Measurement of Moisture Content (Water Content) in Edible Oil Using the Volumetric Karl Fischer Method According to ISO 8534:1996

Abstract

The presented example covers validation, estimation of measurement uncertainty and establishing traceability for determination of moisture content (water content) in edible oils using the volumetric Karl Fischer method.

Appendices

Exercise 1: Establishing Traceability in Analytical Chemistry

1. :

Specifying the analyte and measurand

Analyte	Water
Measurand	Content of water in the sample
Units	mg kg−1

2. :

Choosing a suitable measurement procedure with associated model equation

Measurement procedure	Volumetric Karl Fischer titration
Type of calibration a	Standard curve	Standard addition	Internal standard

1. ^aThe titer of the titrant (mass of water per unit volume of titrant) is determined. This could be considered as two-point standard curve

Model equation

Calculation of water content in sample is carried out according to the following equation (mathematical model):

$$ c_{\text{water}} = \frac{{V_{\text{t}} \cdot T_{\text{t}} }}{{m_{\text{s}} }} \cdot f_{\text{r}} \cdot 1000 $$

(13)

c_water is concentration of water in oil sample [mg/kg], V_t is volume of titrant used for titration [ml], T_t is the titer of the titrant (mass of water per unit volume of titrant) [mg/ml], m_s is the sample mass [g], f_r is the repeatability factor, 1000 [g/kg] is the unit conversion factor. The reason for inclusion of f_r is the following: the amount of injected sample is measured gravimetrically and each time the sample amount is slightly different. Thus, the repeatability components of V_t and m_s are not included in the uncertainties of these quantities but are taken into account separately by the repeatability factor f_r.

3. :

List the input quantities according to their influence on the uncertainty of the result of the measurement (first the most important ones). At this point, your judgement should be based on your previous experience only

1	f r
2	T t
3	V t
4	m s
5

4. :

List the reference standards needed and give also the information regarding traceability of the reference value

For the analyte

1	Name/ChemicalFormula/producer:	Water reference standard solution, e.g. any of the Hydranal standards available from Sigma-Aldrich
2	Name/ChemicalFormula/producer:

For the other input quantities

1	Quantity/Equipment/calibration: e.g. mass/balance/calibrated by NMI, U = xx (k = 2), see also data yellow sheet	Balance calibrated by NMI
2	Quantity/Equipment/calibration:
3	Quantity/Equipment/calibration:
4	Quantity/Equipment/calibration:

5. :

Estimating uncertainty associated with the measurement

Are all important parameters included in the measurement equation?	Yes	No
Other important parameters are:

6. :

How would you prove traceability of your result?

1	By measuring independently prepared reference standards (e.g. other manufacturer) with reference values traceable to the same point of origin
2
3

7. :

Any other comments, questions…

.

.

Exercise 2: Single Laboratory Validation of Measurement Procedures

2.1 Part I: General Issues

1. :

Specify the measurement procedure, analyte, measurand and units

The measurement procedure	Volumetric Karl Fischer titration
Analyte	Water
The measurand	Content of water in the sample
Unit	mg kg−1

2. :

Specify the Scope

Matrix	Edible oils
Measuring range	100–500 mg kg−1

3. :

Requirement on the measurement procedure

Intended use of the results:	Characterisation of edible oil quality
Mark the customer’s requirements and give their values		LOD	No requirement, as this is not trace analysis
		LOQ	100 mg kg−1
		Repeatability	≤8 mg kg−1 expressed as standard deviation
		Within-lab reproducibility	≤16 mg kg−1 expressed as standard deviation
		Measurement uncertainty	≤20 mg kg−1 expressed as standard uncertainty
		Trueness	It is important, but is actually taken into account by measurement uncertainty
		Other-state

4. :

Origin of the Measurement Procedure

		Validation
New in-house method		Full
Modified validated method		Partial
Official standard method		Confirmation/Verification

2.2 Part II: Parameters to Be Validated

5. :

Selectivity/Interference/Recovery (where yes, please give further information e.g. which CRM, reference method)

	CRM/RM: analysis of available CRM or RM
	Further information:
	Spike of pure substance
	.
	Compare with a reference method
	KF titration is in fact the reference method for moisture determination in edible oil. However, interlaboratory comparison (even if the other participants have likewise KF methods) is certainly useful
	Selectivity, interferences
	Selectivity is provided by (1) the chemistry of the KF method and (2) the simplicity and small variability of the edible oil matrix
	Test with different matrices
	.
	Other – please specify
	.

6. :

Measuring range

	Linearity
	Upper limit
	LOD
	LOQ

7. :

Spread – Precision

	Repeatability
	Reproducibility (within Lab)
	Reproducibility (between Lab)

8. :

Robustness

Variation of parameters

The following parameters should be varied:        - Stirring speed;        - Titrant flow rate (i.e. rate of piston movement in the burette);        - Age of titrant.

