Janie Dubois is Product Manager, Analytical Imaging, Malvern Instruments (USA).

Janie Dubois

The full version of this analytical technologies feature can be read in the May issue of our digital magazine: http://www.pharmtech.com/ptedigital0510

NIR chemical imaging is based on the same principles as NIR spectroscopy, but uses a focal plane array detector in place of a single detector element to capture the spectral signature at tens of thousands of spatially resolved positions simultaneously. Some have called it NIR spectroscopy on steroids! The parallel data acquisition results in a very high throughput; one image takes under 3 minutes to acquire, which corresponds to 2 milliseconds per full spectrum. For certain quality control applications, such as monitoring the spatial distribution of an API for example, it is possible to acquire the spectral information at a few selected wavelengths. In the latter case, all spatial positions are interrogated in a few seconds, which means that hundreds to thousands of discrete samples can be imaged in a day.  

The main difference between NIR chemical imaging and traditional NIR spectroscopy is the dimensionality of the measurement, which relates to the purpose of the measurement. Traditional NIR spectroscopy is an averaging technique; a single spectrum is acquired from a fairly large spot and possibly all the way through a tablet. This spectrum represents the average chemical composition of the tablet and is very powerful to determine the concentrations of various organic ingredients. NIR chemical imaging is a detail-oriented technique; a multitude of spectra are acquired, each from a very small spot on the tablet and from a limited depth. Each spectrum represents the chemical composition of a small area, which enables the comparison of all the spots acquired to determine the spatial distribution of various ingredients within one tablet. This spatial distribution is then compared amongst samples and correlated with performance parameters such as dissolution rate and stability. 

NIR chemical imaging is a very powerful tool for the analysis of coated and uncoated pharmaceutical tablets, granules, extrusion cores and transdermal delivery systems. The key is always to aim for an application where it is the chemical heterogeneity present within the sample that is important. Chemical heterogeneity can be deliberate and part of the structure that yields the desired performance, or it may be caused by process changes and negatively impact performance. Such applications abound in various steps of formulation design, analytical services and quality control. Examples include the determination of blending, coating or milling endpoints in formulation design, risk assessment of ingredient selection and changes of suppliers, troubleshooting of dissolution failure, and counterfeit detection and sourcing. If the analysis seeks a comparison of means (for example dose uniformity between tablets), traditional NIR spectroscopy is more appropriate. 

Most analyses performed with NIR chemical imaging on solid dosage forms are done in diffuse reflectance — a sampling technique that simply involves positioning a tablet or other sample on the sample stage and "staring" at it with the camera for a few seconds to a couple of minutes. The main advantage of using diffuse reflectance in NIR imaging with a system providing a wide depth of field is the absence of sample preparation, which increases analysis speed and decreases personnel involvement (and associated cost).  

 Finally, NIR chemical imaging is a great tool for Quality by Design (QbD), laboratory analysis and even at-line measurements, but generally overkill for on-line monitoring. The technology exists to perform blend monitoring on line and automation in data analysis can be implemented, but experience has revealed that it generates an amount of data that often needs to be distilled down to become manageable, to a point that makes it only marginally more informative than a single point, traditional NIR on-line probe would have been. The cost of on-line implementation is also an important consideration. Also, NIR chemical imaging is better suited for solid samples; powders with a particle size in the single or tens of microns range are better interrogated by a particle imaging system. Such systems can rapidly sort through tens of thousands of particles to find the ones of interest based on size and shape; chemical analysis with a spectroscopic approach can then be more efficiently performed strictly for these selected particles. 

Better understanding of processes and products leads to better control, a smaller regulatory burden when the time comes to implement changes and better profitability. In this regard, NIR chemical imaging is fully aligned with the QbD initiative and provides information that bridges the gap between starting material quality assurance and performance verification of finished tablets, transdermal patches and oral wafers. In addition, the ability to analyse intermediates enables the gathering of a knowledge base on the effects of ingredients and modifications at various stages of the process during formulation development, and further link into the consequences for finished products. In short, NIR chemical imaging is a risk assessment tool that provides quantitative data for the determination of critical quality attributes. 

The knowledge gathered during formulation development can save a lot of time and money when having to deal with a performance failure. However, even when it is not used to collect this important information during formulation development, NIR chemical imaging remains a powerful tool for root-cause analysis of performance failure of finished products and intermediates, enabling the rapid implementation of corrective measures. As such, it can have a significant impact on production efficiency and cost. 

The power and limitations of NIR chemical imaging are quite well-understood at this point and the technology already available will satisfy measurement needs. We believe that NIR chemical imaging-derived information will be increasingly requested for new product applications because of the fact that it bridges a gap in product and process understanding that is important for the success of QbD. As more applications include the technical data acquired with this instrumentation, there will be less acceptance of the old gaps in knowledge and assumptions.  

