QAQC Protocols 2007

DRILL CORE SAMPLING AND ANALYSIS PROTOCOLS (QAQC MANUAL)
Charles Beaudry, M.Sc., P.Geo., go 26 August, 2007
QAQC Manual - 2007
Copyright Notice: Charles Beaudry 2007. This manual is copyrighted by the author and by various sources as cited but not necessarily used with permission. Any reproduction of material should include the original citations. No reproduction permitted except for personal use.
Charles Beaudry, M.Sc., P.geo., go
ii
QAQC Manual - 2007 EXECUTIVE SUMMARY In recent years Regulatory Authorities in Canada and other countries have increased reporting requirements for exploration projects. Project reports must include descriptions of measures taken to ensure a minimal level of quality in analytical results and corrective actions taken when results fail to meet requirements. All these procedures come under the general term QAQC which is composed of two parts namely QA or quality assurance which consists of the design or planning of sampling program so as to ensure a minimum level of quality and QC or quality control defined as the active procedures used during sampling to monitor, detect and correct errors to guarantee that results that do end up being used in the project actually meet the design specificiations. The QAQC manager is of crucial importance in an exploration project since he or she is tasked with planning the quality system and delivering the desired level of quality in the database and reported in the QAQC report. The position must be filled by a qualified person or QP as defined in Canada by regulation NI43-101 as a licensed engineer or geoscientist with a minimum of 5 years experience in exploration and some experience in the area of application. Although currently no equivalent definition exists in Brazil this may change in the future. Moreover Brazilian geologists meeting minimum requirements can apply (and some have) to become members of any one of several professional associations in Canada. This manual proposes a staged approach to QAQC on the premise that the requirements of early stage projects are simpler than for advanced projects in which the analytical results will be used for resource estimation and major investment decisions. Three stages are proposed with Stage I primarily applied to grass roots diamond drilling and other types of sampling where the focus is on batch to batch variability of analytical results. Stage II is a transitional stage that is implemented immediately upon discovery of a potentially economic mineralization and Stage III is applied with deposit delineation and definition drilling where the number of control procedures is much greater to ensure all results meet the requirements for mineral resource estimation. Guidelines are presented that cover all procedures from the drill to sample preparation, to analysis and include transportation and chain of custody. Procedures are presented for the interpretation of batch analytical results along with criteria for accepting and rejecting batches. This manual is a general guide that will be applicable in most cases. However for precious metal projects with erratic mineralization, it is possible that special sampling and sample processing requirements will need to be implemented. Special testing is required in such cases that go beyond the scope of this manual and the reader is advised to consult with sampling specialists who can design special sampling protocols adapted to the particular needs of an orebody.
iii
QAQC Manual - 2007
iv
QAQC Manual - 2007 TABLE OF CONTENTS

EXECUTIVESUMMARY .................................................................................................................... III TABLEOFCONTENTS......................................................................................................................V LISTOFFIGURES............................................................................................................................VI LISTOFTABLES.............................................................................................................................VII INTRODUCTION...............................................................................................................................1 TERMSOFREFERENCE.....................................................................................................................2 DESCRIPTIONOFPROCESS...............................................................................................................2 DRILLCORESAMPLING....................................................................................................................3 SECURITY.............................................................................................................................................3 PROCEDURESATTHEDRILL...................................................................................................................... 3 TRANSPORTATION ..................................................................................................................................3 CORESAMPLING...................................................................................................................................4 DRILLCOREDUPLICATES......................................................................................................................... 5 SAMPLEPREPARATION ....................................................................................................................6 CRUSHING ............................................................................................................................................6 PULVERIZING........................................................................................................................................6 DIGESTIONS....................................................................................................................................7 ANALYSIS........................................................................................................................................8 ASSAYVSGEOCHEMICALANALYSIS............................................................................................................ 8 ANALYSISFORDELETERIOUSORSECONDARYELEMENTS................................................................................. 9 DOCUMENTATION.................................................................................................................................9 CONTROLMATERIALS....................................................................................................................10 PURCHASEORPREPARATION?................................................................................................................ 10 INSERTIONOFPULPSTANDARDS............................................................................................................. 11 MaterialsRequired..................................................................................................................... 