|Presentation to DIS T & O meeting
Daniel A. Mayton
June 3, 2003
I spoke about the process of identifying and correcting the most nefarious problem we ever encountered in all the years we have been making Ductile Iron at Urick Foundry.
Urick Foundry was started sometime around 1900. It was a Gray and Ductile iron foundry that made large motor castings until 1985, when it was converted to a Disamatic, 100% ductile iron shop. Presently the melting is in a water walled cupola with channel induction holding. The pattern tooling is plastic and the treatment method is "Delayed Inmold". This method was used because of its simplicity, ability to use automatic pouring, lack of fade, and the high elongation properties due to the late inoculation. Our parent company is Ridge Tool, Elyria Ohio, a subsidiary of Emerson Electric, but we also make castings for several other
customers. The primary trade customers are railroad, truck hardware, valves, and electrical insulators. We also do some work with recreational vehicle hitches, dentist chairs, ATV parts, and compressor crankshafts. Our castings are finished in cells and shipped quickly since the finished goods storage space is limited. We use an 92 AFS base molding sand, primarily for the finish on the pipe wrenches, but also on other castings where cosmetic finish is important. Much of our product is Austempered, and we also coordinate machining and painting with local shops in Erie for those customers who have those requirements.
When we started to see changes in our BHN numbers, we wondered what was going on in our process. It had been running well since 1985. The "soft iron" was a problem that hit us like a tornado and left us stumped.
On February 13, 2002 we saw the first symptoms. I received word that the pearlitic ductile iron we were producing was below the hardness requirements. What normally gave us BHN's of 229 to 255, began to be about 217. We immediately quarantined the castings and checked the chemistries to see what was wrong. The Copper content was correct, and all the other elements were within range. We looked into the Boron content, realizing that back in 1986 we had a similar problem when we had used briquetted cast iron materials. This melt stock was supposed to be made from chips from our own castings. However, we found that the material sent to the
briquette operation was mixed with some Malleable Iron turnings, which had boron in them. Consequently we had a boron buildup in our metal. We stopped using the briquettes and the problem immediately disappeared. The cost was some heat-treating, but the problem was solved quickly. This time, we saw no evidence of Boron in our chemistry so we were stumped. We immediately contacted our alloy suppliers for help, sent samples from castings before and after February 13, and waited for the results. In the meantime, we tried several solutions such as increasing the copper level, with no change. We finally resorted to alloying with Tin and that produced the hardness and
microstructure we needed. It was the first time we had used Tin and had no comfort level with it. According to the Ductile Iron Handbook, the pearlite stabilizing influence of Tin is at least ten times times that of copper, and six times that of chromium, but even it had no effect until we used more than 0.04%.
When the results of the initial investigations from our suppliers turned up without a clear answer, we remained stumped. Their chemical analyses did not show anything amiss, and the Boron contents were practically non-existent. Actually, in one report, the Boron content was higher in the "good"" sample. A literature search produced an article that looked familiar, it was written by Lyle Jenkins in Ductile Iron News, 1995, and talked about Boron causing soft pearlitic iron. He stated that it was caused by new linings in electric furnaces, by wrought iron base alloys, by Boron treated steels, by Malleable iron, and by cast Cobalt
based superalloys. He stated that Boron is difficult to measure in a spectrometer because the spectral lines of sulfur and boron are very close and sulfur can affect the Boron reading. He suggested a need for research to determine if Boron, causing soft castings, also causes reductions in toughness and fatigue strength. About that time, I ran into Professor Carl Loper at an FEF meeting and discussed the problem with him. He told me Boron was the culprit, and to repolish my micros and look for tiny nodules which would be the result of Boron Nitrides. I told him I would but did not really think that was the answer. I subsequently called Professor Wallace and discussed the
problem with him. I asked if there was anything else beside Boron that would cause these problems. Initially I thought our iron was either too clean, since we use about 10% Pig Iron, or some other element was creeping into the melt that we didn't analyze for. He told me that Boron was the probable culprit, but that I should analyze all the materials going into the iron and see if anything had changed. We began an investigation of all the materials, but that took some time to complete.
In early April I called AFS and talked to Norm Bliss, the Technical Director, to see if AFS had any information of what may be the cause. He said he would do a search, but that he had received numerous calls from others having the same problem. We had a meeting at the AFS Casting Congress and hash it out. When I entered the room that Norm scheduled for the meeting and was surprised that it was packed. Several professors, many foundry men and suppliers were all there. It seemed incredible that we all were experiencing the same problem. I remember Bill Powell of Waupaca foundry saying he had never experienced "such an odd phenomena".
He stated that he had to "junk up the iron with tramps" to get the Brinell hardness he needed in his pearlitic iron. The result of that meeting was that there was in fact something in the steel we were buying, probably Boron, which had to be the culprit. We were all going to go back and check our sources and see what we could find. I also learned that this phenomenon occurred in both Gray and Ductile irons, in both Cupola shops and Electric melt shops, and was happening from Kansas to Pennsylvania and beyond. We did however; open up a dialog that was very helpful. It seemed that Boron could indeed be the culprit and that we needed to take a fresh look at our
In early May, we began an in depth study of all the possible changes that might have occurred at Urick in the previous months. Anyone involved with melting was interviewed, as were our suppliers of steel and alloys. We listed several possible causes: the pig, the steel, the sand, the alloys, the cores, and the inoculants. We sent each out for analysis for Boron. We also started to take another look at the Boron content of the pearlitic ductile and the ferritic ductile.
