A Review of Analytical Methodologies in PM Analysis
This paper was presented by Dr. Malkit Basi, Laboratory Manager and Technical Director of Ledoux & Company, at the June 2006 IPMI Conference.
An attempt has been made to review a broad range of methodology currently in use at labs providing services to the precious metals industry. Preference of a particular technique may vary from one lab to another. Very often this choice is based upon the range of equipment available for assay purposes and the training level of its staff. It is important to realize however is that the methodology be appropriate for the type of material being analyzed and the tolerance on the final result be acceptable for commercial exchange.
Few basic Considerations.
Proper sub sampling methods for materials that lack homogeneity.
Ability and knowledge of the Analyst.
1. Representative Sample:
An assayer can only provide a result that at its very best reflects the true contents of the sample itself. Therefore it is important that the sample submitted to the lab is a good representation of the bulk material. Furthermore when the lab sub divides the sample, due care is taken to insure that it is sub sampled in a way that is appropriate for that type of material.
Samples for precious metals assays are often submitted to us in various shapes and forms. We have received circuit boards in their entirety, jewelry items of various sorts and rocks the size of a melon. So-called prepared samples will normally be submitted as:
The assumption is that somebody has drawn a small portion of the entire shipment in a way that the results can be projected to arrive at the true value of that shipment.
Sub-sampling the material for assay purposes is a critical step. Very few materials are uniform enough that you may remove a spoonful, analyze a single portion and report the results. In the real world the material will have to be prepared either before or after it arrives at the lab. This entails pulverizing the sample to a particle size where it can be considered reasonable to take a spoonful, analyze a few portions, report the average and feel comfortable with the end product. Scientific measurements are only true if the results can be replicated.
Picture of swing mill
Picture of the sieves
For umpire samples we would avoid any preparation unless it has been agreed with both parties up front. We tend to analyze many more portions when umpiring than we would for party assays. In some cases we may exhaust the entire sample sent to us if the replicates do not agree. Generally we would analyze 3 – 6 portions depending on the type of material. This is to ensure that we have a representative result.
When preparing samples it must be realized that if one is not careful then we may change the true character of the material. It is well known that soft metals like gold may easily smear inside the grinding equipment. In this way gold may be lost. Furthermore the equipment would have to be thoroughly cleaned before use so that the next lot does not get contaminated.
There are times when samples are either too large and of wide particle distribution or too small to be properly analyzed. Furthermore because of particle size effect and density factors a physical separation of small and large particles can occur. In this case it becomes difficult to extract a representative sample.
Freeport, Tintaya and Alumberera, Grasberg copper concentrates are known to show the gold nugget effect and the material may segregate in the sample bag. Some of the lead concentrates may also show high variability with respect to gold. Refractory material sent for platinum, palladium and rhodium assays behaves similarly. Samples of this type always show high portion to portion variation and the results show poor precision.
Table refractory glass
In this type of situation it is very difficult to provide results that can be considered representative of the bulk. Where applicable we will analyze several portions of the sample and report all the result with a view that when exchange is carried out due consideration can be given to the wide distribution of results.
This type of material should preferably be sub-divided using a rotary splitter and a large part of the sample be consumed in the test so that at least one can safely arrive at a reliable value for the precious metal contents of the submitted sample.
Picture of a Rotary Splitter.
Basically the entire sample is fed by a vibratory feeder and collected in containers that are spinning at a constant rate. In this way collection containers have the same particle distribution as the original sample submitted to the laboratory. If a rotary splitter is not available then the sample should be quartered and coned. This procedure has stood the test of time.
Material processed primarily from recycling electronic scrap and that from the jewelry industry has very often a metal fraction and a fine powder. Coarse and fine fractions may be submitted to the lab separately and knowing the weights of each the final settlement results can be provided on ‘composite basis’.
Picture Powder and Metal mix.
Mixed fractions always produce results with a higher degree of variability and the results are not easy to replicate. If necessary the sample can be sieved and the coarse material analyzed separately in its entirety. Composite value derived form analyzing each fraction separately will provide a result that can claimed to be representative of the lot.