9. :

Quality Control

	Control charts
	Participation in PT schemes

10. :

Other parameters to be tested

	Working range and testing of homogeneity of variances
	R square
	Residual standard deviation
	Standard deviation of the method
	Coefficient of variation of the method

2.3 Part III: Some Calculations and Conclusions

11. :

Calculation of parameters requested by the customer

These parameters are:	LOD	LOQ	Repeatability
	Within-lab reproducibility	Uncertainty	Trueness
Parameter 1: Repeatability	Repeatability sr is found as pooled standard deviation from the data of Table 1: $S_{\text{pooled}} = \sqrt {\frac{{\left( {n_{1} - 1} \right)s_{1}^{2} + \left( {n_{2} - 1} \right)s_{2}^{2} + \cdots + \left( {n_{k} - 1} \right)s_{k}^{2} }}{{n_{1} + n_{2} + \cdots + n_{k} - k}}} \quad \quad \left( {14} \right)$ k            number of data groups s1, s2, …  within group standard deviations n1, n2, …  numbers of measurements in groups sr = 5.5 g kg−1
Parameter 2: Within-lab reproducibility	There are no direct data available on within-lab reproducibility. However, it can tentatively be assumed that s RW  = 2 ∙ s r sRW = 11 g kg−1
Parameter 3: LOQ	Dedicated LOQ determination was not carried out. However, from the available data (Table 1) it is known that the procedure operates without problems at 80 mg kg−1. So, this value is tentatively used as an estimate of LOQ
Parameter 4: Measurement uncertainty	See separate section on measurement uncertainty

12. :

The method fulfils the requirement for the intended use:

Parameter	Value requested by the customer	Value obtained during validation
LOD
LOQ	≤100 mg kg−1	80 mg kg−1
Repeatability	≤8 mg kg−1	5.5 mg kg−1
Within-lab reproducibility	≤16 mg kg−1	11 mg kg−1
Measurement uncertainty	uc ≤ 20 mg kg−1	ISO GUM Modeling: 7 mg kg−1 Nordtest: 15 mg kg−1
Trueness
Other

Yes No

13. :

Calculation of other parameters

Exercise 3: Building an Uncertainty Budget

3.1 Here Only the ISO GUM Modelling Approach Is Addressed

1. :

Specify the measurand and units

Measurand	Water content of edible oils
Unit	mg kg−1

2. :

Describe the measurement procedure and provide the associated model equation

Measurement procedure:

Karl Fischer (KF) titration is based on the following reaction:

(15)

(16)

The reaction is very fast and with strict stoichiometry. Solution of iodine, SO2 and pyridine dissolved the alcohol ROH is the titrant solution. In the classical KF titrant the alcohol ROH is methanol. In modern commercial titrants ROH is often methoxyethanol and pyridine is often replaced by imidazole (both because of potential toxicity). Titration is carried out in ROH or in a mixture of ROH and some other solvent (if samples are not soluble in ROH).

The end point of the titration is indicated by a small amount of unreacted iodine in solution. End-point is usually determined voltammetrically: alternating current of constant strength is applied to a double Pt electrode. Potential difference between the Pt wires is monitored. Even small quantities of iodine lead to a dramatic drop of the potential difference.

Model equation:

See Eq. 13.

3. :

Identify (all possible) sources of uncertainty

	Uncertainty of concentration of the titrant
	Uncertainty of measurements of the titrant volume
	Method bias (taken into account within the other input quantities)
	Matrix effect
	Other: Sample mass
	Other: Overall repeatability
	Other:

4. :

Evaluate values of each input quantity

Input quantity	Value	Unit	Remark
Sample mass, ms	9.7734	g
Titrant volume used for titration Vt	0.720	ml
Titer of the titrant Tt	4.9987	mg ml−1
Repeatability fr	1	–	Repeatability is taken into account using a multiplicative factor as the overall repeatability of the procedure. This means that the repeatability contributions of the other input quantities are not taken into account with those quantities

5. :

Evaluate the standard uncertainty of each input quantity

Input quantity	Standard uncertainty	Unit	Remark
Sample mass, ms	0.00016	g
Titrant volume used for titration Vt	0.0031	ml
Titer of the titrant Tt	0.05	mg ml−1
Repeatability fr	0.015	–	Evaluated as pooled repeatability standard deviation (sr) divided by the actual water content (cwater) in the sample

6. :

Calculate the value of the measurand, using the model equation

c_water = 369 mg kg⁻¹

7. :

Calculate the combined standard uncertainty (u _c ) of the result & specify units

Using: Mathematical solution; Spreadsheet Approach; Commercial Software

Input quantity	Value	Standard uncertainty	Unit	Remark
Sample mass, ms	9.77340	0.00016	g
Titrant volume used for titration Vt	0.720	0.0031	ml
Titer of the titrant Tt	4.9987	0.05	mg ml−1
Repeatability fr	1	0.015	–

u_c = 6.8 mg kg⁻¹

8. :

Calculate expanded uncertainty ( U _c ) & specify the coverage factor k and the units

U = 2 ∙ u_c = 14 mg kg⁻¹

9. :

Analyse the uncertainty contribution & specify the main three input quantities contributing the most to U _c

1	Repeatability fr
2	Titer of the titrant Tt
3	Titrant volume used for titration Vt

10. :

Prepare your Uncertainty Budget Report