Theoretical Framebuilding Part 3: Metals and Heat
The finest steel has to go through the hottest fire.
-Richard M. Nixon
In my days working in science I had a supervisor that always said – do the experiment on paper FIRST and that means putting down the 5 minutes it takes you to walk to the spectrometer with the samples as well as anything else. He was right. If it didn’t work ‘on paper,’ it was NEVER going to work as you intended in practice aka it would fail miserably. This is the final of three part series on why and what makes a custom fit bicycle frame, different (and better) than ‘off the peg’ one. The posts are intended to shed light behing ‘hand picked, custom butted, special shaped tubing mix’ and all the terms you see and hear mentioned about custom bikes. This article is meant to educate and represents a successful experiment ‘on paper’ by an aspiring framebuilder (me). The science, metallurgy, physics and engineering are all researched, the opionions (as harsh as they may be) are all personal. The articles are ranked by order of importance. Part 1 discusses probably the most important aspect – geometry. Part 2 talks about tube forming and butting and its effect on “ride feel.” Part 3 delves in metallurgy and what heat does to metals and describes terms like heat treatment, aging, alloying, material stiffness and strength.
Fighter jets and cold war secrets and what the two have to do with custom bicycle frame building.
Chaos and Crystals
Metals exist in crystal structures with the atoms neatly arranged in a lattice. Each crystal is composed of many small crystallites called grains. Crystal structures are all neatly arranged, however, most pure metals are either too soft (ie pure gold), brittle, chemically reactive (ie rusting/corroding) or a combination of the three for any practical use. Therefore by creating some disorder or changing the neat crystal structures by adding small amounts of other elements, metal alloys can be created. Alloying elements can modify, sometimes, quite significantly mechanical properties such as making the original metal stronger, less brittle, much harder, and resistant to corrosion, or have a more desirable color and luster.
Cycling sidenote #1: Since this is cycling oriented article, I am using the currently available alloys from the major tubing manufacturers; this is done with illustration purposes only; I am not ranking anything or saying how one is better or worse than the other. Everything represents an option. The high technology previously reserved for high tech (military) applications is now widely available to the small framebuilder, making it a perfect time for a steel, aluminum or titanium custom frame.
On the molecular level alloying elements can either directly take place of one of the metal atoms creating a substitutional alloy or if the added elements are much smaller in size, they slot in between the metal atoms creating an interstitional alloy (steel is an example of iron/carbon alloy. A combination of both types of alloys can exist (stainless steel is one such example).
Bring on the Fire!
Joining metals involves heat. In the case of cycling there are 2 techniques – brazing and TIG welding. The former relying on a filler metal with a LOWER melting temperature than the materials being joined that creates a joint, while welding melts the edges of the two parts being joined and fuses them together (with the help of a filler material). Brazing is the ‘classic method’ that has many a great victories and pieces of cycling memorabilia to its name and it still employed today. It is a lower temperture and in soem cases the preferrred option. TIG welding is not a new technique, however, since it can change the metal significantly it was not available for cycling applications until some advances in metallurgy allowed it, such as air hardening steels that actually become stronger post welding – such as the Reynolds 631 and 853 series.
What advances and why were they needed?
Cycling Sidenote #2: Back in the day when Reynolds 531 was what was predominant alloy used for brazing frames, Reynolds released the first heat treated (more on that below) set for bicycle frames – the ultra-thin and light 753 series. Only selected framebuilders were allowed to work with it after certification, since it required a controlled lower temperature process to avoid overheating the already thin tubes. While thicker material allowed some leeway, overheating thinner materials was easier with more dramatic consequences ie significant loss in strength.
The takeaway point here is that heat is energy. Adding energy to neatly arranged crystals gives the atoms freedom (to move around) so in the words of the great physicist Richard P. Feynman atoms start to jiggle and at certain temperatures all the particles in the matrix undergo a change called an allotropic transformation; things get chaotic. The crystal lattice is not the same. The process of heating a metal above its recrystallization temperature is called annealing. Annealing makes metal softer, more workable – like the red hot metal piece that the blacksmith works with the hammer; it clears out any deformations (strains) within the crystal lattice.
When the metal is allowed to cool it recrystallizes back in a nice and orderly manner and if annealing (heating) is allowed to continue the crystals (grains) in the lattice start to grow large and coarse. Grain size depends on the rate of growth, which in turn depends on the rate of cooling. A fast cooling rate forms small grains rapidly and a slow cooling rate produces large grains.
Two things come out of the above. First, irregularities in the metals crystal lattice sometimes are desirable. Remember how I mentioned that introducing some disorder in the neat crystal lattice is a way to improve properties. Cold working/work hardening is one such way. What it does is literally banging the atoms into submission by rolling, folding, bending, etc. the metal – you add deformations. Steel can be cold worked/work hardened to get stronger. Once annealed you lose all that and hence the extra strength; don’t overheat cold worked steel.