11 Procedure...................................................................................................................................12 SampleLog.................................................................................................................................12 BLANKS.............................................................................................................................................13 PreparationofBlanks................................................................................................................. 13 MaterialsRequired..................................................................................................................... 13 Procedure...................................................................................................................................13 InsertionofBlanks...................................................................................................................... 14 MaterialsRequired..................................................................................................................... 14 Procedure...................................................................................................................................14 VALIDATIONOFCONTROLMATERIALSANDLABORATORIES........................................................................... 14 ValidatingStandards.................................................................................................................. 15 ValidatingBlanks........................................................................................................................ 15 SELECTIONOFLABORATORIES.......................................................................................................15 LABORATORYCOMMUNICATIONANDSERVICECONTRACT............................................................................. 15 SETTINGLABORATORYPERFORMANCESPECIFICATIONS................................................................16 STANDARDS ........................................................................................................................................16 BLANKS.............................................................................................................................................19 CRUSHINGANDPULVERIZATION.............................................................................................................. 19 QUALITYASSURANCEPROGRAMS.................................................................................................21
QAQC Manual - 2007

PROJECTSTAGES.................................................................................................................................22 BATCHASSEMBLY................................................................................................................................23 CHAINOFCUSTODYANDSAMPLEINTEGRITY .............................................................................................. 23 SAMPLEPREPARATIONFLOWSHEET......................................................................................................... 24 INSERTIONOFCONTROLMATERIALS........................................................................................................ 26 LOGICTABLEOFFAILURESANDTABLEOFFAILURES..................................................................................... 26 QUALITYCONTROL........................................................................................................................26 INTEGRITYOFSAMPLES......................................................................................................................... 27 CORERECOVERY ..................................................................................................................................28 CHAINOFCUSTODY............................................................................................................................. 28 VALIDATINGANALYTICALBATCHES.......................................................................................................... 28 QUALITYCONTROLFAILURES:REQUESTFORREPEATANALYSES..................................................................... 29 RECORDKEEPINGANDAUDITTRAIL......................................................................................................... 30 VERIFICATIONORCHECKASSAYS............................................................................................................. 30 ADDITIONALCONTROL/MONITORINGMEASURES....................................................................................... 31 DuplicateAssaysandPrecision................................................................................................... 32 ParticleSizeAnalysis................................................................................................................... 34 CONCLUSIONS...............................................................................................................................34 REFERENCES..................................................................................................................................35 APPENDIX1:GLOSSARYOFTERMS................................................................................................36 APPENDIX2:LABORATORYAUDITCHECKLIST..............................................................................41 APPENDIX3:EXAMPLELABORATORYCONTRACT..........................................................................46
LIST OF FIGURES Figure 1: Simplified flow chart of drill core sample processing process. ................................ 2 Figure 2: Frequency distribution of a sample from a normal population. .............................. 17 Figure 3: Example of a copper control chart for a pulp standard. .......................................... 17 Figure 4: Example of Blank Control Chart. ........................................................................... 20 Figure 5: Example of general safety chart providing guidelines for sample grain size reduction and acceptable sub-sampling masses (from ALX Chemex, 2000). ......................... 21 Figure 6: Example of Sieve test control chart. ....................................................................... 21 Figure 7: Methods for preventing (left) and detecting (detecting) tampering of samples (photo from Smee and Associates, 2002). ............................................................................... 24 Figure 8: Example of sample preparation flowsheet. ............................................................. 25 Figure 9: Recommended charts for check assays. A) XY plot with X=Y control line; B) Percent difference between check and original assays vs mean of the two assays (it is recommended to avoid confusion that the original assay always be subtracted from the check or new assay). .......................................................................................................................... 31 Figure 10: Example of absolute difference vs mean of duplicates plot, identifying possibly different ore type (from Smee and Associates, 2002) . ........................................................... 32 Figure 11: Example of Howard and Thompson Precision plot (from Smee and Associates, 2002)........................................................................................................................................ 33 Figure 12: Example of percent absolute difference between duplicate values vs the cummulative percentage. In this case 90% of the samples have a difference of 30% or less between the duplicate values (see blue line). Average Cu_pct values plotted in red for reference. ................................................................................................................................. 33
