job:LAY03 page:40 colour:1 black–text efficiency of the flux is derived from the resulting wetting curve. The solderbath contains a 63% Sn solder (for instance to JSTD-006, Sn63Pb37C), held at a temperature of 250 °C/480 °F. Corrosive action The test for corrosive action is again confined to observing what a flux will do to copper during soldering, or what the residue which is left on the copper will do in a moist atmosphere. In ISO 9455–13 flux residue, left on a copper coupon after having melted a small amount of 60% tin solder together with the flux under test, is stored in a humid atmosphere, at 40 °C/645 °F and 91% to 95% relative humidity, for three days. Corrosion is deemed to have occurred if the flux residue has changed colour, or if white spots have appeared in it. In ISO 9455–5, a drop of the flux to be tested is placed on a flat glass slide, on to which a thin film of copper, with a thickness of 0.05 m/0.002 mil (500 angstrom) has been deposited by an evaporation technique, a so-called ‘copper mirror’. Copper mirror slides are commercially available. The slide with the drop of flux on it is kept in a humidity chamber at 23 °C/73 °F and 50% relative humidity for 24 hours, and then examined. If the copper mirror has disappeared underneath the flux, it is deemed to have failed the test. A flux which passes the copper mirror test is an ‘L-type’ (low activity) flux, which group comprises all R-type fluxes, most RMA, and some R. If some of the copper mirror has gone, it is an ‘M-type’ (medium activity) flux, which may still be an RMA, but is mostly RA and sometimes a watersoluble or a synthetic activated flux. If the copper mirror has disappeared completely, the flux is an ‘H-type’ (high activity). Watersoluble and synthetic activated fluxes fall in that group. An important aspect of flux classification relates to the surface–insulation–resistance (SIR) properties of a flux (ISO 9455–17, not yet issued). Halide content Determination by analysis If a halide-free flux is specified, somestandardsgive a detailed analyticalprocedure for quantitatively determining the halide content of the flux. If this exceeds 0.05% by weight of the rosin content of the flux, calculated as Cl, the flux does not conform to, for example, a BS 5625 halide-free flux. If it exceeds 0.5% calculated Cl on the solids content of the flux, it does not conform to an ANSI/J-STD-004 flux of type LI. Silver-chromate test This is a qualitative yes/no test, and does not indicate a specific halide percentage. Silver chromate (AgCrO  ) is a brick-red substance, which turns white or yellow in the presence of a halide. Silver-chromate impregnated testpaper is commercially available. If such a piece of paper turns white or yellow when a drop of the flux under test is placed on it, halide is deemed to be present, and the flux cannot be 58 Soldering job:LAY03 page:41 colour:1 black–text classed as ‘L0’ or ‘L1’ to J-STD-004. There is a problem, though: certain acids and amines (which may well be free of halide) are also capable of causing the colour of silver-chromate paper to change. Because this test is relatively insensitive, a flux with up to 0.05% halide will still pass it as ‘halide-free’. Beilstein test This test, which is mentioned in ANSI/J-STD-004, is more sensitive than the silver-chromate test, but it is a qualitative test and gives no indication of the actual quantity of halogen present. Its drawback is that it will also respond to any non-ionic halogen in a halogenated solvent, should any such be contained in the flux. The Beilstein test detects the presence of halogen in an organic compound. It requires a small piece of fine copperwire gauze, which is heated in an oxidizing flame (e.g. the blue part of a bunsen-burner flame) until it ceases to turn the flame green. It is withdrawn, allowed to cool, and a small amount of the flux under test is placed on it. It is then put back into the flame. If the flame turns blue-green, the flux contains traces of halide. If not, it is deemed to be halide-free. The Beilstein effect depends on the formation of a volatile copper halide. (F. K. Beilstein, Russo- German chemist, 1838–1906.) Solubility of flux