Issue 101
Where Did the Gold Go? Precious Metal Losses in Refineries
For a precious metals operation, the most important task, after ensuring the safety of workers and community, is control of the precious metals inventory.
Fire Assay (cupellation) © Metalor
In some operations, such as a jewellery fabrication, this is readily accomplished through a system of weight checks and piece counts. In these operations, the form of the material might change, but for the most part, the composition does not.
However, in a refinery, it is more challenging to keep track of precious metals inventory, because the material is changing in form and composition throughout the operation. Feed material is received, and then it is melted, sampled and weighed. The material is blended with other lots and is processed through pyrometallurgical and hydrometallurgical operations that ultimately produce refined precious metals and base metal by-products.
Complicating matters for refineries are the number of by-products produced from the refining operations. By-products, such as salts, sludges, filter cakes, slags, dusts or solutions, contain precious metals in varying concentrations and they have to be recycled and reprocessed to recover the precious metals content. These by-products are often heterogeneous in nature, highly variable in metals content and numerous in form.
An accurate assessment of the precious metals in these by-products can be difficult to achieve due to the challenge of obtaining representative samples. The usual approach of weighing and grab sampling leads to a poor estimate of the metals content.
Physical Inventories
Precious metals control and accounting is accomplished through regular physical inventories. The refiner has a record from the previous stocktake of the work-in-progress (WIP) at that point in time. The amount of incoming precious metals is known from the refiner’s evaluation of received lots and it also has a record of outgoing material from weights and assays of its products. The refiner then applies the following formula (equation 1) to determine the calculated WIP at the time of the new physical inventory stocktake:
WIPt=WIPt-1 + ∑tt-1Ri -∑tt-1Si
Where:
WIPt = Calculated WIP at time t (or the most recent stocktake)
WIPt-1 = WIP determined from previous stocktake
Ri = incoming metal from lots received between time periods t-1 and t
Si = outgoing metal from lots shipped between time periods t-1 and t.
The formula is often summarised as Calculated WIP = WIP from previous stocktake + received precious metals – shipped precious metals.
The calculated WIP value is sometimes referred to as the book number.
PRECIOUS METALS CONTROL AND ACCOUNTING IS ACCOMPLISHED THROUGH REGULAR PHYSICAL INVENTORIES
Gains and Losses
The purpose of a physical inventory is to determine the actual physical WIP in the operations and compare it with the calculated WIPt value from equation 1 and to compute the gain/loss from the following formula (equation 2):
Gain/Loss = Physical WIP – Calculated WIPt
It needs to be appreciated that the computed gain or loss number is not a single definitive value.
Instead, it is the end result of numerous samples, weights and assays, each of which has an associated measurement error. Just as a repeated assay is unlikely to yield the exact same result, a repeated inventory is unlikely to yield the exact same value. Like an assay, the physical inventory result is an average with an associated standard deviation or error range. The reality is that the calculated error range is often much larger than the industry is comfortable with.
Even with this knowledge, refiners, driven by the need to book a financial result, treat the calculated gain/loss number as a definitive value and book a single number. Caution needs to be applied and it would seem to be prudent to only book gains or losses that lie outside the expected range of variability.
Gains and losses should be viewed in context and with an awareness of the ‘noise’ built into the result.
| Metal Accountability | Metal Loss | Gold Losses $/received oz @ 1800/oz Au | Silver Losses $/received oz @ $18/oz Ag |
|---|---|---|---|
| 99% | 1% | $18.00 | $0.18 |
| 99.9% | 0.1% | $1.80 | $0.02 |
| 99.95% | 0.05% | $0.90 | $0.01 |
| 99.99% | 0.01% | $0.18 | $0.002 |
| 99.995% | 0.005% | $0.090 | $0.001 |
| 99.999% | 0.001% | $0.018 | $0.0002 |
Bearing in mind the inherent error built into the gain/loss calculation, the table below shows the dollar value of various degrees of gold and silver losses. In a world without variability, a useful guideline would be that the losses as a percent of the received feedstock should be no larger than 0.01% for Au and 0.10% for Ag. In other words, accountability for gold should be within 99.99% and accountability for silver should be within 99.9%.