vi
QAQC Manual - 2007
LIST OF TABLES Table 1: Some issues when selecting an analytical method (from Beaudry, 2003). ................ 8 Table 2: Example sample log recording control sample insertion batch numbers and dates shipped and validated (from Beaudry, 2003). ......................................................................... 12 Table 3: Recommended control sample instertion frequency for stages I, II and III projects.22 Table 4: Example of Logic Table of Failures for a copper-gold porphyry project. ............... 27 Table 5: Example of Table of Failures for a copper-gold porphyry project. .......................... 27
vii
QAQC Manual - 2007
viii
QAQC Manual - 2007 INTRODUCTION In recent years regulatory authorities have become more demanding in the reporting requirements of mineral resources. In Canada national instrument NI43-101 is the guideline that prescribes the level of documentary evidence required to demonstrate the quality of analytical results. This manual describes the practical procedures that need to be followed to meet the standards of disclosure in NI43-101 and ensure the highest quality of analyses for the purposes of the project. In addition it provides some background information and some theory to help the practitioner resolve unforseen problems that may not be covered in this introductory text. QA/QC consists of two parts and just like any quality system can be divided into planning (quality assurance) and implementation (quality control) phases. The QA part comprises all the measures that need to be planned and materials acquired in advance of undertaking a drilling campaign and QC consists of all those verification procedures that are required during and after the drilling program. The QAQC manager is a Qualified Person (QP) as defined by NI43-101 and must have the following characteristics: - Must be an engineer or a geoscientist; - Must an individual, not a firm; - Must have at least five years of experience; - Must have experience relevant to the current project; and - Must be a member of a professional association. In Canada professional associations are recognized by statute and are self-regulating organizations that admit members primarily on the basis of academic qualifications and experience and require compliance with competence and ethics standards and have disciplinary powers to suspend or expel a member. Non-Canadian geologists can become members of anyone of several provincial associations and numerous associations outside of Canada are also recognized. In Brazil there are currently no professional associations that are recognized as such but this could change in the future. The QAQC manager is responsible for all the activities related to QAQC, including: Preparation and validation of standards and blanks or purchase of standards; Selection and validation of primary and secondary laboratories; Drafting of laboratory contract; Preparation of logic table of failures; Insertion of control materials; Supervision of core sampling and batch assembly; Lab audit(s); Validation of batches and uploading to DDH database; Corrective actions when necessary; Record keeping; Submission of samples for check assays at secondary lab; Submission of coarse crush replicates and core duplicates; Submission of coarse rejects for PSA analysis; Variance studies; QAQC Report.
The QAQC managers job starts long before the drill arrives on the property and ends long after it has left. Ultimately assays are just numbers until proven otherwise and it is up to the
Charles Beaudry, M.Sc., P.Geo., go
QAQC Manual - 2007 QAQC manager to establish that they represent some underlying physical reality. It is not up to later auditors to prove the assays were wrong.
TERMS OF REFERENCE This manual is primarily concerned with drill core samples since diamond drilling is the recommended sampling method for mineral deposits where sample recovery is acceptable. The procedures presented here will be adequate for early stage projects as well as for deposit delineation and definition and from grass roots exploration to feasibility studies and even currently producing mines. However the same or similar procedures can be applied to other types sampling (e.g. RC drilling, environmental sampling, etc.) or to samples with special requirements (e.g. indicator heavy minerals, highly nuggety gold, etc.) with some modifications as many of the procedures will be similar. The requirements of a QAQC program are different whether one is drilling a geophysical conductor and only trace or anomalous concentrations of metals or systematically drilling of a multi-million tonne mineral deposit. The concept of project stage is introduced to allow the complexity of QAQC to increase as the needs of the project increase from pre-discovery (Stage I) to discovery (Stage II) to delineation and definition (Stage III). In addition to the control of pay metals this manual also provides for early characterization of deleterious elements or complex mineralogy in order to provide the metallurgist with early warning of potential problems.
DESCRIPTION OF PROCESS Sampling drill core is done for the purpose of obtaining concentrations of certain elements or compounds. It is a process in the Six Sigma sense that it has inputs (drill core), outputs (assay results) and can be defined as number of sequential steps. An additional Six Sigma feature of the process is that it has a customer who is the sole arbitor of what constitutes an acceptable product. A simplified flow chart of the process looks like figure 1. Each one of these steps is subject to errors which are of three basic kinds, namely accuracy, precision and sample switching.
Split Core
Crush
Split & sub-sample
Pulverize
Sub-sample
Digest
Analyze
Figure 1: Simplified flow chart of drill core sample processing process.
The whole point of our QAQC exercise is to implement controls on the process in order that bias (accuracy error) is null, variance (precision error) is minimal and sample switching is detected and corrected. Sound simple? Its not. However application of simple procedures can go a long way in preventing the appearance of bias if for no other reason that batches that stray too far from the base line get re-analysed and essentially brought back to the mean.
QAQC Manual - 2007 DRILL CORE SAMPLING At the foundation of collecting core samples is good communication with the drilling company and crew. The drilling company should be informed of the strict requirement to collecting quality core samples prior to awarding of a contract. Geologists on site must communicate with and monitor the drill crew to ensure attention to acceptable practices is maintained at all times. This relates not only to the drilling and sampling, but also to health, safety and environmental issues. Prior to a drill program, materials that will interfere or contaminate the core (i.e. lubricant contamination of the samples) must be identified and avoided in the program.
