Raman is an advanced spectroscopic instrument for molecular identification and is increasingly used in qualitative analysis for rapid material identification, often at the point of use or delivery.
The high resolution of Raman extends its application to structural elucidation and quantitative analysis. When compared to other molecular spectroscopic instruments, Raman enables measurement of aqueous solutions.
This article discusses the application of the portable i-Raman Plus and BWIQ chemometric modeling software from B&W Tek to create quantitative calibration model for determination of the concentrations of glucose, sucrose, and fructose in a tertiary aqueous mixture.
BWIQ, a multivariate analysis software package for analysis of spectral data, is useful in exploring the correlations between spectra and response data or spectra and sample classes.
BWIQ couples conventional chemometric techniques such as Principal Component Analysis (PCA) and Partial Least Squares Regression (PLSR) with new techniques such as Support Vector Machine (SVM) algorithms and adaptive iteratively reweighted Penalized Least Squares (airPLS).
The i-Raman Plus is the most sensitive handheld Raman spectrometer available in the market, which is the reason for the selection of this device for this analysis. The use of a fiber optic probe enables collecting spectra through glass scintillation vials.
Experiment and Results
The experiment involved preparing a total of 31 standard samples with known quantities of glucose, sucrose and fructose in distilled water, followed by an experimental design to develop a calibration curve. The total concentration in every sample was 0.4M of sugar. The collection of Raman spectra was performed over the range of 176–3200cm-1 using a 785nm excitation wavelength set to roughly 300mW power output and an integration time of 50 seconds. Figure 1 shows the representative spectra of the individual sugar solutions.
Figure 1. Raman spectra of aqueous sugar solutions
A set of 62 spectra (each sample scanned twice) were acquired with different concentrations of the three sugars and fed into BWIQ software for subsequent analysis. Fifty spectra were selected as the calibration set, while the remaining spectra were utilized as a prediction set. The data was utilized without any preprocessing.
Using PLS1 regression, a calibration curve was developed on the data over the spectral range of 250 -1500cm-1. Figure 2 shows an overlay of samples with varying concentration.
Figure 2. Overlay of Raman spectra of tertiary solutions of glucose, fructose and sucrose
A three factor PLS model was created for each of the three sugars. The accuracy of the model was demonstrated by plotting the concentrations. Figure 3 shows the results for glucose, for which the R2 value was more than 0.999 for both predicted and measured concentrations. Similarly, the RMSEC and RMSEP were estimated to be below 0.008.
Samples that were not involved in the calibration were employed as a prediction set utilizing the calibration model created in the BWIQ software. The concentration in these samples was successfully determined by using the models.
Figure 3. Measured vs. predicted plot of concentration percentage for Glucose 3 Factor PLS model
The study results clearly demonstrate the successful use of the i-Raman Plus portable Raman spectrometer in conjunction with the BWIQ software to create a quantitative model for determination of the glucose concentration in tertiary aqueous mixtures of glucose, fructose and sucrose.
This information has been sourced, reviewed and adapted from materials provided by B&W Tek.
For more information on this source, please visit B&W Tek.