Good Delivery Rules
Annex D: Proactive Monitoring – Procedures and Criteria
The system of monitoring is designed mainly to provide reassurance to purchasers about the quality they can expect from GD Bars.
The main method of monitoring requires the Refiner to submit a dip sample from a commercial melt for check-assaying by two of the Referees. The dip sampling and Bar casting operation must be witnessed by a Supervisor. An alternative way of monitoring the assaying ability of “four-nines” gold Refiners is described in Section 4 of this annex. All Refiners (including Referees) will be monitored once every three years.
Evidence of GD Bar casting must be witnessed by the Supervisor. Exceptions to this requirement will include Refiners regularly supplying GD Bars to the London bullion market or other such recognised physical market, evidence of which must be submitted.
1. Notice to Refiners about monitoring
The LBMA will send a letter to the LBMA contact at the Refiner concerned (with a copy via email) informing them that proactive monitoring of its gold and/or silver production is to take place within a period of one month. LBMA is willing to be flexible on the time allowed for arranging the monitoring operation, if for example this is affected by holiday periods or other enforced shutdowns.
2. Dip sampling
2.1 Appointment of Supervisor
A Refiner being monitored by the dip sampling method should, in the first place, appoint a Supervisor to witness and report to LBMA about the sampling operation. The list includes the internationally recognised assaying and inspection companies which are within LBMA membership.
These companies have local representatives or laboratories around the world. The costs and expenses of the Supervisor must be paid by the Refiner. The Supervisor will charge a fixed fee (see LBMA website for all fees) for each dip sampling operation witnessed, unless specifically agreed otherwise, plus travelling and subsistence expenses incurred by the Supervisor’s representative. Thus, the expenses chargeable by the Supervisor will depend on the locations of their representative offices relative to that of the Refiner.
2.2 Witnessing of dip sample
The melt from which the dip sample is taken should have a fineness in the range of 999 or above for silver and between 995.0 and a maximum of 999.0 for gold. The sample should be taken from a normal production melt and the operations leading up to the taking of the sample must be witnessed by the Supervisor. The Refiner should be confident about what the melt contains and that it is homogeneous before taking the dip sample. The dip sample should be taken at the final stage of production, that is, just before casting.
The purpose of taking the dip sample is to provide sufficient homogeneous material to provide the samples to be assayed by the Refiner and the Referees, together with enough spare samples in case of various eventualities (such as a sample being lost in the post).
The actual method of taking and casting the dip sample can be either of the following:
- The Refiner may use a standard LBMA mould (which will be brought to the Refiner by the Supervisor). This consists of a two-part cast iron mould which produces a casting with dimensions of:
- for silver: 60 mm in width, 6 mm in thickness and 100 mm in height;
- for gold: 60 mm in width, 6 mm in thickness and 50 mm in height.
The Refiner should have a guillotine or shear available which can be used to crop 5 mm from each edge. In the case of silver, the cropped casting should then be cut into 8 pieces of approximately 25 x 22.5 mm each (giving a sample weight of around 35 grams). In the case of gold, the guillotine should be used to cut off eight samples of approximately 10 grams each.
- The Refiner can use its normal method of dip sampling, provided that this will produce the necessary samples for fire assay (in the case of gold) and, in the case of silver, for spectrographic analysis, including by spark OES which requires a plate type sample of dimensions approximately 25 x 25 mm).
The Supervisor will report to LBMA using a standardised format including information on:
- the use to which the refined metal will be put,
- the raw materials used,
- the processes leading up to the sample being taken,
- the method of dip sampling employed and,
- in the case of Bars which are to be numbered, the numbers of the Bars produced.
2.3 Treatment of the dip samples
Two of the eight samples will be sealed and sent by the Supervisor to LBMA. One will be left with the Refiner for assaying and five will be sealed by the Supervisor and left with the Refiner as reserves.
The sample left with the Refiner by the Supervisor should be assayed by corrected fire assay or appropriate spectrographic technique in the case of gold and by an appropriate spectrographic method of analysis in the case of silver. The number of individual fire assay trials to be carried out is not specified by LBMA but is instead left to the Refiner, according to its normal practice. For gold fire assays, the report should include the individual trial results expressed to five significant figures of fineness and the mean of the trial results, also to five figures.
