Copper: conductivity, oxygen and phosphorus

Kautny ranked copper second only to iron in industrial importance, valued for its ductility, its finish and its conductivity. The same conductivity that makes it valuable in heat exchangers and electrical work makes it hard to weld: the joint loses heat so fast that it resists fusion, as Chapter 2 showed. Copper also reacts with oxygen and hydrogen when hot in ways that can leave a weld porous or the metal beside it brittle, and its alloys bring their own problems, from the zinc that boils out of brass to the hot cracking of some bronzes.

Kautny’s chapter on copper contains one of the most durable insights in his book: that phosphorus rescues copper from its oxide. This chapter follows that thread to the modern grades of copper and fillers, and plans the repair of the cracked joint at the cooling coil’s header block.

Heat: the first obstacle

Copper conducts heat about 7.4 times as well as carbon steel and spreads it about eight times as fast. Kautny’s copper melted at 1,084 °C, the same figure used today. For a thin tube, the conductivity is manageable, because there is little metal to carry the heat away. For a thick section it dominates, and TWI’s guidance is that sections above about 5 mm need preheat for a fluid weld pool and to avoid fusion defects, with thick components needing as much as 600 °C. Chapter 2 showed that preheating the 12 mm header block to 400 °C needs only about 267 kJ, about the same as a steel block of the same size, but that the heat must be delivered fast and to the whole block, or the block acts as a sink and the joint never fuses.

Oxygen and the burnt copper

Kautny described what happens to copper heated in air. It combines readily with oxygen, at first on the surface and then within the metal, forming copper oxide. Copper so affected, he wrote, is said to be burnt: its fracture changes from a light salmon red to a dark brick red, a change he noted could be seen across badly brazed joints. He warned against letting the flame’s bright inner cone touch the molten copper, and that a flame with surplus oxygen burns the gases completely, leaving steam that releases oxygen into the copper. The neutral flame, and protection of the pool from air, were therefore essential.

He also described copper’s capacity to absorb gas, especially hydrogen, when molten, and to release it as it solidifies in bubbles that burst at the surface. If the surface freezes before the bubbles escape, they remain as pores. Copper founders, he noted, knew the phenomenon well.

Kautny offered a practical rule for avoiding burnt copper. An experienced welder, he wrote, can tell by the brightness of the heated area when the metal is about to combine with the air, because it turns a dazzling white. If the joint can be united at a temperature just below that, immediately below the melting point, burning is avoided. Several firms of his day, he noted, had built methods for welding copper and its alloys on exactly that principle: clean the parts, lay them together, heat them with the flame to just below melting on a heated support and consolidate them by hammering. That is a solid-state weld, the idea Chapter 10 returns to.

The modern reading of his rule is that every extra degree and every extra second at high temperature gives copper more opportunity to take up oxygen and hydrogen. Efficient heat input, a short time at temperature and protection of the pool, whether by flux, filler deoxidants or inert gas, all serve the same end he described.

Tough pitch, deoxidised and oxygen-free copper

Modern coppers are classified partly by how they deal with oxygen, and the classification matters greatly for welding. TWI’s job knowledge article on copper distinguishes three groups.

• Tough pitch copper contains a small amount of oxygen, less than about 0.1%, present as stringers of copper oxide. It has high electrical conductivity and good mechanical properties in the wrought condition, which is why it is widely used for busbars and electrical work. In fusion welding, TWI notes, it suffers embrittlement of the heat-affected zone and porosity in the weld metal. The oxide and hydrogen form steam inside the metal.
• Phosphorus-deoxidised copper has had its oxygen removed by phosphorus and is more weldable, although TWI notes that residual oxygen can still cause porosity in welds made without filler, especially where hydrogen is present.
• Oxygen-free copper is also easier to weld.

Figure 1Copper grades and fusion welding

	Oxygen	Fusion welding
Tough pitch	up to about 0.1% as copper oxide	HAZ embrittlement and porosity
Phosphorus-deoxidised	removed by phosphorus	more weldable; use deoxidised filler
Oxygen-free	very low	easier to weld

Kautny’s phosphorus

Kautny explained why phosphorus helps, a century before the grades were standardised in their modern form. Phosphorus has the particular ability, he wrote, to reduce copper oxide back to copper, and makes the molten metal flow as freely as water. In welding, it is introduced through copper filler rods containing a small amount of phosphorus, which he limited to no more than 0.05%, because more makes the copper brittle. The phosphorus reacts with the copper oxide to form a compound that spreads over the pool as a thin, tough film, protecting it from the air, so that for a skilled welder the filler itself replaces the flux. He still advised beginners to use a flux as well.

That is essentially the principle of modern copper fillers. TWI notes that residual oxygen in phosphorus-deoxidised copper can cause porosity in welds without filler, and that the remedy is a filler containing deoxidants. Its table lists Cu 1897 and Cu 1898 fillers for coppers. Kautny’s phosphorus has been joined by other deoxidising elements, and inert gas shielding has largely replaced the protective film, but the idea that the filler must clean up the oxygen is his.

Brasses and bronzes

Copper alloys behave very differently from pure copper and from each other.

Brasses, alloys of copper and zinc, lose zinc when welded. Zinc boils at a temperature below the welding temperature of the alloy, so it volatilises from the pool, producing white zinc oxide fume and porosity. TWI considers only low-zinc brasses fusion-weldable and recommends a zinc-free filler, Cu 6560 or Cu 6180, with high travel speed; it does not recommend fusion welding of high-zinc brass of about 40% zinc, nickel silver or leaded gunmetal. Zinc oxide fume is a health hazard, and the work needs the ventilation and protection that Safe Work Australia’s model code of practice on welding processes describes.