Automation of both data acquisition and processing will definitely see the most innovations over the next decade as the technique moves into routine measurements for formulation development and quality assurance. It will become more of a "walk-up" instrument, utilised by non-spectroscopists and providing a numerical, or even a pass/fail type of answer. Standard operating procedures and validated methods will be key to routine deployment, especially to address the needs of ever-shrinking analytical laboratory staff in many pharmaceutical companies.  

Articles

The author assesses the power and limitations of NIR chemical imaging and its future.

When and why should the pharmaceutical industry use NIR chemical imaging?

NIR chemical imaging is based on the same principles as NIR spectroscopy, but uses a focal plane array detector in place of a single detector element to capture the spectral signature at tens of thousands of spatially resolved positions simultaneously. Some have called it NIR spectroscopy on steroids! The parallel data acquisition results in a very high throughput; one image takes under 3 minutes to acquire, which corresponds to 2 milliseconds per full spectrum. For certain quality control applications, such as monitoring the spatial distribution of an API for example, it is possible to acquire the spectral information at a few selected wavelengths. In the latter case, all spatial positions are interrogated in a few seconds, which means that hundreds to thousands of discrete samples can be imaged in a day. 


NIR chemical imaging is a very powerful tool for the analysis of coated and uncoated pharmaceutical tablets, granules, extrusion cores and transdermal delivery systems. The key is always to aim for an application where it is the chemical heterogeneity present within the sample that is important. Chemical heterogeneity can be deliberate and part of the structure that yields the desired performance, or it may be caused by process changes and negatively impact performance. Such applications abound in various steps of formulation design, analytical services and quality control. Examples include the determination of blending, coating or milling end‑points in formulation design, risk assessment of ingredient selection and changes of suppliers, troubleshooting of dissolution failure, and counterfeit detection and sourcing. If the analysis seeks a comparison of means (for example dose uniformity between tablets), traditional NIR spectroscopy is more appropriate.

Most analyses performed with NIR chemical imaging on solid dosage forms are done in diffuse reflectance  a sampling technique that simply involves positioning a tablet or other sample on the sample stage and “staring” at it with the camera for a few seconds to a couple of minutes. The main advantage of using diffuse reflectance in NIR imaging with a system providing a wide depth of field is the absence of sample preparation, which increases analysis speed and decreases personnel involvement (and associated cost).  Finally, NIR chemical imaging is a great tool for Quality by Design (QbD), laboratory analysis and even at-line measurements, but generally overkill for on-line monitoring. The technology exists to perform blend monitoring on line and automation in data analysis can be implemented, but experience has revealed that it generates an amount of data that often needs to be distilled down to become manageable, to a point that makes it only marginally more informative than a single point, traditional NIR on-line probe would have been. The cost of on-line implementation is also an important consideration. Also, NIR chemical imaging is better suited for solid samples; powders with a particle size in the single or tens of microns range are better interrogated by a particle imaging system. Such systems can rapidly sort through tens of thousands of particles to find the ones of interest based on size and shape; chemical analysis with a spectroscopic approach can then be more efficiently performed strictly for these selected particles. 


Better understanding of processes and products leads to better control, a smaller regulatory burden when the time comes to implement changes and better profitability. In this regard, NIR chemical imaging is fully aligned with the QbD initiative and provides information that bridges the gap between starting material quality assurance and performance verification of finished tablets, transdermal patches and oral wafers. In addition, the ability to analyse intermediates enables the gathering of a knowledge base on the effects of ingredients and modifications at various stages of the process during formulation development, and further link into the consequences for finished products. In short, NIR chemical imaging is a risk assessment tool that provides quantitative data for the determination of critical quality attributes.

The knowledge gathered during formulation development can save a lot of time and money when having to deal with a performance failure. However, even when it is not used to collect this important information during formulation development, NIR chemical imaging remains a powerful tool for root-cause analysis of performance failure of finished products and intermediates, enabling the rapid implementation of corrective measures. As such, it can have a significant impact on production efficiency and cost.The power and limitations of NIR chemical imaging are quite well-understood at this point and the technology already available will satisfy measurement needs. We believe that NIR chemical imaging-derived information will be increasingly requested for new product applications because of the fact that it bridges a gap in product and process understanding that is important for the success of QbD. As more applications include the technical data acquired with this instrumentation, there will be less acceptance of the old gaps in knowledge and assumptions. 