Second, excessive grain growth causes the microstructure to coarsen (big irregular grains) which in most cases causes loss of strength and the grain boundaries can become stress risers where cracks can form (especially under cyclic loads, aka pedaling a bicycle). For the most part minimizing grain growth (by adding various elements) is a goal behind alloys that will see serious heat (ie welding).
Cycling Sidenote #3: Scandium, Zirconium and Niobium
Scandium in 0.3-0.5% is added to some aluminium alloys (called Scandium paradoxically…) to dramatically reduce grain growth at the heat affected welded area. First advantage, less stress riser for cracks to form as the joint is welded/cooled and subsequently loaded (aka hot cracking). Second, the alloy is stronger than regular 7005 aluminium (UTS 550 vs 450 Mpa (+22%!) and Yield Strength 510 vs 380 Mpa (+33%!!!), allowing for parts to be made thinner/lighter. Soviet MiG-21 and MiG-29 fighter jets were among some of the applications for ScAl when it was first made in the 1960-70s. This was a guarded Soviet Cold War secret, coupled to the fact that the major(and only) Scandium stockpiles were mined in Ukraine. The technology was ‘transferred’ by the Ashurst company (now non-existent) in the 1990s and Easton sports saw use in athletic equipment (baseball bats, bicycle frames) and struck a deal. The now discontinued Easton Sc7000 tubes were among the results of the collaboration.
Similarly, when another rare (though less rare/cheaper) earth material – Zirconium was added in concentrations of 0.1% to ScAl, you needed less Scandium to achieve comparable results in the welded area. 7xxx series ScAl has both Scandium and Zirconium added to it. The Columbus XLR8R aluminum tubeset (also now discontinued) was labeled as Zirconium…
In my opinion, ALL scandium that made it to the bicycle industry shared the same basic alloy (Russian classification 1970), whatever the marketing managers decided to call it…the reason being that no athletic company has the money to better the huge budgets of military and automotive/aerospace which are the primary users of ScAl. Nonetheless Sc+Zr, allows for thin walled aluminum tubes (down to 0.6mm!, as compared to 0.9-1.0mm) reducing total frame weight by 10-15%! while keeping overall strength the same, if not a bit stronger. One of the pioneers of butted aluminum tubing for bicycle use – Gary Klein, also developed a Zirconium alloy – the ZR9000. The Dedacciai U2 tubeset (now discontinued) was ScAl. Currently Dedacciai Adversa series is a ScAl tubeset available for purchase.
Niobium, is another potent grain refiner, though it is used in steel. Again paradoxically the alloys are called Niobium (see a pattern here). The Columbus Spirit, Columbus Life and SL tubesets all share the same alloy.
Result – much thinner walled tubing (down to 0.38mm !!!! as compared to typical 0.6-0.7 mm) and hence reduced weight (see another pattern here=) ).
Almost all metals can be annealed so that they are easily worked and ‘reset’ by annealing. Annealing is a heat-treatment. HOWEVER, only aluminum, titanium, copper, magnesium and steel can be made STRONGER by heat treatment and this is what is usually meant by the term. Steel is particularly a nice example to illustrate the principle of how it all works.
Steel is the alloy that you are most likely familiar with – it is formed by iron and carbon. Up to 5% of carbon it is called steel, while anything above that is cast iron. Cast iron is used for engine blocks and other strong and massively heavy stuff, etc. and cannot undergo some of the desirable changes that steel can as such I will not focus on it. Cast iron cannot be hardened
Heat Treatable Steel
What you see below is the phase diagram of steel. Don’t be put off. Steel is for the most parts iron and carbon, therefore as mentioned above the amount of heat/energy that goes into it causes the atoms in the crystal to undergo changes. At room temperature the carbon and iron of steel are heterogeneous – carbon exists in the form cementite and the iron in ferrite.
When heated to its annealing temperature the iron and carbon diffuse and form a solution called austentite (the diamond shaped part in the left of the graph), everything is fine and in equilibrium. If you let it cool slowly the carbon would precipitate out of the solution until we are left with the iron+ ferrite (aka pearlite) as before – two distinct structures. Annealed steel is soft and easy to work with. However if you drop the hot steel in water….
Quenching and Tempering
When quickly cooled (quenched) the carbon doesn’t have time to precipitate out and is stuck inside the iron crystals. This is called martensite – mega extremely hard, however brittle due to the internal stresses (the carbon was literally scrambling to get out, however, you shut the door).
The steel is hardened, however, brittle is rarely a desirable quality, so as you remember heating it again, kind of secondary annealing, and cooling it in a controlled manner increases the toughness – that second process is called tempering.