Security Core must be secured from outside inspection and interference, or accidental internal interference. Chain of custody must be strictly maintained during transportation, sample collection, shipping and preparation to avoid tampering or inappropriate release of privileged information. Assay results must be kept confidential and only released to those on a need to know basis, at least until the results are published. Public release of results will only be conducted through a news release approved by head office. Project staff must be made aware of the need to maintain the confidentiality of both assay and drill results.
Procedures at the Drill Core boxes are labeled, and arrows drawn so that the core is systematically laid in the box; The core box is placed away from any source of contamination; A wooden or plastic marker is placed in the core box after each run and the meterage/footage is written on the marker; Transfer of the core from the core barrel to the box should be done as carefully as possible. No core is allowed to fall onto the ground. Core is directly placed in the core box and a plastic mallet is used to loosen core in the core tube. Breakage of core will produce inaccurate geotechnical measurements; Photographs of the core at the drill site will be beneficial for geotechnical analyses in a number of circumstances where transportation may result in degradation of the core. Document intervals of ground core and immediately address inaccuracies in depth labeling in the core boxes. A rod count must be conducted immediately to accurately measure the depth of the hole (this is at the drill company's cost if poor attention to labeling is indicated); Rod counts should be done at each bit change as a matter of course; If the core is being orientated, the core is systematically marked to indicate base of hole; As soon as a core box is full, a lid is properly secured so that no accidents occur; Poor quality or broken core boxes must be discarded.
Transportation Sample collection and transportation procedures may vary between projects. Differences are commonly related to access, climate, local infrastructure and region. Sample collection and transportation procedures must be documented and made available to field staff. These procedures are recommended to address the following issues where applicable:
Charles Beaudry, M.Sc., P.Geo., go 3
QAQC Manual - 2007 Transportation of core from the drill site to the logging facility must be in a manner to minimize or eliminate shifting of material in the core boxes; Transportation and storage of cut or split core must ensure that the remaining core does not shift and that marked sample intervals remain intact; Appropriate measures must be taken to eliminate the possibility of sample tampering through proper chain of custody management.
Core Sampling The QAQC manager is responsible for the handling of drill core at the core logging and sampling facility, as well as the sampling process. Accurate sample selection, collection and documentation are the responsibility of the geologist. The health, safety and environmental aspects of core cutting must be enforced such as protection for eyes, ears and hands. As well, in certain circumstances, air filtration pressure helmets and washable or throwaway paper overalls may be required. Cuttings that may be high in sulphides, and cutting oils must be disposed of in a proper disposal facility. Upon arrival, ensure core boxes are intact. Document all problems, Organize and open boxes with care Measure depth intervals in each box and document any lost core or depth inaccuracies. Immediately report depth inaccuracies to drill foreman and initiate a rod count, Tag boxes with metal tags listing hole name and the interval. Care should be taken with friable core or poorly indurated core or poorly bound mineralization (fractures, etc.), Geotechnical analyses for Stage II and III projects must be conducted before samples are selected. Photographs may be necessary. Physical property measurements can be conducted at the same time (i.e. magnetic susceptibility, conductivity), Align core to be sampled by matching broken pieces, A geologist is responsible for selecting material and intervals to be sampled, Intervals are selected on the basis of mineralogy (e.g. oxide vs. hypogene, pyrite vs. pyrrhotite, including significant grade variations), textures (grain size, banded vs. massive, disseminated, stringer, folded intervals) and concentrations of specific minor minerals, Sample intervals should not cross geological boundaries. Intervals should not exceed 2 meters for porphyry deposits in mineralized zones and 1 meter for massive sulphide deposits. Intervals should be marked with a grease pencil on the core, and a metal tag or other permanent tag fixed to the core box in case the pencil mark is lost in cutting/ splitting. The geologist marks the core within each interval with a line down the middle of the core to guide cutting/ splitting. In addition arrows pointing down core often help to ensure that the core is replaced in the box in the correct orientation. This line must be selected to reflect the best equal distribution of grade, mineralogy and texture in both halves of the core. A diamond saw should be used unless washing action, inherent in the sawing process, could degrade friable or highly fractured samples. For Stage I projects where volume of sampling may be small a screw splitter may be used. After cutting, both halves are washed and placed back in the core box and again aligned. Care must be taken with this according to the nature of the core and mineralization.
