Delamination in pharma glass vials – 3. Flakes

Welcome to Part 3 of an ongoing series concerning a failure mode known as glass delamination.  Part 1 introduced the basic concept of delamination and Part 2 explored the multi-stage sequence of events that culminate in the fracturing of an altered interior surface layer.  In today’s post, we’ll focus on the flake-shaped particles produced by glass delamination, otherwise known as lamellae.

First, let’s be clear about what I mean by a “flake”.  Delamination flakes are generally thin (on the order of 1 micron or less) with lateral dimensions that can be 100’s of microns.  There is no single defining shape for these particles although the headline image at the top of this post of a fragmented, icy surface of a frozen pond is remarkably similar to glass delamination flakes observed under a microscope.  Figure 1 is a scanning electron microscopy (SEM) image of an actual glass delamination flake obtained by filtering the liquid contents of a glass vial (the background texture in the image is the porous filter).  A key distinction between glass delamination flakes and the broken ice that I used as a visual analogy is rigidity.  The relatively thin delamination flakes can sometimes be observed to curl, and so don’t assume they will always appear perfectly flat.  Delamination flakes will also not necessarily have a “clean” linear edge – they can be more jagged and irregular in appearance.

Next, it’s important to recognize that a flake is a necessary but not sufficient condition for identifying glass delamination – we have to consider both morphology and composition.  Glass delamination particles are created when a silica-rich reaction layer fractures away from the container wall, and so the flakes will principally consist of silicon and oxygen (see Footnote 1).  A great example of why this is important is provided by Ogawa et al. in a published study from 2013 (see Footnote 2). The authors demonstrated the formation of flake-like particles when storing a phosphate buffer solution (16 mM, pH 7 with 0.9% sodium chloride and 0.2% polysorbate 80) in 51 expansion borosilicate glass vials at 40°C for 6 months.  The visual appearance of the sample images shown in the study look identical to what can be observed in a vial that has undergone delamination. Given that SEM is often used to observe the shape of suspected delamination flakes, it’s often convenient to also use energy dispersive X-ray (EDX) spectroscopy to determine the elemental composition.  Ogawa et al. performed EDX on the observed flakes and found they were actually an aluminophosphate-based material.  Using a combination of other techniques, they concluded that the flakes were actually the product of a reaction between the phosphate buffer and aluminum extracted from the glass vial over the course of the stability test.

It’s also worth considering the reverse scenario.  Let’s assume you’ve identified a silica-rich particle in your drug product.  Does that mean it’s glass delamination?  Not so fast – once again, we need to consider morphology as well.  This is demonstrated nicely in a case study published by Ratnaswamy et al. in which particles were observed in a placebo formulation after going through a freeze-thaw cycle (see Footnote 3).  Results from FTIR spectroscopy performed on the particles indicated a mixture of silica-rich material and polysorbate 20 (an excipient used in the formulation).  Visual assessment of the particles suspended in solution with the unaided eye could potentially be confused with glass delamination.  However, closer inspection with SEM revealed a more globular morphology instead of the flake-like shape typically associated with delamination.  In this case, the authors concluded the silica-rich particles were created when silicon dissolved from the glass vial was concentrated during the freezing process.  Freeze concentration of silicic acid (the soluble form of silicon) resulted in precipitation of the observed particles.

One closing thought – all of this discussion implies that a flake-shaped, silica-rich particle must be found to positively say that glass delamination has occurred.  I think that’s a generally correct conclusion, although we can consider edge cases.  The first example might come from studies used to assess glass delamination risk (see Footnote 4).  These studies are typically performed at elevated temperatures to accelerate the sequence of reactions leading to delamination.  However, the solubility of silicon is also increased at elevated temperatures, and so it’s possible that a relatively small amount of particles generated by delamination can be solubilized before being detected.  Secondly, while we are on the lookout for flake-shaped particles in the drug product, I always encourage people to still examine the interior surface of the vial (particularly the heel region).  You may find instances in which the surface is just beginning to show signs of delamination without shedding detectable levels of particulate.  Figure 2 is one such example.  Is this technically delamination if there are no particles? – perhaps not, but I would still be very concerned if I found something like the surface observed in Figure 2.

Questions or comments? – please leave them below or feel free to directly contact me.

Footnotes

1.      To clarify a bit further, let’s consider the term “silica-rich”.  The word “silica” on its own is jargon used by materials scientists and chemists in referring to a solid that roughly conforms to the formula SiO₂.  It doesn’t however specify a particular form of SiO₂ – e.g., quartz, cristobalite, vitreous silica, etc.   “Silica-rich” is then my way of indicating a material that is mostly SiO₂.   What else might we expect to see when analyzing a glass delamination flake?  In terms of the flake itself, it’s not uncommon to also see trace amounts of aluminum that are original to the parent glass and remain behind in the altered surface layer.  I would also generally expect to see varying amounts of carbon due to: 1) residual organics from the drug products and/or 2) the filter material (if used) upon which the flake was collected prior to analysis.  Finally, it’s common to see varying amounts of other elements such as sodium, potassium, chlorine, phosphorus, etc. that were also present in the drug product.

2.      Reference: Ogawa T, Miyajima M, Wakiyama N, Terada K (2013).  Effects of phosphate buffer in parenteral drugs on particle from glass vials.  Chemical and Pharmaceutical Bulletin, 61: 539-545.

3.      Reference: Ratnaswamy G, Hair A, Li G, Thirumangalathu R, Nashed-Samuel Y, Brych L, Dharmavaram V, Wen Z-Q, Fujimori K, Jing W, Sethuraman A, Swift R, Ricci MS, Piedmonte DM (2014). A case study of nondelamination glass dissolution resulting in visible particles: Implications for neutral pH formulations.  Journal of Pharmaceutical Sciences, 103: 1104-1114.

4.      I plan to cover examples of these accelerated studies in a future post, but you can check out the USP <1660> General Chapter for an introduction to the topic.

About the Author

Matthew Hall is Technical Affairs Director for Corning Pharmaceutical Technologies , a manufacturer of primary glass packaging for parenteral drug products.  Based in upstate New York, Dr. Hall serves as a technical expert supporting business operations, sales, and marketing and educating customers on pharmaceutical glass packaging.  He is a member of the Parenteral Drug Association and the International Society of Pharmaceutical Engineering.