Temperature and Time
How fast and for how long you quench steel has profound effects on what happens with the crystal structure (it is not always just dunking a hot part in water….). Below is the time-temperature transformation curve (TTT) of steel. The important point is that there is a Martensite start (Ms) temperature at which martensite begins to form. If the steel is cooled very fast you get 100% martensite (green and yellow lines). If it is cooled slower you get a transformation to various crystal structures such as coarse/fine pearlite, upper/lower bainite, etc. For example, Reynolds 853 and 631 are both air-hardening alloys (pink line) that make the welding zone stronger. Remember that it is the disorder (ratio of the different crystal structures) and chaos you introduce to the ‘regular’ crystal lattice that gives you (un)desirable properties. By adding alloying elements you shift the transformation curves (black lines) as well as adding some features such as corrosion resistance.
It is the disorder (ratio of the different crystal structures) and chaos you introduce to the ‘regular’ crystal lattice that gives you desirable properties.
The addition of certain alloying elements, such as manganese and nickel, can make the austenitic (annealed) structure stable below 910C, facilitating further heat-treatment. In the extreme case of austenitic stainless steel, much higher alloy content makes the austentic structure stable even at room temperature – the Reynolds 921 alloy is one such example.
Precipitation Hardening Alloys
Most heat treatable metal alloys do NOT behave like steel. The reason is that it is the carbon atoms and the structures I described above (all the –ites…) that create the unique characteristics. With other metal alloys you still anneal and quench, however, what happens with the crystal structure is slightly different.
Aluminum is an example of such an alloy. When you anneal precipitation hardening alloys a solution of all the things in the alloy forms (a process not surprisingly called solution heat treating), just like in steel. When quenched, however, the alloying elements do not do crazy stuff (form martensite) like carbon. You are left with a solid state solution at room temperature. Over time the alloying elements slowly precipitate, however, stick close to the base metal crystals, the structures are almost indistinguishable (unlike with steel). This is called aging and it happens even at room temperature, the alloy does get harder (age hardening). As usual you can also age at different (higher) temperatures (artificial aging) and the crystal microstructures change ever so slightly. Just like with steel tempering, the aging processes makes the alloy stronger. When done at higher than ambient temperature it is called precipitation hardening. Aging and tempering accomplish the same thing, though with precipitation hardening alloys it requires prolonged heating (many hours). Different artificial aging conditions make different tempers. This is why aluminum frames needs to be baked in an oven after welding in order to restore the strength lost during the intense heat used for joining metal to metal. Tempering is designated by the T after the alloy number such as 7005-T6, higher number usually refers to stronger. The tempering conditions to achieve the different T4, T5, T6, etc. are alloy dependent.
Cycling Sidenote #4: For bicycle frame tubing, 6061 and 7005 are the available alloys since they can be heat treated/made stronger, extruded (made into tubes) as well as welded. Therefore the question “6061 or 7005 aluminum which is better?” is a common one that needs to be clarified.
They are both aluminum alloys, with different major alloying elements – Magnesium and Silicon for 6061 and Zinc and Magnesium for 7005. By the numbers 6061 has 15% lower ultimate tensile strength (amount of force to break) and about 65% of the fatigue resistance of 7005; 6061 is about 3% less dense (lighter). Those numbers are almost irrelevant to a properly designed bicycle frame (they don’t just all break after 1000 miles and the other one at 1650…). What I mean to say is, that yes continuously bending a metal, like a paperclip for example, will eventually cause it to break. (Aluminum) bicycle frames do not permanently bend, crashes aside, as such you are not going into stresses that would lead to fatigue failure.
When welded both 6061 and 7005 lose their strength (due to annealing in the heat affected zone of the weld; it takes ~11h to fully anneal 7005 and ~2 h for 6061) as a result of the intense heat of the process. HOWEVER, all parts made from 6061 must be solution heat treated/quenched again (530C+quench) in order to once again artificially age/precipitation harden them, while 7005 only needs to be artificially aged at 100-155C for 8-14h (can be done in a much smaller oven), it is all result of the composition of the base alloy 6061 vs 7005. So the process with 6061, post welding, is as follows: solution heat treatment (very hot, 35min~2h), then rapidly quenched in water and/or glycol (bulky and/or toxic) and aligned before it starts to naturally age (as you can imagine such rapid change in temperature can make a pretzel out of the nicely straight frame) and then aged/tempered (~175C, ~8h). 7005 frames can be repaired and just aged again many times, which is another convenient point.