The assay results should be presented in an Excel or .csv file and submitted by email to the GDL Team within four working days after the dip sampling. The method of assaying must be stated in the report (including the type of spectrographic testing used for silver). In the case of the assaying of silver by spectrographic methods, oxygen and nitrogen should be ignored when deducting the sum of the impurities from 1000 (in other words, these gases should be treated as silver).
When determining the assay of dip samples using spectrographic methods, the applicant is responsible for identifying all impurity elements contained therein which will determine the final assay. LBMA does not prescribe detailed procedures or criteria for assaying by means of spectrographic methods but Annex I lists the elements that Referees will typically determine.
The report on a silver dip sample should include the elemental analysis using the LBMA template which will be provided, as well as the silver assay obtained by difference for all trials.
The LBMA Executive Committee will treat the information provided by the Refiner in strict confidence. In particular, no information which could be used to identify the Refiner will be provided to the Referee(s) that will assay the dip sample. However, at the conclusion of the Proactive Monitoring the assays of the Refiner and Referee will be sent (anonymously via the Executive) to each other.
On receipt of the two samples by LBMA, both samples will be sent, according to a rota, to two of Referees who will be asked to assay the sample they receive to five significant figures. It should be noted that the Referees will not be aware of the identity of the Refiner that provided the samples.
For gold, the Referee will carry out at least 6 trials by means of corrected fire assay and will include the results in the report sent to LBMA. In the case of silver, the Referee will normally use one or other spectrographic analysis method and determine the silver assay by difference (with dissolved gases such as oxygen counting as silver). The Referee will provide to LBMA the elemental analysis of the dip sample as well as the silver assay obtained by difference.
If the assays of the Refiner and Referees fail to agree within the tolerances described below, the Refiner will also be asked to unseal one of the spare samples, carry out an assay on it and submit a new assay report to LBMA within five working days.
The Referees are all Refiners of both gold and silver, who have previously demonstrated to LBMA’s satisfaction a very high level of accuracy in the assaying of gold and silver. They also manufactured sets of reference samples which are free from detectable inhomogeneity and whose assay values were established to high levels of accuracy by means of an extensive programme of cross-checking.
2.4 “Four-nines” gold Refiners
LBMA considers that all Refiners on the Good Delivery Gold List must be able to assay across the full range of Good Delivery alloys (namely a fineness range from 995.0 to 999.9) most of which can only be accurately assayed using the method of corrected fire assay. At the top end of this range, on the other hand, spectrographic methods can provide assays of the necessary precision and accuracy. In that these high-gold alloys can be thus assayed without requiring the use of fire assaying, they cannot be used to demonstrate that the Refiner is able to assay over the full range of Good Delivery alloys. For Refiners where the production technology (as well as the products marketed) only involve gold of fineness 999.9 and above, it is recognised that it would be disruptive and onerous for them to have to produce a special low gold content alloy for the purposes of LBMA monitoring.
A Refiner which, for the reasons described above, is unable to provide a gold dip sample with a fineness of less than 999.0, may instead opt to have an alternative form of monitoring, whereby LBMA will send it a set of six approximately 5-gram reference samples for the Refiner to assay using the corrected fire assay method. On receipt of the samples, the Refiner must submit to LBMA within six working days a report showing the mean assay of each sample to five significant figures.
2.5 Assessment criteria and further testing
The Refiner's mean assay value (in the case of dip samples) and detailed trial results (in the case of the four-nines gold procedure) will be assessed by the LBMA as described below. In cases where the Refiner is deemed to have failed, the mean assays and the standard deviations of the assay results may be viewed, anonymously, by the Referees and/or a technical consultant engaged by LBMA.
The criteria are shown below. The tolerances on assaying shown here are expressed in terms of fineness (parts per thousand). Thus, for instance, ±0.10 for an assay of, say, 998.55 means a range of fineness from 998.45 to 998.65.
