Simón Armando-Text-Control sample insertion rate-is there an industry standard

Control Sample Insertion Rate: Is There an Industry Standard?1 Armando Simón, Ph.D., R.P.Geo.2 Keywords: QA/QC, insertion rate, industry standard, exploration, mining.

Abstract There is no definite consensus in the mining industry about the existence of an industry standard for the QA/QC sample insertion rate, which has a clear economic significance in the budget of an exploration program. If a reasonably sound answer is required, the proper way to approach this problem is to inquire the industry. Four main sources have been consulted during the preparation of this paper: well known international QA/QC consultants, SEDAR3-filed technical reports, information published in the Internet by exploration and mining companies in web sites and press releases, and documents issued by regulatory organizations. This review demonstrates that there is a preference in the mining industry for total insertion rates of QA/QC samples close to 20%, including duplicates, standards, blanks and external checks.

Introduction Quality Assurance/Quality Control in exploration geology is not a new invention. However, according to the author’s experience in auditing exploration and mining projects in various continents, rigorous geological QA/QC programs were uncommon until recent years, even in projects and mines run by major mining companies. After the infamous Bre-X affair, strict policies were implemented by all major regulatory bodies. In spite of the fact that comprehensive QA/QC programs are not abundant yet, junior and major companies are increasingly interested in implementing such programs, particularly when public financing is required. Unfortunately, their initial interest is often followed by shock and, sometimes, even by anger, when project management realizes that the implementation of a QA/QC program involves certain modifications, always upwards, in the exploration budget. The insertion of various types of control samples within the ordinary batches implies a directly proportional increase in the handling, preparation and analytical expenses of up to 20% in average, not mentioning the extra cost for the acquisition of commercial certified reference materials (CRM), or for the preparation of in-house CRM. Although the overall cost increase of the implementation of a QA/QC program is relatively reduced, not exceeding usually 1% to 2% of the total exploration costs, the psychological impact of the first figure cannot be neglected. Reluctance for implementation of QA/QC programs does not arise only from management or from budgetary constraints. Control insertions also require improved organization in the sampling process, database preparation, and data processing. ! "# $ & '( ) * # +, # # (# - . 1 2%34 & *3 #, 2 # 5 $ ,6 2 " ! , 6 " 777# # #

" "

%

'!# $ / #

0

#

# (

#

Project geologists not always understand the need for such additional efforts, and often complain about these supposedly “redundant” control measures. The author emphasizes that the purpose of this note is not to discuss the QA/QC principles, but to review some of the most commonly recommended and applied practices regarding the insertion rate of control samples in the mining industry.

New Regulations on QA/QC Technical reports written under the specifications of Canadian NI 43-101, for example, should specify whether a Qualified Person has verified the data on which this information is based, including sampling, analysis and tests. These reports should also describe in detail the QA/QC program and the nature and limitation of the verification, and should explain any problem encountered during data verification (CSA, 2005a, 2005b). In another example, a Competent Person report following the Australian JORC Code should thoroughly describe the nature, quality and appropriate selection of sampling and analytical procedures, as well as the quality control procedures, including the insertion of certified reference materials (CRM), blanks, duplicates and external checks, and assess the actual accuracy and precision levels attained during the project (JORC, 2004). Nevertheless, the new regulations avoid suggesting precise figures for quality control insertion rates, which remain at the latitude of the Qualified or Competent Person. Consequently, the control sample insertion rate becomes a frequent source of disagreement between auditors and QA/QC specialists, on one hand, and project management and geologists on the other.

QA/QC Insertion Rate: Is There an Industry Standard? The author has not found anywhere a clear definition of a QA/QC industry standard regarding the control sample insertion rate. What is the meaning of industry standard? What can be understood as a best practice? Rogers (1998) discusses the difference between industry standard and best practices: “So-called ‘widely accepted’ or ‘industry standards’ practices most neither be confused with ‘best practices’ nor with standards. There are many procedures widely accepted by the mining industry that are adequate in some situations but inadequate in others. Moreover, such acceptance is not universal.” However, if a reasonably sound answer is required, the proper way to approach this problem is to inquire the industry. Four main sources have been consulted during the preparation of this paper: well known international QA/QC consultants, SEDAR4-filed technical reports, information published in the Internet by exploration and mining companies in web sites and press releases, and documents issued by regulatory organizations.

