Pharma glass properties - Part 7. Chemical resistance

Hello everyone – this is Part 7 of an ongoing series about the properties of glasses used in pharmaceutical packaging.?This post technically completes my tour of the properties typically provided on the technical data sheets for pharmaceutical glasses.?The technical data sheet will also provide information on the glass composition, which isn’t actually a property.?However, glass composition certainly has a strong influence on properties, and so I’ll likely create a post on this topic as well.?There are a number of other glass properties that may not be included on the technical data sheet but are definitely relevant to the performance of pharmaceutical packaging.?I have already touched on a couple these “bonus” properties in prior posts (permeability and strength), but there is more to cover.?In other words, this series of articles isn’t complete yet, and so stay tuned for future content.

As I previously discussed with the practical strength of glass, chemical resistance (a.k.a. chemical durability, chemical stability, corrosion resistance, etc.) is another example of what I’m going to call a “pseudo-property”.?Unlike properties such as refractive index, density, etc., the observed chemical resistance of a given glass composition is highly dependent on the specific details of the experiment or measurement being performed (see Footnote 1).???A piece of glass might be incredibly stable under one set of conditions.??That same glass can then be observed to completely disintegrate during a different test.?Does this mean the glass is not durable??Not necessarily.?You just have to be very specific about the testing methodology (see Footnote 2).?

This leads us to the technical data sheet for a pharmaceutical glass, which often includes the results for multiple chemical resistance tests.?This is done to provide a rough picture of how the glass should behave over a range of conditions.?The below list includes some (but certainly not all) of the testing standards that may be used to evaluate and/or classify the chemical resistance of pharmaceutical glasses used for packaging parenteral drugs:

·??????USP <660> Containers – Glass; Ph. Eur. (3.2.1) -?Glass Containers for Pharmaceutical Use

·??????ISO 720 Glass – Hydrolytic resistance of glass grains at 121°C – Method of test and classification

·??????JP 7.01 Test for Glass Containers for Injection

·??????DIN 12116 Testing of glass – Resistance to boiling aqueous solution of hydrochloric acid – Method of test and classification

·??????ISO 695 Glass – Resistance to attack by a boiling solution of mixed alkali – Method of test and classification

·??????ASTM E438 Standard Specification for Glasses in Laboratory Apparatuses; ASTM C225 Standard Test Methods for Resistance of Glass Containers to Chemical Attack

We can make some generalizations before diving into more details about each test.?For example, if you examine the chemical resistance section of a pharma glass technical data sheet, you will observe that the reported results are not quantitative (Footnote 3).?Instead, the reported results are just informing you that a given glass meets a particular performance classification or complies with a given standard based on specified limits.?This is in contrast to an actual property like density that will have a quantitative value associated with it.?In addition, we can generally place these tests into one of three categories: 1) relatively acidic testing conditions, 2) relatively neutral testing conditions, and/or 3) relatively basic testing conditions (see Footnote 3).

USP <660> and Ph. Eur. (3.2.1)

USP <660> and Ph. Eur. (3.2.1) can be listed together because the previously distinct pharmacopeial standards were harmonized about 10 years ago.?Both standards describe three methods for evaluating chemical resistance: 1) a Surface Glass test, 2) a Glass Grains test, and 3) a Surface Etching test.?Each method is distinguished by the physical form of the glass container at the time of testing and the intent of the method.?However, there are also common features to each method – e.g., water for injection (WFI) is used as the test liquid and a colorimetric titration is used to generate a result.?The standards cover multiple container types (vials, ampoules, etc.), but my focus is on converted tubular vials for the purposes of this post.

Assuming that our primary interest is glass packaging for parenteral drug products, the desired outcome of these tests is demonstrating that a given sample complies with the requirements of a Type I glass – i.e., a glass that falls within the highest category of chemical resistance as defined by the standards.?The definition of a Type I glass currently includes two general components: 1) meeting the limits of specific testing requirements (see below for more details) and 2) a compositional requirement – i.e., it must be a borosilicate glass.?At the time of writing this post, there is a proposed revision in place that would modify the second requirement and allow other compositions to be considered Type I glass, provided that they meet the first performance requirement (see Footnote 4).

The Surface Glass test (also sometimes referred to as Surface Hydrolytic Resistance or SHR) is focused on the chemical resistance of the interior surface of an as-received vial.?The test vial is filled to 90% of its overfill volume (see Footnote 5) and then autoclaved at 121°C.?The leachate (i.e., the water that now contains some amount of glass dissolution products) is removed from the vial and several drops of a methyl red solution (a pH-sensitive colored dye) are added.?Finally, the leachate is progressively titrated with aliquots of hydrochloric acid (HCl) to identify the point at which the dye changes color from red to yellow and the amount of added HCl is recorded (see Footnote 6).?A glass vial can be identified as Type I or II (and not Type III) provided that the amount of recorded titrant falls below a particular threshold limit that is dependent on the size of the vial (see Footnotes 7 and 8).

The Glass Grains test is designed to test the intrinsic chemical resistance of glass and can be performed on tubing (a.k.a. glass cane) and containers.?Instead of using an intact container or tube, the sample is crushed and sieved to achieve glass powder.?Similar to the Surface Glass test, this glass powder is autoclaved in water and the leachate is tested using colorimetric titration.?The results of the Glass Grains test are interpreted by normalizing the volume of recorded titrant by the mass of glass powder used in the test.?A glass may be regarded as Type I if the maximum normalized volume of titrant added is 0.1 mL/g.?The Glass Grains test also provides a limit of not greater than 0.85 mL/g to classify a glass as being Type II or III.?A combination of the Surface Glass Test and the Glass Grains Test can be used to unambiguously distinguish among the various compendial definitions of glass types.

Finally, the Surface Etching test can be used to discriminate between Type I and II glass for containers showing high chemical resistance.?The Surface Glass test is used to measure the baseline performance of the as-received container.?Next, the same samples or a different set of samples are filled with a mixed solution of hydrofluoric acid (HF) and HCl for 10 minutes.?The purpose of this solution is to dissolve away a thin layer of the interior container surface, thereby exposing the underlying bulk glass.?The autoclave and colorimetric titration procedure is then repeated on the etched containers and the results from both tests (Surface Glass and Surface Etching) are compared.?The results should be similar for a Type I glass container, whereas a Type II should require a greater amount of added titrant (i.e., more alkali extracted) for the Surface Etching test.?You read that right – there are no quantitative limits provided here.?Judging the degree of similarity is left up to analyst.?I also don’t know how many laboratories are prepared to handle HF in a safe and environmentally responsible manner.?Its use is not to be taken lightly.

All of this procedural information is fine, but what are we actually measuring??In practice, these tests are mainly detecting the presence of alkali ions (sodium and potassium) leached from the glass into the liquid.?The pH of the water used as the corrosive agent increases with increasing extraction of alkali, which in turn impacts the end-point of the colorimetric titration using HCl (see Footnote 9).?Is there more to it? ?Absolutely – other products of glass corrosion can also have varying degrees of impact on the pH of the leachate, but a complete discussion is beyond the scope of this post.?Just remember that the threshold limits of these tests are semi-quantitative in nature.?For example, a Type I glass vial that passes the Surface Glass test essentially just means that the extracted alkali levels were sufficiently low to meet the limit – it doesn’t tell you by how much.

ISO 720

The methodology described in ISO 720 is essentially a variation on the Glass Grains test from USP <660> and Ph. Eur. (3.2.1).?A similar quantity of similarly sized glass powder is autoclaved in water and a colorimetric titration is performed on the leachate using methyl red and HCl.?The key distinction of the ISO 720 method is that the results are instead classified into one of three categories: HGA 1, HGA 2, and HGA 3.?However, a quick comparison of the limits for ISO 720 (see Footnote 10) and the Glass Grains test show that HGA 1 is equivalent to Type I glass and HGA 2 is equivalent to Type II and III glasses.?Finally, the ISO 720 standard contains an interesting compositional restriction.?It states that:

?“The test method shall not be applied to glasses with extreme low alkaline contents or that are essentially free of alkaline species as this method measures only the alkaline release as the indication for chemical durability.”

Based on this statement, it would appear that fused silica packaging could not be evaluated using this test method.

JP 7.01

The methodology in JP 7.01 is another test intended to measure the extraction of alkaline species from glass.?Provided the container is not to be fused (i.e., not an ampoule) for aqueous infusions with a capacity exceeding 100 mL, the glass to be tested is crushed into powder that is relatively coarse compared to the prior tests.?The glass grains are then boiled in water for 30 minutes.?A colorimetric titration is again performed, although the indicator dye is a bromocresol green/methyl red mixture and the acid titrant is sulfuric acid.?An ampoule tested in this manner is deemed to be acceptable for parenteral packaging if the amount of titrant is less than 0.30 mL; a container that is not be fused is acceptable at a limit of 2.00 mL.

For containers not to be fused for aqueous infusions with a capacity exceeding 100 mL, the intact container is filled to 90% of overflow capacity with water and autoclaved at 121°C for 1 hour.?The leachate is tested using the same colorimetric titration procedure, and the amount of titrant consumed should not exceed 0.10 mL.

DIN 12116

The previously described methods all concerned the interaction of glass with a pH-neutral liquid.?DIN 12116 considers instead the resistance of glass to an acidic environment.?Why is the chemical resistance of glass in an acidic environment different??We’ll get into this more in a future post, but acidic conditions encourage the extraction of alkaline species from the glass through an ion exchange process (e.g., hydronium ions from solution swapping for sodium ions in the glass surface).?This process results in what is sometimes called “incongruent dissolution”, meaning that a portion of the material undergoing corrosion is selectively dissolved and another portion is left behind – in this case, a silica-rich surface.

Bulk glass samples of relatively simples shapes (blocks, open-end cylinders, etc.) having known surface area and mass are suspended in boiling 6 M HCl for 6 hours.?The samples are weighed again after corrosion, and the results are used to classify the glass as S1, S2, S3, or S4 (see Footnote 11).?Glasses of the sort used for parenteral packaging would generally fall into the S1 category.

A unique aspect of this method is its use a weight loss measurement instead of titration.?This is perhaps a technically better approach because it is a purely additive metric – i.e., increasing weight loss is directly correlated with increasing corrosion, unlike pH shift (see Footnote 9).?However, the execution of DIN 12116 is arguably more challenging in practice given some of the detailed sample preparation requirements.

ISO 695

ISO 695 is conceptually similar in execution to DIN 12116 with one major exception – the corrosive agent is instead a basic solution prepared from sodium carbonate and sodium hydroxide.?Sample weights before and after boiling in the basic solution for 3 hours (not the 6 hours prescribed in DIN 12116) are used to classify the glass as A1, A2, or A3 ?(see Footnote 12).?Glasses of the sort used for parenteral packaging would generally fall into the A1 category.

I previously mentioned that acidic conditions favor an ion exchange mechanism resulting in incongruent dissolution.?In contrast, basic conditions tend to favor hydrolytic attack of the glass network.?Given that silica (SiO?) is a major component of the glass composition, corrosion at elevated pH tends to result in “congruent dissolution” – the ratio of elements found within the glass corrosion products (e.g., sodium, potassium, boron, silicon, etc.) are similar to the ratio of elements within the bulk glass composition.?

A more in-depth inspection of the methods described in DIN 12116 and ISO 695 can help provide some context on the impact of acidic versus basic corrosion conditions on pharmaceutical glasses.?The total surface area of samples used in DIN 12116 is more than six times greater than ISO 695 (75 cm2 to 100 cm2 for DIN 12116 versus 10 cm2 to 15 cm2 for ISO 695) – in other words, there is much available surface area available for extraction with the acidic test.?However, a comparison of the glass classification limits provided in Footnotes 11 and 12 show the striking difference in applicable weight loss ranges.?The limits for basic corrosion conditions are about two orders of magnitude greater than acidic conditions, meaning that significantly more material is being dissolved on a per unit area basis.

ASTM C225 and ASTM E438

The ASTM E438 standard provides the guidance for classifying glasses in various types, an aspect of which relies on performing the chemical resistance tests described by the ASTM C225 – as a result, they are paired together in this section.?ASTM C225 describes three methods for testing chemical resistance:

·??????Test method B-A involves containers partially filled with dilute acid and autoclaved at 121°C for 1 hour.?The test is primarily intended for containers that will store products with a pH under 5.

·??????Test method B-W involves containers partially filled with water and autoclaved at 121°C for 1 hour.?The test is primarily intended for containers that will store products with a pH over 5.

·??????Test method P-W involves glass powder in water that is autoclaved at 121°C for 30 minutes.

The leachate from each method is tested via colorimetric titration using a methyl red indicator and sulfuric acid.?At this point, we are starting to see some common themes amongst the various standards.?These test methods are designed to acknowledge the differences in corrosion behavior that can occur as a function of pH and for intact containers versus the bulk glass composition.

ASTM E438 is interesting in multiple glass properties in addition to chemical resistance as part of its classification system .?We have covered all of these other properties in past posts (linear coefficient of thermal expansion, viscosity points, and density), and so I’m going to stay focused on chemical resistance. ?There are three possible categories in ASTM E438: Type I – Class A, Type I – Class B, and Type II.?Type I – A/B glasses have an upper titration limit of 1.0 mL per 10 g of glass; Type II glass has an upper limit of 9.5.?Glasses relevant for parenteral packaging should generally meet the performance requirements defined by the Type I – A/B categories.

Summary

I hope this quick tour of chemical resistance tests for pharmaceutical glasses was helpful.?The most important takeaway is that the chemical durability of a glass is strongly influenced by the conditions of the measurement.?There is no single test that can fully capture the behavior of a glass over the range of conditions that may be presented by parenteral drug formulations.?Nevertheless, these tests provide directionally useful information that can help in the selection of packaging materials.

Do you have any questions or additional information that you’d like to share??Please comment below or feel free to directly message me.

Footnotes

1. This is a general statement that extends well beyond glasses.?It applies to all materials including inorganic metals and organic polymers. ?As a result, there are numerous standards out there for testing the chemical durability of various materials for specific use cases.

2. There are many, many different testing parameters that should in theory be considered when designing a chemical resistance test.?For example, there are the more obvious parameters like temperature and the composition of the corrosive agent.?There are also perhaps less obvious parameters such as the surface area-to-volume (SA/V) ratio of the material to the corrosive agent.?A complete discussion of all these parameters is beyond the scope of this post.?I would just caution you to think critically about the conditions for any proposed chemical durability test and consider how they may inform the results.

3.?Quantitative testing brings us into the world of extractables & leachables (E&L) in which the concentrations of individual species are directly measured under various conditions. ?The E&L behavior of pharmaceutical glass is relevant to chemical resistance.?However, the chemical resistance tests described here are more semi-quantitative by treating glass dissolution products in aggregate, meaning that the total contribution of all components released from the glass is being measured by a single metric.?E&L is a big enough topic on its own that I will not dwell on it any further here – another future post!

3.?You may be wondering about my use of the “and/or” in the list of the three general types of test methods.?Most methods only require one pH condition to be tested, whereas other methods may allow for more than one pH conditions to potentially be tested (in this case, ASTM C225).

4. In the interest of full disclosure, I work for a company that has commercialized a non-borosilicate glass container for parenteral drug packaging (we also offer borosilicate containers).?However, my goal in writing this series has been to provide non-biased, informational articles about pharmaceutical glass packaging that are not overtly commercial in nature.?With that in mind, here is my perspective.?Just because a particular glass container (whether it be borosilicate or non-borosilicate) meets the performance requirements of Type I is not an automatic guarantee that it will compatible with a given drug product.?You might therefore ask – why do even worry about conducting these standardized tests??I think they provide value in at least two ways.?First, the compendial test criteria are directionally useful during the drug development process (particularly for smaller drug developers lacking the in-house expertise that appreciates the finer details of drug/container interactions).?It’s generally reasonable to assume that glass containers with the highest levels of chemical resistance are preferred for parenteral packaging.?However, the results of compendial tests such as USP <660> and Ph. Eur. (3.2.1) are never a substitute for quantitative E&L testing and drug product stability testing under actual storage conditions.?Second, the compendial tests can potentially be a useful monitor of quality once a drug product has been commercialized at a volume that relies on multiple batches of glass containers.