residues The average flux user needs guidance on how to assess the ease with which the residue of the flux he is using, or wants to use, responds to the cleaning method he is using or intends to use. The international standard ISO 9455–11: 1991 (E) is relevant to this problem. This standard describes a method of heating a sample of the flux on a dish-shaped piece of brass sheet up to 300 °C/570 °F for a given time, placing the sample in a humidity chamber for 24 hours and then immersing it in the solvent which is to be used for cleaning. The presence of any residual flux left after cleaning is indicated by the ability of the cleaned test specimen to form an electrolytic cell. Surface insulation resistance (SIR) of the flux residue By definition, the residue from a ‘no-clean’ flux remains on the board. Obviously, not only must it cause no corrosion, but its presence must not interfere with the functioning of the circuitry by lowering the surface insulation resistance (SIR) of the board between adjacent conductors: a leakage current of 10\ A between neighbouring IOs of a high-impedance microprocessor is enough to cause it to malfunction (see Section 8.1.1). A number of tests to measure the SIR after various soldering and cleaning procedures have been devised over the years. They are described in Section 8.6.3. J-STD-004 includes a method for testing the flux residue for its moisture- and surface-insulation resistance. The relevant ISO working group is expected to complete its deliberations on the same subject in about three years’ time (informa- tion from BSI, London, April 1997). Soldering 59 job:LAY03 page:42 colour:1 black–text Tackiness of the flux residue Finally, the residue from a no-clean flux must be dry and not sticky or ‘tacky’ under normal temperature and moisture conditions. Tackiness is tested by applying powdered chalk to a fluxed coupon which has undergone a specified temperature regime. If the powder can be removed with a soft brush, the flux has passed the test. 3.5 Soldering heat Conventional soldered joints are made with molten solder. Hence, the soldering temperature must always be at least above the melting point of the solder, i.e. above 183 °C/361 °F. The immediate environment of the joint, and sometimes the whole assembly, must be brought up to the soldering temperature too. The exact tempera- ture needed depends entirely on the soldering method used. It is rarely less than 215 °C/420 °F and is often much higher. 3.5.1 Heat requirements and heat flow Heat is a form of energy, which is usually measured in one of the following ways. One calorie (1 cal) raises the temperature of one gram of water by 1 K (which is the same temperature difference as 1 °C, Section 5.4.2). One calorie equals 4.187 joule, or in units which are meaningful in the context of soldering, 4.18 watt.seconds (W.sec). Table 3.12 indicates the amounts of heat required in some common soldering situations. In this context, it is useful to know the heat conductivity of the various materials involved, so as to be able to gauge the speed with which the heat input spreads within an assembly (Table 3.13). The figures given in Tables 3.12 and 3.13 are worth studying. Table 3.12 shows that organic substances like FR4 have a much higher specific heat than metals. This has an important bearing on most soldering situations. The greater part of the soldering heat expended in making a joint is not used to heat the metallic joint partners, but to heat the FR4 epoxy board on which the copper laminate sits. Hence the need to preheat the boards before they pass through the solderwave (Section 4.3), but also the benefit of preheating the circuit board, at least locally, when soldering single multilead components (Section 5.7), or before carrying out repair work, i.e. desoldering and resoldering single components (Section 10.3). The list of heat conductivities is equally illuminating. The heat conductivity of epoxy is two orders of magnitude lower than that of the ceramic substrate of a hybrid assembly. Hence the need for taking the thermal management of SMDs, which are mounted on an epoxy board, much more seriously than that of hybrid constructions, which were initially the beginnings of SMD technology. The figures also show how even the narrowest air gap prevents the flow of heat between two hot bodies. Hence the need to have a drop of molten solder on the tip of a soldering iron or thermode, or at least some flux on the joint to bridge that gap (Section 5.7). 