International QA/QC Consultants The opinions of recognized international QA/QC consultants, condensed from papers prepared during the last nine years, are summarized below (Table 1): )

,6 2 " ! , 777# # #

6

"

(

#

• •

•

•

•

•

•

Rogers (1998) recommends an internal control using duplicates, CRMs and blanks in a proportion of 1 in 20 samples, and submitting 5% of the samples to an independent lab. Vallée (1998) stresses the importance of systematic verifications of the subsampling and assaying protocols, and refers to a 10% of reference samples (duplicates, blanks) inserted by many companies in lots sent for assaying, in addition to a ‘somewhat lower figure for rock sampling’. Neuss (1998) refers the Outokumpu’s experience in QA/QC best practices. His recommended program includes 2% to 5% field duplicates, 5% to 10% coarse reject duplicates, 5% to 10% internal duplicates (assayed at the primary laboratory) and 5% to 10% external checks (assayed at the secondary laboratory), in addition to at least one CRM and one blank in every submission. Long (1998, 2000) suggest the following insertion rates: 5% coarse reject duplicates, 5% pulp duplicates, 5% CRMs, one blank per batch (approx. 3% to 5%), and for check assays, a portion of the pulp duplicates (3%?). Sketchley (1999) considers that a batch of 20 samples should include at least one CRM (5%), one blank (5%) and one duplicate (5%). If the laboratory control of the analytical system is suspect and cross-contamination between samples is likely, additional CRMs and blanks should be considered. In addition, he states that all duplicate pulp samples should be also analyzed at an independent facility to monitor accuracy and contamination. Bloom (1999) recommends inserting 5% blanks and 5% CRMs (control samples) and routinely sending 10% of the pulps to an umpire lab (check samples). In addition, she recommends the use of field duplicates, but does not suggest any particular insertion rate. In her view, a minimum quality control program would consist of a 15% insertion rate (5% control samples + 10% check samples). She also quotes a study prepared by ACA Howe International on ten exploration programs, for which the direct assaying costs contributed to 3.4% to 14.5% of the total budget, averaging 7%. On this basis, Bloom (1999) estimates that the implementation of a minimum QA/QC program should represent in average approximately 1% of the overall program. Lomas (2004) recommends that for a twenty-sample batch in a copper project, one coarse reject duplicate (5%), one blank (5%), one CRM (5%) and one pulp duplicate (5%) should be inserted. In addition, she recommends that 5% of the pulp duplicates be submitted for control to an external lab.

Other authors (Davis, 1998; Voortman, 1998) are not specific about the insertion frequency of control samples, but advocate for the insertion of field and pulp duplicates, CRMs, coarse and pulp blanks, and the submission of check samples to reference labs. As observed in Table 1, a general agreement seems to exist between international QA/QC consultants about recommending in average overall insertion rates of control samples close to 20%.

$

Table 1. QA/QC Programs: Suggested Insertion Rates by Various Authors Source

Details

Rogers (1998)

Duplicates, CRMs, blanks: one in twenty; external checks: 5% 10% duplicate plus CRMs, a ‘somewhat lower figure’ for rock sampling (?) 2%-5% field duplicates, 2%-5% coarse duplicates, 5%-10% internal pulp duplicates, 5%-10% external pulp duplicates, plus one CRM and one blank in every submission 5% coarse reject duplicates, 5% pulp duplicates, 5% CRMs, one blank per batch (approx. 3%), check assays, a portion of the pulp duplicates (3%) In a twenty-sample batch: one blank, one CRM, one duplicate; in addition, all pulp duplicates should be re-assayed at check lab In a twenty-sample batch: one blank, one CRM; in addition, sending one in ten sample pulps to an umpire lab In a twenty-sample batch: one blank, one CRM, one coarse duplicate and one pulp duplicate; in addition, 5% of the pulps should be re-assayed at check lab (including CRMs)

Vallée (1998) Neuss (1998)

Long (1998, 2000) Sketchley (1999) Bloom (1999) Lomas (2004)

Suggested Proportion of Control Samples Approx. 20% Approx. 15% (?) Approx. 19% to 25% Approx. 21% Approx. 20% Approx. 20% Approx. 25%

