submerged surface without any particular difficulty, etching and the making of a replica require a water-free environ- ment. For this purpose a small fixture was designed, fabricated and tested in the laboratory to see how well it can etch and replicate underwater structures. The paper also describes this fixture. 2, Replication techniques Over the years, a number of replication techniques [3-5] which are adapted primarily to SEM observation have become popular. These can be broadly divided into single-stage and two-stage methods. Among the single-stage methods the follow- ing are the more important. 2.1. Plastic replicas A dilute plastic solution, most commonly 0.1% to 0.3% polyvinyl formaldehyde in chloroform, is allowed to dry on the surface so that over “valleys” of the sample relief, a thicker layer is formed and over “mountains”, a thinner layer of plastic is formed. The plastic film formed is usually stripped dry using an adhesive tape In another variant of this technique, which is more popular, a plastic sheet (cellulose acetate is the most commonly used) is laid across the prepared specimen surface, which is then wetted with a solvent (usually acetone) and pressed firmly into contact with the specimen using the thumb. The solvent partially dissolves and softens the plastic sheet, thus allowing it to penetrate into all recesses of the specimen surface. On drying, the sheet is stripped by prising one corner free. 2.2. Carbon replicas In this technique, carbon is vaporized within a vacuum coating unit and deposited on to the specimen forming a film on it. For separation of the carbon film, dry stripping by an adhesive tape can be employed. Sometimes, an acetone softened cellulose acetate film can be laid on the carbon-coated specimen. On drying, the carbon film can be stripped off together with the cellulose acetate film. Occasionally, S. Jana|Journal of Materials Processing Technology 49 (1995) 85-114 87 for smooth specimens the carbon-coated sample is immersed in a water bath at an angle, when the film separates and floats up. Carbon replicas generally produce details with better resolution and with less artifacts and distortions. 2.3, Two-stage replicas In preparing two-stage replicas, a thick primary impression of the sample surface is, prepared and a thin replica of this primary impression is then made, thus producing positive replicas that are facsimiles of the original surfaces. Indirect examination of a specimen through a positive replica is much pre- ferred compared to a negative single-stage replica, which latter more often than not presents difficulties in interpretation, this being due to the negative replicas being reverse of the original, both in directionality and topography. Positive replicas, however, produce lower resolution as fine details are likely to be lost in the two-stage process. Plastic/carbon replicas are the most widely used two-stage technique. The prepared surface is first replicated using the single-stage softened acetate sheet as earlier described in Section 2.1. The negative replica is inverted, taped to a glass slide and coated with 10-20 nm of carbon using a vacuum-coating unit. The thin evaporated replica produced is then separated from the base acetate impression by dissolving the latter in acetone, 3. Experimental results 3.1. Replication of selected specimens using acetate sheet A number of materials in different conditions were replicated using Struers’ Transcopy replication kit which uses a green coloured acetate sheet with self- adhesive reflecting aluminium backing and ethyl acetate solvent. Following the manufacturers’ recommendations, 3 to 6 drops of solvent were applied on the green side of the acetate sheet and allowed to flow evenly on the surface. In 10-30 s, when the solvent had mollified the plastic and nearly all the solvent had evaporated, the green side was pressed hard against the prepared (polished and etched) surface for 60-90 s. The replica was then loosened carefully and left alone to dry further, or the aluminium backing was peeled off and the replica mounted on a glass slide. Using the above technique selected samples from the following specimens as shown in Table 1 (Table 1 lists all specimens used in the investigation) were replicated to see how well different features of various metallographic specimens replicate on a