60 Soldering job:LAY03 page:43 colour:1 black–text Table 3.12 Heat required to raise the temperature of a substance from 20 °C/68 °F to a soldering heat of 250 °C/482 °F 1 g copper 88 watt. sec 1 g solder 102 watt. sec (including heat of melting) 1 g FR4 338 watt. sec A soldered joint (volume 1 cub. mm) 0.7 watt. sec A circuit board 27 kw sec 23.3 cm ; 16 cm 9.2 in ; 6.3 in (‘Europa’ format) 1.2 mm/47 mil thick Table 3.13 Some heat conductivities in watt/cm °C Copper 3.9 Aluminium 2.2 Brass 1.2 Steel 0.5 Solder 0.5 Ceramic (alumina) 0.25 FR4, rosin 0.002 Air 0.000 000 002 3.5.2 Heating options Equilibrium and non-equilibrium situations The basic aim of every heating process is the transfer of heat from a heat source to the heat recipient, i.e. from a hot body to a colder one via a heat transfer medium. There are two basic heating situations: equilibrium and non-equilibrium systems. In equilibrium situations, the temperature of the heat source is the same as the soldering temperature which must be reached. The time within which the joint reaches its soldering temperature depends on the efficiency of the thermal coupling between source and joint. The joint cannot be overheated, i.e. it cannot get too hot, but it can be ‘overcooked’, i.e. it can be heated for too long a time. The latter carries the risk of excessive growth of the brittle intermetallic compound, and thus an unsatisfactory joint structure and the risk of a shortened joint life-expectancy. In non-equilibrium situations, the temperature of the heat source is higher, often very much so, than the soldering temperature itself. Whether the correct soldering temperature is reached or exceeded is a matter of timing the heat exposure. The higher the temperature of the heat source, the steeper is the temperature rise of the solder joint, and the more critical becomes the precise control of the duration of its heat exposure. Overheating may not only endanger the joint and its properties, but in severe cases it can damage the assembly itself (Figure 3.14). Wavesoldering, vapourphase soldering, hot air or gas convection soldering, impulse soldering and handsoldering with a soldering iron present equilibrium Soldering 61 job:LAY03 page:44 colour:1 black–text Figure 3.14 Equilibrium and non-equilibrium heating situations heating conditions. Infrared soldering, laser soldering and flame soldering are non-equilibrium systems. Heat sources A thermostatically controlled electrical resistance heater is the most common primary heat source. This transmits its heat to the heat-transfer medium, whether it 62 Soldering job:LAY03 page:45 colour:1 black–text be the drop of solder on a soldering iron or the solderwave in a wavesoldering machine. The reader may be amused to learn, though, that the first few wavesolder- ing machines were gas heated. Small, pointed butane- or propane-gas flames are used for soldering individual joints in awkward locations. Equipment using a very hot, needle-shaped hydrogen– oxygen flame is also commercially available. These flames, which represent extreme cases of non-equilibrium heating, may be hand-held, but more often are manipu- lated by programmed robots, and then of course equipped with controls which prevent overheating. Laser beams present the ultimate in non-equilibrium heating. To speak of the ‘temperature’ of a laser source makes no real sense; what matters is the extreme energy density of the spot of laser light, which impinges on the joint surface, and which may reach 10 kw per square millimetre (Section 5.6). A very precise energy dosage is of the essence, to avoid destruction of the joint and burning a hole into the substrate. Exposure times are measured in milliseconds. Heat transfer mechanisms The soldering heat can be transmitted from the heat source to the joint by any one of three basic mechanisms: conduction, convection