For representative analysis a completely homogeneous sample is required. Preferably the sample should be ground to -200 mesh. Most of the time we receive prepared samples of high uniformity. This type of material can be mixed in the sample package and 2 - 10 g portions can be drawn and analyzed.
Metal samples on the other hand present a different type of problem. Very often they arrive at the door as pins or buttons.
Picture Pin sample
Picture button sample
Pin samples sometimes may contain glass occlusions that may not be part of the sample meant to be assayed. If that happens to be case, then care need to be exercised so that the glass particles are not weighed as the part of the portions to be assayed. Glass occlusions can be removed easily by physical impacting such as hammering or rolling. In such cases just taking the pin, chopping a few pieces and analyzing is not adequate.
Button samples may also have some slag attached to them. We may receive silver, copper, iron and lead buttons. The buttons may have to be chopped or drilled so that fractions can be drawn for assay work.
Aqua regia leaches.
Sometimes sedimentation occurs during storage, especially if the sampling was carried out at an elevated temperature. On cooling solids may fall out.
Proper storage containers should also be used. Elements such has gold will reduce on to surfaces of plastic containers.
Solution volumes are very much dependent on the temperature. As a matter of consistency it is always best to carry out assays on weight basis. Alternatively density of the solution can be recorded at the same time.
It should also be remembered that the solutions should always be stored in appropriate containers for storage and transportation. Unless properly sealed, containers often leak when transported by air due to pressure variations.
2. REPORT BASIS.
For powder samples it is normal to report the results on dry basis. Drying temperatures and times will vary by sample type. Customers may wish to receive results on:
As Received basis
Typical drying temperature is 105 C overnight and the ignition temperature is 1000 C for 4 hours
Alternative temperatures and times may be selected if appropriate. Drying and ignition loss may change due to oxidation of sample components and due to volatilization losses.
Table -- LOI 4hrs vs 1 hr.
Gold is a commonly exchanged commodity. Results from proof corrected fire assay usually form the basis of commercial exchange. Rarely customers will request for the results to be reported on uncorrected basis. Such assay do not account for the losses to slag and cupel absorption and will not reflect the true value of the gold present in the sample.
Unless asked otherwise we will report in percentage.
We are often asked to report in the following units:
Parts per thousand
It is very important to establish correct sampling and reporting protocols for the assay result to be meaningful.
2. TYPE OF METHODOLGY
2.1 FIRE ASSAY
There are many books on this subject. No discussion is complete without mentioning the works of Edward E. Bugbee8 and Ernest Smith9
The fire assay method is most suitable for Gold and silver analysis. This has remained the standard for commercial transactions. An excellent collection of literature exists. Although in use since antiquity the actual procedure still remains somewhat of a mystery. Both American Society for Testing Materials and International Organization for Standardization for Standards have tried to introduce unified methodology for testing some of the materials.
· ASTM E1335 Standard Test Method for DETERMINATION OF gold in Bullion by Cupellation
· ISO 11426 Determination of gold in gold jewelry alloys - Cupellation method (fire assay)
Under the ASTM E1335 laboratories worldwide can participate in the proficiency tests conducted every six months. When the procedure was written initially it was meant to handle relatively simple gold/silver alloys with low base metal loading. Since then it has been applied to a wider range of alloys. When such materials are introduced into the proficiency program suggestions are often made how the procedure can be logically extended to handle such materials.
Examples include -
· Samples containing higher amounts of zinc needs additional lead for cupellation.
· As the nickel contents of the alloys increase, a higher cupellation temperature is necessary
Cupellation procedure can also be used for silver assays to arrive at a commercially acceptable results. In this case total precious metal beads are produced. Cupellation temperature is used where silver will not be lost. The precious metal bead must be tested for the presence of other metals so that a true silver figure can be given out. This bead may contain the following elements:
Au Pt Pd Bi Pb
Retentions of these elements may sometimes be dependent on the concentrations of other elements. An excellent example was given by Ronald McCloskey10 IPMI Newport Rhode Island. He has summarized how the presence of platinum and palladium affects the lead retention.
Fire assay procedures may undergo some or all of the following steps:
At each stage there is a risk that the precious metals may be lost. At the same time the inquarted element may in fact be retained. Therefore it is imperative that properly prepared proofs be run alongside to work out the gains or losses.