plastic film. Fig, 1(a) is the microstructure of specimen 1. There is difference in shade between the grains and the boundaries appear as thin black lines. The dark skeleton pattern inside the grains represents coring or unevenness in composition, that is, in the present specimen, an unequal distribution of tin in copper. Deep etching has resulted in some 88 ‘S.Jana/Joumal of Materials Processing Technology 49 (1995) 85-114 Table 1 List of specimens replicated Specimen 1 Copper/4% Tin alloy, sand cast condition Specimen 2 Copper/4% Tin alloy sand cast and annealed for 2h, Specimen 3 Copper/8.4% Phosphorus eutectic alloy in sand cast condition Specimen 4 _Aluminium/4% Copper alloy in sand cast condition Specimen S 0.35% Carbon steel, normalised from 870°C Specimen 6 0.35% Carbon stee! water quenched from 870°C (1 h heating) Specimen 7 Brass (60% copper/40% zinc) hot extruded moderately cold drawn Specimen 8 Brass (80% copper/20% zine) hot extruded, cold worked, annealed for 1h Specimen 9 Brass (80% copper/20% zine) hot extruded and cold drawn corrosion pits. Fig. 1(b) is the acetate replica of specimen 1. The grain boundaries are replicated clearly but the coring effect cannot be seen. The corrosion pits appear as inclusions due to the reverse topography. Fig. 2(a) shows the microstructure of specimen 2. Owing to annealing the coring effect has disappeared. The acetate replica of specimen 2 (Fig. 2(b)) captures the features quite well. Microstructural features of specimen 3 are shown in Fig. 3(a). The grains are composed of colonies of fine lamellar eutectic structure, high-power examination being necessary to resolve the structure properly. Free pieces of copper phosphide dendrites (white in appearance) and a copper-rich solid solution (appearing as dark spherical depressions) can be seen clearly also. The acetate replica, (Fig. 3(b)), was unsuccessful in reproducing most of the features except for the spheroids of copper rich solid solution. As will be seen later, other replicating materials were more successful Fig. 4(a) illustrates the microstructure of specimen 4 showing cored solid solution grains and a speckled eutectic pattern in the regions containing Cul phase. Some of the eutectic occurs at the grain boundaries and some within the grains. The acetate replica, Fig. 4(b), reproduces the features reasonably faithfully. Figs S(a)-7(a) respectively illustrate the microstructures of specimens 5-7, showing normalized and water quenched medium-carbon steel and the fibrous two-phase structure of brass. Their acetate replicas, Figs. 5(b)-7(b), meet with varying degrees of success. The ferrite—pearlite phases of the normalized steel reproduce reasonably well, Fig. 5(b), although at higher magnifications the contrast of the replicated features appeared to be poor. ‘The martensitic structure of the water quenched steel was too fine to be replicated, Fig. 6(b). The fibrous two-phase structure of brass replicates well. The faint indications of strain markings in the lighter phase of the micrograph are, however, absent in the replica, Fig. 7b). S. Jana|Journal of Materials Processing Technology 49 (1995) 85-114 89 Fig. 1. (a) Microstructure of specimen 1; (b) an acetate replica ( 268), %0 ‘. Jana Journal of Materials Processing Technology 49 (1995) 85-114 Fig 2. (a) Microstructure of specimen 2; (b) an acetate replica (> 66). S, JanajJournal of Materials Processing Technology 49 (1995) 85-114 o Fig. 3. (a) Microstructure of specimen 3; (b) an acetate replica (x 66). 2 S. Jana/Journal of Materials Processing Technology 49 (1995) 85-114 Fig. 4, (a) Microstructure of specimen 4; (b) an acetate replica (x 132) S. Jana{Journal of Materials Processing Technology 49 (1995) 85-114 93 Fig. 5. (a) Microstructure of specimen 5; (b) an acetate replica (x 264). 