and radiation. Conduction relies on a direct physical contact between a hot solid body or liquid and the surface of one of the joint members. The efficiency of heat transfer depends critically on the close fit between the heating and the heated surface. Any airgaps between them fatally affect the heat transfer. Molten solder is the best heat-transfer medium available: being a liquid, it conforms perfectly to whatever surface it has to heat. This is the virtue of the solderwave, as well as of the drop of molten solder on the tip of a soldering iron, which will come in very useful with repair soldering (Sections 10.2 and 10.3). Strictly speaking, convection comes into the heat-transfer mechanism of wavesoldering as well, because the solderwave consists of a body of moving solder. By contrast, dipsoldering in a stationary bath relies on heat transfer by conduction only, like a soldering iron. 3.6 Solderability 3.6.1 Wetting and dewetting Wetting The term ‘wetting’ describes the behaviour of a liquid towards a solid surface with which it comes into contact. In our case,we are naturallyconcerned with the way the molten solder behaves toward the substrate. Though everyone knows instinctively what is meant by ‘wetting’, it will be useful to examine in detail what is involved in wetting in the context of soldering, and how it can be quantified and measured. Section 3.3.1 described the soldered joint as the result of a surface reaction between molten solder and a solid metallic substrate, and it was explained why an intimate contact between the two is a precondition for the joint to form. We must now amplify this by saying that wetting is the precondition for this intimate contact. Soldering 63 job:LAY03 page:46 colour:1 black–text Figure 3.15 The wetting angle Wetting is not a ‘yes or no’ situation; there is a scale of wetting quality between total non-wetting and complete wetting. The yardstick for measuring the quality of wetting is the ‘wetting angle’, which is formed between the surface of the solid and that of the liquid along the line where they meet (Fig. 3.13 and 3.15). A wetting angle of 180° is a sign of total non-wetting, while an angle towards zero denotes complete wetting. In the context of soldering, a wetting angle of less than 60–75° is normally, but arbitrarily, considered acceptable; anything up to 90° is doubtful, and beyond 90° definitely bad. Whether and when ‘bad’ can or should be equated with ‘non-acceptable’ will be discussed in Section 9.3. The wetting or contact angle between the molten solder and the substrate is the result of the opposing forces of the surface tension of the solder, which tries to pull it together into a globule (somewhat flattened by gravity), and the interfacial tension between the solder and the substrate, which tries to pull the solder across its surface, so that as much of the solder as possible can come in contact and react with it. The wetting angle can be interpreted in terms of the three surface energies involved: that of the molten solder, of the solid, and of the interface between the two. Klein Wassink provides a detailed discussion of this aspect. In practical terms, the significance of wetting can be stated very simply. Good wetting helps the solder to get to all the places where it ought to be; doubtful and bad wetting prevent the solder from entering a joint. Dewetting ‘Dewetting’ is not the same as ‘non-wetting’. As the term implies, in a dewetting situation the molten solder did get to where it ought to be, but it does not stay there. Instead, it pulls back and forms separate islands of solder, with areas of exposed intermetallic compound in between. This situation can occur in dip-tinning, e.g. in the hot-air levelling process for circuit boards (HAL), or in wavesoldering, but only rarely in reflowsoldering. Dewetting is caused by local, untinnable spots of surface contamination, such as oxide particles, or surface dirt like traces of silicones or fingerprints. Non-metallic inclusions in galvanic coatings, for instance embedded colloids caused by unsuitable or badly controlled plating baths for copper, nickel or gold, can cause dewetting too. Surfaces which dewet are at first completely covered with molten solder, which 64 Soldering job:LAY03 page:47 colour:1 black–text Figure 3.16 Dewetting and non-wetting bridges the untinnable spots. Before it can solidify, its surface tension pulls the still liquid soldercoating apart, and away from the discontinuities (Figure 3.16). 