Testing of industrial products - Aggregates for construction

Results of the cross-testing experiment on the Chloride test

Repeatability and reproducibility

Estimates of repeatability and reproducibility are given in the Tables. The standard deviations are also shown in Figure 1. Note that in Figure 1 the standard deviations shown for Levels 1 and 3 are values of Sr1 and SR1, whereas those for Level 2 are values of Sr and SR.

Note also that these standard deviations are estimates that can be influenced by individual results. In particular, the value of SR1 from Level 1 in the French experiment would be smaller (0.0040 instead of 0.0064), and close to the corresponding value for the European experiment, if the high results from Laboratory A were not included.

With this in mind, Figure 1 can be seen to provide evidence that the repeatability and reproducibility of the full procedure as used in Levels 1 and 3 (involving sample division, extraction, and the Volhard titration), are substantially higher than those of the Volhard titration alone, as used in Level 2. Hence the first stage of the test method - sample division and extraction - must have a significant influence on the precision of the test.

The lines shown on Figure 1 were calculated using the standard deviations for Levels 1 and 3, but, with only two points per line, the results from the European and French experiments alone cannot justify the use of the functional relationships shown in Figure 1. However, Figure 2 shows the results from an earlier experiment carried out in the UK (Sym R, 1986. Precision of the methods for determining water-soluble chloride, water-soluble sulphate, and total sulphate in BS 812 Parts 117 and 118. Cement and Concrete Association, Crowthorne.) involving four aggregates (a 14mm sea-dredged gravel from the English Channel, a 28mm colliery shale, a 14mm limestone from Bahrain, and a 5mm crushed limestone and shell mixture, also from Bahrain). This provides some justification for the use of straight lines. Hence it would be reasonable to quote the functional relationships given in the Tables (for the European experiment) in the European Standard test method as a statement of the precision of the determination of water-soluble chloride salts in aggregates.

Comparison of Figures 1 and 2 suggests that the precision achieved in the European experiment is better than that achieved in the earlier British experiment. Volhard titration was used as the method of analysis in both cases, and the procedures for test portion preparation were similar, but the method of extraction used in the British experiment was to place the test portion of aggregate in an equal mass of water in a bottle and:

mix the contents by occasional shaking or rolling for a minimum of 24 hours

Thus it was left to the operator to decide how often the contents should be agitated. The method of extraction required by the draft European Standard method is a more closely-controlled process:

agitate the bottles continuously for 60 minutes by means of the shaker or roller

It is possible that this accounts for the improvement in the precision.

Assessment of the precision of the determination of chlorides.

The proposed draft European Standard specification for aggregates for concrete (CEN/TC 154 committee paper N 232, January 1994.) contains the following information about limits for the chloride content of aggregates.

The requirements of 5.5 of ENV 206 (European Prestandard ENV 206 Concrete - performance, production, placing and compliance criteria. March, 1990.) are usually achieved when the water-soluble chloride ion content of the sand and coarse aggregate, combined in the proportions intended for a particular mix, do not exceed the values given in Table A. The values given in Table A are intended for the guidance of aggregate producers.

The calculation of the total water-soluble chloride ion content of the concrete should be made using measured values of chloride ion content for each component of the mix (see 5.5 of ENV 206).

The calculation of the chloride ion content of the concrete may be based on the use of the nominal chloride contents of the constituents (see ENV 206, 11.3.12 case d) if the chloride contents of the aggregates do not exceed the values in parenthesis in Table A.

Table A. Maximum water-soluble chloride ion content of aggregate for concrete
Type or use of concrete Maximum chloride content
expressed as percentage of
water-soluble chloride ion content
by mass of combined aggregate
Prestressed concrete0.03 (0.02)
Reinforced concrete0.06 (0.04)
Plain concrete0.15

It has been argued (J»rck, Sym and Powell, 1994. A study of mechanical tests of aggregates. Green Land Reclamation Ltd Report GLR 3036/03a.) that the reproducibility standard deviation of a mechanical test, when expressed as a coefficient of variation, should be no more than about 8%, if the test method is to be used to assess the compliance of aggregates with specifications. The argument used to justify this value applies equally-well to non-mechanical tests when their results are to be compared with upper limits in specifications, so it can be applied to the determination of chloride contents.

The functional relationship for SR1, obtained from the European experiment, gives the coefficients of variation CR1 shown in Table B. These results show that the reproducibility of the determination of chloride contents is adequate when assessed by the above criterion, for chloride contents down to 0.01 %. However, there is a need for laboratories to check that they do not produce erratic results, using Proficiency Testing.

Table B. Reproducibility coefficients of variation
for the determination of water-soluble chlorides
Chloride content
expressed as percentage of
water-soluble chloride ion content
by mass of combined aggregate
coefficient of variation
(from Table 2)
CR1 = 100 (0.044 + 0.0002/X)

Rounding of chloride determinations

The draft European Standard that describes chemical tests for aggregates (CEN, 1993. Proposed prEN Tests for chemical properties of aggregates. Part 1 Chemical analysis. CEN/TC 154/TG 8 committee paper.) requires the results of chloride determinations to be rounded to the nearest 0.01%. However, chloride contents in the range 0.02 to 0.15% are of interest when the test is used to assess the compliance of concrete with a limit on the total chloride content of the concrete. If chloride contents are rounded to the nearest 0.01% there is a risk that significant variations in the chloride content of the aggregate may be concealed.

At the highest level of chloride content examined in these experiments (Level 1, for which the general average of all the results in the European experiment was 0.084 %), the repeatability standard deviation is Sr1 = 0.0010%. The repeatability standard deviation is smaller at lower chloride levels. These results therefore justify reporting chloride contents to the nearest 0.001%, and it is recommended that this should be required in the European Standard test method.