Tag Archives: Trials

Internal Oxidation of Cu-1%Al Alloy

This project came from work that prof. Houbaert had been doing for Pepsi in Mexico before he joined prof. Dilewijns’ laboratory as the professor for physical metallurgy. This involved improving the wear resistance of copper mounds used for the production of glass cola bottles.

The idea was that you enriched a copper melt with a touch of aluminium to the tune of 1%. After you cast the alloy into the correct shape (in our case that was a mere solid cylinder) you then exposed it to an oxidising atmosphere which converts the aluminium (but not the copper) to its oxide and thereby makes the surface layer of the copper shape more wear resistance.

I was tasked with getting all the equipment together, consisting of two tubes inside their respective furnaces, the first one containing Cu2O which at the correct temperature dissociates to CuO and releases oxygen gas into the argon-nitrogen carrier gas. The second furnace contains the Cu-1%Al sample to be oxidised, and the combination of the oxygen content of the carrier gas and the temperature of the second furnace ensures that the aluminium oxidises but not the copper.

I then ran several trials, first of all to see whether the concept worked as intended (it did), and secondly to see what the effect on the properties would be. This involved hardness profiles, metallography and SEM analysis, and in my opinion I had covered sufficient ground in preparation of a thesis student to have a running start in the coming academic year.

The only thing that went somewhat awry was when we were trying to remelt the alloy, hoping to economise on raw materials, and during the remelt process smoke started to appear from the back of the furnace. We quickly tipped out the contents of the furnace, and discovered that the aluminium in the copper had started to eat away at the basic lining -we were just in time to avoid the melt from spilling on the shop floor.

Unfortunately prof. Houbaert hadn’t realised that by the time the thesis student would start on this project, I would be in the army, meaning that the student didn’t have the guidance of the person who had done all the advance work. At the end of all this I received a copy of the student’s thesis and was disappointed that it hardly covered any ground that I hadn’t already done in my preparatory work.

Apart from that this could have been a job that I could have been proud of. As it is it felt a bit like an opportunity missed. When I rejoined the laboratory in 1983, there was no sign of this project, and I can only assume that it had been quietly dropped in the meantime.

Internal oxidation set-up


Hand Beader

At Tinplate R&D we used to have a hand beader on loan from Impress. Before I continue I should explain what beads signify in connection with food cans: they’re the sort of ribs you see indenting the cylindrical body of a food can. Their main function is to give the can body some degree of rigidity when it goes through the heating and cooling cycle when the food is being cooked inside the can. We did have an MB80 industrial beader to give our cans a standard bead, but if you want to investigate the effect of different bead geometries on the can properties, then you need to have a process where you can apply a variety of geometries rather than one standard one. That’s where the hand beader comes in.

Anyhow, to get on with the story, at some stage Impress said they wanted the hand beader returned, which left us without the means to implement our research programme just as we were about to launch into a extended study of how different bead profiles affected the can properties. Fortunately we had the drawings, and knew what needed to be done to produce a copy of the machine that had gone walkies.

The only thing we now had to overcome was the inertia of the British Steel ordering system, where the concept of approved suppliers had been introduced fairly recently, and which could have set us back several months unless we managed to find a back door solution. A good thing that Norman Leah knew of a small workshop in Ammanford which was an the approved supplier list, and was in the line of work that could supply us with most of the framework. Somehow along the line I had the impression that Norman might know the person in charge and was pushing him some jobs whenever he could as a favour.

Still, that was only a vague impression, and if it helped speed up the project, who was I to complain or dig any deeper for possible ulterior motives. Another way of getting things done quickly was to buy standard parts with cash, which could then be reclaimed from expenses. Anything to avoid going through a system of getting three approved suppliers, get a quote from each of them, and then leave it up the purchase department to come up with a decision, which might or might not take several weeks.

In the end we managed to have the hand beader rebuilt in about a month, from the moment the original one was returned until the moment that we had a fully working replacement. Basically it put our research programme back on track with only a minor delay, whereas it could otherwise easily have been derailed. It was a very satisfying moment, a rare occurrence during my time when I worked for David Jones.

Slab Stacks and Hydrogen

In an earlier blog I described how I failed to recognise the effect of a gas shroud whilst pouring on reducing the hydrogen content of liquid steel (Hydrogen Shroud). There was, however, another experimental part to my hydrogen project, which was aimed at assisting the Finite Element Analysis (FEA) calculations.

You see, in order to be able to calculate the diffusion of hydrogen out of a slab, I had to have some idea of how fast the slab would be cooling until for all intents and purposes no more hydrogen diffusion would take place. This was fairly simple for a single slab surrounded by and atmosphere at ambient temperature: all it required was to perform an FEA calculation for temperature first, and then use the calculated temperature at a given time to calculate the speed of the hydrogen diffusion.

Another method for assisting hydrogen diffusion from slabs was to place them in soaking pits normally used for reheating ingots prior to slab rolling. In this case knowing the slab temperature was simple, since this should merely reflect the temperature of the soaking pits themselves.

There was, however, one other method that I investigated as an alternative to the use of soaking pits, and that is to place a small number of critical slabs inside a stack, surrounded by other slabs once they had come off the slab mill and still were close to white heat (around 1100°C if I remember correctly). This way, the cooling speed could be reduced substantially over that of a single slab.

The way I measured the temperature was using a thermocouple placed inside a stainless steel hollow tube, which I positioned on top of one slab (with the aid of a long metal bar, which also prevented the thermocouple from being crushed) as the stack was being formed, and then had it covered by the subsequent slabs in the stack. On the other end the thermocouple fed into a data logger, which was placed inside a changing room locker which was shielded from the radiating heat from the stack by some cardboard.