3.6.2 Capillarity and its effects A capillary is a very thin hole or a narrow gap (from capillus, Latin for ‘hair’). If the surfaces of the hole or gap are wettable, interfacial tension quickly pulls the liquid solder into it with considerable force, often against the force of gravity. On the other hand, if the inner walls of the gap are untinnable, the surface tension of the solder prevents it from entering it. If one of the members forming the gap is movable, like the gullwing legs of an SMD during reflowsoldering, the interfacial tension pulls the walls of the gap towards one another, which means it pulls the flat end of the leg into the middle of its footprint. If, on the other hand, one or both are untinnable, the surface tension of the solder pushes them apart (Figure 3.17). The consequences of capillarity for soldering are important: 1. If the joint surfaces are wettable, capillarity pulls the solder into the joint, against the forceofgravity if necessary.Ifthey wetbadly ornot atall, thesolder cannotget into the joint, even if gravity would tend to pull it into the gap. 2. If one of the joint members is mobile, as is the case in reflowsoldering, and if both are wettable, interfacial and surface tension pull the joint members together. If one of them is unwettable, they are pushed apart. In reflowsoldering, capillary forces are the cause for the self-alignment of BGAs and small SMDs, but also for ‘tombstoning’ and the floating of chips or melfs on badly designed layout patterns (Sections 6.4.2 and 11.2.2; also Figure 3.18). 3.6.3 Capillarity and joint configuration Capillary joints and open joints The way in which the solder flows into a wettable joint depends on the soldering method and on the shape of the joint itself. Basically, there are two types of joint: ‘capillary joints’ and ‘open joints’ (Figure 3.19). With a capillary joint, or lap joint, two flat and essentially parallel surfaces face one another, and the joint forms a Soldering 65 job:LAY03 page:48 colour:1 black–text Figure 3.17 two-dimensional gap. Tubular joints, like through-plated holes, are a special form of capillary joint, where the gap is cylindrical. With an open joint, or butt joint, one or both of the joint members are not flat, and they touch one another along a line or just in one spot. With capillary joints, the escape route for the air and flux in the joint can get blocked if the molten solder closes all the edges around the joint gap before all the air and flux inside the gap have been pushed out by an orderly, frontal advance of molten solder into the gap from one (or at most two) sides only. With an open joint, there is no such problem: from whichever direction the solder enters an open joint, the escape routes for air and flux cannot be blocked. With wavesoldering, all capillary joints, flat and tubular, fill from one side only. Both types will normally be sound, especially the latter, unless air or water vapour escape from the walls into the hole after the solder has entered it (blowholing), which is a matter dealt with in Sections 9.5.3 and 11.2.2. 66 Soldering job:LAY03 page:49 colour:1 black–text Figure 3.18 The effects of capillarity in reflowsoldering. (a) Self-alignment of components; (b) tombstoning Figure 3.19 (a) Capillary joints and (b) open joints Soldering 67 [...]... 