Insights from industry

Figure Out Any Particle with Automated Counting, Sizing, Morphology, and Identification

insights from industryLisa KrapfField Automation ScientistUnchained Labs

In this interview, AZoM talks to Lisa Krapf, Field Automation Specialist at Unchained Labs, about how automated counting, sizing, morphology, and identification can be used to figure out any particle.

Where is particle analysis used?

Particle analysis is ubiquitous throughout product development to manufacturing. It is used early on in development to determine characteristics such as morphology and size distribution, formulation component aggregation propensity, or API quantities in complex mixtures. These properties are important indicators for candidate selection because they can impact bioavailability, performance, and stability.

At a later stage, particle identification can be used to inform on product purity and monitor particle changes in formulation mixtures. This information can help fast track process development, monitor process variability, and prepare the batches before release.

There are many forms of particle analysis used to analyze and characterize particles to allow biologics development. For example, it is important to characterize aggregation content to understand mechanisms for formulation stability. Also, API particle size and shape may change as processes are transferred to manufacturing. 

Another example of particle analysis application is inspecting raw materials to investigate how their quality impacts the quality and safety of the finished product. These are just a few examples of why people study particles and biologics development.

What conditions are needed for successful particle analysis?

There are different types of particles and different reasons to analyze them. The technical requirements to assess particles are high. To thoroughly examine the particle, we need to have the ability to describe the particles by count, size, and shape. But the quantity of particles and their quality does not tell a full story, especially if the particles are in mixed populations. Because of this, the identification of particles is also an important consideration for particle analysis.

To get statistically significant data to draw the right conclusion, it is also essential to be able to characterize a large number of particles in an experiment. The technique should be easy to use and have a minimum impact on particles during sample preparation to prevent introducing changes to particle number or morphology.

What benefits does the Hound offer in particle analysis?

The Hound is the only instrument that covers every particle analysis need throughout the life cycle of a drug product. It is the only particle analyzing instrument that can characterize particles by count, size, and shape, while also being able to perform chemical or elemental identification on the same platform. Hound is also 21 CFR part 11 compliant, making it a fitting instrument for regulated labs.

The particle size distribution is a critical quality attribute throughout the life of a drug product. It is important to report the particle size distribution since this directly impacts product safety and drug absorption.

Hound is a fully automated microscopy system that will automatically acquire images and analyze them for you. It does count, shape, and size distribution for a wide variety of sample types, whether the particles are in solution or a dry dispersion.

Hound can automatically scan and stitch a large area when it images the sample. It detects the particles on each image, counts the particles, and stores size and shape information for each particle. It also tracks the coordinates of the particle so that you can go back to find the particles of interest at any time to do further analysis. Hound can also analyze particles in a size range from 2 micrometers to 15 millimeters, and the user-friendly software lets you set up custom size bins so you can focus on particle sizes that are important to you.

Sample preparation for Hound analysis is also very straightforward. If you have particles in solution, you can filter them on a gold-coated membrane filter and directly analyze the immobilized particles on the filter rounds. If you have powdered samples, you can directly disperse the particles on the filter round and analyze them right away.

Particle analysis is particularly important in drug products. Could you talk us through an example of how Hound has benefitted the analysis of drug products?

We have used Hound for particle counting with regards to drug products, for example, inhalers. It is important to analyze foreign particles from an inhaler device to assess the safety of the device. However, collecting all the particles generated from an inhaler spray can be cumbersome. With Hound, we demonstrated an innovative way we capture particles from an actuated inhaler and an easy way to analyze the insoluble foreign particles.

Two actuations of a metered-dose inhaler were collected in a particle-free and air-tight collection device that you can see on the second picture from the left. To be able to quantify foreign particles, the soluble API and excipient particles needed to be removed. To dissolve them, 70 milliliters of solvent was added. The liquid was then filtered using our 0.8-micrometer filter round to isolate the insoluble particles on the filter. The particles were then set to dry out and directly mounted on Hound for automated count and sizing.

