MEASURING WATER ACTIVITY (a_w) USING THE NOVASINA a_W CENTER

MFLP-63
January 2006

---

Health Products and Food Branch
Ottawa

Patti Wilson
Canadian Food Inspection Agency
P.O. Box 1060,
1992 Baffin St.
Dartmouth, N.S.
B2Y 3Z7

E-mail: wilsonpa@inspection.gc.ca

1. Application

This method may be used to determine the water activity (a^w) of foods to determine compliance with the requirements of Sections 4 and 7 of the Food and Drug Act. This revised method replaces MFLP-63, dated April 1999.

2. Description

The Novasina a_w Center consists of a measuring station and a temperature controlled a_w -box with 3 sensors. Studies have shown that the Novasina a_w Center produces satisfactory results for the measurement of water activity in foods (8.2, 8.4 and 8.6).

3. Principle

Water activity plays an important role in the preservation of food and the control of microbial growth, especially pathogens. The Novasina a_w Center measures the relative equilibrium humidity in % rh which is directly correlated with the water activity in accordance with the following formula: a_w = ERH/100.

4. Definition of Terms

4.1 See Appendix A of Volume 3.

4.2 water activity (a_w) : is the ratio of the water vapour pressure in any food system to the water vapour pressure of pure water at the same temperature.

4.3 equilibration: a change in the a_w reading of ≤ ± 0.001 a_w in a 10 minute period.

4.4 humidity standard: is a saturated solution of a salt of a known % relative humidity at a designated temperature.

5. Collection of Samples

5.1 See Appendix B of Volume 3.

5.2 Keep samples sealed to prevent loss or gain of moisture during transport.

6. Materials and Special Equipment

6.1 Novasina a_w Center (available from Novasina AG, a division of Axair Ltd., Pfäffikon, Switzerland or JEB Controls Inc., Kirkland, QC),

6.2 Novasina humidity control standards (available from Novasina AG, a division of Axair Ltd., Pfäffikon, Switzerland or JEB Controls Inc., Kirkland, QC),

6.3 a_w disposable bowls (available from Novasina AG, a division of Axair Ltd., Pfäffikon, Switzerland or JEB Controls Inc., Kirkland, QC),

6.4 Pre-filters (available from Novasina AG, a division of Axair Ltd., Pfäffikon, Switzerland or JEB Controls Inc., Kirkland, QC).

7. Procedure

7.1 Handling of Samples Units

7.1.1 In the laboratory prior to analysis, except for shelf-stable foods, keep sample units refrigerated or frozen, depending on the nature of the product. Thaw frozen samples in a refrigerator, or under time and temperature conditions which prevent loss of moisture.

7.1.2 Analyse sample units as soon as possible after their receipt in the laboratory.

7.2 Calibration

7.2.1 Before calibrating prepare the Novasina by:

1. Turn the power switch on and leave the instrument for a minimum of 2 hours.
2. Set the temperature dial to 190 (25°C) and ensure that the temperature is holding.
3. Set the RTD-24 switches to "S".
4. Remove the pre-filter from each sensor (see Maintenance Procedures).

7.2.2 To decide which humidity standard to use for the upper and lower points, consider the anticipated range of the sample and choose a humidity standard range which will include this value. Only samples within the range of the upper and lower points can be reported as a numerical value rather than > the upper point or < the lower point.

For routine analysis of food, use SAL-75 as the lower point and SAL-98 as the upper point.

If using more than one humidity standard with the same water activity, dedicate one humidity standard to one sensor and label it.

7.2.3 Calibration can proceed two ways: by using the built-in dummy values (see the Novasina Operation Manual) or by using the D-Pot the procedure (see Appendix 4). With only one D-Pot cable there will be some connecting and disconnecting to calibrate all three sensors at the same time.

7.2.4 Adjustment of the low range is optional and should be done after the medium and high range adjustments are made. Do this using SAL-53.

