Titanium to Steel two
Introduction: The Goal of the Research
This research is focused on the challenge of diffusion bonding titanium to steel to create a titanium-clad high-carbon steel material. The motivation stems from a desire to combine modern and historical techniques to produce a unique and functional material for blades that serves as an exciting canvas for creativity. Bonding titanium to steel is particularly difficult due to the formation of brittle intermetallic compounds at the interface, which compromise bond strength and structural integrity.
Background: Understanding the Problem
Titanium and steel react at high temperatures to form brittle intermetallic compounds at the joint interface that result in weak bonds. Achieving a successful bond requires a balance between temperature, pressure, and time to minimize these reactions. One approach is to use interlayers—thin sheets of other metals that act as diffusion barriers to reduce the formation of intermetallic compounds. In my previous experiments, we demonstrated that copper was effective as an interlayer, while nickel also showed promising results in maintaining adhesion to titanium
Scope of the Experiment
The objective of this experiment was to expand on previous findings by:
Testing various interlayers (nickel, copper, zirconium, and niobium) with different thicknesses to analyze their bonding performance.
Test bonding through power hammering instead of hydraulic pressing.
Exploring the viability of direct bonding between titanium and steel, inspired by research suggesting that strong bonds can be achieved without interlayers under controlled conditions.
Methods
Interlayer Experimentation
Billets of titanium and steel were prepared with interlayers of varying thicknesses (0.05 mm to 2 mm).
They were arranged in this order: Grade 5 Titanium, Grade 2 Titanium, Interlayer, 1080 Steel, Interlayer, Grade 2 Titanium, Grade 5 Titanium.
Interlayers tested included nickel, copper and both zirconium, and niobium which were chosen for their lower coefficients of thermal expansion.
The billets were enclosed in argon-purged 1.5mm wall mild steel canisters to prevent oxidation, heated to 850°C in an electric kiln for 10 minutes, and violently forged under flat dies on a compressed air driven steam hammer for approximately 20 seconds. They were then cooled to room temperature slowly over the course of 30 minutes.
Direct Bonding Experiment
This experiment was conducted because of insights and research provided by Dr Amir Shirzadi, his paper Gallium-assisted diffusion bonding of stainless steel to titanium; microstructural evolution and bond strength shows that a direct bond of titanium to stainless steel results in a stronger tensile strength than nickel interlayers. He suggested minimising time at bonding temperatures by immediately quenching the billet after bonding.
This meant Test 1, an unhardened bend couldn’t be performed as it had already been hardened.
For the direct bond, the titanium and steel billet was preheated to 700ºC in a hot gas forge, finally brought to a controlled 850°C in an electric kiln, and then hammered under the full force of my steam hammer for 5–10 seconds.
The billet was then immediately quenched in oil while still at approximately 800–820°C to minimize time at bonding temperatures.
Test 1: Crude Bending Force Measurement
All samples were machined to 15x57x6.7mm with the steel core remaining central.
The first test involved measuring the bending strength of the as-bonded samples in a straightforward setup. Each sample was clamped in a vise with 33 mm of the material exposed. A 15-inch adjustable spanner with a crane scale attached was used to apply force to the sample. The material was bent to 90 degrees, and two measurements were taken:
Yield Force: The force at which the material began to bend.
Peak Force: The highest force recorded during the bending process.
Test 2: Bending Hardened Samples
For test 2, samples were heated in an electric kiln to 815°C, Quenched in oil and Tempered at 250°C. The zirconium and niobium interlayer samples experienced immediate failure after quenching.
The purpose of test 2 was to evaluate the mechanical strength of the hardened and tempered samples. Using the same setup as Test 1, the samples were clamped with 33 mm of material exposed and bent using a 15-inch adjustable spanner with a crane scale attached. As before, the yield force and peak force were measured.
This test revealed a significant difference in performance compared to the first test. Many of the samples fractured under the applied force, highlighting the increased brittleness introduced by the heat treatment.
Test 3: Edge Testing
The third test focused on evaluating the edge durability of the final samples. The process involved the following steps:
Samples were heat-treated (hardened at 815°C and tempered at 250°C).
A bevel was ground onto the sample edges using a belt grinder, followed by a small secondary bevel.
The sample was used to cut through a 1/8” bronze TIG filler rod three times while clamped in a vise.
The edge was subjected to three hammer blows with a 500 g hand hammer against a 3 mm thick mild steel edge.
Observations were recorded for any rolling, cracking, or deformation of the edge to assess the material's suitability for high-stress applications.
Bending forces Un-Hardened vs. Hardened.
Observations from 10 samples.
Pictures of all 10 samples, test order L-R = 1-3 (minus sample 10 where test 1 was not carried out).
Final thoughts
To explore the direct bond method further, I need to learn more about steel heat treatment. I am unsure if it is even viable full-stop to heat, forge and quench in one go and be able to acheive the proper hardness in the core material. Most properly done heat treatment (as needed for a high end blade) requires careful thermal cycling after forging before the quench is done.
I will potentially need to explore higher hardness steels that can withstand a slower quench but still get the hardness needed. If I can’t achieve the hardness required with this method I should explore quickly reheating and quenching the remaining direct bonded sample I have to see if this reheat causes issues from more time at temperature.
The bonds between grade 2 and grade 5 titanium failed quite often in the tests. I was surprised by this.
I hope in sharing these observations some of you can have a head start playing with these materials and in turn share what you are learning.
Please get in touch with any feedback or ideas below.