Hound is fast at counting; it can generate a complete size distribution report of more than 6,000 particles in 2 minutes with automated counting and sizing analysis. In the inhaler example, Hound's counting capability was used to evaluate foreign particle load by looking at particles of 2 micrometers and larger.

The need for characterized particle size distribution will never go away in the drug product life cycle. But, it will become increasingly challenging to get the correct particle size distribution in the later development stages as the composition of formulation becomes more complicated than just a single active ingredient.

Could you explain how using Raman spectroscopy alongside laser-induced breakdown spectroscopy helps users identify particles?

Besides providing particle characterization, such as counting and sizing, Hound can also perform chemical and elemental ID with the Raman and laser-induced breakdown spectroscopy (LIBS). It has three laser options to ID almost anything. The three lasers are Raman 785 and 532 nanometers and laser-induced breakdown spectroscopy.

The dual Raman lasers in Hound can perform chemical fingerprints on a wide range of particles. The red 785-nanometer laser can help to ID fibers and contaminants such as a lab wipe, and the green 532-nanometer laser is ideal for identifying protein aggregates. Both lasers can identify particles down to 2 micrometers and use a built-in and customizable reference database to compare the collected spectra and find a match for the material of the particle.

Raman can identify a wide range of organic and inorganic material down to 2 micrometers based on their chemical structure. It can identify cellulose, proteins, different types of polymers or even the polymorphs of other compounds. Unlike FTIR, Raman is inert to water, so identifying biologics of particles in solution is not a problem for Raman.

Hound uses gold-coated filter rounds as the gold surface is Raman inactive, which results in no noise and high-quality spectra. Hound uses LIBS to identify metal and other inorganic particles down to as small as 20 microns. For laser-induced breakdown spectroscopy, a high-energy pulsed laser is focused on the surface of the particle, ablates a small volume of the particle, and creates a plasma. When the atoms in that plasma fall back to their inactive state, they emit element-specific spectra that are measured for material identification.

Like Raman spectroscopy, LIBS identifies the particle materials by matching a known spectrum to built-in or custom-added spectra in the reference database. It is a great addition to Raman spectroscopy, which cannot identify Raman-inactive metal particles, or glass, which typically emits low Raman signals. But, these materials typically provide a great and very material specific signal when doing LIBS. The other great thing about LIBS is that it only needs one second to identify a particle.

No additional sample preparation is required for LIBS. It is very sensitive and can distinguish minor variations in material, like discrimination of field particles, or glass fragments from different sources, by their elemental composition. Lastly, as LIBS ablates a small volume of material on the particle surface, it can also be used to perform a layer analysis by taking several measurements on the same position and observing the changes in elemental composition, while acquiring spectra from different depths.

How does Hound make the identification of particles easier? Is it possible to identify particles in complex mixtures?

Besides using the gold-coated filter round to filter out the solution or dispersed dry powder samples on the round, users can also directly apply samples in different formats, such as a gel or slurry on the gold-coated wet rounds without additional dilution or preparation. If someone wanted to study particles in solution, users can simply load the colloidal or biologics in solution on wet rounds to analyze the sample in situ, without worrying about losing particles, shearing, or breakage during the sample preparation process by filtration.

Take an example that includes API distributions in a topical cream without additional sample preparation using wet rounds. In this topical cream example, it was very challenging to separate two similarly sized and shape API just with count, size, and shape characterization alone.

We have been able to show that we can use Hound's Raman spectroscopy capability to obtain the particle size distributions of the two different API in a mixed sample. This helps to understand the composition of the topical cream formulation and can, for example, be used to compare batch-to-batch variation.

Using a Raman inactive gold-coated wet round, the topical cream can be directly smeared on the slide and analyzed with Hound without any additional sample suspension or dissolving steps.

The API particles in the topical cream were similar in size and shape. It would be impossible to distinguish the API population by just size, count, and morphology alone, which is all the most particle analysis instruments would be able to do. Knowing the APIs are adapalene and benzoyl peroxide, we created custom reference spectra for both APIs and then we used Raman spectroscopy on Hound to analyze enough particles to get statistically meaningful size distributions for both API. Using fully automated Raman analysis mode, Hound can identify and distinguish over 4,000 API particles in less than three hours.