QAQC Manual - 2007 Prior to collecting half core samples, the geologist reviews the sample intervals with the sampler. The same half of the core should be systematically taken. Sample numbers to be used for quality control standards are not used for drill core samples. The sampler labels plastic bags with the appropriate sample numbers, saws the core along the marked line and places the drill core for the appropriate intervals in the sample bags. Every effort is made to select a representative sample of drill core with poor coherence. The designated geologist, sampler or QAQC manager prepares the quality control standards to be included with the sample shipment and completes the sample log.
Exceptions to the procedures recommended above may apply to specific projects. The recommended procedures can be customized for each project but it is necessary to establish a detailed procedure to avoid selection bias. In any event all procedures should be documented. Drill Core Duplicates The equivalent of field duplicates, at the drilling stage, is the collection of the second half of the drill core for analysis. Drill core duplicates are recommended for Stage II and III projects. Generally drill core duplicates are introduced for less than one in 100 samples and they should be carefully selected to represent different ore types, alteration styles and rock competency. At this rate however nearly 3,000 samples are require to obtain the minimum 30 samples for to statistically prove that bias is present. There is generally considerable reluctance to submit the second half of the drill core for analysis. Regulatory authorities require that one half of the core be preserved until the end of a feasibility study. Removal of the second half of the drill core eliminates its use as a reference or library sample, and the intersection cannot be used for metallurgical studies. The purpose of analyzing the second half of the drill core is to compare the result with that of the primary sample. This comparison will provide an understanding of variability introduced by selecting one half of the drill core versus the other. In essence core duplicates record the total variance of the system and corresponds to the zero distance variance on the variogram. Moreover and perhaps more importantly core duplicates will allow to determine if selection bias is occurring at the core sampling stage. The study of drill core duplicates may lead to the conclusion that alternative drilling or sampling techniques are required, such as larger diameter drill core. It is not effective to sample quarter-core, or one-third of the core, as this does not reflect the procedures that are to be tested. The decision to collect core duplicates is serious and often controversial. It is really only relevant to advanced projects and is only valid once enough samples have been collected (>30), in other words, there is no point in collecting just a few selected duplicates. It can be very useful when drilling deposits with erratic mineralization where differences between the core halves can be important and the possibility of selection bias is high. Remember however that the greater the variance between the core halves and the greater the selection bias will need to be for it to be detectable with statistical (hypothesis testing) tools.
QAQC Manual - 2007 SAMPLE PREPARATION Sample preparation, or at part of it may be done in the field but if conditions permit is often left to the laboratory. The decision to carry out full or partial sample preparation is generally based on infrastructure and volume of samples. For small projects the overhead of establishing a sample preparation facility is usually not worth the trouble. However for advanced projects in remote locations with large throughput of samples field sample preparation can significantly reduce transportation costs. The decision should be made by the project manager in conjunction with the QAQC manager after careful study of alternatives. Crushing 3.4.1. Crushing and Splitting
Samples should be crushed to achieve a minimum of 80% passing a 10 mesh screen (less than 2 mm) for base metals and >85% for gold and precious metals. A riffle splitter or, more rarely, a rotary splitter is used to select a sub-sample for pulverising. The amount of sample to be pulverised must be specified to the laboratory. Variations of crush size may be required on a project basis. The most important factor for controlling variance in coarse crush sub-sampling is particle size distribution. The protocol requires that at the start of a project 5% of coarse crush reject samples be submitted for particle size analysis (PSA) until about 30 samples are tested and thereafter about 1% of samples. A variety of crushing equipment is available commercially. The most commonly used is the Rhino jaw crusher and the roller crusher. However because of inherent occupational health and safety issues related to the roller crusher this type of equipment should not be used for routine assays. The Rhino crusher, which provides excellent productivity and minimal contamination, has historically been plagued by poor PSA performance, especially on the first pass where samples typically achieve about 75% -10mesh (2mm). Even after several passes the samples often do not attain 85% passing -10mesh. For base metals this is often acceptable but for gold and other precious metals a better crushing may be required. Some innovative solutions have been proposed such as a vribating sieve attached to the crusher permitting only the oversize fraction to be recrushed a second or even thirs time. However any use of seggregating equipment should be monitored and the samples carefully re-homogenized with a Jones splitter or on a rolling mat. Some commercial laboratories may be equipped with other types of crushers. It is up to the QAQC manager to ensure that the equipment that is to be used is both unbiased and introduces acceptable variance. Pulverizing Crushed material should be pulverized to achieve a minimum of 95% passing a 150 mesh screen (106 microns). Optimal particle size, for a particular style of mineralization, is determined by conducting studies of multiple splits of the coarse crush reject. Both the size of the sub-sample and the grinding time can be varied. The sample size and grind characteristics impact on equipment selection. In the event of excess variation these parameters must be changed.