Automation of both data acquisition and processing will definitely see the most innovations over the next decade as the technique moves into routine measurements for formulation development and quality assurance. It will become more of a “walk-up” instrument, utilised by non-spectroscopists and providing a numerical, or even a pass/fail type of answer. Standard operating procedures and validated methods will be key to routine deployment, especially to address the needs of ever-shrinking analytical laboratory staff in many pharmaceutical companies. 

NIR chemical imaging is based on the same principles as NIR spectroscopy, but uses a focal plane array detector in place of a single detector element to capture the spectral signature at tens of thousands of spatially resolved positions simultaneously.

When and why should the pharmaceutical industry use NIR

What do mapping and imaging techniques add to traditional spectroscopic analysis for pharmaceutical applications?

With bulk or traditional single point spectroscopy, one spectrum is collected for each sample using a single element detector with relatively short measurement times. For tablets, this provides one averaged measure of composition for either the whole tablet or a selected portion. Mapping approaches and chemical imaging, on the other hand, record multiple spectra for each sample, providing insight into the distribution of components — most usefully the API and key excipients — across the tablet's surface.  

Knowledge of the spatial location of different chemical components is valuable for a wide variety of applications, but is particularly useful for the analysis of tablets because it enables manufacturers to understand the links between raw materials, process parameters and a drug's critical quality attributes — the parameters that dictate clinical performance. By identifying the physical distribution of chemical components within the tablet, mapping and imaging techniques allow an appreciation of the relationship between structure and functionality. This helps with the more precise and sensitive definition of critical quality attributes and with the development of understanding of how these can be effectively controlled during manufacture. In this way, these systems closely support the knowledge-led approach enshrined in Quality by Design because they provide the insight necessary to gain a more complete model of product behaviour.  

Which applications require the extra spatial detail provided by imaging technology?

Assessing the information required for a particular application is crucial when selecting a spectroscopic system and the need (or not) for spatial detail is key. While case by case consideration is advised, the following examples review some common objectives and information goals. 

Quality assurance or control often requires only the average chemical composition of a tablet, together with perhaps, averaged measures of moisture content and coating thickness. All of these can be estimated using bulk measurements. Mapping or imaging is needed where spatial resolution is essential, for example in determining chemical and moisture distribution within a sample, and/or coating uniformity. This is the case when distribution rather than just average composition critically impacts clinical performance.  

For troubleshooting purposes, bulk measurements provide adequate information about chemical content (dosage) variation over time in a manufacturing run or between batches, or to study differences in properties such as averaged coating thickness between tablets. Spatial distribution information, on the other hand, can help to locate problems with coating integrity and thickness variation across a tablet's surface or at the apex of angled tablets, for example, which may lead to dissolution test failures. 

For more in-depth investigation, such as root cause analysis of dosage variation and the underlying reasons for a failed batch, or batch-to-batch variability, spatially resolved measurements provide much more information and have become routine in many companies. This application demands careful consideration of every piece of available information and is consequently an area where imaging techniques have proven their value. Near-infrared chemical imaging, in particular, is favoured for troubleshooting because the field of view of a single image can encompass a whole tablet surface, or even multiple tablets, which makes it very fast and appropriate for the analysis of hundreds of tablets per day. This higher throughput is facilitated by automation and is essential to obtain enough information to confidently make decisions about processes.  

Poor counterfeits can be detected easily with single-point spectroscopy because they are often made with different ingredients, or even with no API. A closely copied fake, however, may be extremely similar to a genuine medicine, with differences undetectable by bulk analysis. With spatially resolved information, it is possible to see the size of API and excipient domains across the tablet surface and their relative distribution, both of which are artifacts of the manufacturing process. The positioning of ingredients in a tablet is equivalent to a fingerprint of its formulation and process; for counterfeit investigations, this fingerprint not only demonstrates the fact that it is a fake, but provides more sensitive differentiation that can even help with determining the likely source of a counterfeit product.  

The latest guidance from the FDA and the ICH encourages an approach where quality is built into a product (Quality by Design) rather than tested for post-production. This relies on a thorough understanding of product performance and the manufacturing parameters that control it. Spatially resolved chemical composition, blend heterogeneity, preferential aggregations, moisture migration and hydration state changes are common examples of product parameters requiring mapping or imaging techniques. Bulk measurements, on the other hand, may be used to monitor dose uniformity, total moisture and sometimes hydration state changes, provided these are large enough to be detected when 'diluted' in the spectrum of the whole tablet. 

How do the different spectroscopic systems compare in practical terms?

While all spectroscopic techniques require a certain level of operator proficiency, some systems offer greater user-friendliness or higher levels of automation — both of which can extend the number of users able to operate the system and/or enhance productivity.  