THE FOCUS OF THIS ARTICLE IS THE TECHNICAL LOSSES THAT ACTUALLY RESULT IN THE LOSS OF PHYSICAL METAL
When discussing precious metals losses, it is important to distinguish between transactional and technical losses. Transactional losses are those caused by incorrect entries into the lot tracking, accounting or ERP systems. They tend to be referred to as accounting or data entry errors and occur through the transposition of numbers, double-booked entries or overlooked entries. These errors can, with diligence, often be found and corrected.
The focus of this article is the technical losses that actually result in the loss of physical metal.
These are the more problematic losses and can be grouped into two categories: known losses and unknown losses.
Known Losses
Known losses are losses that can be calculated and accounted for even though they are largely unrecoverable. These losses occur due to the deportment of precious metals into other products of the operation or into by-products. A well-known example is the loss of gold in the refined silver output of a refinery. For example, a typical quantity of gold in refined silver bullion is 5 ppm. If the operation is a large silver refinery producing 10 million oz of Ag per year, a 5 ppm gold content results in an annual loss of 50 oz of gold for which the refiner is never compensated. Another example occurs in the sale of gold bullion.
The refiner is only compensated for the stamped purity of its bars, for example, 99.99%, even though the purity of its refined material might be higher, e.g. 99.996%. In a refinery with a large output, small differences between actual purity and compensated purity can be considerable.
The reasons for unknown losses are numerous and include:
- Precious metals below the payable limits in subcontracted materials
- Contracted retentions of metal in subcontracted materials
- Trace amounts of precious metals in wastewater or airborne discharges from a refinery
- Weight giveaways on final products due to rounding rules and minimum weights.
REFINERIES SPEND AN ENORMOUS AMOUNT OF EFFORT IN TRACKING DOWN THE CAUSES OF UNKNOWN LOSSES AND WORKING TO MINIMISE THEM
Unknown losses
In inventory accounting, the correct approach is to include known losses in the ∑tt-1Si term from equation 1. When this is done, equation 2 yields the value for unknown losses or gains. It is the amount and the value associated with unknown losses that are the single largest cause of distress and management turnover at any refining operation. Even when the losses are small and tolerable, they can be a source of frustration as their cause might not be obvious. As a result, refineries spend an enormous amount of effort in tracking down the causes of unknown losses and working to minimise them.
The reasons for unknown losses are numerous and include:
- Theft
- Incorrect values from the most recent stocktake
- Incorrect values for WIP from a previous stocktake
- Incorrect evaluation of incoming material
- Incorrect values for outgoing or shipped materials.
Each of these possible losses are discussed below.
Equipment including reactors, piping and ductwork, all of which require regular cleaning
1 Theft
The security operations at most refineries are designed with two purposes. The first is to limit outside access and the second is to keep the precious metals within the operation.
Even in the most highly secured operations, theft from the inside is a continual threat.
When working with large amounts of high-value material that is not immediately traceable, the temptation to try sneak small amounts out of a facility is ever present.
Outside pressures on employees, driven by addiction, financial hardship or even blackmail, can drive the most well-intentioned individuals to make poor choices. Thefts often start off small but then grow in amount and frequency, and the ingenuity applied in illegally removing precious metals from a facility is often astounding. Even though most operations have a separate security department whose employees serve as gatekeepers, theft from an operation often involves the collusion of these same gatekeepers with operators in the facility who have hands-on access to the precious metals.
2 Incorrect values from the most recent stocktake
One of the most important tasks taken on by a refinery is the periodic determination of the precious metals physical inventory in an operation, also known as a stocktake. In most refining operations, this is done several times a year and it involves the stopping of the operations for a few days. All the equipment and reactors are cleaned out, all the WIP material is assembled and consolidated into lots, which are weighed, and samples are taken for analysis. For some difficult-to-sample materials, the lots would be separately processed and tracked, with care being taken not to commingle them with materials produced after the inventory cut-off date.
These special lots would be treated to recover the precious metals content, which can then be weighed and assayed. A common example is the treatment of gold-bearing solutions by precipitating out the gold, which can then be melted, weighed and analysed.
There are many ways that a physical inventory evaluation can go wrong, leading to incorrect results. The most common are:
- Equipment or operations incompletely cleaned out
- Incorrect weights assigned to lots
- Poor sampling practice.