SEDAR-filed Technical Reports A second source for commonly used insertion rates in the mining industry is represented by technical reports published in the SEDAR system. The author reviewed a random selection of recently published technical reports (16) resulting from placing Google® queries5 for “technical report”, “insertion rate”, “qa/qc” and “43-101”, with no preference for region, size of the company or type of mineral. Not all of the consulted reports had definite figures to describe the QA/QC protocols, but those with detailed data are listed below: • • • • • •

3

6

Porcupine Project, Canada (GoldCorp): Coarse rejects: 5%; coarse blanks: 5%: pulp duplicates: 5%; CRMs, 5%; check samples: 5%. Total: 25%. Source: AMEC (2006) Modder East Load Project, South Africa (sxr Uranium One Inc. and Aflease Gold Ltd.): Coarse blanks: 2%; CRMs, 9%; pulp duplicates: 11%; check samples: 2%. Total: 23%. Source: SRK (2007). Perama Hill Project, Greece (Frontier Pacific Mining Corporation): Duplicates, 10%; other control samples: 9%. Total, approx. 19%. Source: RPA (2004). Nuestra Señora, Mexico (Scorpio Mining Corporation): Coarse duplicates, 2.5%; CRMs+blanks, 2.5%; pulp duplicates, 5%; pulp check samples, 5%; coarse reject check samples, 2.5%. Total: 17.5%. Source: CAM (2006). Twangiza Project, Congo (Banro Corporation): 2% coarse blanks, 8% CRMs; in addition, check assays (proportion not specified). Total: 15% (?). Source: Skead (2006). Mirador Project, Ecuador (Corriente Resources): Coarse duplicates: 5%; pulp duplicates: 5%; CRMs: 5%. Total: 15%. Source: MDA (2006).

!

8

2'

9

4#

)

• •

HC Property, Nevada (J-Pacific Gold): 5% twin samples, 5%; coarse duplicates, 5%; pulp duplicates, 5%. Total: 15%. Source: Durgin (2005). Pueblo Viejo, Dominican Republic (Placer Dome): 10% of CRMs and blanks. Source: AMEC (2005).

A general trend for using a 4% to 5% insertion rate for each type of control samples (blanks, duplicates, CRMs, check assays) can be observed, although in some cases particular sample subtypes are ignored. The insertion rate is less than 17% only when check assays are not included. An acceptable average is approximately 18%, with minor differences in some particular types of samples. The lack of duplicates in the Pueblo Viejo program invalidates it as an element for comparison. In many of the studied examples, only one CRM was included in the QA/QC program. When various CRMs were considered, sometimes there was no correlation between the grade levels of the CRMs and the actual sample grades. Not infrequently, the author has reviewed projects where CRMs have below cut-off values, or even close-todetection-limit levels.

Information from Exploration and Mining Companies This information has been obtained mainly from press releases published in the Internet by a random selection of exploration and/or mining companies. The selection resulted from placing Google® queries6 using “exploration”, “mining”, “qa/qc” and “insertion rate” as key words. Unfortunately, most companies do not offer details of their QA/QC protocols in the press releases, but the author could find some examples: • • •

• • • •

Carpathian Gold (Colnic, Romania): Coarse blanks, 5%; CRMs, 4%; Check samples, 20% (before AMEC’s Technical Report in 2006). Total: 29%. Source: www.carpathiangold.com/site06/images/CheckCode.pdf. African Copper (Dukwe Project, Botswana): approx. 20% control samples. Source: www.mineweb.net/co_releases/302480.htm. Aurelian Resources, FDN epithermal Au-Ag: CRMs, duplicates and blanks, 15%; in addition, samples from significant drill intercepts are sent to two reference laboratories. Total: 18% (?). Source: www.aurelian.ca/dynamic/press/pr-2006-08-21.pdf. GlobeStar Mining. Regular practice: Duplicates: 4%; CRMs, 4%; blanks, 4%; check samples, 4%. Total: 16%. Source: www.globestarmining.com/content/standards.php. Cambridge Mineral Resources: blanks, 5%; duplicates, 5%; CRMs; 5%. Total: 15%. Source: www.cambmin.co.uk/?page=press_releases&num=61. Scorpio Mining Corporation (Nuestra Señora, Mexico): Coarse duplicates, 2.5%; CRMs, 2.5%; check assays: 5% pulps, 2.5% coarse rejects. Total: 12.5%. Source: www.scorpiomining.com/i/pdf/QAQC-NS.pdf. Belvedere Resources: 12% control samples (only CRMs, blanks and duplicates). Source: www.belvedere-resources.com/rss/

The same general trend for using a 4% to 5% insertion rate for each type of control samples (blanks, duplicates, CRMs, check assays) is observed. With the exception of the Nuestra Señora Project, the insertion rate is less than 16% only when check assays are not included. An acceptable average is approximately 20%, with minor differences in some particular types of samples.