Fire Assay on its own or in combination with other wet chemical procedure may also be applied to various type of powders and complex sweeps. Use of proper flux and the fusion temperatures become very important in this case. Limitation of lead fire assay is reached when samples with high alumina, iron and chrome have to be assayed. In
Particle size reduction can be carried out for easy to grind samples. Where the possibility of the sample components smearing in the grinding container it should be avoided. For the correct sample type it is easy to fuse ˝ AT. For difficult to fuse sample the portion size may have to be reduced to 1 – 2g. It brings with itself some problems of its own. For many type of samples it may not be appropriate to take few 1-g portions out of 1-oz sample and then try to project the results for a valuation of shipment worth few hundred thousand dollars.
Labs with limited capability only, there may not be any other options open. It is much better to opt out for wet chemical analysis at this stage where much larger samples 10 – 20 g can easily be handled providing a better estimation of the lot as a whole.
2.2 Classical Wet Chemistry.
Classical procedures are often labor intensive and time consuming and may involve complex separations. Thorough knowledge of sample composition and chemistry is needed if reliable results are to be produced. However this type of methodology provides a standard benchmark test. That is why they are often used for referee purposes. Examples are various type of gravimetric and titrimetic procedures. For titrimetry it is essential to analyze standard reference standards alongside. This is only way the reliability of the procedure be tested.
Vast amount of literature exist on this subject.
Analytical Treatise by Kolthoff and Elving
Gravimetric methods rely on the isolation of on element or it compound and then weighing it. Weights can easily be recorded 6-digit accuracy or better. Therefore the measurement errors are negligible. This type of methodology works best in the hands of trained chemists.
2.3 Instrumental Methods.
The trend over the second half of the 20th century has been to move towards more and instrumental methods. Introduction of the Spectrograph in the 40s and the 50s was followed by Atomic Absorption Spectrometry in the late 50s and the 60s. It was a step in the right direction. For the precious metal chemist the impact was minimal. Breakthrough came with the plasma spectrometry in the 70s and the 80s. In the very beginning DC Plasma was promoted but it has by now almost became extinct and has been superseded by Inductively coupled Plasma Emission Spectrometry (ICP-ES)..
Techniques such as solution photometry, atomic absorption and spectrographic analysis at one time very popular but now have become almost obsolete. Other techniques such as GDMS have emerged as powerful tools.
Following type of instrumentation may often be found in precious metal lab.
· Inductively Coupled Plasma Emission Spectrometry. Commonly used in quantifying metals in solution.
· Inductively Coupled Plasma Mass Spectrometry. Most useful for quantifying impurities in refined materials.
· X-ray Fluorescence Spectrometer. Where suitable standards available it is a rapid method for analyzing alloys. Otherwise in the Semiquant mode it is an invaluable tool for the chemist for determining the approximate composition of the material for selecting appropriate methodology and proceeding with the full assay.
If the sample is uniform in nature and can easily be dissolved, direct instrumental analysis can sometimes provide reasonable results at a reduced cost. Assay times can also be shortened immensely. However it must be realized it is a comparative technique and can be easily influenced by matrix interferences.
When using instrumental methods of analysis it becomes necessary to isolate precious metals and then quantify them for an accurate analysis. Otherwise incompatibility of various elements to stay in solution at the same time and interference problems may easily provide erroneous data. Fire Assay is very often used for isolation of precious metals. Various metals can be used as collectors.
Where available reference materials should be analyzed at frequent intervals as a part of the quality control procedure. The methods must be shown to work by showing spike recoveries.
Performance based methods must be shown to work on well characterized materials.
Instrumental methods are rapid. Accuracy depends on the availability of good standards.
The methodology must evolve with time to accommodate accuracy requirements and commercial pressures as to the turnaround times and the cost of analysis.