Bronzes vary. TWI identifies silicon bronze as probably the easiest to weld, advises against welding phosphor bronze without a filler because of porosity, and regards gunmetal as unweldable because of its susceptibility to hot cracking. Aluminium bronzes and cupro-nickels are generally weldable with suitable fillers, and cupro-nickels need no preheat.

Figure 2Copper alloys and fusion welding

	Main issue	Typical approach
Low-zinc brass	zinc fuming, porosity	zinc-free filler, fast travel
High-zinc brass, about 40% Zn	heavy fuming	fusion welding not recommended
Silicon bronze	few	probably the easiest bronze to weld
Phosphor bronze	porosity	weld only with deoxidised filler
Gunmetal	hot cracking	generally considered unweldable
Cupro-nickel	conductivity close to steel	no preheat required

Kautny’s alternative for brasses and bronzes was a hammering method: heat the parts to just below melting on a heated anvil and work them with small hammers, about 8 mm square on the face, so they weld in the solid state without the zinc boiling off. He also described brazing, which he called hard soldering. Brazing remains the preferred joining method for many copper alloy parts, because the filler melts well below the parent metal and the alloy’s troublesome elements are never boiled out.

Welding, brazing or braze welding

For any copper joint, three options exist. Fusion welding melts the parent metal and gives the strongest joint with the highest service temperature, at the cost of the heat and metallurgical problems above. Brazing joins with a filler that melts below the parent metal and flows into a close-fitting joint by capillary action. Braze welding uses a similar low-melting filler but builds it up in a groove like a weld. Brazing and braze welding avoid melting the copper and so avoid most of the problems of heat, oxygen and zinc, but their joints are weaker at high temperature and the filler’s corrosion behaviour must suit the service.

Figure 3Three ways to join copper

	Parent metal melted?	Main trade-off
Fusion welding	yes	strongest joint; heat, oxygen and zinc problems
Brazing	no; capillary joint	needs close fit; weaker at high temperature
Braze welding	no; built-up groove	tolerant of gaps; filler corrosion must suit service

Working safely with hot copper

Copper work brings hazards beyond fume. A header block preheated to several hundred degrees looks no different from a cold one, and a burn from touching it is a common injury; preheated parts should be marked and handled with tools. Copper and copper-alloy fume, and above all zinc oxide fume from brasses, need local exhaust ventilation and, where required, respiratory protection under the controls in Safe Work Australia’s model code of practice on welding processes. Fluxes used for brazing and gas welding give off their own fumes and leave residues that must be cleaned off, both because they can be corrosive and because they hide the joint from inspection.

Testing copper joints

Copper joints in fluid systems are usually tested by pressure or leak testing, which finds through-wall defects but not porosity or embrittlement that has not yet broken through. A bend test on a trial weld made with the same procedure, material and filler shows whether the heat-affected zone has been embrittled, which is the test that would have caught the tough pitch failure described below before it reached a customer. Where the joint is critical, radiography or sectioning of a trial joint checks porosity. As with every metal in this book, the trial piece is cheaper than the failure.

The coil header repair

This is an illustration. The cooling coil is phosphorus-deoxidised copper, confirmed from the coil maker’s data. The crack is at the joint between a 15 mm, 1 mm-wall tube and the 12 mm thick header block, caused by vibration. The original joint was a fusion weld. After the failed first attempt described in Chapter 2, the workshop’s plan became:

1. Remove the cracked weld and prepare the joint, cleaning back to bright metal.
2. Support the header on insulating firebrick, not the steel bench.
3. Preheat the whole header block evenly to about 400 °C, within TWI’s range for thick copper, checked with a contact thermometer because infrared readings on bright copper are unreliable.
4. Weld by tungsten inert gas with a deoxidised copper filler from TWI’s listed types, directing the heat mainly at the block and letting the thin tube follow.
5. Support the tube against vibration with a clamp, so that the cause of the original crack is addressed as well as the crack. A repair that restores the joint but leaves the vibration in place would simply crack again at the same point.
6. Pressure test the coil and inspect the joint.

A failure: the wrong copper bar

This is an illustration. On a similar coil a year earlier, the header block itself had been replaced, machined from a length of copper bar found in the workshop’s offcut rack. After welding, the coil failed its pressure test with a leak beside the weld, and a section showed porosity in the weld and a brittle, cracked band in the header metal next to it.

The bar turned out to be tough pitch copper, an offcut from an electrical busbar. Its oxide had reacted with hydrogen during welding, exactly as TWI describes, embrittling the heat-affected zone and making the weld porous. The fault was not the welding but the material: copper bar looks the same whatever its grade. The workshop now buys header material against a stated deoxidised or oxygen-free grade with a certificate, and marks offcuts with their grade, so that the rack can no longer supply an anonymous piece of copper. Where a grade is in doubt, a short bend test on a welded scrap of the same bar takes minutes and shows at once whether the heat-affected zone embrittles, which is cheaper than a failed pressure test on an assembled coil.

Exercise: a hotter preheat

The header block weighs about 1.8 kg. Calculate the heat needed to raise it from 20 °C to 600 °C, TWI’s upper figure for thick copper, using 0.39 kJ/kg·K, and compare it with the 267 kJ needed for 400 °C. Then explain why a higher preheat is not automatically better for this joint.

Heating to 600 °C needs about 1.8 × 0.39 × 580 ≈ 407 kJ, about 53% more than the 267 kJ for 400 °C. A higher preheat makes fusion easier, but the header is joined to a thin tube that overheats easily, the hotter block takes longer to bring up and cool, copper oxidises more readily the hotter it gets, and the welder must work close to a very hot part. The right preheat is the lowest that gives a fluid, fully fused pool for this section and process, established by a trial weld and written into the procedure, not the highest figure in the guidance.