QAQC Manual - 2007 The laboratory's procedures to clean pulverizer bowls between samples must be investigated. The use of silica sand cleaners before each sample is recommended to be specified in cases where samples are high in sulphides or clay content, to avoid sample cross-contamination. Commercial laboratories typically use vibratory ring pulverizers. The most common are the LM1 and LM2 and variants and copies thereof. The basic difference between the two is size as the LM2 can handle up to 1Kg whereas the LM1 can hold only 250g of material. For base metals the LM1 is generally fine but for gold the larger LM2 results in lower sub-sampling variance. A larger mill the LM5, which can contain up to 4 Kg was introduced on the argument that processing the total sample processing eliminates sub-sampling variance. Altghough undoubtedly true from the perspective of the physics of sub-sampling, nervertheless one must consider whether the larger mill actually produces as homogeneous a sample. In fact some experiments done in-house at Noranda indicated that the LM1 actually produced more homogenous pulps than the LM2 in the case of gold analyses. The LM5 can provide a good sample but care must be used to collect all the material from the bowl and properly split the material for sub-sampling. Because the LM5 is too heavy to manipulate manually in many cases the pulp is simply scooped from the bowl instead of being properly split with Jones or riffle splitter. Pulverizing specifications are generally 95% or more passing 150mesh. In practice this specification is easy to reach with LM1 and LM2 pulverizers. This level of performance however is more difficult to attain with the LM5 pulp which if used directly could lead to incomplete digestion. It is therefore recommended for flowsheets using the LM5 that a split be re-pulverized with a LM1 or LM2 mill or something similar. DIGESTIONS Once a sample has been pulverized to a stated size specification (>95% -150mesh) it is ready for digestion. Usually about 0.25 to 1.0 g of pulp material is weighed out to three decimal places and submitted to a digestion prior to analysis by atomic absorption (AA) or ICP-ES instruments. An experiment showed that there is no improvement in precision from the use of larger sample weights for digestion. In most cases a pulp weight of around 0.5 g is sufficient but sometimes as little as 0.1g may be used if the average grades are very high such as with VMS mineralization but a weight should not exceed 1.0 g except for fire assaying or bulk leaching of samples. Larger weights are generally not recommended as this increases digestion time and may increase matrix effects leading to bias. These recommendations are for the analysis of major components such as bases metals and do not extend to gold or other "trace" elements. In the case of gold, silver and PGE's a much larger sample is required, typically up to 50g in order to minimize variance and ensure a statistically representative number of metal-bearing grains are present in the charge. Numerous methods have been developed to dissolve constituents of a rock prior to instrumental analysis. The most common method however is by acid attack, especially for rock samples. In Canada and other temperate climates hot aqua regia, a combination of heated nitric acid and hydrochloric acid provides acceptable assay results, particularly when economic elements form sulfide or carbonate mineral species. In warmer climates such as Australia and South America and elsewhere it is generally reported that aqua regia digestion under reports the amount of metal in samples (negative bias). In these tropical environments a mixture of hot hydrofluoric acid with one or more of nitric, perchloric or hydrochloric acids are used. In this case glass equipment cannot be used and are usually replaced by special test tubes made of teflon as it may be heated up to 240C. Such multi-acid procedures are more expensive however and require special ventilation huts to prevent the accumulation of
QAQC Manual - 2007 explosive gases and precipitates. In addition certain elements such as As, Sb, and Ge are known to form volatile fluoride complexes which can lead to negative bias for these elements. Typical digestion times range from 1 to 3 hours depending on the type of material, the acid mix, and the size of the sample. The digest is usually heated to dryness or near dryness in a hot water bath (aqua regia) or directly over a hot plate (multi-acid) before being re-dissolved to constant volume in beakers (for geochemical analysis) or volumetric flasks (for assays). Depending on elements being analysed the digestion parameters can be modified. Lab personnel should be consulted and possibly some testing carried out before committing to a particular procedure. Whatever method that is finally selected should be rationalised and properly documented in the QAQC section of the project report.
