Determining the Best Suited Applications for the Ocean MZ5 Mid-IR Spectrometer
The Ocean MZ5 is a miniature, attenuated total reflection (ATR) spectrometer with measurement capabilities from 1818-909 cm-1 (5.5-11 µm). This fully self-contained instrument — including sample interface, light sources and detector — provides a compact, fast and scalable alternative to traditional FTIR spectroscopy. Uses include chemical discrimination, food and flavorings analysis, environmental testing and scientific research. Some common applications of ATR-MIR spectroscopy are shown in Table I.
Table I. Example MZ5 MIR Applications
|Fuel Analysis||Ethanol spiking in gasoline|
|Fatty acids content in biodiesel|
|Octane level rising|
|Materials identification||Biomaterial analysis|
|Food & Agriculture||Agricultural measurements and monitoring|
|Food and beverage QC|
|Anti-counterfeit/Authentication||Identification and authentication of essential oils|
|Testing and qualification of milk and adulterated substances|
With so many opportunities for MIR spectroscopy, how can a researcher determine whether the MZ5 is a good fit for their application? Will the MZ5 provide the necessary data or is a larger, more expense FTIR system required? Prof. Dr. M Jäger and his students from Robin Legner’s laboratory at Hochschule Niederrhein, University of Applied Sciences in Krefeld, Germany, had exactly these questions. So they spent three weeks evaluating the MZ5 through several student-led projects. Their findings are summarized here.
MZ5 for the Determination of Ethanol Content in Wine
This student-led project investigated the use of the MZ5 to determine ethanol content in multiple varieties of white, rosé and red wines. The MZ5 spectra for four white wine varieties are shown in Figure 1. The spectrum for pure ethanol with a strong characteristic peak at 1065 cm-1 (CCO stretching vibration) is included for comparison. This ethanol peak is observed in the MZ5 spectra for the white wines.
Figure 1. MIR spectra for white wines with the spectrum for pure ethanol included for comparison. The strong ethanol peak at 1065 cm-1 observed in the spectra can be used to determine ethanol content in the wine.
Additional wine measurements were made with the MZ5 by spiking alcohol into red and white wine varieties (standard addition). Multivariate analysis using principal least squares regression (PLS) was done to compare the predicted ethanol versus the known alcohol concentration. In Figures 2 and 3, the ethanol concentrations predicted from the MZ5 spectra agree well with the known ethanol concentrations, indicating that the MZ5 spectral data can be used to determine ethanol content of wine.
Figure 2. Determination of ethanol content in white wine using PLS. The predicted and known ethanol concentrations show good agreement for the white wine samples prepared by standard addition.
Figure 3. Determination of ethanol content in red wine using PLS. The predicted and known ethanol concentrations show good agreement for the red wine samples prepared by standard addition.
The students concluded that the MZ5 yielded precise and robust results for the determination of ethanol in wine. They also found that univariate analysis was not sufficient to quantify ethanol concentration in their samples prepared by standard addition. Multivariate analysis was required.
Finally, with spectral data for a range of wine varieties, principal component analysis (PCA) was applied to the spectral data to group and classify the wines using their spectral signature. The results of the PCA are shown in Figure 4, where grouping of the wines based on whether they are white, rosé or red wine is clearly observed. For this data set, 96% of the variation among the MIR spectra can be explained with three principal components.
Figure 4. PCA demonstrating clustering of MZ5 spectra based on wine type.
Note that these results for wine add to our body of MZ5 measurements around wine and spirits. Our MZ5 application note on Decoding Dangerous Drinks with a Spectral Sensor used the MZ5 to detect very low levels of denaturant in spirits. In this study, the MZ5 was able to detect denaturants at 5x below the required detection limit.
Investigation of Edible Oils
The second project investigated the MZ5 spectra for a range of edible oils. MIR spectra were collected for rapeseed oil, olive oil, sunflower oil, safflower oil, linseed oil, sesame oil, pumpkin seed oil and pistachio oil. The spectra acquired for these oils is shown in Figure 5.
Figure 5. MIR spectra for edible oils. The spectral shape is very consistent regardless of the source of the edible oil.
PCA analysis was applied to the spectra to evaluate whether the MZ5 could be used to classify the oils into groups. As shown in Figure 6, the various oils clustered well based on type, with 76% of the variation among the MIR spectra explained with three principal components.
Figure 6. Multivariate analysis using PCA successfully groups the MZ5 spectral data by edible oil type.
The students concluded that, like the wine spectra, the edible oil spectra required multivariate analysis to distinguish the different oils from one another. Univariate analysis did not enable oil classification.
Even though the students were used to very high-resolution data acquired with their high end MIR-FTIR instrument, they were very happy with the quality of the spectra they measured with the much more affordable MZ5. When compared to traditional methods using an FTIR spectrometer with ATR accessory, the MZ5 is more compact, fitting into an easy carrying case small enough to be a personal item on an airplane; is more robust, with no moving parts or mirrors that can go out of alignment; and is more affordable.
The team reached the following conclusions from their three-week sprint with the MZ5:
- Determination of ethanol content in red and white wines worked well with the MZ5, providing a much more affordable system compared with the larger MIR-FTIR system.
- Similar investigation of sugar content in wine (data not shown) found that the sugar concentration was too low to enable classification of the different wine varieties.
- Edible oils were easily differentiated based on the MZ5 spectra and multivariate analysis.
- Fuels were also differentiated based on ethanol content (data not show).
In general, the students found the measurements with their aqueous samples were straightforward and very easy to perform using the MZ5. They achieved precise results when they combined their MZ5 spectra with multivariate data analysis. Also, the team mentioned some challenges with their volatile fuel samples with the samples evaporating during the measurements, so they recommended applying fresh samples to the MZ5. We describe similar challenges in our application note Solving the Mystery of the “Vanishing” MIR Data along with additional ways to deal with volatile samples.