Consideration of assays from first dip samples
The Refiner's and Referee’s assay results on the two dip samples provided by the Refiner will be assessed as follows:
- Full pass – In the case of gold, agreement between the Referee’s and Refiner’s assays within ±0.15 will be regarded as a full pass with no further testing being required. In the case of silver, different criteria apply depending on whether the sample’s fineness (as assayed by the Referee) is above or below 999.5.
Above 999.5 agreement within ±0.05 will be regarded as a full pass, while below 999.5 agreement within ±0.15 will be regarded as a full pass.
- Borderline failure – i.e., agreement in the range ±0.16-0.25 (or for silver samples of fineness of 999.5 and above, agreement in the range ±0.06-0.15). This will require that the Refiner be asked to assay one of the spare samples which have been sealed and left at the Refiner by the Supervisor. On receiving the assay results from the Refiner, the LBMA Executive Committee will compare all of the results once again and, if necessary, taking technical advice, decide on whether the results are acceptable. If they are not, the Refiner will be asked to arrange for a new dip sample to be witnessed within one month and provide a further two samples for testing by the Referees.
- Fail – i.e. a divergence of >0.25 (or for silver samples of fineness of 999.5 and over, a divergence of >0.15). In this case, the Refiner would be required to provide a further two samples from a new witnessed dip sample within one month.
Cases where a second dip sampling operation is required In general, two different Referees will assay the second pair of samples compared to those that assayed the first samples. LBMA will assess the results based on the criteria described above but taking into account all the assay results provided by the Refiner and the Referees.
If necessary, after taking advice from a technical consultant, LBMA will then decide on one of the following courses of action:
- The Refiner will be informed that it has passed the monitoring test.
- The Refiner will be asked to assay a set of LBMA reference samples (under similar conditions as for a new applicant for GoodDelivery accreditation).
In the latter case, LBMA will assess the assay report subsequently provided by the Refiner and decide whether:
- The Refiner has satisfied the criteria and will therefore be informed that it has succeeded in passing the monitoring test; or
- The Refiner will be required to undergo a full re-application for Good Delivery accreditation.
In the latter case, except in cases of gross failure, the Refiner will normally continue to be listed until the results of the re-application are available.
Criteria for assays provided by “four-nines” gold Refiners
In the case of the “four-Nines” gold Refiners which opt to be monitored by means of assaying a set of six LBMA reference samples, the criteria for passing the test are the same as those applicable in the case of new applicants for listing, except for the allowed divergences:
- Assays of 999.5 and above should agree to ±0.05; for example, the Refiner’s assay on a sample assaying 999.84 according to the Referee would have to fall within the range 999.79 to 999.89;
- Assays below 999.5 should agree within ±0.15 provided that no significant bias is apparent; for example, the Refiner’s assay on a sample assaying 996.73 according to the Referee would have to fall within the range 996.58 to 996.88.
However, it will be deemed acceptable if there is not more than one divergence provided that this is not greater than ±0.25.
2.6 Conclusion of monitoring
The GDL Team will inform the Refiner of the outcome of the assay comparisons as soon as they have been reviewed by the Chief Technical Officer. A table showing the comparison of the anonymised mean assay values will be provided to the Refiner and the Referees which participated. LBMA will provide the Refiner with a certificate confirming the success of the PAM process.
2.7 Provision of comparisons of assay results
For silver dip samples, the LBMA Executive Committee will provide guidance to the Refiner about the differences between its analysis and that of the Referee by highlighting any elements which are found in significantly different concentrations.
2.8 Charges for reference samples and re-testing
In cases where the comparison of the Refiner’s and Referees’ dip sample assay results suggests the need for the Refiner’s assaying ability to be more thoroughly checked by means of it assaying a small set of reference samples, as described above, there will be an additional charge as outlined on the LBMA website, under the section Provision of Self Testing Samples to Bona Fide Applicants.
The cost of shipment of these samples to the Refiner will be payable in addition.
The additional charge for a complete reapplication and re-test of the Refiner’s assaying ability and Bars would be the same as for new GD applicants.
2.9 Completion of monitoring
Refiners must complete the Proactive Monitoring Programme in full (including bar casting) within six months from receipt of the notification letter.