Using Raman spectroscopy combined with image analysis for size distribution, we could get distinct API size distributions from the two different APIs of similar size and shape in the mixed sample in this topical cream example. In this topical cream, about three times more benzoyl peroxide particles were found than adapalene particles. We also got information on the size distribution of each particle type. For example, benzoyl peroxide had a D90 value of only 23.7 micrometers, while the D90 value of adapalene was 7.6 micrometers.

Using the automated Raman spectroscopy feature on Hound, precise API characterization including count, size, and chemical identification of the topical cream was completed quickly. It is a perfect tool to quickly give answers to minor changes in particles in order to fast-track formulation and process development.

Hound's image analysis also performs morphology measurements for rectangularity, elongation, equivalent circle diameter, and fibrosity for particle classification. The morphology grouping combined with identification makes Hound an efficient and rapid identification tool.

In another example of how Hound streamlines the process of working with complex and mixed samples, we assessed the size distribution of 1,000 API particles in a nasal spray. The API was mixed with high filler content to ensure product integrity. The challenge was that the API content ratio was so low in the nasal spray formulation that looking for 1,000 API particles would be like looking for a needle in a haystack.

In this example, an initial analysis determined that the relative number of API particles in the sample was 2%, while the cellulose filler content was 98%. From comparing reference samples, we learned that the elongation factor of API particles was always below 2.5, in contrast to the cellulose particles that were typically slightly elongated and had an aspect ratio of up to five.

A secondary selective analysis method was then set up to automatically direct Raman identifications towards particles with elongation factors less than 2.5. This morphologically directed Raman analysis increased API particle detection rates from 2% to 34%. As a result, only 3,000 particles needed to be identified to achieve 1,000 API identifications, as opposed to 45,000 particles needed without morphology classification beforehand. With the morphology classification to home in the API population, the identification of 1,000 API particles was achieved 15 times faster.

In this nasal spray example, the average API size was 5.2 micrometers with a D90 value of 12.5 micrometers. By adjusting the particle selection parameters to perform morphology directed identification, it was possible to only focus on particles of interest without wasting time screening through all of the particles in a sample. This makes Hound great for the precise characterization of particles of interest in complex mixtures.

How can the Hound help with the identification of contaminants in samples?

One major reason for batch failures, in parenterals, for example, is when visible particles are found in the product. Particles can come from a variety of sources, including materials that are intrinsic to the product, such as protein aggregates, or they can be external contaminations from any step in the manufacturing process.
When these contaminations occur, you usually want to find out how critical they are, where they come from, and hopefully eliminate the root cause of the particle contamination. Hound can help you by quantifying the particle contamination and even identify the root cause using Raman and LIBS to figure out the particle material and compare it to your lab or production-specific possible particle sources.

For example, we found fibers in a product. These fibers caused three batches of the product to fail. Therefore, there was an urgent need to find out where those fibers came from and to eliminate the source of the particles. The Raman spectrum clearly showed that it was a cellulose fiber, but in addition to knowing that it was cellulose, we wanted to know what was the particular source of this cellulose fiber. The Raman spectrum showed an additional peak at 1,600 wavenumbers, which is not present in most cellulose spectra. As a result, we could use this peak to find the specific source of the contaminant.

An investigation was started to find out where in the process those fibers appeared. First, common lab equipment and supplies were rinsed and the wash liquid was filtered to find out where the fibers appeared, and I can give examples from three of the samples that were taken.

The rubber stoppers used for vial capping, the clean-in-place equipment, and the tubing that was used in the process were analyzed. The filters prepared from those rinses were analyzed on Hound, and some fibers were found. But, when taking Raman spectroscopy, it turned out that none of those fibers had the characteristic peak at 1,600 wavenumbers, which was seen for the contaminant.

The next step was to collect all types of cellulose materials in the lab, take spectra of those, and compare them with the contaminant spectrum. The range of materials that were collected included different clothing fibers, lab wipes, and fibers from the autoclave bag. None of these materials had an extra peak at 1,600 wavenumbers. Also, the overall matching rank between the contaminant spectrum and those spectra was poor. That was a hint that none of those was the actual contamination source.