4 S. Jana}Journal of Materials Processing Technology 49 (1995) 85-114 Fig. 6. (a) Microstructure of specimen 6; (b) an acetate replica (x 264) S. Jana)Journal of Materials Processing Technology 49 (1995) 85-114 95 Fig. 7. (a) Microstructure of specimen 7; (b) an acetate replica (x 132) 96 S. Jana} Journal of Materials Processing Technology 49 (1995) 85-114 3.2. Effect of technique parameters on acetate replica quality A high fidelity replica is one with high accuracy and resolution, the latter being the ability of the replica to reproduce two features in close proximity as separate and distinct entities. Accuracy, on the other hand, includes the dimensional accuracy and the completeness of the reproduction. The quality of acetate replicas produced earlier is mainly dependent on three parameters in the procedure involved in producing it, these being: (i) the amount of solvent applied; (ii) the amount of pressure applied to the softened plastic in contact with the specimen surface; and (iii) the time allowed for the replica to dry before being stripped from off the specimen. Although, general guidelines exist in the metallographic replication procedure [6], the significance and the relationship of the parameters involved have not yet been extensively tested and documented. Experiments were, therefore, performed to study and optimise these parameters. Since the quality of replicas depends both on dimensional accuracy and resolution, itis difficult to quantify. For the basis of comparison, the “completeness” of reproduc- tion of surface feature was used. For this the image analyzer was used and the number of features on a particular location of the specimen was first counted and recorded. Thereafter, a replica of the specimen (at the previous location) was made and the number of features replicated counted again. The difference between the two sets of readings is the number of features lost. (It is assumed no artifacts were produced to add new features to the replica). feature count — deviation Percentage of replication completeness : a ua feature count x 100%, where Deviation = feature count on specimen — feature count on replica Due to its distinct features of grain boundaries and inclusions that are easily counted by the image analyzer, specimen 2 was chosen as the standard for compa son. Two perpendicular lines were scratched into the surface using a sharp needle. By positioning the image analyzer's measuring window against the perpendicular lines, it ‘was ensured that the same location is used for feature count both on the specimen and on the replica. The three parameters were each assigned two levels of conditions: this is shown in Table 2. The experimental plan consisted of taking an observation at each one of the possible combinations that could be formed for the different levels of the factors. At each treatment combination, a replica was made and the “replication completeness” values calculated through image analysis. To minimize errors, a few sets of readings were obtained using different magnifications and at different locations and their mean was recorded, the results being shown in Table 3. Next, the orthogonal array L4(2°) (ie., 2 levels, 3 parameters) for Taguchi's method was constructed: this is shown in Table 4. S. Jana|Journal of Materials Processing Technology 49 (1995) 85-114 7 Table 2 Parameters and the levels assigned 4 Ao 4 Amount of solvent used 3 drops (Just sufficient to cover, 6 drops (Flood the surface) the surface) B By B, ‘Amount of pressure applied Medium (slight thumb pressure) High (forced down by hand) c Co c Stripping time 0s 1208 Table 3 Replication completeness ‘Treatment Mean replication combination ‘completeness (%) AoBoCo 685 A,BoCo 746 AaBiCo m2 AiBiCo 76.1 ABC na ABoCt m2 ABC n2 ABC. 785 Table 4 Orthogonal array for Taguchi's method Expt, Factors Replication no, completeness AB Ce 1 0 0 0 68s 2 o 1 1 22 3 10 1 72 4 1 1 0 764 Using Taguchi’s statistical method to analyse the results, the optimum combina- tions of parameter levels were calculated, and are shown in Table 5. It was thus concluded that the best combination of parameters*to use is (i) a greater amount of solvent (6 drops in this case); and (ii) a longer time for the replica to dry before stripping (1205s in this case) The amount of pressure applied during replication was considered insignificant, although a certain amount of pressure is still required to be applied to help the softened plastic to take up the features of the specimen surface and to expel trapped air. 98 ‘S. Jana} Journal of Materials Processing Technology 49 (1995) 88-114 Table 5 Optimum combination of parameter levels. Factors and levels Total Replication com- pleteness (%) Ao 7035 Ay Test By 7385 B, TAs Ce 7230 C 75.70" 3.3. Other factors affecting acetate replica quality 3.3.1. Degree of etching Deep etching, which causes greater feature relief in specimens, inevitably results in replicas with better contrast. The features replicated were observed to be more pro- nounced and distinct. However, excessive etching is not recommended as it tends to disfigure particle morphology through attack at the particle-matrix interface, making interpretation more difficult. Also, if the height difference between features exceeds the depth of field of the optical microscope, examination of the replica would be difficult. 