Subscribe to our newsletter to receive the latest news and events from TWI:

Subscribe >
Skip to content

Teething troubles ... dental engineers given a helping hand

Connect, no. 134, January/February 2005, p.6

Fig.1. The specimen was subjected to a pull test
Fig.1. The specimen was subjected to a pull test

Failures in healthcare products take on a seriousness and urgency of their own, often not attributed to more benign engineering structures.

So when TWI was asked to investigate a medical service weld failure it rose to the challenge by immediately mobilising specialists in the fields of both resistance welding and structural integrity.

The product in question was a teeth brace. Or, as it is known in the world of specialist orthodentic products, a molar band-buccal arrangement.

The investigation formed part of a product and process review for Joining Forces, Yorkshire and Humber and was followed by a feasibility study.

The molar band-buccal structure is fabricated in medical standard stainless steel and involves spot welding a bracket some 5mm long to a narrow band which embraces each molar tooth.

It was the integrity of these tiny welds which was thrown into doubt when a patient returned their broken brace.

The objective was simply to establish possible causes of failure of the assembled orthodontic band.

First, a dozen static tensile tests were carried out on specimens provided by the client, to determine failure load. The load was applied at 90 degrees to the welds at a rate of two millimetres per minute. The failure loads varied widely from 254 to 441N. The service loads, endured by the structure inside the patient's mouth, were not known.

Second, the assembled bands were examined metallographically. Two specimens were sectioned horizontally and ground and polished to a one micron diamond finish.

And third, for reasons of comparison, the bands which had failed in service were also examined metallographically.

Fig.2. The red areas show successful spot welds but the blue area indicates a failed weld
Fig.2. The red areas show successful spot welds but the blue area indicates a failed weld

Examination of the bands using twenty five times magnification under a binocular microscope revealed four to five electrode indentations on each side of the buccal. In several cases small deposits of electrode material were found at the edge of the indent, suggesting that damage to the electrode tip had occurred.

Examination of the fracture surfaces also revealed that failure had involved pulling a plug or button from the parent band material.

There also seemed to be a difference in appearance between the tested failures and the service failure offered by the client.

A closer look using a scanning electron microscope revealed microvoid coalescence in the tested weld, a certain indicator of ductile failure. However the service failures showed crack arrests across the fracture surface indicating a progressive failure. This was borne out by the presence of contamination found at the start of the failure.

It is likely that final failure of the band occurred following overload of a partly broken joint.

TWI was able to conclude that the service failures of the bands could be attributed to cyclic loading and a progressive failure culminating in overload. The joint strength of the bands appears to be limited by the endemic fatigue performance of spot welds in general. Their geometry inherently includes a notch, or crack initiation point.

Increasing the size of the weld may not make any difference.

As a result of this product and process review the manufacturer was advised about the two loading types contributing to failure, and TWI recommended a review of its welding procedure, with particular respect to the weld positioning and quality control. An alternative joining process, perhaps adhesive bonding, should also be considered.

More information may be obtained from Andrew Low. E-mail:

For more information please email: