Tuesday, April 15, 2014

Alchemy in a Modern Age

Alchemy has long been pursued by humanity in the hopes of transmuting common metals into precious gold. However, alchemy can be seen in the broader sense as the inexplicable transmutation of something common into something unique in its properties and the mysterious process by which it arrived there. The use of fire to turn rocks into steel is about as close as anyone has come, and perhaps even in the ancient times was revered for its intrigue. Even now, hundreds of years after the first instance of this incredible transformation, the mysteries that surround it are no less spectacular.

A third method of producing steel, iron, or cast iron (the first two previously discussed being bloomery smelting and hearth melting) is the reduction of material in a crucible. Wootz, a term which has come to mean a few different things in recent years, has become synonymous with crucible steel. Anyone interested in the subject should read this article by J.D. Verhoeven et. al. which describes much more than my own knowledge can try to explain. In the modern sense of the word, wootz has come to mean a high carbon (>1% C) crucible steel that displays dendritic patterns of banded carbides. This is a result of the chemo-physical process that melts the steel and what happens to it while it cools.

I again recently had the privilege to witness Jeff Pringle's 'atomic marshmallow' that he uses in the melting of crucible steel, and now that I have a better understanding of what is happening, I will detail the process here. Wootz holds a special beauty, and the fundamental simplicity behind the process (although it takes a far deeper understanding and years of experience to do this reliably and predictably) encourages me to try this on my own.


Creating a simple melting furnace is as easy or complicated as you want to make it. However, in its most reduced form, it can be as little as a blanket of kaowool bound with wire, resting atop a few firebricks. Here, Jeff uses an old acetylene tank as a form to size the volume of the chamber, wrapping the blanket around twice so it is 2 inches thick. The important part of this is to have the inside end curl in the direction of the fuel/air inlet flow so it is not forced out of place and the walls separated.


A little extra kaowool on the side that bears the brunt of the blast helps it from forming a thin spot. In the above picture, it is easier to see what I mean by the direction of the blanket wrap. The inlet will be placed on the upper left side so the flow circulates counter clockwise.


This will be the third melt of the day. Into the crucible goes a collection of bloomery bits, hearth steel, and a little iron sand amongst other things.


Covering the surface is crushed charcoal. This helps absorb the excess oxygen that is not combusted with the propane as well as adding carbon to the metal.


Atop the charcoal goes a few handfuls of slag glass from previous melts. Once the glass liquefies, it creates a continuous surface over the metal, sealing it off from the environment. This layer prevents oxygen from migrating through to the steel and prevents carbon from escaping. Slag also helps collect impurities in the steel and bring them out to the surface where they are easily removed. Slag is mostly composed of silicates, and as a result forms a beautiful glass (which is more or less useless other than reuse in later melts).


Finally, new glass from a bottle is placed on top to increase the volume of the slag glass.

Into the side of the furnace is fitted a pipe, cut through the wool blanket so it is an inch or two off the bottom. This is connected to a propane tank and a blower, making a simple burner. Adjusting this to have a good fuel/air mixture is unique to every burner and every furnace, so it takes a little tweaking until it runs smoothly.


After the furnace is lit, the top is fitted in place. This is the upper 2~3 inches of the blanket cut into a strip and rolled in the fashion shown above. Wire pierced through the roll keeps it together, with a vent in the centre. All furnaces need to breathe, vice the back pressure and stagnating mixture will cause a dramatic loss in efficiency (if not cause it to fail altogether).



Once the steel has fully liquefied, the crucible sits for a few more minutes before disengaging the fuel supply. Periodically throughout the process, the contents of the crucible are stirred with a long iron rod (not galvanized or otherwise plated!!). It is easy to tell when the steel has gone liquid, and to feel any solid pieces left in the crucible.

Because the kaowool is rated to, depending on the grade, ~2300(F), the ultra high temperature melts degrade the layers if left unsealed. Some coat the inner chamber with castable refractory, although it is not strictly necessary if longevity of the furnace is built with acceptable losses in mind.


This is the crucible from the second melt (don't have any of the third) as it cools. The slag in this crucible began to rise and bubble nearly to the top, quite a strange sight to watch.


