Partager l'article
Whether you are new to the technique, a seasoned veteran, or merely just curious about near-infrared spectroscopy (NIRS), Metrohm is here to help you to learn all about how to perform the best analysis possible with your instruments.
In this series, we will cover several frequently asked questions regarding both our laboratory NIRS instruments as well as our line of Process Analysis NIRS products.
The following questions will be answered in the rest of this post (click to jump to the topic):
- What is the difference between IR spectroscopy and NIR spectroscopy?
- NIR spectroscopy is a «secondary technology». What does this mean?
- What is a prediction model, and how often do I need to create/update it?
- How many samples are required to develop a prediction model?
- Which norms describe the use of NIRS in regulated and non-regulated industries?
- How can NIRS be implemented in a production process?
- How can product quality be optimized with process NIRS?
1. What is the difference between IR spectroscopy and NIR spectroscopy?
IR (infrared) and NIR (near-infrared) spectroscopy utilize different spectral ranges of light. Light in the NIR range is higher in energy than IR light (Figure 1), which affects the interaction with the molecules in a sample.
This energy difference has both advantages and disadvantages, and the selection of the ideal technology depends very much on the application. The higher energy NIR light is absorbed less than IR light by most organic materials, broadening the resulting bands and making it difficult to assign them to specific functional groups without mathematical processing.
However, this same feature makes it possible to perform analysis without sample preparation, as there is no need to prepare very thin layers of analyte or use ATR (attenuated total reflection). Additionally, NIRS can quantify the water content in samples up to 15%.
Want to learn more about how to perform faster quality control at lower operating costs by using NIRS in your lab? Download our free white paper below.
White Paper: Boost Efficiency in the QC laboratory: How NIRS helps reduce costs up to 90%
The weaker absorption of NIR light leads to using long pathlengths for liquid measurements, which is particularly helpful in industrial process environments. Speaking of such process applications, with NIR spectroscopy, you can use long fiber optic cables to connect the analyzer to the measuring probe, allowing remote measurements throughout the process due to low absorbance of the NIR light by the fiber (Figure 2).
For more information, read our previous blog post outlining the differences between infrared and near-infrared spectroscopy.
2. NIR spectroscopy is a «secondary technology». What does this mean?
To create prediction models in NIR spectroscopy, the NIR spectra are correlated with parameters of interest, e.g., the water content in a sample. These models are then used during routine quality control to analyze samples.
Values from a reference (primary) method need to be correlated with the NIR spectrum to create prediction models (Figure 3). Since NIRS results depend on the availability of such reference values during prediction model development, NIR spectroscopy is therefore considered a secondary technology.
For more information about how Karl Fischer titration and NIR spectroscopy work in perfect synergy, download our brochure below.
Read our previous blog posts to learn more about NIRS as a secondary technique.
3. What is a prediction model, and how often do I need to create/update it?
In NIR spectroscopy, prediction models interpret a sample’s NIR spectrum to determine the values of key quality parameters such as water content, density, or total acid number, just to name a few. Prediction models are created by combining sample NIR spectra with reference values from reference methods, such as Karl Fischer titration for water content (Figure 3).
A prediction model, which consists of sufficient representative spectra and reference values, is typically created once and will only need an update if samples begin to vary (for example after a change of production process equipment or parameter, raw material supplier, etc.).
Want to know more about prediction models for NIRS? Read our blog post about the creation and validation of prediction models here.
4. How many samples are required to develop a prediction model?
The number of samples needed for a good prediction model depends on the complexity of the sample matrix and the molecular absorptivity of the key parameter.
For an «easy» matrix, e.g., a halogenated solvent with its water concentration as the measurement parameter, a sample set of 10–20 spectra covering the complete concentration range of interest may be sufficient. For applications that are more complex, we recommend using at least 40–60 spectra in order to build a reliable prediction model.
Find out more about NIRS pre-calibrations built on prediction models and how they can save time and effort in the lab in our blog article below.
5. Which norms describe the use of NIRS in regulated and non-regulated industries?
Norms describing how to implement a near-infrared spectroscopy system in a validated environment include USP <856> and USP <1856>. A general norm for non-regulated environments regarding how to create prediction models and basic requirements for near-infrared spectroscopy systems is ASTM E1655. Method validation and instrument validation are guided by ASTM D6122 and ASTM D6299, respectively.
For specific measurements like RON and MON analysis in fuels, standards such as ASTM D2699 and ASTM D2700 should be followed.
For further information, download our free Application Note.
6. How can NIRS be implemented in a production process?
Chemical analysis in process streams is not always a simple task. The chemical and physical properties such as viscosity and flammability of the sample streams can interfere in the analysis measurements. Some industrial processes are quite delicate—even the slightest changes to the process parameters can lead to significant variability in the properties of final products. Therefore, it is essential to measure the properties of the stream continuously and adjust the processing parameters via rapid feedback to assure a consistent and high quality product.
Curious about this type of application? Download it for free below!
Inline moisture analysis in fluid bed dryers by near-infrared spectroscopy
The use of fiber optic probes in NIRS systems has opened up new perspectives for process monitoring. A suitable NIR probe connected to the spectrometer via optical fiber allows direct online and inline monitoring without interference in the process. Currently, a wide variety of NIR optical probes are available, from transmission pair probes and immersion probes to reflectance and transflectance probes, suitable for contact and non-contact measurements. This diversity allows NIR spectroscopy to be applied to almost any kind of sample composition, including melts, solutions, emulsions, and solid powders.
Selecting the right probe, or sample interface, to use with a NIR process analyzer is crucial to successful process implementation for inline or online process monitoring. Depending upon whether the sample is in a liquid, solid, or gaseous state, transflectance or transmission probes are used to measure the sample, and specific fitting attachments are used to connect the probes to the reactor, tank, or pipe. With more than 45 years of experience, Metrohm Process Analytics can design the best solutions for your process.
Click below to find a selection of free Application Notes to download related to NIRS measurements in industrial processes.
7. How can product quality be optimized with process NIRS?
Regular control of key process parameters is essential to comply with certain product and process specifications, and results in attaining optimal product quality and consistency in any industry. NIRS analyzers can provide data every 30 seconds for near real-time monitoring of production processes.
Using NIRS process analyzers is not only preferable for 24/7 monitoring of the manufacturing process, it is also extremely beneficial for inspecting the quality of raw materials and reagents. By providing data in «real-time» to the industrial control system (e.g., DCS or PLC), any process can be automated based on the NIRS data. As a result, downtimes are reduced, unforeseen situations are avoided, and costly company assets are safeguarded.
Furthermore, the included software on Metrohm Process Analytics NIRS instruments has a built-in chemometric package which allows qualification of a product even while it is still being produced. A report is then generated which can be directly used by the QC manager. Therefore, the product quality consistency is improved leading to potential added revenues.
Do you want to learn more about improving product quality with online or inline NIRS analysis? Take a look at our brochure here.
In the next part, we cover even more of your burning questions regarding NIRS for lab and process measurements:
Frequently asked questions in NIRS analysis – Part 2
Want to learn more about NIR spectroscopy and potential applications? Have a look at our free and comprehensive application booklet about NIR spectroscopy.
Your knowledge take-aways
Optimizing Chemical Recovery in Pulp Mills with online spectroscopy
NIR spectroscopy: The efficiency boost for QC labs
Inline moisture analysis in fluid bed dryers by NIRS
Monograph: A guide to near-infrared spectroscopic analysis of industrial manufacturing processes