Previously the makers of 6061 frames sent the finished product for the whole process to the tubing providers (ie Columbus) and for example Orbea had their own treatment facilities; when you run a big operation (most frames were aluminum at one point in time, not carbon) with large production, the whole thing becomes economical. The, now discontinued, Columbus Starship tubeset was a 6061. Therefore unsurprisingly 6061 as far as bicycle frame tubing (for the small custom builder) has largely given way to 7005, especially if we take into account Scandium 7xxx alloys (mentioned earlier). Though 6061 WAS very popular and is still widely avaialble alloy. Why? While about 10-20% weaker than 7005, it is more easily manipulated to make butted/shaped tubing (it bends easier without cracking) as well as it is readily weldable, and it is (significantly) cheaper than most 7005. Couple that with the 3% lower density and some of the lightest aluminum frames ever were made with 6061. Also, for welding 6061, the TIG fillers used – 4043 and particularly 4643 aluminum – make welds that are heat treatable and as such post aging are as strong as the base alloy. In the case of 7005 the compatible TIG filler wires 5183/5356 are NOT heat treatable and the weld is only 60-70% of the strength (breathe, your 7005 frame is ok); quite simply tubing needs to be made thicker(heavier) in order to account for that. So overall 6061 tubing can be made thinner and lower weight (see that same pattern again=) ). Though rare, heat treatable filler wire for 7005 series aluminun does exist (5180) and again you see how Scandium/Zirconium is an absolutely amazing advancement for joining (the strongest/7xxx series) aluminum.
In my opinion 7005 (scandium/zirconium) and aluminum in general is still an extremely viable material for high performance racing frames, since they can be made to measure for both rider proportions and weight/power and the bike still easily meet the 16.8kg/15lbs UCI weight limit if not below.
Stainless steels are high alloy (with a lot of elements) steels, as such they can behave like precipitation hardening alloys. The super strong Reynolds 953 is a maraging (martensitic + aging) steel that is extensively heat treated. The properties (superior strength) are derived mainly from the rapid precipitation of the alloying elements together with martensite.
Titanium exists in alpha and beta (crystal structure) alloys depending on the elements added (not surprisingly called alpha and/or beta stabilisers). Alpha alloys are readily weldable while beta ones are easily machinable. Alpha alloys are NOT heat treatable; beta ones can be solution heat treated and aged. The alloys used for cycling (Ti 6Al-4V or Grade 5, and Ti3Al-2.5V or Grade 9) represent an alpha-beta or an in between mixture that can be heat treated for increased strength. Grade 5(6Al4V) titanium is the most widely used Ti alloy worldwide, however, for bicycle frame building Grade 9 (3AL-2.5V) represents a good option that combines ease of production (hence lower raw material cost) as well as good weldability. Titanium possesses extreme!!! corrosion resistance and can be quite hard wearing on tooling. As compared to steel and aluminum, titanium needs to be properly welded or the welds will be very short lived. Even small contaminations from oxygen in air, grease from fingers etc. can cause the welds to crack. Cleanliness and proper technique is required. An important point here is what is left for making bicycle components is the scraps and leftovers of the production for the aerospace industry as such choices in sizes and such are not as many (though more than sufficient) than for example steel.
Conclusions (and Decisions)
All of the above demonstrate principles and show how metallurgy can provide plethora of options to meet the requirements for virtually all types of bicycle frames and metal fabrications. Of course they also act as a reminder that proper welding/brazing technique is key, otherwise all of the expensive manipulations can be negated/lost and even the safety of the frame compromised. Speaking of cost, the fact that one tubeset is more expensive than another does not mean it is better. The widely available mild steel is plenty strong for bicycle frame, therefore the fact that something is heat treated and ‘by the numbers’ is stronger does not immeditely point to the fact that it is better per se. Better than what? What are you going to use it for? All the alloys and modifications made to them were created to fulfill a goal.
In general, cost is a reflection of the amount of work (research, heat treatment, etc.) or expensive alloying materials that have gone into production. The more stuff you do/add to a metal, you add value, and it is usually done in order to make it stronger and/or corrosion resistant and/or weldable/workable. With stronger material you need less of it, so you can get lightweight. Though using little material can make things not so resistant to crashes and hard use. In the end the final product and how it is going to be used is what should guide decisions, rather than fads, numbers, shiny stickers and improperly named alloys. It’s unsustainable in the grand scheme of things (ecology) and adds unecessary expense to use the most expensive material for a product which can be more than adequately served by something else – your daily driver car can be made of the latest aerospace alloys, though steel and aluminum are more than adequate for the vehicles intended use. As a final note, it is refreshing that in the age of carbon fiber there is an ever increasing list of high tech metal tubing, that was once classified military use only, to the small builder.
Theoretical Framebuilding Series.
- Part 1: My Frame is Bigger than Yours: Geometry
- Part 2: I Like Big Butts: Tube Sizes, Butting and Ride Feel
- Part 3: Metallurgy and Materials
For further information check the ever-increasing Reading List
I welcome comments, however, before asking a question please visit the Frequently Asked Question (FAQ) page.
All images are copyrighted to their original owners, I do not own any of them: