John Bak

Particulate matter by light obscuration (LO) testing quantifies particles in suspension by their projected shadow onto a photo sensor. This is the preferred method of sub-visible particle quantification (generally < 100 µm), as it has a high testing throughput and robustness compared to other means, such as membrane or flow imaging microscopy. However, not all preparations are suitable for testing by LO. In some cases, the issues may prevent usable results from being generated (such as an instrument being unable to maintain a flow rate due to viscosity issues or unable to measure particles in an opaque solution), and in other cases the results generated may be inaccurate (such as an instrument undercounting clear particles due to clarity issues or under-sizing aspherical particles due to shape issues). This article discusses both cases, with a focus on those where inaccurate results are generated because the added risk of the inaccuracy was not discovered.

	The most apparent symptom of a viscosity issue is the instrument’s inability to pull solution at the predefined flow rate. This almost always prevents results from being generated, though in borderline situations the solution may still flow through the instrument and simply generate errors at a higher-than-normal frequency. Errors should be reviewed carefully to determine if they are related to an increased load on the motor (or any other system responsible for pulling the solution through the instrument); such errors are a secondary symptom of flow rate difficulties. This and other viscosity issues can often be mitigated through dilution.

	Coloration issues, a variation of clarity, can present symptoms similar to viscosity issues, such as being unable to generate results or generating results with a higher-than-normal frequency of errors in borderline situations. If the solution has significant coloration, well discernable to the naked eye, it may be the root cause. Coloration errors are distinguishable from viscosity errors as they stem from the bulk solution blocking too much light to resolve particle shadows, and generate errors relating to sensor contamination. This is because the instrument observation is the same as if there were a stuck particle in, or a film built up on, the sensor. This issue, similar to viscosity issues, can often be mitigated through dilution.

	Any symptom-alleviating dilution should be carefully executed to perform its function without being unnecessarily extreme (1,2). High dilution factors decrease overall sensitivity to the test solution while increasing sensitivity to lab technique. Because of this, it is possible that symptoms are severe enough that they are not mitigated prior to generating new issues. In such cases, check to see if a different sensor type may be utilized (likely still in conjunction with dilution). Manufacturers typically have multiple types of sensors available, each with their own specialty, that may be of aid. However, if dilution cannot be made sufficient, particles may need to be quantified by another technique.

	Inaccurate results generated-variability by instrument

	Contrast limitations affect preparations exponentially as the refractive indices of particles and solution approach each other, which can lead to otherwise identical instruments generating noticeable differences in their results due to slight variation in their construction/assembly. Such an unexpected variation should be evaluated to determine if it can be attributed to an instrument error, or if it may be a contrast issue. The contrast of a particle in suspension is defined by the absolute value of the difference in refractive indices (Δn) between the particle and surrounding fluid. Particle counts begin to be noticeably affected when Δn is less than approximately 0.10, with more significant effects when Δn is less than approximately 0.05 (3,4,5).

	If the chemical composition of a particle population is known, the bulk refractive indices of each may be used as rough approximations to determine if contrast issues are likely. While the refractive index of a fluid often can be easily determined or well-approximated, a particle’s refractive index can be more challenging and laborious to obtain. Additionally, a single refractive index is rarely representative: even if a particle population is both isotropic and chemically homogeneous, the effective refractive index may be a function of particle size due to edge effects. Because of this, it is advised to reactively test for any suspected contrast issue rather than determining refractive index value(s) for particles in advance. Preliminary testing can be performed by slightly altering the solution refractive index, and thus particle contrast, between analyses and checking for an amplified difference in particle counts.

	To illustrate how instrument variability can be attributed to contrast issues in even a relatively mild case, see 

. Three instruments equipped with the same model of sensor that would typically be considered identical are compared, with one producing noticeably different results. A 30-µm standard preparation (Thermo 9030 glass standard: 29.5 µm diameter, 2.44 g/cm3 density, 1.52 refractive index) was suspended in a water–glycerol solution (Δn of approximately 0.11) and analyzed on each LO counter (HIAC Royco Model 9703 or 9703+ Liquid Particle Counting System with a 

 [USP]  <788> calibrated HRLD 400 sensor). Compared to counters B and C, the results obtained from counter A were subtly, though noticeably, different: both the mean and standard deviation of the particle distribution increased. Thus, counter A, though otherwise identical to counter B and counter C, must have a different sensitivity to contrast issues.

	Inaccurate results generated-variability by flow rate

	An aspherical particle can flow through an instrument sensor either in line with the surrounding laminar flow and appear smaller than actuality, or be rotating, or “tumbling,” and appear larger than actuality (6). If results are sensitive to flow rate, it may be a symptom of aspherical particles, as increased flow rate can increase the likelihood of tumbling. It is also likely that different instrument types affect the incidence of tumbling. Sensitivity to flow rate (as a symptom of aspherical particles) may appear subtly as a broadened size distribution, but may indicate that further evaluation is needed. While certain other types of particles, such as proteins, are also sensitive to flow rate, evaluation for aspherical particles should always be performed if they are at all a consideration. Depending on the level of accuracy required, an additional calibration curve or other means of correlation may be warranted. However, any LO result generated with an additional calibration curve should only be considered a worst-case scenario, as counts and sizes may be biased high for particles less aspherical than those represented by the additional calibration curve.

	The simplest and easiest way to check for aspherical particles is microscopy, though any technique that incorporates visual analysis may be applicable. Observed particles do not need to be highly aspherical to indicate an impact on results, as symptoms may be noticeable for particles with an aspect ratio (the ratio of the minor to major axis) up to approximately 0.5, with more significant symptoms appearing if aspect ratios are approximately 0.25 or less. However, aspherical particles typically only noticeably impact results when they are a large portion of the particle population. Therefore, if they are not observed upon visual analysis it is highly unlikely they are a root cause of any symptom.

	To illustrate how the apparent size of an aspherical particle is impacted by its orientation, see 

. It is likely that only a portion of a particle is within the sensor’s field of view at any given time, as LO instruments are configured to reduce particle coincidence by only viewing a thin slice of solution (increasing the maximum analyzable particle concentration). As the field of view begins to approximate a cross-section for larger particle sizes, the apparent size of an aspherical particle will become orientation dependent. An aspherical particle that stays in line with the surrounding laminar flow will appear relatively small, while a tumbling aspherical particle may pass through the sensor sideways and appear much larger.

	In summary, a variety of issues can arise while performing LO testing, causing symptoms that either make results unable to be generated, or generate inaccurate results. If results cannot be generated, dilution is typically an appropriate next step, though it must be done with care to ensure it does not cause any additional issues. If results generated appear inaccurate, further evaluation should be performed specific to the situation. When diagnosing any symptom, it is important to be aware of both the chemical and physical characteristics of any expected particles, as they can give critical insight. Symptoms that generate inaccurate results should be evaluated especially carefully due to the risk associated with their potential to be overlooked. In most cases, measures can be taken to ensure testing per LO is suitable, though in certain situations other techniques should be considered.

 104, 1899–1908 (2015).
	2. T. Werk, D.B. Volkin, and H.C. Mahler, 

 Sci 53, 95–108 (2014).
	3. Z. Hu and D.C. Ripple, 

 119, 674–682 (2014).
	4. P. Oma, D.K. Sharma, and D. King, 

 36, 311–320 (2010).
	5. D.C. Ripple and Z. Hu, 

 33, 653–672 (2016).
	6. R.E. Cavicchi, et al., 

	Brent Denton is scientist, John Bak is principal scientist, and Jonathon Salsbury is associate director; all at PPD Laboratories GMP Lab.


	Vol. 44, No. 5
	May 2020
	Pages: 52–55

	When referring to this article, please cite it as B. Denton, J. Bak, and J. Salsbury, “The Limits of Light Obscuration,” 

Articles

Light obscuration testing is the preferred method of sub-visible particle quantification but is not suitable for every preparation.