Undergraduate research · UNSW Sydney (Summer 2026) & Harvey Mudd College · Ongoing
Lead-Free Carbon-Bronze Alloys
Lead makes bronze easy to machine and gives sliding parts the lubricity they need, but it’s being legislated out of bushings and bearings for health and environmental reasons. Earlier work in the lab pointed to carbon as a replacement. The catch: carbon barely dissolves in copper. So the real problem is finding compositions and heat treatments that keep lubricating graphite in the microstructure. Over summer 2026 at UNSW Sydney, our team of three made and characterized 21 experimental carbon-bronze alloys. I calculated compositions, prepared charges, and ran melting, heat treatment, sample preparation, optical microscopy, and hardness testing, and I improved the data visualization in our MATLAB analysis.
Role: alloy processing & characterization · Team: Samara Baidwan, Katie Cheng, Joshua Tanksley · Advisors: Prof. Lori Bassman (HMC), Prof. Kevin Laws (UNSW Sydney) · Tools: arc melter, optical microscopy, SEM/EDS, MATLAB
Methods
From raw elements to friction tests:
- Charge preparation: Pure elements cut and weighed into a Cu–Mn–Ni–Al bronze base with added carbon.
- Melting & casting: Alloys combined in arc melters and an induction furnace, then cast as small buttons for analysis and large buttons for friction testing.
- Heat treatment: Held at up to 900 °C in a muffle furnace and water-quenched, to approach equilibrium microstructures.
- Grinding & polishing: Sections polished to a 1 µm finish.
- Optical microscopy: Stitched images of each polished surface.
- Image analysis: A MATLAB thresholding algorithm measures carbon within selected regions.
- SEM & EDS: Electron imaging and composition mapping to identify each phase.
- Hardness testing: Trends across compositions and heat treatments.
- Pin-on-disk friction testing: Friction and lubricity of the most promising alloys.
Measuring the carbon
The question for every alloy was how much carbon survived, and in what form. The lab’s MATLAB pipeline thresholds the optical images inside hand-drawn regions of interest that represent the bulk alloy, excluding large carbon masses that never dissolved during melting. The detected carbon area estimates each alloy’s carbon content, and particle counts per mm² are sorted into four size groups for each heat treatment. I improved the pipeline’s data visualization, reformatting the histograms with clearer legends, layout, and carbon-percentage labels so results could be compared at a glance, and adapted it to process image sets stored on an external drive.
Preliminary findings
EDS line scans confirmed that the black inclusions are graphite, not hard carbides, which is the form of carbon a self-lubricating bearing needs. Several of the 21 alloys retained significant carbon, and like a graphite pencil they leave a carbon trail when smeared across paper. That’s an early sign of the self-lubricating behavior bushings and bearings depend on, but a mark on paper isn’t by itself a test of bearing performance. We then selected the most promising alloys and cast larger samples in a bigger arc melter for pin-on-disk friction testing.
What’s next
Our results so far are mixed, so I’m now analyzing the full dataset to find which composition and heat-treatment variables control how much carbon each alloy retains, using statistical and machine-learning models. I’ll be attending TMS 2027.