3. METHODOLOGY SELECTION
3.1 Refined materials.
American Society for Testing Materials in discussion with various refiners has defined Specifications when resting for refined Silver, Gold, Platinum, Palladium, rhodium, Iridium and ruthenium
ASTM B413 - Silver - Standard Specifications
ASTM B562 - Gold - Standard Specifications
ASTM B561- Platinum - Standard Specifications
ASTM B589 - Palladium - Standard Specifications
ASTM B616 - Rhodium - Standard Specifications
ASTM B671 - Iridium - Standard Specifications
ASTM B717 - Ruthenium - Standard Specifications
The impurities can be measured by any proven technique. No particular method for testing has been defined. It has been left to the discretion of the Assay Lab. Traditionally it has been DC arc. Nowadays with the impurities can also be effectively be measured by alternative techniques such as GDMS and ICP-ES and ICP-MS or GDMS. Gold, platinum and palladium can easily be dissolved in aqua regia. With a combination of ICP-ES and ICP-MS the all the common impurities can easily be measured.
There is definitely lack of commercially available reference materials for checking the validity of the procedures mentioned above.
Support materials and washcoats have been changing over the coarse of years to keep up to date with the commercial pressures and clean air requirements. Pellets catalysts with mainly alumina supports gave way to silica and alumina base. Nowadays we are coming across more exotic materials such as Silica carbide. Using a single procedure it is not possible to cover them effectively. Washcoats also may have various amounts of Zirconia and Ceria.
Alternative 1. Nickel Sulfide Collection. Precious Metal Separation. ICP.
Nickel Sulfide 3 - It is an effective way of extracting precious metals from a variety of difficult to handle marices. Method is particulary effective for recovering Platinum, Palladium, Rhodium and Ruthenium. Effective gold recovery is dependent on a number of factors. Smith and Gerrard 6 compared nickel sulfide vs lead collection and showed quite clearly with the fluxes they were using that Nickel sulfide was very effective in collecting gold, platinum, palladium and rhodium and that lead collection was unnecessary. At Ledoux and company we use nickel sulfide collection for collecting platinum, palladium and rhodium in difficult to handle matices. I realize it will effectively collect Iridium and rhodium too. Sample should preferably be ground to -200 mesh for achieving good results. Sample sizes 5 - 10 g can easily be accommodated.
Jain et al 7 have effectively used this procedure to collect low levels precious metals including gold. Final quantification was by ICPMS
Alternative 2. Tin/Tellurium Separation 11
It is an effective way of determining precious metal loading of the autocatalysts. In the initial step the sample is fused with sodium peroxide. By the very nature of this exercise this limits the portion size to at the maximum approximately 2-g. Therefore the sample will have to be very finely ground. We prefer the sample be ground to –200 mesh.
Alternative 3. Wet dissolution precious metals and support materials with a mixture of mineral acids. Recovery of precious metals in the insoluble by fire assay. Precious metals separate through classical separation and then quantified by solution spectrometry.
Alternative 4. Gold Collection through fire assay.
Alternative 5. X-ray Fluorescence.
Unless matrix matched standards are available quantification is at the best is only semi-quant. Furthermore even if the standards are chemically matched particle size distribution is impossible to match. By the very nature of the technique it plays a significant role in the accuracy.
Lack of Reference Standards. NIST 2555 and 2556. Both are very different from the materials in use today. Need for performance based methods. Some companies have conducted RRs. On unusual materials wide variations between labs.
3.3 Reforming catalysts:
Alternative 1. Acid dissolution of precious metals and support material. No separation. Platinum Colorimetric. ASTM D 4642-92.
ASTM committee D-32 drafted standard test method for Platinum in Reforming Catalysts13. It was only applicable to non zeolite fresh alumina based catalysts. No matrix separation was required. Dissolution in dilute hydrochloric acid was considered sufficient. In those days instrumentation such ICP was still in its infancy. Therefore colorimetric finish using stannous Chloride was selected. It is an effective way of quantifying platinum. Still the procedure is used highly effectively by various labs across the globe. The only flaw in the procedure is that when the used catalyst are analyzed there may be residues that may not dissolve that easily. Further treatment of the residue is necessary before results are considered exchange quality.
Work by Silve Kallmann14 formulated a philosophy that still forms the basis of the approach taken by Ledoux and Company. Total dissolution of the sample is achieved. Precious metals in any residue are extracted by fire assay using silver inquart and combined. Since then the only change to this procedure was introduced in the early 90’s with the introduction of ICP. The instrument itself can achieve a precision of 0.1 – 0.25%. Exchange quality results can easily be produced.