7.2.5 To check the linearity of the calibration, measure the a_w of 4 humidity standards with each sensor and calculate a regression line (see Appendix 2). The salts used should include the upper and lower points plus two in between (see Appendix 3 for a selection of salts). Dedicate one humidity standard to one sensor if you have more than one humidity standard with the same water activity.

Use the Calibration Record form in Appendix 1 to record the calibration and regression line data.

7.2.6 Install fresh pre-filters in front of each sensor.

7.3 Humidity Standards Validation

7.3.1 To determine reproducibility or variation of the humidity standards, take 20 daily a_w measurements of each humidity standard for each sensor, using the standard dedicated to that sensor, and record it on a Novasina Control Chart. For Novasina supplied humidity standards, the upper and lower limit should not exceed the labelled value ±0.005 which is the Novasina stated claim.

7.3.2 To determine the accuracy of uncertified humidity standard material, use a method that is independent of the calibration methodology in order to have a meaningful check on the standard material.

There are several options:

1. Send blind samples to another CFIA facility to be checked on the same type of aw aw apparatus.
2. Test blind samples on a type of apparatus different than the one than you are using.
3. Send blind samples to an external calibration laboratory to certify the standards.
   
   Certified humidity standards are best if available.

7.4 Preparing Control Charts

7.4.1 Prepare quality control charts using graph paper or a software package. Determine the acceptable limits of the humidity standards (see 7.3.1).

7.4.2 Determine the upper and lower limits of the humidity standards to be used to determine readiness (usually SAL-90 and SAL-98) using the data generated in 7.3.1. After 20 readings have been taken, calculate the mean and standard deviation for each sensor and each humidity standard. Use this data to determine the upper and lower limits where the upper and lower warning limits equal 1.5 standard deviations and the upper and lower control limits equal 2 standard deviations.

7.5 Determining Readiness of Aw Center

7.5.1 The humidity standards used to determine if the a_w center is ready to take sample readings must not be the same standards used to calibrate the unit.

7.5.2 Before taking sample readings, check the validity of the calibration by taking a reading of the humidity standard nearest to the anticipated water activity of the sample(s), usually SAL-90 or SAL-98 for most foodstuffs.

Correct the reading using the regression line or regression formula of the last valid calibration and plot this value on the appropriate control chart for the sensor and the standard. Proceed with the sample measurement if the value falls between the upper and lower limits determined in 7.4.2.

7.5.3 If the humidity standard reading is outside the upper or lower limits, repeat the reading 3 times allowing the sensor to return to ambient humidity between readings.

If the 3 readings fall within the upper and lower control limits, then no adjustment is necessary and continue with the sample readings. If the 3 readings are still abnormal, then recalibrate.

If the problem persists do a 4 or 5 point calibration check. Results which are in a different range than previous calibration checks but are linear indicate an equipment malfunction. Results which are not linear indicate a problem with one or more humidity standards which will need to be replaced.

7.6 Sample Measurement

7.6.1 Before taking sample readings on a given day, prepare the Novasina by:

1. Turn the power switch on and leave the instrument for a minimum of 2 hours.
2. Set the temperature dial to 190 (25°C)
3. Set the RTD-24 switches to "S".
4. Turn the printer on and select "1" to record only one channel or "6" to record all three channels (optional).
5. Determine the readiness of the instrument by measuring humidity standards as described in the Quality Control section (see 7.5).
6. Install fresh, appropriate pre-filters.
7. Ensure that the a_w calibration range includes the anticipated reading of the samples material.

7.6.2 Samples should be analysed in triplicate.

7.6.3 Prewarm refrigerated and frozen samples to 25°C before opening to prevent condensation on the sample and to reduce the equilibration time. Care should be taken to avoid evaporation of water from the sample.

Crush or cut up the sample material to homogenize. Fill the plastic sample bowls to the rim but the material must not protrude beyond the rim. Take care not to directly handle the sample material.

7.6.4 Samples which are not analysed right away can be stored by sealing the filled plastic sample bowls in plastic bags and storing them at an appropriate temperature. Again, prewarm refrigerated and frozen samples to 25°C to prevent condensation on the sample.