94–98 30 Zado, F M (19 83) Increasing the Soldering Efficiency of Noncorrosive Rosin Fluxes Western Electric Eng., 27(1), pp 22–29 31 Klein Wassink, R J (1989) Soldering in Electronics, 2nd ed., Ch 5.5 .3 Electrochemical Publications, Ayr, Scotland 32 Lea, C (1992) After CFCs? loc cit., p 30 1 33 Manko, H H (1979) Solders and Soldering, McGraw-Hill, NY, p 31 3 job:LAY 03 page:65 colour:1 black–text Soldering. .. State of the Art Soldering & SMT (Ayr, Scotland), No 4, pp 27 38 , No 5, pp 56–66, No 6, pp 18–27 23 Engelmaier, W (19 93) Reliability of Surface Mount Solder Joints; Physics and Statistics of Failure Proc Intern Conf Softsoldering, Munich DVS Report 1 53, Duesseldorf, Germany, pp 149–160 24 IPC (1992) Guidelines for Accelerated Reliability Testing of Surface Mount Solder Attachments IPC Document IPC-SM-785... McGraw-Hill, NY, p 31 3 job:LAY 03 page:65 colour:1 black–text Soldering 83 34 Klein Wassink, R J (1989) Soldering in Electronics, 2nd ed., Ch 2 Electrochem Publ., Ayr, Scotland 35 Wise, E M (1948) Metals Handbook, ASTM, Cleveland, Ohio, p 1111 36 Klein Wassink, R J (1989) loc cit., pp 217–218 37 Strauss, R (1988) Wavesoldering v Reflow -soldering: Metallurgical Consequences of the Choice between them Proc... Canad J Chem., 33 , p 511 15 Schmitt-Thomas, K G., Lang, H.-P and Moedl, A (19 93) Metallurgical Examination of Thermally Stressed TAB Outer-lead Bonds Conference on Soldering, Science and Practice’, Techn Univ Munich, March 19 93 DVS Rep 1 53, Duesseldorf, Germany (in German) 16 Strauss, R (1988) Wavesoldering v Reflowsoldering – The metallurgical consequences of the Choice of Method Brazing & Soldering, ... SMD Technology, Brazing & Soldering (UK), Spring 1986 12 Bader, W G (1969) Dissolution of Au, Ag, Pd, Pt, Cu and Ni in a molten Tin–Lead Solder Welding J., 12, 48, pp 551–557 13 Steen, H A H and Becker, G (1986) The Effect of Impurity Elements on the Soldering Properties of Eutectic and Near Eutectic Tin–Lead Solders Brazing & Soldering (UK), No 11, pp 4–11 job:LAY 03 page:64 colour:1 black–text 82 Soldering. ..job:LAY 03 page:50 colour:1 black–text 68 Soldering The penetrating speed of molten solder into a flat capillary gap between two copper surfaces, 0.09 mm /3. 5 mil apart, has been measured for various solders and at various temperatures by McKeown (see Reference 2) At 2 43 °C/470 °F, using a 63% tin solder and a concentrated zinc–ammonium chloride flux, McKeown measured a penetration speed of 3. 5 sec over... Table 3. 14 This listing assumes clean, though not oxide-free, surfaces Solderability is governed by several factors, among them the following: 1 The ease with which surface oxides or sulfides are dissolved by a flux 2 The surface energy of the metal surface (which means its readiness to react with whatever comes in contact with it), metals with low surface energies being more difficult to solder 3 The... result Dwell time in solderbath 3 sec 2 sec 5 sec 30 sec job:LAY 03 page: 63 colour:1 black–text Soldering 81 The test parameters depend on the soldering method by which the components are to be used They are tabulated in Table 3. 16 Having been dipped, every specimen is examined for signs of dewetting, under a magnification of about ;5 If more than 95% of the surfaces of every specimen are covered with a... Wantage, UK 3 Raynor, G V (1947) The Pb–Sn Equilibrium Diagram Met Abstr (London), 19, p 150 4 Earle, L G (1946) The Pb–Sn–Ag Equilibrium Diagram J Inst Met (London), 72, p 4 03 5 Smernos, S and Strauss, R (1984) Low Temperature Soldering, Circuit World (Ayr, Scotland), 10 (3) , pp 23 25 6 Vianco, P T and Frear, D R (19 93) Issues in the replacement of Pb-bearing Solders Journal of Materials, July 93, pp 14–19... measuring the solderability of a metal surface, using a flux of standardized fluxing power, or the fluxing efficiency of a flux, using a copper specimen with a surface of standardized solderability, coated with the flux under test (see Section 3. 4.9) By now, it has reached a Figure 3. 22 The solder meniscus and its assessment job:LAY 03 page:58 colour:1 black–text 76 Soldering high degree of technical perfection . (Figure 3. 14). Wavesoldering, vapourphase soldering, hot air or gas convection soldering, impulse soldering and handsoldering with a soldering iron present equilibrium Soldering 61 job:LAY 03 page:44. watt. sec A circuit board 27 kw sec 23. 3 cm ; 16 cm 9.2 in ; 6 .3 in (‘Europa’ format) 1.2 mm/47 mil thick Table 3. 13 Some heat conductivities in watt/cm °C Copper 3. 9 Aluminium 2.2 Brass 1.2 Steel. with which the heat input spreads within an assembly (Table 3. 13) . The figures given in Tables 3. 12 and 3. 13 are worth studying. Table 3. 12 shows that organic substances like FR4 have a much higher