5.?It’s important to understand the difference between nominal fill volume and brimful capacity.?For example, you might think about a 2R ISO format vial as being designed to hold 2 mL of liquid.?However, the ISO 8362-1 standard for converted tubular vials also states that a 2R vial has a brimful capacity of 4±0.5 mL.?You would never actually fill a vial to this brimful capacity, but it can become relevant when you’re looking to potentially squeeze an extra dose of drug product into the container.?USP <660> and Ph. Eur. (3.2.1) describe a straightforward method for determining brimful capacity by experimentally measuring the mass of six filled containers, calculating a mean value, and then multiplying the result by 0.9.?

6.?I am glossing over a lot of details in describing the various test methods.?As I mentioned at the beginning of this article, being highly specific about methodology is critical when conducting and interpreting chemical durability tests of glass.?However, my goal here is just to provide you with a general sense of how the test is performed, discuss how the results are interpreted, and highlight what I consider to be important points along the way.

7.?The below table specifies the titration limit values to classify glasses using the Surface Glass test from USP <660> and Ph. Eur. (3.2.1).?Why do the limits change as a function of container size??Imagine that you have 2 mL and 50 mL vials made from the same glass composition and with nominally identical interior surface chemistries.?Performing the Surface Glass Test on these two containers would produce different results because of the different SA/V ratios, one of the influential factors in chemical durability testing that I mentioned in Footnote 2.?As the container size increases, the amount of glass surface area in contact with a given volume liquid is decreasing.?The titration limits are adjusted to account for this effect.

8.?In case you don’t already know or need a quick reminder – a Type I glass has the highest chemical resistance both in terms of the interior container surface and the bulk glass composition.?As such, Type I glass can be regarded as a preferred packaging material for parenteral drug products.?A Type III glass has moderate chemical resistance for the interior surface and the bulk glass – it’s generally not preferred for parenteral drugs.?Finally, a Type II glass is the result of a Type III glass being treated to achieve high chemical resistance at the interior surface (at least within the context of the compendial test), although the bulk glass still possesses moderate chemical resistance.

9.?Is there more to it than just extraction of alkali? Absolutely, yes – other products of glass corrosion can have varying degrees of impact on the final pH of leachate that is produced during these compendial tests.?For example, the network formers commonly used in pharma glass packaging (i.e., SiO? and/or B?O?) dissolve to produce species that are both weakly acidic, meaning that the increased alkalinity of extracted sodium and potassium ions can be offset to some extent.?This starts to raise interesting questions about what the titration-based compendial tests are actually measuring and reinforces a comment that I made in Footnote 4.?Compendial tests of the sort described in this article are not a replacement for quantitative E&L tests and drug stability tests.

10. The below table specifies the titration limits for classifying glasses according to ISO 720.?HGA means “hydrolytic resistance of glass grains according to the autoclave test method”.?Now that’s an efficient acronym.

11.?The below table specifies the weight loss limits for classifying glasses according to DIN 12116.?The seemingly odd numerical basis for the limits (i.e., “half the mass loss”) is derived from an older version of the standard that specified boiling in HCl for 3 hours instead of 6 hours.?Subsequent testing has confirmed that halving the mass loss after 6 hours of boiling properly normalizes the data to what would effectively be achieved after 3 hours.?However, I will add that this normalization likely only works because we’re working with bulk glass samples.?I suspect an equivalent mass of glass grains of the sort used in USP <660> and Ph. Eur. (3.2.1) would show significant non-linear behavior that would not allow for this simple mathematical hack because of increased surface area.

On a different note, I’m always curious about the origin of naming schemes – these things are often not explained in the standards.?Why do the categories start with the letter “S”??I’m assuming it's related to the word “sauer” because it’s a DIN standard.?I look forward to comments from any readers that know the correct answer.

12. The below table specifies the weight loss limits for classifying glasses according to ISO 695.