4

6

!

8

2'

9

4#

3

Regulatory Bodies In 1999 the TSE and the Ontario Securities Commission prepared a document which later became the basis of NI 43-101 (CSA, 2005a). The document, named Setting New Standards (TSE-OSC, 1999), recommended that in a sample batch of 20 samples there should be a duplicate sample, a coarse blank, and a CRM. The document also recommended that previously assayed pulps be re-submitted to the same lab (rate not stated) and to another lab as check assays (rate not stated). The first three control samples would represent an overall 15% insertion rate, and the additional pulp reassays (internal and external to the primary lab) would probably take the total to a figure close to 20%.

A Typical Example A comprehensive QA/QC program should allow assessing the sampling, sub-sampling and analytical variances, as well as analytical accuracy and possible contamination during preparation and assaying. The QA/QC programs should be tailored to the specific needs of each project. Whereas an overall insertion rate of 20% can be in principle recommended, the individual proportions of the various types of control samples should reflect the problems with higher probability of occurrence. With the advance of the QA/QC program and the identification and correction of those problems, the amounts and relative proportions of control samples can be adjusted accordingly. For an initial core drilling QA/QC program, AMEC would recommend the following control sample insertion rates (Table 2): Table 2. Core Drilling QA/QC Programs: AMEC’s Suggested Insertion Rates Sample Type

Sample Sub-Type Twin Samples* Duplicates (6%) Coarse Duplicates* Pulp Duplicates* CRMs (6%) CRMs* Coarse Blanks* Blanks (4%) Pulp Blanks* Check Samples (6%) Check Samples**

QA/QC Label TS CD PD CRM CB PB CS

Insertion Rate 2% 6% 2% 2% 6% 6% 2% 4% 2% 4% 4%

* To be assayed by the primary laboratory; ** To be assayed by the secondary laboratory. It should be emphasized that the external check batches should also include pulp duplicates, CRMs and pulp blanks in appropriate proportions, so that precision, accuracy and possible contamination at the secondary laboratory could be independently assessed.

Conclusions •

•

A general agreement appears to exist between the consulted sources (international consultants, SEDAR-filed technical reports, published company information and documents from regulatory bodies) about an average recommended insertion rate of control samples of approximately 20%. Most companies do not differentiate the duplicates subtypes (twin samples, coarse and pulp duplicates) or blank subtypes (coarse and pulp blanks), all of 4

•

•

them with different functions in a comprehensive and properly conducted QA/QC program. In many of the studied examples, only one CRM was included in the QA/QC program. When various CRMs were considered, sometimes there was no correlation between the grade levels of the CRMs and the actual sample grades. A common situation was using CRMs with below cut-off values, or even with close-to-detection-limit levels. The implementation of a comprehensive QA/QC program would represent, in average, an increase of 1% to 2% of the total exploration costs.

Recommendations •

•

The QA/QC programs should be tailored to the specific needs of each project. Whereas an overall insertion rate of 20% can be in principle recommended, the individual proportions of the various types of control samples should reflect the problems with higher probability of occurrence. With the advance of the QA/QC program and the identification and correction of those problems, the amounts and relative proportions of control samples can be adjusted accordingly. It is essential that QA/QC programs are comprehensive, including all types and subtypes of control samples, namely twin samples, coarse and pulp duplicates, coarse and pulp blanks, CRMs corresponding to relevant grade values, and external check samples, so that precision, accuracy and possible contamination at the various points in the sampling-preparation-assaying sequence are properly assessed.

