Technical

INDEX

This section of the website contains useful technical information and it will grow over time, eventually becoming a best practice guide. At the moment it contains a mix of definitions, product and shipping information under the titles below.

1. Assay bias
2. Assay quality assurance (clients duty)
3. Certified reference materials
4. Certification
5. Compliance
6. Control rules
7. Control samples
8. Critical systematic error
9. Data quality objectives
10. False rejection rate
11. Limits
12. Manufacturing
13. National Nuclear Regulator
14. Packaging
15. Prices and dispatch
16. Reference materials
17. Resource audits
18. Uranium standards, shipping-export-import regulations.
19. Use of reference materials

1. Analytical bias

Analytical (assay) bias results from differences in methods, techniques, equipment and calibrations and is an issue in all branches of analytical science (chemical, biological etc).

In exploration and mining, where major decisions are based on assay results, low or high bias analytical results may cause inaccurate and materially unacceptable numbers.

It is a property that can only be measured by inter-laboratory testing and can be detected with a well designed assay quality assurance program using control samples, data quality objectives and control rules. These programs measure assay results for accuracy and precision and assure the required analytical quality at a minimum cost.

What is an acceptable bias? This depends on the absolute affect of the error on the operation, and we can use the financial definition of materiality as a guide. Anything over 5% (+-2.5%) needs to be reported.

Therefore process control and ROM sample bias of <2.5% is probably not material. This should be tighter for concentrate and bullion samples; and could be wider for low grade or tailings samples.

2. Assay quality assurance (clients duty)

A well set up quality assurance program for assay data should more than pay for itself through significant savings from process improvement, or additional value from auditable mineral resource categorization.

If your economic element control samples show a <0.1% bias, a CV <1%, >90% detection of critical systematic error and a false rejection rate of <5%; then the company might be advised to save money by increasing the ratio of samples to controls. Alternatively if you have no control results, or control samples are showing that analytical results are poor; then you need to place a higher priority on error prevention. You will probably discover, by looking at real numbers, you can stop chasing some ghosts You might even be able to start some method improvement.

Anybody sending samples to an assay lab has to set up a quality assurance (QA) program over and above any QA program run by the lab. Both the customer and the lab QA program have one goal (to check the lab) and both programs must be auditable.

The customer QA program should comprise at least:

1. Prior testing the primary and secondary labs for accuracy and bias.
2. Submission of routine QC samples, including reference materials and blanks, either as field samples or inserted in the lab as pulps.
3. Use of random secondary reference materials to check the labs treatment of the primary reference material.
4. Monthly QC meetings with the lab.
5. Actioning of QC failures.
6. Reporting of QC actions.

3. Certified Reference Materials (CRM's)

A CRM is a reference material, characterised by a metrologically valid procedure for one or more specified properties, accompanied by a certificate that provides the value of the specified property, its associated uncertianty, and a statement of metrological traceability.
The difference between a CRM and an AMIS reference material is solely in the last two requirements (uncertainty and traceability). From May 2009 new AMIS certificates will fulfil the CRM requirements by including the following information:

Combined standard uncertainty: The samples used in this certification process have been selected in such a way as to represent the entire batch of material and were taken from the final packaged units; therefore all possible sources of uncertainty (sample uncertainty and measurement uncertainty) are included in the final combined standard uncertainty determination. The uncertainty measurement takes into consideration the between lab and the within lab variances and is calculated from the square roots of the variances of these components using the formula:

Metrological Traceability: The values quoted in our certificates are based on the consensus values derived from statistical analysis of the data from an inter laboratory measurement program. Traceability to SI units is via the standards used by the individual laboratories the majority of which are accredited and who have maintained measurement traceability during the analytical process.

4. Certification

Certification of the RM is based on a measurement campaign (round robin), so allocation of property values takes place on the basis of agreement among the independent measurement results and, not necessarily with direct traceability to Standard International units. Two of the ISO assumptions that allow this are; that there will be enough capable labs and that the results from each lab will be statistically compatible (ISO Guide 35 Clause 9.2.3 and Clause 10). To overcome the effects of this "inter-laboratory issue", and to achieve "a property value having satisfying uncertainty" (ISO Guide 35 Clause 10.2.2) RM producers raise the minimum number of laboratories involved to as many as possible (in our experience, ideally, at least 14 labs per method) and scrutinize the data with the aid of outlier treatment techniques. This should lead to a very accurate measure for a given method; notwithstanding the underlying assumption that what the good inter-laboratory labs reported was accurate. However, an amount of bad data may have an effect, resulting in limits which may be too broad for effective monitoring of a single laboratory or production process. This is a major concern to the RM producer; and by simple extension, the RM consumer. RM consumers may wish to set their own limits, (within the RM producers limits) based on their own data quality objectives and control measurements.

AMIS round robin laboratories are selected from a list of 50 laboratories (currently) around the world. Platinum round robins typically involve 20-30 laboratories, gold round robins up to 20. Each laboratory is given eight or nine selected packages of sample taken from throughout the batch. A sample of a different reference material may be included for QC purposes. Results from the labs that respond are used for the determinations.