Like any other methods ICP is also prone to interferences effects and this could limit the accuracy. An example is the presence of Molybdenum. This element shows significant interference at the primary emission lines 265.945 and 299.xx nm.
Alternative 3. Fire assay collection through silver or gold inquartation followed by ICP finish.
Direct extraction of platinum by silver or gold inquart through lead assay is also prevalent. Like any other fire assay procedure small losses may occur during the whole process. Matching proofs cannot really be prepared and Standard Reference Materials are really available for checking the accuracy. Therefore the method becomes as good as the analyst itself. Sample size is rather limited. Method is better suitable to zeolite base.
Alternative 4. Sodium Peroxide Fusion. Tin/Tellurium Separation. ICP finish.
Portion sizes are limited.
This type of material is often unclassified. Sweeps can virtually contain anything you can think of. Silve Kallmann15 in his paper published in 1983 had outlined a general approach. The basic concepts are still true. For simple sweeps fire assay plays a central role. Gold is analyzed by parting and weighing. Platinum and palladium is analyzed in parting acids after silver separation. Silver is very often reported by difference. This approach fails miserably in complex matrices. Presence of constituents like large amounts of alumina, chrome and iron limit the portion sizes that can be fused by fire assay. Therefore enough gold may not be recovered for weighing for an accurate assay. Furthermore the presence of other precious metals may complicate the situation more.
Alternative 1. Sodium peroxide fusion followed by PGM precipitation with Sodium Formate or Tin/Tellurium.
Alternative 2. Nickel Sulfide Collection.
Nickel Sulfide is an excellent collector for Rhodium. Small losses may occur during the fusion step for Platinum and Palladium. Therefore empirical correction must be applied for the data to comparable with other methods. Gold recovery very often happens to be low. The method can be applied to a wide variety of matrices, Merits of this technique are many. However it requires that the lab be equipped for this type of work. Besides an adequate furnace for fusion a proper scrubber system needs to be in place for removing the hydrogen sulfide fumes generated during the dissolution of the nickel sulfide buttons.
Alternative 3. Copper or Copper Sulfide Collection16
Alternative 4. Lead button Collection.
Depending on the metal contents both precious metal and base metal contents a scorification may have to be carried out. Then resulting lead buttonis dissolved and precious metals determined gravimetrically or instrumentally as appropriate. The method is highly suitable for sweeps containing high lead and silver.
Alternative 5. Silver bead method.
5 – 15 g sample can be fused depending on the sample type. 0.20mg – 0.25g precious metals can be collected this way. Silver beads can then be treated to isolate precious metals. Final quantification can be gravimetric or instrumental as appropriate.
3.5 Dental Alloys12
By the very nature the alloys contain high amounts of gold and palladium and smaller concentrations of platinum and silver.
Alternative 1. Fire Assay.
Fire assay can be used in a limited way. High concentrations of platinum can sometimes contaminate the parted gold. Therefore the purity needs to be checked. Also the dental grindings are not be that uniform in nature. Therefore large sample weights are preferred and this can be be easily accommodated by the wet chemical procedures.
Alternative 2. Wet Assays.
Silver is analyzed gravimetrically by precipitation as AgCl. Gold is reduced to metal and weighed. Palladium when present above 2% is best determined gravimetrically by DMG precipitation.
Gold and Palladium is impregnated on to Silica or Alumina supports. Approach varies considerably.
Alternative 1. Gold by Fire Assay palladium by DMG precipitation.
Gold in silica based VAMs can easily be determined through fire assay by parting and weighing. Since the palladium to gold ratio is high therefore the purity of gold should always be checked. Proofs must always be run alongside. Palladium can be quantified in the parting acids by precipitation with dimethylglyoxime and weighing either as its complex or by reduction to metal.
As above but palladium is quantified by Inductively Couple Plasma.
Gold and Palladium collected in silver inquart. After nitric acid dissolution silver is removed as its AgCl. Then both gold and palladium quantified by ICP.
Alternative 4. Gold by hydroquinone and palladium by DMG precipitation.