Once the crucible cools and the steel solidifies again, the slag glass can be chipped and broken away to expose the crucible steel below.

As it slowly dropped from over 1600(C)/2900(F) degrees, we cut the pucks from the previous two runs. Using an angle grinder to half the steel, breaking it in the middle to expose the natural grain, the results yielded confirmation that the two were both extremely solid, relatively homogeneous pucks.



Grinding these smooth, polishing the surface, and etching it reveals the dendritic patterns natural to wootz.

Photo Credit: Buffalo River Forge
These branching patterns are formed by the carbide clusters, which are more resistant to an etchant than the darker steel between, acting almost like a natural damascus (which uses two or more contrasting alloys of homogeneous steel [or layers of contrasting bloom/hearth steel, which is less homogeneous yet still capable of achieving contrast in the differential carbon/other alloying element] to create a pattern).


Finally, after a whirlwind morning of melting, here are the results of the three melts. The one on the back is the last, the two middle halves the second, and the frontmost half from the first. Once you witness the creation of steel, there is no doubt that it is truly alchemy in its purest form.

Thursday, April 10, 2014

The Fires of Ancient Industry

Long before the Bessemer process, steel was, and still is to some degree, made through bloomery furnace. I had the pleasure of observing a smelt run by Jesus Hernandez and Dennis McAdams through which a mixture of crushed taconite ore and hematite in the form of iron sand. After the fashion of ancient ironworkers, the only fuel used in turning ore into steel is charcoal and oxygen. Lee Sauder has a more detailed view of the furnace construction on his website here. However, below the furnace is a slightly modified design using a three part construction, sectioning the stack into 'rings' that are placed one atop the next for a more preserving means of removing the bloom at the end of the smelt.

A quick point of nomenclature. There are melts and smelts, both of which result in the creation of iron/steel/cast iron. However, smelts are the process of turning ore into metal whereas melts condense or refine pre-existing metals into a reduced, consolidated, or carbonized form. By this definition, the recent post on hearth furnaces are melts, not smelts. In the same line of thinking, bloom is only the product of a smelt, never a melt. Melts produce pucks or ingots (or similar words).

Now that that is out of the way, here is a dance with the fire gods.


Before the smelt can commence, the stack must first be pre-heated. This is usually done with brush, green wood, or anything that is not charcoal. The reason being is charcoal is more valuable than the generic fire wood. Also, green wood does not burn as hot as charcoal, so using it to bring the furnace up to temperature is a more efficient use of resources. The tuyere is also placed in a different location than it will be for the smelt. Notice the bricks on the right side of the stack. That is where it will later be moved to. It is on the front because that hole is lower, forcing the air up through every part of the furnace rather than coming in a few inches above the bottom (which allows for different chemistry inside the furnace).


While the stack heats, the charcoal must be prepared. Much of the work in a smelt happens before/after the actual reduction of ore into steel. Here, Dennis and Derick break apart bags of hardwood lump charcoal into smaller, more consistent pieces.

In addition to the charcoal, the ore must be prepared. To the left is the roasted, crushed ore. Only the first few charges were ore, the remainder being iron sand depicted below. The iron sand has also already been cleaned and sorted.

While the prep commotion is happening, the ore charges are carefully measured and weighed, and sorted for the smelt. How much ore can be fed into the smelter depends on the size of the furnace amongst other things. With the addition of the tuyere window, it is easier to see if the bloom is growing or has stagnated, and if the slag needs to be tapped.

Unfortunately (or fortunately, depending on how you see it), this smelt did not need to be tapped, as there was almost no slag at all.

Throughout the duration of the smelt, 46 charges of ore went into the stack, each weighing in at 750g  plus 4 or so at 500g each as the stack began to reach its capacity. A total of 36kg of ore was used. Paired with the ore, 1kg of charcoal in each charge made 50kg of fuel burned by the end of the smelt.



When the furnace has been burning for a few hours, the tuyere is moved to the actual position and supported with brick, sealed with clay. Before the first charge of ore, everything is sealed as best as possible, patching any holes around the three pieces of the stack.