Results are compiled into a database, and then everything is sent to an independent geochemist who checks the data and calculates the consensus values and limits. The final limits are calculated after a three step examination of the data, first removing incompatible data outside a spread normally expected for similar analytical methods done by reputable laboratories. Then, data from any one laboratory is removed from further calculations, if the mean of all analyses from that laboratory failed a t-test of the global means of the other laboratories. Next, data that falls outside of the 2 standard deviations isremoved. The mean and standard deviations are then re-calculated.

Analytes with an RSD of near or less than 5 % are reported as "Certified Concentrations" with limits at two "Between Laboratory" standard deviations. Those with RSD's of between near 5 % and 15 % are reported as "Provisional Concentrations" with limits at two "Between Laboratory" standard deviations. Those with RSD's over 15 % are reported as "Informational Values" ("Indicated Values" on certificates prior to April 2009).

This method is different from that used by Government agencies in that the actual "between-laboratory" standard deviation is used in the calculations. This produces upper and lower limits that reflect actual individual analyses rather than a grouped set of analyses. The limits can therefore be used to monitor accuracy from individual analyses, unlike the Confidence Limits published on other standards.

Data, which can be verified by the customer, is compiled onto a"Certificate of Analysis" that is avaialble from this website.

A more detailed proficiency report is circulated to the managers of participating laboratories. This report contains all of the data, graphs, a description of the methods and a laboratory ranking (based on the z-scores). This report is confidential and is circulated only to the lab managers.
LABORATORY MANAGERS INTERESTED IN JOINING AMIS ROUND ROBINS - PLEASE CONTACT US

5. Compliance

Reference materials have always been used for assay lab quality control (QC) and quality assurance (QA) but their use has increased and been more prominent following tightening of various codes, legislation and regulations worldwide for the recording and reporting of mineral resources and reserves. This followed several prominent poor corporate governance and fraud cases. South African and Canadian stock exchange rules actually specify QA/QC from exploration through to the mineral resouce statement. It will also be required for an independent sign-off of exploration data or a mineral resource. They have become necessary since SOX for mining company internal audit procedures. Although corporate compliance requirements still fall short of what has become best practice or what is necessary.

There is little doubt, that in the competition for investment money, especially in a high risk industry like exploration and mining, the company with the best set of books will attract the best premium. While audit trails and quality assurance are a very small part of the business, they are an important cornerstone.

6. Control samples

Control samples are normally samples, of a matrix-matched reference material (RM), made to check the accuracy and precision of runs of analytical results for samples being analysed by a specific analytical method. To behave the same during analysis control sample characteristics should be the same or similar to the sampled material. Property values are typically based on a measurement campaign (round robin) using a number of independent laboratories. The recommended concentrations and limits will reflect the average results and the cumulative precision produced by the participating laboratories.

7. Control rules

Control rules are the criteria used to accept or reject an analytical run. The rules must be set up to provide a high enough percentage of true alarms with a low enough percentage of false alarms and they must be documented. A single rule procedure may be sufficient for robust analytical procedures, or to use as an indicator, but multi rule procedures are more commonly used, particularly for more complex methods. Sample rules are:-

Reject a run if, for the major economic element:

1. One control is outside 3SD (standard deviations)
   OR
2. 2 consecutive controls are outside the same 2SD
   OR

1. Reject a run if there is a range of >4SD between 2 controls
   OR
2. 4 consecutive controls are on one side of the mean and >1SD from the mean
   OR
3. 10 consecutive values are on the same side of the mean.

8.Critical systematic error

The critical systematic error is measured from blank and duplicate sample failures and should be very high, probably up to 90%. That is, 90% of the blank samples submitted come back with nothing in them.

9. Data quality objectives

These data quality objectives need to be reported:

1. Analytical bias (say <2.5%)

2. Coefficient of variation (say<5%)

3. Detection of critical systematic error (say >90%)

4. False rejection rate (say <5%)

Data quality objectives are realistic operating specifications giving allowable levels of inaccuracy (and imprecision) for different grades of material, and for each process related analytical or operating quality requirement. Objectives have to be set and must take into account cost implications if limits are set too tight or too loose. These will result in too many QC failures (true and false). There must be realistically high probabilities for error detection and realistically low probabilities for false alerts. Primary and secondary laboratories have to be checked for compatible equipment, methods and detection limits and obviously, the success of the program relies on having appropriate grade and matrix control materials.

10. False rejection rate

The false rejection rate is measured on the reanalysis of QC failures and shoud be very low, say <5%. That is, <5% of the failures, when reanalysed, came back with the same results. Or 95% of the failures you picked up were genuine QC failures.

11. Limits

Certification of the RM is based on a measurement campaign, so allocation of property values takes place on the basis of agreement among the independent measurement results and, not necessarily with direct traceability to Standard International units. Two of the ISO assumptions that allow this are; that there will be enough capable labs and that the results from each lab will be statistically compatible (ISO Guide 35 Clause 9.2.3 and Clause 10). To overcome the effects of this "inter-laboratory issue", and to achieve "a property value having satisfying uncertainty" (ISO Guide 35 Clause 10.2.2) RM producers must raise the minimum number of laboratories involved* and scrutinize the data with the aid of outlier treatment techniques. This should lead to a very accurate measure for a given method; notwithstanding the underlying assumption that what the good inter-laboratory labs reported was accurate. However, an amount of bad data may have an effect, resulting in limits which may too broad for effective monitoring of a single laboratory or production process. This is a major concern to the RM producer and it should be an issue (by simple extension) to the RM consumer.