3.3.2. Repetitive replication Repeated replication of the same location of a specimen helps to differentiate between replicating artifacts and replicated microstructure. In most cases, due to incomplete cleaning of surfaces, in the initial replication, “pseudo-replicas” are for- med. Subsequent replication shows fewer artifacts and foreign particles due to the resultant cleaner surface. 3.3.3. Reflective aluminium backing The aluminium backing present on the transcopy’s acetate sheet is meant to reflect light and improve the observed contrast of the features replicated. Although the aluminium backing does, in some cases, reflect the reliefs on the replica giving a virtual increased depth, more often than not the strong reflected light blinds off most of the topographical contrasts, thereby worsening the situation. An example is shown in Figs. 8(a) and (b) which illustrate the replica photograph of specimen 8. The alumi- njum-backed replica shows reduced contrast and the micro-strain lines become visible only in the absence of aluminium backing. The aluminium backing can be removed easily by rubbing and washing off and removal should be done whenever necessary. 3.4. Alternative replicating materials A replicating material should be chosen on the basis of availability, economy, accuracy, minimization of artifact, ease of use and reproducibility. Although cellulose S. Jana|Journal of Materials Processing Technology 49 (1995) 88-114 9 Fig. 8. An acetate replica of specimen 8: (a) with aluminium backing; (b) with reflective aluminium foil backing removed (132), 100 ‘S Jana}Journal of Materials Processing Technology 49 (1995) 85-114 acetate is the most widely used replicating material, field investigations throw up situations where other available alternatives or “make-do” materials have to be used. It was also felt that alternative materials should be explored to see if any other material can out-perform cellulose acetate. For this purpose, replicas of selected specimens were produced using eight different replicating materials and their suitability and the accu- racy and resolution of the replicas produced from these materials were evaluated. 3.4.1. Glue The glue used was the common clear liquid type of paper glue. The glue is water based and becomes translucent and rubbery when dry. The replicating technique consisted of spreading evenly the easy flowing glue onto the surface of interest, letting the glue dry and then peeling off the rubbery replica with a pair of tweezers. Leaving a greater amount of glue to dry resulted in thicker replicas which, though tougher and easier to detach from the specimen, required a much longer time to dry. Thinner-glue replicas generally gave better resolution with less artifacts but were very fragile and thus prone to damage and deformation. Thin-glue replicas adhered easily to glass slides and were thus convenient to mount for microscopy. To hasten the drying process a hair dryer or a lamp was employed. However, experience showed that natural drying produced glue replicas of higher quality. Results of glue replicas are shown in Figs. 9b), 10(c) and 11(a). Replicas from other replicating materials are also displayed in subsequent figures for the purpose of compari- son. The quality of the glue replicas was surprisingly good considering that it is such a cheap everyday material. Some of the limitations of glue should, however, be noted: is slow drying, Gi) the specimen surface to be replicated should be horizontal, as otherwise the glue will simply flow off, (ii) artifacts in the form of bubbles are quite common. 