All in all this worked fine, but to stand within 5 yards of slabs at 1100°C was far from comfortable, even when wearing protective gear. In the end I did two different configurations for stacking the slabs and called it a day after that, thinking that they hopefully would be representative for this type of slab cooling. In the report I made it seem like I had performed several more stacks, but had chosen the two that I had actually performed as representative samples.

Getting the data went without a glitch, and I had my temperature trace for calculating the dehydrogenation of the slab. As expected it worked better than a single slab, but not so well as slabs placed in a soaking pit. Besides, the formation of the stack would be depending on how the operator built it, whether there were any delays at the slab mill whilst the stack was being built, and on top of that the cooling would be affected to some extent by the ambient temperature, which during the winter nights could go well below freezing.

In short, the soaking pits allowed a far better control over the slab temperature, and that was then my advised method for reducing the hydrogen content of a critical slab. In addition, someone asked me whether giving ingots a similar soak would help – I knew this wouldn’t be the case given the far higher volume-to-surface area ratio of an ingot, but did the FEA calculation anyway to prove the point. Besides, since an ingot is a cast structure it would contain a lot more spaces between the crystals, and at times complete gas bubbles, which would trap the hydrogen and completely scupper the diffusion of hydrogen out of the ingot.

To be honest, a thought experiment probably would have come up with a similar answer, without the need for any experiments, but being able to show actual temperature traces and calculated hydrogen content graphs was somehow far more convincing to my audience than if I had merely stated “trust me, I know what I’m talking about”.

When Is a Project Finished?

That was the question asked in all earnest when I was at Iscor’s Steelmaking Technology department. The situation was thus : for every major project, you had to write a project brief, stating the initial problem setting, the desired outcome and the steps proposed to achieve the latter. So far so good.

However, there must be something especially bureaucratic n the Afrikaner mind, because once you had stated that you were going to take steps 1 to 14 to resolve a problem, you were expected to complete all those steps. Never mind that it’s not always clear in advance what steps will have to be taken – you could, for instance, stumble upon an intermediate finding that looked so promising that it took on a life of its own and generated a whole new set of action steps. Or a finding quite early on could make it clear that steps 6 to 9 are no longer required because of something you’ve learned in step 5.

To the Afrikaner mindset the first instance, where one action step generates a whole new subset of its own, would require you to write a whole new project brief, in essence creating a completely new project. In the case of the latter, it is more likely than not that steps 6 to 9 would still be undertaken, or at least you would have a fight on your hands convincing your boss that it was indeed a good management of resources to drop them.

But the instance I’m thinking of was a project on the cleanliness of steel, where one action step made such a difference that the goal stated in the project brief was already achieved. This is what caused the question to be asked : now that we’ve achieved the aim, do we continue pouring resources into this project, or do we cut it short ? I can’t quite remember what exactly was decided in the end, but I suppose you could make a case for both types of action.

Cutting the project short once you’ve achieved your aim frees up finite resources in manpower and time on the production lines for other projects. Whereas continuing your project as originally planned makes it possible (at least in principle) that you’ll discover further means of improving the steel cleanliness.

In a way, the hydrogen project I described in an earlier blog was a victim of this type of box ticking. I had done my trial attempting to enrich the steel with hydrogen, had failed, and that was that action step taken care of. If anybody at this stage had said “hey, that’s funny …” maybe further steps could have been added to what was officially a completed project brief, but somehow that opportunity was overlooked.

Presumably as much my fault as the system’s (after all, I had by then started working on inclusions in DWI tinplate, and had lost interest in the hydrogen-in-steel issue), but a project leader more inclined to follow his nose rather than completing the paperwork might just possibly have found an answer to a question that hadn’t been asked, which is “How do you keep the hydrogen content of your steel low in the first place?”

Salt Spray Test

As I may have mentioned in an earlier blog, the topic of my thesis was the investigation of weak acidic chloride solutions for electrogalvanising a steel substrate, and examining the respective corrosion behaviours of zinc layers produced using different coating parameters. I also used different types of chromate treatments in an attempt to improve their corrosion resistance.

So far so good, but how to evaluate how well the zinc coating was performing ? Obviously there was the possibility of a visual examination, and cross sections could be made to see how the coating thickness varied across the sample. But most important of all was the sample’s corrosion behaviour, and this was evaluated using the laboratory’s salt spray test chamber.

The idea was that different samples with varied coatings and chromate treatments would be hung up in the test chamber until first white rust (i.e. corrosion of the zinc) or red rust (corrosion of the underlying steel substrate) occurred, or until a predetermined maximum time had elapsed. For some of the better chromate-treated samples the test duration could last as long as 320 hours (about 2 weeks).

Obviously, the test had to be interrupted at regular intervals in order to examine the samples, and here I must admit that I had my doubts whether I performed the test properly : you were supposed to wash the samples clean prior to re-inserting them in the chamber. I’m fairly certain I didn’t do this. I also don’t remember whether I discussed the test with my supervisor in order to make sure that I followed a standard way of operating the equipment.

I do remember having a discussion with my predecessor, who had performed similar tests using alkaline solutions, plus I had the benefit of being able to look up stuff in his thesis.

Still, the literature states that the salt spray test should not be used to predict actual corrosion behaviour in a natural environment, but instead compare the efficacy of different samples under controlled conditions. I suppose that’s what I did, and since I used the same methodology throughout, I suppose my data must have had some meaning, if only through its internal consistency, but comparison with other data sets might be more problematical.

To be honest, I never had any feedback from the company who supplied the chemicals for the weakly acidic solution or the chromate treatments, so not really sure how much practical use the outcome of my thesis had.