RM consumers may wish to set their own limits, (normally within the RM producers limits) based on their own data quality objectives and control measurements. For example, limits probably don't need to be as tight, for tailings or low geochem grade levels, as for high ore and concentrate grade levels (financially).

*we find, practically, that at least 14 labs should report per method

12. Manufacturing

Manufacture begins with canvassing potential customers about desired characteristics for new standards. Ore with these characteristics is sourced from a mine or exploration project. The ore is crushed, then dry-milled and air classified to 100% <54µ. This fine powder is mixed in a blender for 14 hours and then split down into numbered 1 kg tubs. These lots are sampled for quality control and for round robin analysis. Quality control will typically comprise sampling 30 tubs selected from the whole stream. Round robin samples are selected the same way, so that one laboratory will receive samples from the beginning, end, and from throughout the batch.

13. National Nuclear Regulator

African Mineral Standards is licensed by the South African National Nuclear Regulator to produce, distribute and export reference materials containing uranium according to Nuclear Authorisation COR-198 (see section below "Uranium standards, shipping-export-import regulations")

Download the attached guide for more information.

14. Packaging

African Mineral Standards packaging is designed to protect the contents from moisture, oxidation, ordinary segregation, electrostatic segregation and tampering.

Laboratory Packs contain 1 kg of material sealed into 2.5 litre capacity bottles (space to shake). These are delivered sealed in barrier foil pouches with an oxygen reducer.

Explorer Packs contain between 50 to 250 gm of material (whatever you want). The Explorer Pack is reference material in a standard, unmarked, paper geochem envelope , vacuum sealed into a protective barrier foil pouch. Explorer Pack's are designed by a geologist, especially for the busy field geo and will withstand abusive treatment.

15. Prices and Dispatch

Prices are available on request. We normally dispatch within a week of receipt of order and we ship to any where in the world.

Delivery within Johannesburg is free.

16. Reference Materials (RM's)

Reference materials, commonly referred to as "standards", are an important element in exploration and mining best practice. They are samples, with a known grade, made from a specific ore, used as control samples to check results for specific methods from chemical analysis.They have been tested and found fit for their purpose, which can include the calibration of a measurement system, assessment of a measurement procedure, assigning values to other materials and quality control.

They are used to check on:

• accuracy from individual analyses
• method bias
• sample custody procedures
• sample traceability
• assay laboratory quality assurance
• the integrity of the mineral resource database
• the integrity of the mineral resource estimate

17. Resource audits

Mining professionals have to plan very carefully, for future resource and reserve audits, right from the commencement of exploration. Resource audits are similar to financial audits. They follow a sequence of paperwork set out in a logical path, to validate that best practice has been followed from geological sample to the database and into the mineral resource model. A sample audit trail, as an example, will establish that a specific set of assay results used for a mineral resource or a production efficiency calculation can be traced back to a specific sampler, a specific sample and a specific analysis.

18. Uranium standards, shipping-export-import regulations

Shipping of AMIS uranium bearing reference materials to countries such as Canada and Australia should not be a problem for two reasons. Firstly AMIS uranium standards contain such small quantities of uranium and thorium they generally do not classify as radioactive. Secondly African Mineral Standards is licensed by the South African National Nuclear Regulator to produce, distribute and export reference materials containing uranium according to Nuclear Authorisation COR-198.

The regulations (IAEA Safety Standards Series No. TS-R-1) are written in officialese, which makes them a bit inaccessible. Our guide summarizes the relevant bits. Notwithstanding, the ultimate responsibility to let the goods through will in any event rest on local customs officials. We will give them the paperwork filled out according to the international conventions; but customers are still advised to pass our paperwork by their customs officials first.

Download the attached guide for more information.

19. Use of reference materials

A simple example of how reference materials are used is as follows. They are placed in the middle of a stream of similar samples being submitted for assay. The reference materials should report results within set confidence limits. Their results are used as a check of the overall accuracy and precision of results in the final database or resource calculation. They will also verify that other individual assays match the correct samples in a batch.

This is one of the ways assay laboratory quality control and quality assurance is managed. It is also a standard auditing procedure to check on the efficiency of mine sampling procedures and metallurgical processes and, it is one of the things an exploration manager has to do to generate a bankable mineral resource.

The bankability of mineral resources is especially important. Audit trails are needed when due diligences or reviews are undertaken on a mine or exploration project. These happen when money is being raised or if the project is being sold or, simply if production and metallurgical process efficiencies are being audited as a management check. Audit trails provide a high degree of assurance (to share holders, financial institutions, exploration project management and mine management) that mine or exploration sampling meets the standards required by best practice and by the compliance codes for stock exchanges around the world (NI 43-101, SAMREC, JORC etc).