3.4.2. Elmer's glue This is a “general purpose” glue and was earlier recommended by Yunus et al. [7] as a replicating material. The viscous liquid glue is white coloured and when com- pletely dry is translucent and hard. Similar procedure as earlier outlined for glue was followed for this material also, Good quality replicas were obtained, as can be seen in Figs. 9(c), 10(4) and 11(b). The replicas obtained were tougher and less susceptible to damage than those obtained with glue. 3.4.3, Polystyrene A dilute solution of 10% polystyrene in acetone was tried out as a replicating material. The solution was deposited onto the surface of the specimen and allowed to evaporate, leaving behind a thin film of polystyrene. Dry stripping with a pair of tweezers was used. As with glues, thinner replicas gave better results. Figs. 9d), 10(e) and 11(c) show the replicas obtained and their fairly good quality. The main problem encountered with this material was the difficulty in separating the replica from the specimen, In cases where the specimen surface was rough, the polystyrene replicating, material required some forceful prising and peeling before separation could be achieved, which usually resulted in some damage. The material is also quite brittle and cracks occurred frequently. 5. Jana}Journal of Materials Processing Technology 49 (1995) 85-114 101 Fig. 9. Specimen 2: (a) an acetate replica from a different location to those of Fig. 2} (b) a gue replica; (c) aan Elmer's glue replica; (d) a polystyrene replica (x 66) 10 ‘S.Jana{Journal of Materials Processing Technology 49 (1995) 85-114 Fig. 9. Continued, S. Jana}Journal of Materials Processing Technology 49 (1995) 85-114 103 Fig. 10. Specimen 9: (a) microstructure (b) an acetate replica; (ea glue replica; (d) an Elmer's glue replica; (6) a polystyrene replica; (f)a polyester replica; (g) a gelatin replica (x 264, 104 S. Jana Journal of Materials Processing Technology 49 (1995) 85-114 Fig, 10, Continued, S. Jana/Journal of Materials Processing Technology 49 (1995) 85-114 105 Fig, 10. Continued. 106 ‘S. Jana{Journal of Materials Processing Technology 49 (1995) 85-114 Fig. 10, Continued, 3.4.4, Polyester Polyster pellets softened by acetone to a tacky consistency was tried out. Softened polyester in putty consistency was pressed flat against the specimen surface and left to dry. On drying, the thick polyester replica was prised free using a blade. The replica can be mounted on plasticine for observation under microscope. Formation of bubbles and some haze when under optical microscope were two problems encoun- tered. Fig. 10(f) shows a replica obtained from the material. 3.4.5. Gelatin This is the commonly available “starch”. Gelatin was smeared on the specimen surface to form a thin layer about 1-2mm thick. On drying, the film could be removed easily. The replicas, although giving fairly good resolution, were not capable of providing consistent results and tended to be full of artifacts. The brittleness of the replicas caused them to often break into small pieces. Also these were extremely sensitive to moisture. Fig. 10(g) shows a gelatin replica. 3.4.6. Wax ‘Two types of wax, e.g. paraffin wax and microcrystalline wax were used. Pieces of wax were softened by heating with a specimen dryer and pressed hard against the specimen surface. Once solidified, the wax could be prised free and mounted on S. JanajJournal of Materials Processing Technology 49 (1995) 85-114 107 Fig. 11, Specimen 3: (a) a glue replica (from a different location to those of Fig. 3} (6) an Elmer's glue replica; (¢) a polystyrene replica (> 132). 108 S. Jana/Journal of Materials Processing Technology 49 (1995) 85-114 Fig. 11. Continued plasticine for examination. The time allowed for the wax to solidify appeared to be very critical, since insufficiently solidified wax would result in parts of it remaining stuck to the specimen, while a fully solidified wax would not detach without cracks forming on it. The replicas obtained using both kinds of wax were of generally low quality giving poor resolution, and the presence of micro-cracks obscured the features replicated. 