Figure 2: Actual and predicted methanol concentrations using PLS regression model.
|
Partial least square (PLS) regression method was used for calibration and prediction model for methanol. The spectral region
from 950–1200 cm-1 was chosen for developing the calibration model. The actual vs. predicted concentration value of methanol is shown in Figure
2. It is shown that the measured data and PLS prediction match very well with correlation coefficient R2 @ 0.997.
Conclusion
An affordable, high sensitivity ProRaman instrument was used with PLS method to successfully analyze low concentration methanol
in 40% alcohol. Based on our findings, the detection limit for methanol in 40% of ethanol is much better than 50 ppm and reliable
quantitative determination using PLS prediction could reach 50 ppm of methanol in 40% alcohol.
References
(1) F. Bindler, E. Voges, and P. Laugel, Food Addit. Contam.
5, 343–351 (1988).
(2) W.M. Mackenzie and R.I. Aylott, The Analyst
129, 607–612 (2004).
(3) L.M. Reid, C.P. O'Donnell, and G. Downey, Trends in Food Science & Technology 17, 344–353 (2006).
Enwave Optronics, Inc.
18200 McDurmott St., Suite B, Irvine, CA 92614
tel. (949) 955-0258, fax (949) 955-0259
Website:
http://www.enwaveopt.com
|