References AMEC (2005): Pueblo Viejo Project, Province of Sánchez Ramírez, Dominican Republic. NI 43-101 Technical Report. Report prepared by AMEC Americas Limited for Goldcorp Inc. Project # 147107, October 16, 2005. Published at www.sedar.com. AMEC (2006): Review of Porcupine Joint Venture Operation, Ontario, Canada. NI 43-101 Technical Report. Report prepared by AMEC Americas Limited for Goldcorp Corporation. Project # 152236, August 28, 2006. Published at www.sedar.com. Skead, M. (2006): Twangiza Project, South Kivu Province, Democratic Republic of the Congo. NI 43-101 Technical Report. Report prepared for Banro Corporation. March 30, 2006. Published at www.sedar.com. Bloom, L. (1999). The role of economic geologists in evaluating assay data quality. Manuscript. CAM (2006): Nuestra Señora Project, Sinaloa, Mexico. NI 43-101 Technical Report. Report prepared by Chlumsky, Armbrust and Meyer, LLC. for Scorpio Mining Corporation. Project # 047251, April 17, 2006. Published at www.sedar.com. CSA (2005a). National Instrument 43-101, Standards of Disclosure for Mineral Projects, Canadian Securities Administrators. CSA (2005b). Companion Polity 43-101CP to National Instrument 43-101, Standards of Disclosure for Mineral Projects, Canadian Securities Administrators. Davis, B. (1998). What is a sample? What does it represent? In: More Meaningful Sampling in the Mining Industry. Australian Institute of Geoscientists, Bull. Nr. 2, 1998, pp. 39-45. Durgin, D. (2005): 2004 Year-End Technical Report Hc Property Eureka County, Nevada. Report prepared for J-Pacific Gold, Inc. March 17, 2005. Published at www.sedar.com. JORC (2004). Australasian Code for Reporting of Mineral Resources and Ore Reserves - The JORC Code 2004 Edition. The Joint Ore Reserves Committee of the Australasian Institute of Mining and Metallurgy, Australian Institute of Geoscientists and Mineral Council of Australia. Lomas, S. (2004). QAQC Program. General Discussion. AMEC Internal document. Long, S. (1998). Practical Quality Control Procedures in Mineral Inventory Estimation. In: Quality Assurance, Continuous Quality Improvement and Standards in Mineral Resource Estimation. Exploration and Mining Geology, V.7, Nr. 1 and 2, 1998, pp. 117-127. Long, S. (2000). Assay Quality Assurance-Quality Control Program for Drilling Projects at the Prefeasibility to Feasibility Report Level (3rd. Ed.). MRDI, Internal report. MDA (2006): Technical Report Update on the Copper, Gold, and Silver Resources and Pit Optimizations, Mirador Project, Ecuador. Report prepared by Mine Development Associates for Corriente Resources Inc. Project # 775-856-5700, May 18, 2006. Published at www.sedar.com. Neuss, I. (1998). Sampling “Legends” – What Can We Learn from Busang? In: More Meaningful Sampling in the Mining Industry. Australian Institute of Geoscientists, Bull. Nr. 2, 1998, pp. 109-118. Rogers, R.S. (1998). Forensic Geology and Mineral Exploration Projects. In: Quality Assurance, Continuous Quality Improvement and Standards in Mineral Resource Estimation. Exploration and Mining Geology, V.7, Nr. 1 and 2, 1998, pp. 25-27.

:

RPA (2004): Report on Perama Hill Gold Deposit Mineral Resource Estimate, Greece. NI 43-101 Technical Report. Report prepared by Roscoe Postle Associates Inc. for Frontier Pacific Mining Corporation. May 13, 2004. Published at www.sedar.com. Skead, M.B. (2006). NI 43-101 Technical Report, Twangiza Project, South Kivu Province, Democratic Republic of the Congo. Report prepared by Banro Corporation, March 30, 2006. SRK (2007): Dominion Uranium Project North-West Province, Republic of South Africa. NI 43-101 Technical Report. Report prepared by SRK Consulting for sxr Uranium One Inc. Project # 370896, March 2, 2007. Published at www.sedar.com. TSE-OSC (1999). Setting New Standards: Recommendations for Public Mineral Exploration and Mining Companies. Toronto Stock Exchange-Ontario Securities Commission Mining Standards Task Force. Vallée, M. (1998). Sampling Quality Control. In: Quality Assurance, Continuous Quality Improvement and Standards in Mineral Resource Estimation. Exploration and Mining Geology, V.7, Nr. 1 and 2, 1998, pp. 107-116. Voortman, A.G.W. (1998). Avoiding Over- and Underestimation of Grade by Using Correct Sampling. In: More Meaningful Sampling in the Mining Industry. Australian Institute of Geoscientists, Bull. Nr. 2, 1998, pp. 3945.

*