The first two can be overcome by good operating practices, such as clear stocktake protocols that provide explicit instructions as to how equipment should be cleaned out and procedures for ensuring weighs are checked, rechecked and recorded correctly. It takes work to find all the precious metals-containing material in a refinery and, as a result, equipment, ductwork, piping and air scrubbers have to be disassembled, cleaned out and reassembled. A well-designed refinery considers that periodic equipment clean-out is necessary and clean-out ports, inspection hatches and manways are incorporated into the equipment design.
ALL MATERIALS SHIPPED FROM A REFINERY SHOULD BE HOMOGENISED WEIGHED AND SAMPLED INTERNALLY
Poor sampling practice
The most common cause for error is due to poor sampling practice. There can be an over-reliance on solution assays in order to avoid the work associated with the precipitating out and processing of the precious metals in solutions. If the solutions are concentrated and not well mixed, they will yield poor samples.
Sampling errors are particularly common with intermediate and lower-grade WIP materials such as sludges, filter cakes, slags, spent crucibles, etc. There can be large volumes of these materials in a refinery and it takes a significant amount of work to correctly homogenise and sample these materials. Sometimes, these materials are directly subcontracted, in an unprocessed, non-homogenous form, to another operation specialising in these materials.
This leaves the refiner dependent on the determination of the precious metals value by the subcontractor. Errors do occur in these situations and to avoid these, all materials shipped from a refinery should be homogenised, weighed and sampled internally.
One of challenges with treating WIP lots after a stocktake is ensuring their segregation and tracking. It is good practice to minimise inventories of by-products at the time of stocktake. This reduces the amount of material that has to be treated after stocktake and also lessens the evaluation errors that accompany these materials.
XRF © Metalor
3 Incorrect values for WIP from a previous physical inventory
If a counting, weighing or analytical error similar to those described above was made during the previous stocktake, this would create an error in the previous precious metals physical inventory result. If the error is not repeated during the current stocktake, an over-evaluation of the previous inventory would yield a loss once the current inventory numbers are included in equation 1. An under-evaluation in the previous inventory would, of course, lead to a gain after the current inventory results are included in equation 1.
4 Incorrect evaluation of incoming material
Errors made in evaluating the precious metals content of incoming lots is a common reason for precious metals losses. If errors are made in the evaluation process, through incorrect weights, poor samples or incorrect analysis, this will result in the over- or under-evaluation of the precious metals content.
For those experienced in the precious metals industry, it is understood that coming to a final agreement with the customer of the precious metals value in a lot is ultimately a negotiation bound by the settlement terms outlined in the refining contract and underpinned by the statistics of sampling and analysis. Ultimately, there is no ‘true’ value for the precious metals assay in a lot – there is simply a mutually agreed to, or negotiated value.
Reiss and Taylor1 have examined these negotiations and they note that in the case where simple averages are offered (= measured assays), both sides give up a little when they settle. The refinery pays out a bit more than its average result and the customer receives less as well. Because the incentives exist to reduce these losses, Reiss and Taylor have used game theory and Monte Carlo simulation to provide guidance for negotiating parties to minimise their losses in settlement negotiations.
For the refiner, the final settled value is often a little bit higher than the average that came out of the assay laboratory and it is important to account for these small and cumulative differences when calculating the total received precious metals over a specified time period.
Gold grain weighing © Metalor
A – Sampling errors
Refiners deal with two basic types of material. High-grade metallic materials and lower-grade materials that are often non-metallic in nature, such as sweeps. Each of these present their own sampling challenges and once a representative sample is taken, it then needs to subsampled correctly to produce the analytical aliquot. In the steps from sampling the lot to splitting the sample to subsampling in the lab, there are numerous potential sources of error.
Refiners need to understand and apply good sampling practice. It takes knowledge, experience and effort to produce samples that are representative of the received materials. Poor samples can result in an under- or over-evaluation of the received materials, which can result in unexpected losses or gains.
B – Analytical errors
Even if representative samples of materials are taken, mistakes can occur in the analytical step. The mistakes range from mixing up lots, poor preparation of the analytical samples, weighing errors, incorrect analytical procedures, faulty calibrations, poor proof corrections and miscalculations of the final results. Most good assay labs have rigorous quality assurance and control protocols in place to ensure that these errors are minimised and when they occur, they can be identified. Many labs also have programs where they periodically run standard reference materials or even just their own internal standards to ensure that reliable and consistent results are being produced.