Jesus Hernandez, one of the leaders in bloomery experimentation and research, is weighing more of the iron sands and carefully recording the weights added and the time between each charge. Mixing the charcoal and whatever is being fed into the furnace to make bloom from ensures repeatability of results. Alternating layers of charcoal and ore/iron sand are added to the top of the stack whenever the level burns low enough to add another.


As the charcoal burns down and the ore/iron sand heats, it liquefies and runs to the bottom of the stack, collecting carbon on the way. The depth of the stack, the time in the furnace after it reaches the bottom, and the heat at which it burns (the amount of oxygen pushed through the tuyere) are the three main contributing factors, and changing any of them can have dramatic differences in results.


Keeping the time between charges constant helps eliminate part of this uncertainty.


Dennis watches over the furnace as the charcoal burns down. For this type of furnace construction, the pieces can be removed sequentially. After the final charge of ore is introduced and the charcoal reaches the bottom of each section, it can be easily removed. This is important later when the bloom is removed from the stack.


As the furnace nears the critical point where the bloom can be removed, everyone gathers around for the initial consolidation. Hammering on the raw bloom helps break off slag, condense the relatively spongy structure of the steel, and reveal the character of the final product. Cast irons will be difficult to forge and crumble easily, while lower carbon steels will compact more readily.


When the stack has burned low enough, the bricks of the plinth are pulled carefully away. Instead of breaking the walls of the furnace, this design allows for the bloom to be pulled from the bottom and the lower third of the stack to be lifted away.


There is a critical moment when the last remaining coals flood out with the bloom in a deluge of fire. At night, this is a spectacular show rivalled only by tapping the slag that pours out in a winding river.


Guided out by a shovel and lifted onto a stump for consolidation, the bloom sits like a meteor.


Consolidation is an arduous process. Teams with sledge hammers and mauls strike like clockwork, compacting the unshapen mass into a more manageable shape. Bits of the bloom, faults and slag inclusions and regions of higher carbon content flake away, leaving behind a solid ancony (worked bloom).


Through the consolidation, it was immediately obvious that the bloom was closer to cast iron than mid to high carbon steel. The large pieces cracked easily and were reluctant to forge without splitting.


From large to small, most of the bloom showed consistent results.


In all, the smelt yielded just over 13kg of bloom, an enormous amount of material for the size of the smelter. This was the product of over 25kg of ore and iron sand.


A quick spark test by Mark Green shows that it is indeed a cast iron bloom. What will be done with it is anyone's guess, but it can be worked carefully as cast iron, or remelted and decarburized into a high carbon blade steel, used as contrasting layers for pattern welded steel, or a number of other things.

Saturday, April 5, 2014

Airfoil Flow Visualization


I recently had to the opportunity to do something a little different from what you might be used to seeing around here. As an aeronautical engineer by education, it is finally time to share some of that and take a break from the more traditional side of craftsmanship (but not for long!)

Although the pictures had to be modified heavily to capture the contrast of the flow, here are a series of visualizations of a NACA 0012 airfoil cross section using helium bubbles in the airstream. A quick background on what this means and why it is important. This particular airfoil (read, wing) is symmetric, meaning if you were to draw a straight line from the leading edge to the trailing edge, the two halves would be the same.

As you vary the angle of attack, α, a number of things happen. Without going too deep into what is happening aerodynamically, generated lift changes and the amount of air that adheres to the surfaces of the wing changes. At lower angles of attack, more air will 'stick' to the wing. The higher it goes, a turbulent section begins to form until virtually the entire wing fails to keep the flow attached, and thus generate significantly less lift.

For the photos below, the airfoil varies from 0 to 24 degrees (max on the armature), and in the end of the video changes from 24 to -8 degrees where you can see what happens and when the flow reattaches.

Unfortunately, several of the images needed to be taken at a slight angle to eliminate the glare from the wind tunnel window. In the video, the same angles of attack are used as in the photos.




α = 0 degrees



α = 4 degrees


α = 12 degrees


α = 20 degrees


α = 24 degrees