3.4.7. Silicone rubber For high topography work silicone rubber was earlier successfully tried by Scott [8] and Kusy and Leinfelder [9]. However, its effectiveness for low topography field microscopic samples is not well established. Self-curing silicone rubber sealant was, therefore, experimented with as a replicating material. A layer of silicone rubber was, left on the surface to cure by itself. Upon drying it could be peeled off easily: however, the material has a tendency to adhere strongly to the specimen surface, which caused the replica to be damaged usually, during detachment. The replicas produced were of poor quality, 3.4.8. Comparison of replicating materials Using the same method described earlier, the replicating completeness was deter- mined for each material. The results are shown in Table 6, which indicates also the time required to prepare (dry) the replica. S. Jana/Journal of Materials Processing Technology 49 (1995) 85-114 109 Table 6 ‘Comparison of replicating materials Replicating material Replication completeness Time required for “ preparation ‘Transcopy acetate sheet 79.2 <3min Glue 94 >ih Elmer's glue 530 pth Polystyrene 20 > 30min Polyester 556 > 30min Gelatin : > 30min Wax = < 10min Silicone rubber >2hr The results for gelatin, wax and silicone rubber were unattainable due to excessive artifacts appearing on them. Also to be noted is the very good result for common everyday glue. 3.5, Two-stage replication Two combinations of replicating materials described here were found to be capable of producing positive replicas of fairly good quality. 3.5.1. Acetate (— ve) /glue (+ ve) The earlier-described glue was applied onto acetate replicas and left to dry, the dried glue (positive replica) being detached simply by peeling it off. Fig. 12a) shows a positive replica of this kind. Acetate ( ~ ve)/Elmers glue ( + ve) was also tried, Fig. 12(b) showing the result. 3.5.2. Glue( — ve)/Polystyrene( + ve) Polystyrene was applied over inverted glue replicas. On drying, the glue was washed off in running water leaving the polystyrene positive replica behind. Elmer's glue was found to be unsuitable as the negative replica because of chemical interaction between the two materials. Fig. 12(c) shows a glue( — ve)/polystyrene( + ve) replica. Generally speaking, the acetate( — ve)/glue( + ve) combination produced better qual- ity replicas than the glue( — ve)/polystyrene( + ve) combination. In the latter case, due probably to chemical interaction, the positive replicas tended to be full of defects and artifacts. 3.5.3, Comparison of two-stage replicating material combinations ‘Using the method described earlier, the replication completeness of these positive replicas was determined and shown in Table 7. 110 ‘S. Jana{Journal of Materials Processing Technology 49 (1995) 85-114 Fig. 12. Two-stage positive replication of specimen 2: (a) acetate (— ve)/glue (+ vek (b) acetate ( — ve) Elmer's glue (+ vel (c) glue (— ve)/polystyrene (+ ve} (66) ‘S. Jana|Journal of Materials Processing Technology 49 (1995) 85-114 un Fig. 12. Continued Table 7 ‘Comparison of two-stage replicating materials Replicating materials Replication completeness ( Acetate ( — ve}iglue (+ ve) 768 Acetate ( — ve}/Elmer’s glue (+ ve) 725 Glue (— veypolystyrene (+e) 68.5 Acetate ( — ve}! 792 * Acetate single-replca is included here for comparison 3.6, Fixture for underwater replication In designing a fixture capable of replicating underwater structures, the following considerations served as the basis: (i) polishing of the underwater structure would be done separately and the fixture would not perform this function; (i) the fixture should bbe light-weight and easy for the diver to handle; (ii) it should be water-proof and be able to withstand water pressure up to, say, 70 m depth; (iv) it should be reusable. nz , Jana}Journal of Materials Processing Technology 49 (1995) 85-114 SECTIONS 5° Fig. 13. Assembly drawing of the replicating fixture: (1) housing; (2) plunger, (3) rectangular block; (4) plunger base; (5) magnetic base; (6 and 7) ‘O" ring: (8) stud; (9) hexagonal nut Fig. 14. Photograph of the prototype fixture showing the plunger, housing, check valve, magnets, ai-line and syringe, 'S.Jana(Journal of Materials Processing Technology 49 (1995) 85-114 3 The assembly drawing of the replicating fixture is shown in Fig. 13, whilst a photo- graph of the fixture is shown in Fig. 14. The fixture is to be placed on the underwater polished surface in such a way that the air chamber sits directly over the area of interest. The two magnets are then activated, where upon the fixture seats firmly on the metal surface. Compressed air is then forced into the air chamber creating a water-free environment inside. The air pressure is then reduced to about I atmo- sphere by manipulation of the valve. Next, the air chamber is flooded with etching solution using the syringe