Another step was taken to investigate possible contamination sources upstream in API manufacturing. Again, several samples were taken from different cellulose materials and reference spectra acquired. This time, it turned out that there was a lab wipe that gave us a spectrum that had a very specific peak at 1,600 wavenumbers. It became clear that this was probably the source of contamination.

So, there was a very good idea about what the source of the contamination was, but it was not clear how it got into the product. To figure out that last question, the tanks that were used in manufacturing were rinsed, and the rinse was filtered and analyzed with Hound. Hound captured images of fibers that were found in the tank. The Raman spectra matched the contamination fiber and the sample taken from the lab wipe. With this information, it was possible to figure out the pathway, and how particles got into the product. The tanks were cleaned with these lab wipes that released fibers, and those fibers ended up in the product. Knowing this pathway, the particle source could be easily removed.

Hound can also be used for root cause investigation for metal contaminants using laser-induced breakdown spectroscopy. In another example, we used Hound to identify a particle originating from a crimp cap as aluminum. However, simply knowing the material was aluminum could be too broad to understand where the shard came from.

By creating a complete custom referenced spectrum library, precise matching can be achieved to determine the exact root cause of where a particle comes from. In this example, a precise match down to a name brand of crimp cap could be achieved by adding a custom crimp cap reference spectrum.

This crimp cap test was sensitive enough to catch the subtle difference in the 403-nanometer peak to identify the contaminant culprit not only as aluminum but an aluminum crimp cap. This illustrates the power of the customizable reference library that can give you a specific contaminant characterization.

Hound comes with over 150 Raman reference spectra and 50 LIBS references. To get exact matches for the unique particle of your interest, you can add a spectrum of materials specific to your process into the reference database for future matching. It takes only 15 minutes to load the samples, obtains the spectrum from your materials, and add it to your custom reference database. The more specific references you add to the database, the more precise matches may be.

Could you summarize the main benefits of the Hound?

Hound is a microscopy-based particle analysis tool that can provide quick, manual, or automated particle counts, as well as information on size and morphology, in only minutes. Hound can be used for a variety of applications that need particle characterization and identification. It can provide quality and quantity information of the direct product through particle characterization, and it is the only particle analyzer equipped for Raman and LIBS identification capability to properly monitor process and batch variability and contamination events.

The easy-to-prepare samples retain sample integrity since it can often be measured in their native form. It can also rapidly identify particles in mixtures by using morphology-directed Raman or LIBS identification. Hound is made for a regulated lab environment as a 21CFR part 11 compliance tool.

Hound will help when you find yourself in a situation where you cannot easily distinguish particles of the same size and similar shape, and you need parameters other than size to distinguish between the two.

About Lisa Krapf

Lisa KrapfLisa Krapf is a field automation scientist at Unchained Labs. She graduated with a Diploma degree in Physics. Lisa has 5 years of experience working on particle characterization using microscopic and spectroscopic tools in different application areas with a focus on the pharmaceutical industry.

Disclaimer: The views expressed here are those of the interviewee and do not necessarily represent the views of Limited (T/A) AZoNetwork, the owner and operator of this website. This disclaimer forms part of the Terms and Conditions of use of this website.


Please use one of the following formats to cite this article in your essay, paper or report:

  • APA

    Unchained Labs. (2020, October 28). Figure Out Any Particle with Automated Counting, Sizing, Morphology, and Identification. AZoM. Retrieved on December 07, 2022 from

  • MLA

    Unchained Labs. "Figure Out Any Particle with Automated Counting, Sizing, Morphology, and Identification". AZoM. 07 December 2022. <>.

  • Chicago

    Unchained Labs. "Figure Out Any Particle with Automated Counting, Sizing, Morphology, and Identification". AZoM. (accessed December 07, 2022).

  • Harvard

    Unchained Labs. 2020. Figure Out Any Particle with Automated Counting, Sizing, Morphology, and Identification. AZoM, viewed 07 December 2022,

Ask A Question

Do you have a question you'd like to ask regarding this article?

Leave your feedback
Your comment type