Suite 103, 56 Kitchener Pde

Bankstown, NSW, 2200

Ph: 02 9708 5068

Fax: 02 9708 5537

 

 
 
NIR Technology Australia    

 

         

 

 

 

Application Note 07: Determination of Ethanol Content in Water

Samples Using a 38-element diode array detector NIR Transmission Instrument.                                                              April 2000 

Introduction

 

This study was performed to evaluate the use of a 38-element silicon diode array NIR spectrophotometer to develop a rapid alcohol analyser for potential use in the wine industry. The sampling procedure involved the use of a liquid cell in transmission mode and the calibration developed was tested for reproducibility, repeatability and temperature stability.

 

Description

 

A preliminary calibration was developed using a set of ethanol solutions made up to the concentrations listed in table 1 prepared in Milli-Q water and buffered with 2.5g/L potassium hydrogen phthalate. Samples were injected into a liquid cell with a pathlength of 30mm and scanned on a NIT 2000 near infrared transmission analyser in the wavelength range 720-1100nm The calibration was optimised and tested for temperature stability with samples equilibrated between 10 and 30oC. This temperature-stabilised calibration was tested for temperature stability, reproducibility and day-to-day repeatability.

 

Results

 

1)      Calibration

 

A preliminary calibration was developed using Unscramber™. Principal Component Analysis (PCA) indicated that in the alcohol range 10-20%, there was a good cluster of spectral results. The calibration statistics for the unstabilised  calibration are given in table 2.

 

Table 1: Concentrations of alcohol samples for preliminary calibration

Sample

Concentration (%)

Vol EtOH required (ml)

Mass KHP (g)

1

10.0

5.00

0.125

2

10.5

5..25

0.125

3

11.0

5.50

0.125

4

11.5

5.75

0.125

5

12.0

6.00

0.125

6

12.5

6.25

0.125

7

13.0

6.50

0.125

8

13.5

6.75

0.125

9

14.0

7.00

0.125

10

14.5

7.25

0.125

11

15.0

7.50

0.125

12

15.5

7.75

0.125

13

16.0

8.00

0.125

14

17.0

8.50

0.125

15

20.0

10.00

0.125

 

 

To test the temperature stability of this calibration, standard alcohol samples were incubated at 10oC and these samples were continuously measured on the spectrometer until the sample temperature reached 30oC. Temperature was monitored using a thermistor in direct contact with the solution. The results of these experiments are shown in figure 1.

 

 

Table 2: Statistics for the alcohol calibrations

 

Preliminary Calibration

Temperature Stabilised Calibration

Number of PC's

5

5

Elements

71

110

RMSED

0.087

0.101

SED

0.087

0.102

Slope

0.999

0.998

Correlation

0.999

0.999

 

 

 

 


 


Figure1: Variation of predicted alcohol vs. temperature for a 10% and a 15% alcohol standard.

 

Figure 1 indicates that the predicted alcohol results vary linearly with temperature and therefore the calibration needed to be corrected for temperature. This was achieved by adding temperature stabilisation samples to the calibration. The statistics are also shown in table 2.

 

 

2)      Temperature Stability

 

Figure 2 shows the predictive ability of this calibration with a 13.5% alcohol standard in the temperature range 22.0-30.5oC. Table 3 provides measures of the accuracy, AAD and Standard Deviation for the temperature stability shown in figure 2.

 

Table 3: Relevant statistics for calibration temperature stability

[ethanol] (%)

NIR Predicted

Alcohol (Avg)

Std Deviation

AAD (%)*

13.50±0.1

13.45±0.1

0.05

0.05

*AAD (absolute Average Difference) between consecutive scans

 

 

 


Figure 2: Predictive ability vs. temperature for the stabilised calibration.

 


3)      Predictive Ability

 

Figure 3 shows the predictive ability of the temperature stabilised calibration for the following ethanol concentrations, 10.0%, 11.0%, 12.0%, 13.0%, 14.0%, 15.0%. All samples were equilibrated to 24.0oC Relevant statistics are also given.

 


 


Figure 3: Predictive ability of the temperature stabilised calibration.

 

4)      Reproducibility/Repeatability

 

a)      Reproducibility (between scans): See table 3.

 

b)      Reproducibility/Repeatability (between refills). Table 4 shows the calibration performance between refills and on a day-to-day basis

 

Table 4: Reproducibility/Repeatability Results

Sample

Temp (oC)

Day 1

Day 2

1

2

3

1

2

3

Tap Water

24.0

0.07

0.04

0.05

0.03

0.01

0.04

12.0%

24.0

12.05

12.03

12.04

12.06

12.01

12.00

15.0%

24.0

14.97

15.02

14.93

14.98

14.93

14.95

 

Reproducibility

AAD between sample refills = 0.03%

Std Dev =0.008%

 

Repeatability

AAD day-to-day = 0.02%

 

Conclusions

 

a)      The 38-element silicon diode array NIR instrument can be calibrated to measure alcohol content of a set of ethanol standards.

 

b)      Temperature stabilisation may be achieved within the calibration by the addition of samples measured at the expected sample temperature extremes (i.e. 15-30oC).

 

c)      Reproducibility and Repeatability on a day-to-day basis were found to be within specification (i.e. Within ±0.1%).

 

d)      Although the prediction of actual samples with the ethanol calibration was not successful, the results indicate that the development of a red and a white wine calibration is possible and on the basis of the results in this report, should provide accurate prediction data.