We noticed for the first time that the Boron content of the pearlitic ductile was higher than that of the ferritic ductile. We thought that was strange since all the materials going into the charge were the same, except for the copper. We filed this away for further study. In the interim, we started receiving Boron analyses from the outside labs. Finding nothing, we began to really look at our chemistries, downloading all the numbers to an excel worksheet to allow us to graph them.
Finally, about May 7 or so, we discovered that we had an increase in Boron over time. When we graphed all the data, there was a peak in boron coinciding with the times we ran pearlitic ductile. We found out that we were standardizing with a different standard, because of the copper, and that the system did not recognize Boron in our standard iron. Since we ran pearlitic iron once per week at most, we had not recognized the trend, until it was separately graphed. See spreadsheet graph of B concentration by date.
When we analyzed the historical chemistries for Boron, we noticed that the jump was a very small one. We went from about .001% to .002% with our standards and on our spectrometer. This seemingly small jump however, is exactly where the problem occurs for us. Below .0012% or so the iron seems to act normally. Above this number, the effect is enormous, until you reach about .01% where the effect reverses itself and Boron Carbides form causing loss of elongation, machining problems, etc. This small amount can come from anything; refractory material, steel, alloys, even the bark of trees used to make paper bags that we use to add alloy. We even
wondered if the Boron Nitride disc in the spectrometer was the culprit. The graphs showed that clearly this phenomenon occurred over time and had a distinct jump on February 13. The problem with trying to discover the source of Boron is the variation in results from one lab to another. We checked steels, both at our lab, at our source lab, and at outside labs. Most of the time the results were inconclusive, but over time we found a pattern emerging.
We found that a certain type of slitter scrap seemed to have more Boron in it than punching scrap. We also found traces of Boron in other materials such as Copper, discs we use in our inmold process and in the FeSi and FeMn we use, and even in the CaC2 we use to desulfurize. All these however, were so low that they could not affect the overall Boron chemistry when one looked at their total contributions on a mass balance level. We even looked at our sands, cores, and outside purchased cores, only to find the same thing. We needed something significant, something that was high in Boron, and something that was a large part of the charge. Until
we found the main source, we started controlling where we bought our steel and by the end of May, we noticed a drop in Boron content of about .0005% to .0008% per day. This was encouraging, but we realized that since Boron was in our sprue and that we only used the pearlitic sprue once per week, it would take a long time to flush it out of the system. Things progressed and by the end of June, we were finally in a position to stop adding Tin. We did notice that our Boron levels plummeted when we changed SiC suppliers. We didn't know if this was significant or was caused by a water leak in the cupola, but we were at the .006 % B level, a place we hadn't been since February.
The results were short lived however, as Boron came back with a vengeance in mid July and August. Finally, on August 16, six horrible months after the start of the problem, we found that one day (while we were running low Boron levels), the Boron spiked. The concentration went from .0005% to .0011% and then to .0032%, all in two hours span.
Fortunately the lab and the melt department were looking for anything like this and they immediately went to the Cupola and discovered a lone skid of SiC that was left over from the previous month was brought out and used during the night. We immediately stopped using it and the Boron returned to normal. We had our culprit, and all that was left to do was find out why the various SiC sources varied in Boron Content. Our suppliers analyzed their components and found that the source appeared to be a filler used in the SiC. It was made from crushed SiC crucibles, which contain Boric Acid, the same material that causes problems in recently
sintered coreless furnaces. One of our suppliers went so far as to analyze all his SiC materials, both domestic and Chinese. He found that the Chinese 50% SiC had the highest value, while the 88% had the lowest. He also found that the domestic material was low, both the 45% grade and the 85% grade. He found "Plate scrap" to be the highest (.0059%) vs. Chinese 88% and domestic 45% to be the lowest at .001%. Our own analysis from several outside labs showed similar results, although results often varied widely. In my last discussions with our current supplier, I suggested that they market a "Low Boron" grade of SiC similar to "Low Sulfur"
graphite. They said the cost for analysis was high but they would think about it.
Although Tin can is used to increase hardness, it must be handled with care. Tin causes a decrease in machinability at the same hardness value, and a drop in impact resistance. It is important to discuss this with customers so they know what the options are. Both heat treatment and Tin additions affect the overall costs negatively, and must be used only when necessary.
Conclusions and Recommendations
- Be sure that your Boron channel on your spectrometer is accurate. Round robin test it against other labs to be sure it is giving you the information you need.
- Discuss SiC chemistry with your suppliers and determine where they are with Boron. You may have to use several independent labs to find one you can trust. Work with your suppliers to get the lowest Boron content materials you can. Higher Boron SiC can be from the source as well as additional ingredients added.
- Work with your steel suppliers to find trusted sources. Avoid spot buys and short-term suppliers. Test all incoming steel. Watch not only for Boron, but other tramps as well. Titanium, Manganese, and Chrome are other tramps that frequently occur. Add Boron to the list of elements you chart. If Boron climbs above .001%, take action to reduce it.
- And a final note; until a magic pill can be found to neutralize Boron, the only practical way to deal with it in Ductile Iron is to eliminate it. It has been found that increasing the Titanium content will tie some of it up, but this doesn't work well in Ductile. The base Titanium in your iron should be a known value however, because changes in Pig iron may alter long-standing relationships with Boron and Titanium in your base iron, negatively affecting the hardness values. If the Titanium is lower than normal, Boron will go after Nitrogen, forming Boron Nitrides, increasing the count of very small nodules, thereby making the iron virtually
self-annealing. The answer seems to be total control of two of the most important raw materials in the system, Steel and Silicon Carbide. If either one begins to show increasing amounts of Boron, take immediate action to eliminate it.