ANALYSIS Assay vs Geochemical Analysis The request for analysis should include the method code or quotation number that will identify a specific analytical procedure. Analytical methods are selected to achieve acceptable precision for the anticipated grade range. Cost savings may be achieved by using multi-element techniques, however detection limits need to be carefully selected in order not to miss trace concentrations that may be key to further targeting of the exploration process. In general, assay determinations for commodity elements provide more precise data than geochemical determinations for sub-economic or economic ore grades. There is a continuum of procedures available that may not be clearly identified as being specifically assay or geochemical determinations. It is recommended that laboratories be asked to specify potentially interfering elements, elements that are volatilized (i.e. lost) during digestion or fusion, and minerals that might not be dissolved by the procedure. Specialised procedures may be required to characterise the ore, such as "acid soluble" techniques that preferentially dissolve copper present as oxides. The precision of these techniques is not typically at the same level as "total metal" assays but is used to assess the metallurgical performance of particular ore types in a deposit. Some of the technical issues to consider when selecting an analytical method are summarised in table 1.
Table 1: Some issues when selecting an analytical method (from Beaudry, 2003). Analytical Procedure Geochemical Analysis Assay Sample Weight 0.2 0.5 g 0.25 1.0 g Sample Dissolution Selective extractions HCl+ HNO3+ HClO4 HF Aqua regia digestion Alkaline fusion Elements Compromises used to achieve Optimized for single elements maximum no. of reported elements Dilutions Usually imprecise and performed Precise and performed in in test tubes volumetric flasks Upper Limit of Detection Precision is poor at upper limits No upper limit of detection of detection Lower Detection Limit Generally less than 1 ppm or Usually 0.01% for base metals 0.0001% for major elements Charles Beaudry, M.Sc., P.Geo., go 8
QAQC Manual - 2007

Instrumentation Specialized Methods A.A.S. , I.C.P.-O.E.S., I.C.P.M.S., neutron activation Includes analysis of water, biogeochemical samples, gases, MMI, Enzyme Leach, etc. A.A.S., I.C.P.-O.E.S., XRF, fire assay Colorimetry and gravimetric techniques may be used for high grade samples or concentrates
Analysis for Deleterious or Secondary Elements Minor or trace elements present within a mineral deposit may have positive or negative impact on the economics of the mining project. They may concentrate with the economic minerals in the concentrator and lead to either unsellable concentrate or the imposition of penalties by the smelter or can represent a potential environmental, safety, or health (ESH) issue during mining, processing, or tailings and waste disposal. Alternatively, secondary elements may be present in sufficient minor quantities to positively impact the NSR under favourable metallurgical conditions (secondary pay elements). In any case, the presence of these elements may have a profound positive or negative effect on the economics of a deposit and must therefore be thoroughly researched early in the project. In some cases these elements can be present in the bulk rock sample in very low concentration, perhaps even below the detection limits of many analytical techniques. However, under certain processing conditions they may be concentrated by a factor of ten or more in the final concentrate. For example, if selenium was solely present in the mineral sphalerite and its bulk concentration in the core sample was 0.02%, it could be concentrated a minimum of six times in the sphalerite concentrate where its concentration would be 0.12%. This would represent significant increased costs through recovery problems and treatment charges. The elements and their threshold concentrations that may impact on the economics of a mining project vary according to the minerals that contain the elements, the type of concentrate produced and to which smelter the concentrate is destined. The upper threshold limits for key elements are usually determined on a project basis by undertaking preliminary metallurgical tests. Once the elements are identified, a program is formalised to determine if those elements will be present in concentrates at levels that cause difficulty. The final assessment of a potential problem must be made through discussions with a metallurgist or a mineral processor with experience in the specific deposit type. There is a wide variation of possibilities for the occurrence of deleterious or potential secondary pay elements in the minerals being concentrated for refining. These elements may in fact be present only in waste minerals and would therefore report to the tails. As such the presence of a deleterious element in the bulk chemical analyses may have no effect on the recovery of a desired element. It is however important to know what is reporting to the tails for treatment and environmental reasons. Optical mineralogical and microprobe analyses will determine in which minerals potentially deleterious or secondary pay elements occur.
Documentation A brief description of the methods is included in a project report and detailed methods included in appendices. Where possible methods should be specified using the method codes. Most laboratories will supply detailed method descriptions if requested or they may be available from some laboratories' web sites. Documentation will be specific to each project. It may be necessary to change procedures during the course of a drilling program. The reasons for these changes and which samples are affected must be documented in detail.
QAQC Manual - 2007
CONTROL MATERIALS The choice of control materials is one of the most important decisions for project QAQC as the quality of the assays is be directly related to the quality and relevance of the control materials. On early stage projects the issue is not as critical and all that is required is to establish that the lab is providing accuracy analyses for elements that are at anomalous levels and in many cases a single standard may be sufficient. However with advanced projects several standards may be required and multiple elements may need to be controlled. The choice of standards in these cases is not simple and poor choice of standard can lead to problems in resource estimation. For blanks it is usually necessary to supply the material because you cant normally buy it. In some instances this can be a problem, especially for remote projects that has not had any previous drilling.