When fire assay facility does not exist the alternative is to remove silica by volatilization with hydrofluoric acid. Then gold can be precipitated by reagents such as hydroquinone and weighed. Palladium is determined by demethylglyoxime.
Providing the samples show a high degree of homogeneity chemical procedure can be devised to handle almost any type of material and produce results that are a close reflection of the metal contents. Methods have to be selected that are fit for the purpose and are not too expensive to carry out.
Various routes will lead to the same destination. Whether one chooses to drive, fly or sail one must ensure that one is properly equipped and has the sufficient expertise. Selecting chemical methodology is no different. Correct execution is only possible if proper knowledge base exists and trained personnel are available.
It must be remembered though that one route may be more cost effective than the other. This will definitely affect profit margins and productivity. Standardizing procedures is very difficult though because each laboratory is equipped and manned in its own way.
1. Frank Smith, Mark Payette, John Bozic, David Maskery and Zbigniew Waszczylo. Alternative to the Fire Assay Procedure for the Preparation of Automobile Catalyst Samples Prior to Analysis. Canadian Journal of Analytical Sciences and Spectroscopy Vol 44, No. 5, 1999
2. Silve Kallmann and Everett W. Hobart. Determination of Silver, Gold and Palladium by a combined Fire-Assay atomic absorption procedure. Talanta 1970, Vol 17, pp 845-850.
3. Report No. 1371 Concentration of Noble Metals by a fire assay technique using nickel sulfide as the collector. National Institute for Metallurgy, Johannesburg, South Africa.
Report No. 1705 The effects of various matrix elements on the efficiency of fire assay procedure using nickel sulfide as the collector. National Institute for Metallurgy, Johannesburg, South Africa.
4. Analysis of Automobile exhaust emission control catalysts, Silve Kallmann , Talanta Vol. 27, No. 10, pp. 827-833. Pergamon Press 1980.
5. The determination of platinum in reforming and emission control catalysts by differential spectrophotometry, Silve Kallmann, Talanta, vol. 23 pp. 579-583, Pergamon press 1976.
6. Analysis of Gold and PGMS by Nickel sulphide Collection and ICP. P.D. Smith and P.W. Gerrard. IPMI Proceedings. Pp 353 -360. Proceedings of the Eleventh International Precious Metals Institute Conference Brussels. Belgium 1987.
7. A Procedural Modification of Enhanced Recovery of Precious Metals (Au, PGE) Following Sulphide Fire Assay and Tellurium C-Precipitation: Applications for analysis of Geological samples by Innductively Coupled Plasma Mass Spectrometry. Min Sun, Jinesh Jain, Mefu Zhou and Robert Kerrich. Canadian Journal of Applied Chemistry Vol 38, No.4, 1993
8. Edward E. Bugbee. A Textbook of Fire Assay. Legend, Inc.
9. Ernest .A. Smith, The sampling and Assay of Precious Metals Met-Chem Research.
10. Ronald McCloskey- Corrected Silver Assays and Lead Retention in dore bead-Analytical Seminar 84 IPMI Newport Rhode Island pp 23 - 28
11. Louis N. Shenouda and William A. Millard, The determination of Precious Metals in Automobile Catalysts and Other Low-Grade Materials using Tellurium Collection.26the Conference Priceedings of the IPMI, Miami, Florida.
12. William H. Swanger, Analysis of Dental Gold alloys. Department of Commerce. Scientific Papers of the Bureau of Standards. Volume 21. United States Printing Office. Pp 209-239.
13. Standard Test Method for Platinum in Reforming Catalysts by Wet Chemistry. ASTM D 4642-92.
14. The Determination of Platinum in Reforming and emission Control Catalysts by Differential Spectrometry. Silve Kallmann, Talanta. Vol 23 pp 579-583 Pergamon Press 1976.
15. Referee Analysis of Precious Metal Sweeps and Related Materials. Silve Kallmann and C. Maul. Talanta Vol 30, No. 1, pp 21-39, 1983.
16. Analysis of Sweeps. The Cuperous Sulphide Collecting System. Silve Kallmann. Talanta, Vol 33, No. 1, pp 75-83, 1986