Endoscope Window Bonding: After the Initial Voids Were Controlled

In the revised specimens, the initial voids were brought under control. No leakage has been detected through 60 sterilization cycles, but probing the outer adhesive edge revealed peeling in layers.

The leak

The first sign of trouble was a continuous stream of bubbles from the edge of a sapphire window. I was running an underwater vacuum leak test after low-temperature hydrogen peroxide plasma sterilization. The bubbling appeared at different positions around the window perimeter. After further sterilization cycles, I also saw fogging on the inner surface of the sapphire window.

The test showed where bubbles emerged from the window assembly. It did not show what had happened inside the adhesive layer. I needed to examine that layer directly.

Inside the adhesive layer

I examined the window perimeter closely and sectioned regions that looked like bubbles. Those regions were hollow inside. Where the adhesive had fully filled the region, the section was solid.

Figure 2. An early window specimen (left) and a section through a bubble-like region (right).

The sections confirmed that cavities were present. I still needed to understand how they had formed.

I reviewed published work on bonding and repeated sterilization. An Olympus patent application checked adhesive layers for bubbles and cracks after low-temperature hydrogen peroxide plasma treatment and separately measured post-treatment bond strength.[1] I focused on what those observations and tests actually recorded.

Comparing curing conditions

I needed to distinguish the effects of curing conditions and groove geometry on the specimens.

I compared mechanical test results for two window specimens made under different curing conditions, labeled A and B, with one specimen per condition.

The maximum recorded loads were 83.7 N for condition A and 357.4 N for condition B. These values refer to the complete window assemblies.

In the geometry comparisons reported here, I held the curing parameters constant. Had the curing conditions also varied, any improvement would have been difficult to attribute to the groove change.

Revising the groove geometry

One explanation I considered was that cure shrinkage and redistribution of the liquid adhesive left localized cavities. The groove defined both the adhesive layer's geometry and the volume it could hold. I revised the groove geometry under fixed curing conditions and checked whether the earlier voids still appeared.

The initial voids seen in the early specimens were no longer observed in the revised specimens, and repeated checks gave consistent results. This result applies only to the geometry and process conditions evaluated here.

Figure 3. Revised specimens: window-perimeter overview (left) and lower-perimeter detail (right). These photographs illustrate adhesive appearance; they are not presented as a record of the 60-cycle inspection.

I did not directly record how the cavities formed, so the explanation involving cure shrinkage and adhesive redistribution remains unconfirmed. I have also not established whether those cavities connected to the locations where bubbles emerged during the original leak test.

The outer edge after sterilization cycling

At the time of writing, the current specimens have completed 60 sterilization cycles, and no leakage has been detected in the underwater vacuum leak test.

During further inspection of the window perimeter, I lifted and peeled away the outermost adhesive in layers with the tip of my tweezers.

Figure 4. Probing the outer adhesive edge with a tweezer tip (left) and a detail view of the edge (right).

The earlier sections had shown cavities inside the adhesive layer. This inspection revealed peeling at the outer edge. The leak-test results had not revealed this condition, so I turned my attention to how this outer layer related to the remaining bonded region.

Investigating the outer adhesive edge

I began to consider the local geometry and exposure of this outer layer. The same Olympus application also notes that exposed adhesive at an endoscope tip is relatively susceptible to degradation during sterilization.[1]

I suspected embrittlement at the outer edge. What I directly observed was peeling in layers during manual probing; a change in the adhesive's brittleness has not yet been independently confirmed.

I still need to locate the separation: within the adhesive or at the substrate interface. I also need to establish whether it is confined to the outer edge and how it relates to the remaining bonded region and subsequent leak-test results.

I have demonstrated a way to control the initial voids in these specimens under the conditions evaluated here. Whether the revised window assembly can withstand the target number of sterilization cycles remains to be verified.

References

[1] Olympus. US20160222261A1. Adhesive composition for medical devices and medical device. Description: Effects; Evaluation for Sterilization Resistance; Measurement of Adhesive Strength.

https://patents.google.com/patent/US20160222261A1/en