Macro, programmable or SOP-type options are necessary tools for the deployment of measurements to a larger group of users. Some degree of sample preparation, presentation and instrument set up is inevitable, but this varies considerably from technique to technique and should be carefully considered at the instrument selection stage. Generally, spectroscopic data acquisition is easy and fast, but data manipulation to extract relevant and valuable information is more tedious. As such, this is the portion that benefits most from automation tools.  

Table 1: Concentration distribution estimated from the surface of a whole tablet and from two selected areas, a and b.			

Once a mathematical model is in place, averaged compositional analysis by conventional spectroscopy is a simple, cost-effective method. Data acquisition is rapid and easy, with any operator able to generate results; productivity is in the order of thousands of samples per day. Usually an entire tablet is analysed, but if the edges and centre are compositionally identical, then a proportion of the tablet may be sufficiently representative. 

 gives a concentration distribution for a tablet based on (a) whole tablet image and (b) two smaller portions of the same tablet. 

 shows the chemical images of the three areas. Together, these underline the importance of selecting a sampling method appropriate to the measurement objective when employing a bulk or averaged measurement technique.  

Figure 1: Chemical images of the surface of a whole tablet and two sub-areas displayed on the same colour scale.			

Compared with single point measurement, linear mapping is much more manually intensive and time-consuming. With mapping systems, a computer-controlled stage moves the sample relative to the detector, allowing measurement at different points across the designated area. Individual points at which spectra are collected can be selected to cover either the entire tablet area or a representative portion; pixels in a mapped image may arise from areas bordering each other or not. Subsequent analysis of the spectra reveals the identity and distribution of the chemical species across the surface of the sample. Mapping systems demand a high level of flatness that often requires sample embedding and/or sectioning, and data acquisition rates are low; high resolution mapping productivity is in the order of up to just tens of samples per day.  

As with mapping, chemical imaging enables the accurate visualisation of almost all chemical compounds in the tablet, determining both the amount present and its location. However, chemical imaging collects thousands of spectra simultaneously from an entire field-of-view, without moving the sample, reducing measurement times to about 1 min. Focusing on a few key wavelengths cuts these further to just seconds, which means hundreds of samples can be measured in a single day. In general, camera-based imaging systems deal well with raw samples, such as whole tablets, granules and transdermal patches, or even whole blister packs, and analysis can be conducted through the packaging.  

The camera field-of-view can encompass either the sample surface or 'zoom in' to an area of interest and the pixels of the camera always probe a contiguous area. Since an automated sample changer is invaluable for high throughput projects, selecting an instrument that can be equipped with one and that supports automated data acquisition will significantly streamline analysis. 

How can I evaluate the cost-benefit ratio of different systems? 

Tablet analysis is conducted for a variety of reasons and the chemical information sought depends on the overall objective; therefore, the cost–benefit ratio is different in every case.  

The basic equipment cost associated with spectroscopic technology increases from conventional to mapping systems, with chemical imaging systems requiring the largest upfront capital investment. However, this analysis ignores ongoing cost. On the basis of cost per spectrum, or cost of analysis per tablet, single-point spectrometers remain the cheapest solution, but imaging systems are the next least expensive, with mapping becoming the most costly option because of low productivity and the maintenance requirements for moving parts systems. Finally, the cooling mechanism is also a consideration. Cooling systems that use liquid nitrogen may require filling during a working day and there is also the need for maintenance of the dewar itself and all seals. The stability of the cooling system and its effects on spectral noise should be carefully considered with relation to the throughput required of the instrument and the environment in which it will operate. 

Deciding how much money to spend on instrumentation also depends on the financial savings it will deliver. A failed batch that requires reprocessing is a costly fault. Being able to pinpoint the cause and avoid a recurrence requires complete information and the flexibility to access and scrutinise data for hundreds or thousands of tablets. Here, the potential financial savings may well outweigh the initial investment in an instrument capable of solving this problem.  

In summary, the decision as to which spectroscopic technique is best suited for a given tablet application rests on a number of important factors. Beyond selecting the most appropriate analytical approach based on application range and the type of information required, choosing the right spectroscopic tool demands a clear understanding and consideration of key issues such as analysis speed, number and expertise of users, potential time and cost gains. 

Single point conventional spectroscopy is a well-established, inexpensive method that offers fast averaged composition data for applications where this information is sufficient. Chemical imaging, on the other hand, is the most cost-effective approach over the long term where there is an ongoing need for spatially resolved chemical analysis to support knowledge-led development, effective troubleshooting, root cause analysis and the most sensitive quality control/counterfeit detection. 

Product Manager, Analytical Imaging, Malvern Instruments (USA)

Vibrational spectroscopy, which encompasses near-infrared, mid-infrared and Raman spectroscopy, is an important analytical tool in the pharma industry.