5 Incorrect values for outgoing or shipped materials
The products leaving a precious metals refinery are a combination of refined metals in various forms, subcontracted refining lots shipped to other refiners, as well as plant trash and scrap. Possible sources of error associated with shipped materials include:
A – Incorrect shipment weights
It is possible that incorrect weights are assigned to material shipped out of a facility. This is not normally an issue with refined materials in ingot form, as the shipments tend to be rigorously checked and they are accompanied with bar lists, which include bar counts, weights and serial numbers. It is more likely to occur with refined materials in grain, powder or sponge form, but many refiners have instituted procedures that involve weight checks by different personnel to prevent this from occurring.
B – Evaluation errors
There is generally little room for error in the shipment of refined materials as the various materials have been subjected to extensive analysis to ensure they meet the required purity standards. However, materials generally subcontracted, such as sweeps, slags, spent crucibles or sludges, are prone to evaluation errors if good sampling protocols are not followed. As noted previously, sometimes refiners choose to send out these materials in an unprepared form and rely on the receiving party to prepare, homogenise, sample and assay these materials. This is a risky endeavour and it leaves the sending party poorly prepared for the settlement process. Instead, the materials should be evaluated internally so that they are correctly prepared for the settlement process.
C – Plant trash and scrap
Most refiners do a good job at ensuring that internally generated plant trash, such as rags, wipes, used safety equipment, filters, etc., are processed internally through incineration. However, sometimes food-related waste is segregated and is simply consigned to an outside dumpster. This is also the fate of broken equipment such as filters, pumps and reactors. Sometimes this equipment is poorly cleaned out and inspected, and this can be a cause for precious metals losses. In the case of one refining operation, the food waste placed in the dumpster was a conduit for theft.
Concluding Remarks
It is important to appreciate that just as there is variability in weight and assay determinations, there is variability in the precious metals gain/loss number that is calculated after a physical inventory determination. When booking gains or losses, caution needs to be applied and it is prudent practice to only book gains or losses that lie outside the expected range of variability.
Due to the built-in variability of the results, even the best-run refining operations will at times experience losses or gains. As with large losses, large gains are a cause of great concern and indicate a fundamental problem. They need to be diligently investigated so that the cause can be determined and corrected.
Refining operations make substantial investments in security equipment, personnel and material control software packages to monitor the flow of precious metals within a refining operation and to prevent losses.
All these are important and necessary but need to be coupled with the general operating guidelines shown in the table below, which go a long way to understanding and preventing losses.
| Guidelines for Reducing Understanding and Reducing Precious Metal Losses |
|---|
| 1 Gain an understanding of the built-in variability of the physical inventory results using standard statistical techniques. |
| 2 Understand and apply the Theory of Sampling to improve the sample reliability. |
| 3 Process by-products internally and in a timely fashion. |
| 4 Keep WIP levels low – this way, there is less to count, sample and weigh, leading to a smaller chance of errors. |
| 5 Watch your receipts – ensure good sampling and assaying practices, and invest in quality assurance and control programs in the melt shop, lab and assay lab. |
| 6 Watch your shipments – institute multiple weight checks, be rigorous about plant trash checks, and do not send out non-homogenised and unsampled materials. |
Acknowledgements
The author wishes to thank Arthur Taylor and Mike Hinds for their comments on early drafts of this article.
1Reiss, M., and Taylor, A, ‘MMC Assay Exchange Program White Paper’, 2019, Private Communication
Mike has Master’s degrees in Chemistry and Business and a PhD in Metallurgical Engineering (University of Utah) and he has spent most of his career in the extractive metallurgy and precious metal business starting as a gold miner in South Africa. He has served as VP Production and General Manager for Metalor Technologies USA and has published over 40 papers. He is also the holder of two US Patents for precious metal recovery.
He is the President of Argo Advisors International, a consultancy and engineering company that specialises in the precious metals and extractive metallurgical industries. He has worked and consulted for major metallurgical research organisations and metals operations around the world.
When he is not consulting, he is a Professor in the MBA program at Franklin Pierce University, New Hampshire where he coordinates the Energy and Sustainability program. He is a member of the Mineral, Metals and Material Society and Co-President of the New England Chapter of the International Precious Metals Institute. During the summer of 2009, he was a Visiting Scholar at the Federal Reserve Bank in Boston. In 2015 he was awarded a Fulbright Scholarship by the US Government for energy studies in Botswana.