7.6.5 Place the plastic sample bowls into the stainless steel measuring bowls by pulling up on the appropriate sensor knob and using the tongs to install the stainless steel measuring bowls under the water activity sensors making sure that the fit is proper and the seal is snug.

Close the chamber door securely.

7.6.6 When equilibrium has been reached, Note the water activity value to three places after the decimal and the temperature reading for the appropriate channel.

7.6.7 Correct the a_w values with the latest calibration regression line for each channel. Report sample a_w values by first determining the mean of multiple readings and then reporting 2 places after the decimal, rounded as necessary. Record sample a_w reading on the appropriate lab sheet.

7.6.8 After every reading, dry the sensors by removing the bowls and allowing the sensors to return to ambient humidity. Clean and dry the stainless steel bowls and filter as required and discard the pre-filter.

7.7 Maintenance Procedures

7.7.1 Humidity Standards

Humidity standards with a labelled water activity should be purchased from a reputable source. Certified humidity standards would be ideal.

For humidity standards above room humidity, if the they show signs of drying out, rejuvenated them with distilled water so that the water is just visible on the surface. To store these humidity standards, seal them in their containers with tape and store in an airtight container when not in use.

For humidity standards below room humidity, if they fill up with water, use an absorbent tissue to remove the water until it returns to the normal level. These standards should be stored in sealed containers in a dry area such as a desiccator.

The humidity standards should be replaced if they become contaminated or if all of the salt has dissolved. Replacing the humidity standards occasionally is recommended.

7.7.2 Pre-Filters

Over time, exposure of the sensors to fatty, acidic and other foods can lead to sensor failure. As a preventative measure, pre-filters should be used for every sample measurement.

A selection of pre-filters are available from the Novasina dealer. Except in cases where special pre-filters are required, chemically-neutral, fast-flowing paper filters can be used.

To install or remove the pre-filters, lift the gold-coloured retaining ring and while lifting the sensor with the knob, grasp the sensor and pull down firmly to disengage the sensor. Turn the sensor upward, remove the C-clip and turn the sensor downward to drop the metal screen out. DO NOT TOUCH THE BUTTON SENSOR. Remove or replace the filter, replace the metal screen and C-clip, reengage the sensor and replace the retaining ring.

7.7.3 Sensor Care

Do not submerge sensors in water or any other liquid.
Do not leave sensors in draft.
Do not spray with liquid or blow on sensor.
Do not subject sensor to changes in temperature which allow condensation to form.
Avoid physical shocks.
Protect sensors from strong substances such as acid or alkaline vapours.
Do not cool or heat sensors beyond specified limits.
Do not place sensors in a humid environment.
For long-term storage, place sensors in a sealed container.
Use the proper filters.
Avoid direct contact between the sample and the sensor.
Clean the stainless steel filter frequently.
Protect sensor from contamination by dust or particles.
Allow sensors to return to ambient humidity after each use.

8. References

8.1 Gaudreault, France. 1994. Determination of Water Activity of Foods Using the NOVASINA. Procedure # MM-INS13-001. HPB, Quebec.

8.2 Kiltic, D., Favetto, G., Chirife, J., and Resnik, S. 1986. Measurement of water activity in the intermediate moisture range with the Novasina Thermoconstanter humidity meter. Lebensmittel Wissenschaft und-Technologie 19: 297-301.

8.3 Novasina AG. 1988. Novasina aw center Operation Manual. Pfäffikon, Switzerland.

8.4 Stoloff, L. 1978. Calibration of Water Activity Measuring Instruments and Devices: Collaborative Study. JAOAC 61: 1166-1178.

8.5 Stroup W.H. and Peeler J. T. 1987. Evaluation of Precision Estimates for Fiber-Dimensional and Electrical Hygrometers for Water Activity Determinations. JAOAC 70: 955-957.

8.6 Troller J.A. 1977. Statistical Analysis of a_w Measurements Obtained with the Sina Scope. Journal of Food Science 42: 86-90.