and the plastic tubing. This is followed by forcing out the etching solution, again with compressed air. The metal surface is now etched and ready for replication. Solvent is then injected into the air chambers using the same syringe and plastic tubing. The plunger which holds the acetate sheet is now pressed down gently but firmly. After about 120s, the plunger is pulled up. The magnets are deactivated, the fixture brought outside the water surface and the acetate replica detached, dried and mounted on glass slide for observation. The diver then has to fix another acetate sheet on to the plunger and the fixture is ready for another trip underwater. Details of the fixture are available in Refs. [1] and [2]. The prototype, made with an aluminium body, functioned satisfactorily and could replicate structures placed inside a water tank in the laboratory. 4. Conclusions 1. Acetate sheets were found to be eminently suitable for replicating field metallo- graphic features. 2. Although acetate replicas can reproduce faithfully distinct microstructural fea- tures such as grain boundaries, strain lines, etc., their reproducibility does not encompass very fine features such as matersitic structures, coring, etc. 3. Through experiments and statistical analysis, it was found that the quality of acetate replicas was dependent on the following: (i) the amount of solvent used to soften the acetate sheet, where using a greater amount of solvent (6 drops in the present investigation) produced better results; (ii) the time allowed for the replicas to dry before stripping off from the specimen, a longer time (greater than 120s) being recommended. 4. The amount of pressure applied to the back of the acetate sheet during replica- tion was found to have no significant effect on the quality of replicas produced, the reflective aluminium foil backing being found to be redundant (sometimes a hin- drance) in most cases. 5. Amongst the different alternative replicating materials tried out, the common liquid glue has the potential to be the “next best” material. [ts major limitation is the length of time required for replication, which is very long (taking about 45 mins to dry). 6. Acetate ( — ve)/Glue ( + ve) two-stage replication material combination seemed to be very promising for field metallography. 7. The unadaptability of the replication technique to underwater structures has been overcome by designing and fabricating a prototype fixture which can success- fully replicate submerged structures. 14 5. Jana) Journal of Materials Processing Technology #9 (1995) 85-114 Acknowledgements ‘The paper is based on the work carried out by Messrs Cham Tud Yinn, T. Mohan and T. M. Premnath as part of their final year projects at Nanyang Technological University toward partial fulfilment of their Bachelor’s in Engineering degrees. Their contribution is gratefully acknowledged. References [1] S.Jana and CH. Quek, Non-destructure metallography and its potential application underwater, Eng J. Singapore, 141) (1987) 22-29, (2] S.Jana, Insitu metallography and its potential application underwater, Proc. Ist Pacfic-Asia Offshore Mechanics Symp. Vol. II, Seoul, Korea, 1990, pp. 131-136, [3] OS. Cranskshaw, Instruction of Replica Techniques for Scanning Electron Microscopy, Vol. 1V, ‘Scanning Electron Microscopy, SEM Inc, AMF O'Hara Chicago, 1984, pp. 1731-1757, [4] CH. Premeijer, Replica Techniques for Scanning Electron Microscopy, Scanning Electron Microscopy, Vol II, SEM Inc, AMF O'Hara Chicago, 1978, pp. 831-836. [5] ME. Blum, On-site Metallographic Investigations Using Replica Technique, Microstructural Science, Vol 12, North-Holland, Elsevier, 1985, pp. 143-150, [6] ASTM Designation ES 12-87 (1987), Emergency Standard Practice for Production and Evaluation of Field Metallographic Replicas ~ ASTM Annual Book of Standards, Vol 01.01, 1987, [7] M. Yunus, K. Kulshreshtha, P. Nawani, and K.J. Ahmad, Surface replica by Elmer's glue ~ a new imprinting material, Microsc. Acta, 85(3) (1983) 255-258. [8] EC. Scott, Replica production for scanning electron microscopy: A test of materials suitable for use in field settings, J. Microsc, 125(3) (1982) 337-341. [9] RP. Kusy, and K-F. Leinflder, In-situ replication techniques: I. Preliminary screening and the negative replica technique, J. Dent, Res, 56(8) (1977) 925-932.

You might also like