8.7 Association of Official Analytical Chemists. 1990. Water Activity of Canned Vegetables. In: Official Methods of Analysis - AOAC - 15th Edition. pp. 987-988.

Appendix 1

Calibration Record

Date:
Analyst:

Are the Dummy Temperatures OK? Y N

	Temperature °C	Calibration References
		Medium	High
Std Used		SAL-	SAL-
Sensor #1
Sensor #2
Sensor #3

Linear Regression

Date:

Salt	y	Sensor #1 x1	Sensor #2 x2	Sensor #3 x3
SAL-53	.529
SAL-75	.753
SAL-84	.843
SAL-90	.901
SAL-98	.980
a
b
r

where:

a = y-intercept
b = slope
r = correlation coefficient

To correct aw values: y = a + bx (where x = reading)

Appendix 2

Calculation of a Regression Line

Salt Std		y-values	x-values
SAL-11	.113
SAL-53	.529
SAL-75	.753
SAL-90	.902
SAL-98	.980

Input the x-values by taking a_w readings of the salt standards.

Calculate:

Σy

Σy²

Σxy

Σx

Σx²

n =5

b (slope)= Σxy - Σx Σy/n / Σx² - [(Σx)²/n]

a (y-intercept)= Σy - bΣx / n

Then: y = a + bx

Appendix 3

Table I. Water Activity (a_w) of Reference salt slushes at 25°C

Salt	aw
MgCl2	0.328
K2CO4	0.432
Mg(NO3)2	0.529
NaBr	0.576
CoCl2	0.649
NaCl	0.753
KBr	0.809
(NH4)2SO4	0.810
Kcl	0.843
Sr(N03)2	0.851
BaCl2	0.902
KNO3	0.936
K2SO4	0.973

Appendix 4

Calibration using the D-Pot Procedure:

1. Perform calibration by inserting the D-pot into the junction of each wire between the conditioning chamber and the control apparatus.
   
   
2. Install SAL-75 and leave it for about 2 hours to reach equilibrium.
   
   
3. Set the D-Pot selector button on "S" or "sensor", depending on your unit, and the channel selector button to the channel to be calibrated.
   
   
4. When equilibrium has been reached, if there is a difference between the SAL-75 value (0.753 at 25°C) and the digital readout on the aw center, adjust the value by turning the screw of trimmer potentiometer "M".
   
   
5. Set the D-pot selector button on "D" or "POT", depending on your unit, and adjust the D-pot setting using the thumb wheels until the digital readout on the aw center is 0.753. Record the D-pot value on the calibration sheet (Appendix 1).
   
   
6. Set the D-pot selector on "S" or "sensor".
   
   
7. Record the time at which the final readings were taken, the temperature, the channel number and the humidity source (calibration sheet, Appendix 1).
   
   
8. Replace SAL-75 with SAL-98, marking the time at which the salt was placed under the sensor.
   
   
9. Wait about 2 hours for equilibrium to be reached.
   
   
10. Place the selector button on the channel to be calibrated.
    
    
11. When equilibrium has been reached, if there is a difference between the SAL-98 value (0.980 at 25°C) and the digital readout on the aw center, adjust the value by turning the screw of trimmer potentiometer "H".
    
    
12. Place the D-pot selector button on "D" or "POT" and adjust the D-pot setting using the thumb wheels until the digital readout on the aw center is 0.980. Record the D-pot value on the calibration sheet (Appendix 1).
    
    
13. Adjust the thumb wheels of the D-pot to the value recorded while adjusting "M" and turn the "M" trimmer potentiometer screw so that the digital readout on the aw center is 0.753.
    
    
14. Adjust the thumb wheels of the D-pot to the value recorded while adjusting "H" and turn the "H" trimmer potentiometer screw so that the digital readout on the aw center is 0.980.
    
    
15. Repeat steps 13 and 14 until the 0.753 and 0.980 values appear on the digital readout.
    
    Moving the "M" and "H" trimmer potentiometers results in a reciprocal value change. Through a series of adjustments, these values become increasingly accurate.

The apparatus is now calibrated