Purchase or Preparation? In general early stage projects will use either purchased standards or standards supplied by another project. There is generally no mineralized material in sufficient quantity to manufacture a standard in any case and the requirements of the project are modest. All that is required is a standard with anomalous or maybe near cut-off concentration for the principal sought after pay element. Once a discovery is made it is usually necessary to continue with purchased standards until enough reject samples are available of ore grade material to blend into standards. It may in fact be necessary to purchase additional standards during the interim period to provide accuracy control around the average grade and near the highest grades of the evolving discovery. Enough material should be purchased to bridge the period of time required to manufacture and carry out a round robin survey and to overlap the introduction of the new project standards. When manufacturing standards you should plan to make enough material to reach production. However because of shelf life issues or because the quantity required would be too great (often the case with gold deposits), it may be necessary to manufacture standards all through the delineation phase of the project. For example, for base deposits only 10 grams is sufficient for a standard pulp aliquot (this will include enough material for the original analysis and 1 or 2 repeats). For a 50gram Au fire assay the standard should be at least 200 grams, that is 20 times more than for a base metal standard. This means that for the same insertion rate you will go through the standard 20 times faster. To put it another way 50kg of a base metal standard will be sufficient to analyze 100,000 samples whereas the same gold standard will only analyze 5,000 samples. Larger pulp standards are possible but much more expensive because of the homogenization issues associated with large masses of powdered rock. Once the material for the control standards is received at a laboratory it is likely to take 4-6 weeks to complete the preparation and analytical stages. An approximate cost to prepare a 50 kilograms control standard is in the order of four to five thousand dollars per pulp standard, depending on the number of variables being determined. Typically, half the cost is related to submission of the material for determination of "acceptable" values and measurement of homogeneity. This becomes a significant expense and the time delay is not always manageable.
10
QAQC Manual - 2007 In certain cases, it may be difficult to prepare a control standard that has sufficient homogeneity, especially if the deposit being drilled contains erratic or nuggetty gold veins. It can be particularly difficult to prepare a control standard homogeneous with respect to gold or other metals that are distributed as nuggets or discrete grains. The insertion of control standards where metal concentrations cannot be anticipated reduces the effectiveness of a quality control program. It may be preferable to use purchased control standards in such cases. The principal drawback for the use of purchased control standards is that the mineralogy of the reference material may not match that of the samples which may lead to biased results. It is preferable in many circumstances to prepare control standards from locally available materials. This is particularly important for gold projects where the flux used for the fire assay may have to be adjusted according to concentration of sulphides, oxides, carbon, etc. Similarly, for copper-oxide, nickel-laterite and other projects, where a variety of specialized digestions are used to predict metal recovery, it is important to monitor the effectiveness of the digestions based on reference materials built from sub-economic and economic ores. A series of control standards at different grade ranges is required and possibly materials with different characteristics. Where necessary, control standards are recommended to be prepared that are representative of different ore types, and possibly oxide vs. sulphide mineralization. Ore types should not be mixed when selecting materials for control standards. It may be possible to use control standards prepared for other projects in the same region and geologists are encouraged to coordinate preparation of control standards with other interested parties. A total of 3 to 5 control standards are recommended; one for cutoff grade, one for average grade and one for high grade. Special ore types may also require standards. These control standards can usually be developed from drill core, outcrop, excess metallurgical samples or other sources. They must be carefully blended, crushed and pulverized, well homogenized, split and then submitted to at least 5 laboratories to determine the range of "acceptable" values. For certain analytical methods this last requirement is actually difficult to attain. Pulp standards that are rich in sulphides require special handling. The oxidation of these materials may alter the analytical results, particularly using hydrochloric and nitric acids for the determination of base metals. It is preferable to store these materials in vacuum-sealed bags and, if possible, in a nitrogen environment to maintain the stability of the sulphides. Control standards developed for a specific project must be strictly monitored with respect to a reasonable shelf life and/or the effects of oxidation or degradation over time. They should also be stored in a cool dry place to minimize degradation.
Insertion of Pulp Standards This procedure deals with the insertion of control standards or purchased reference materials to monitor laboratory performance for Stage III projects.
Materials Required Fully-prepared (crushed, pulverized, homogeneous and certified) control standards Small sample bags (approximately 10 x 16 cm) Sample log (see Note below)
11
QAQC Manual - 2007 Procedure Regularly spaced sample numbers ("-25", "-50", "-75" and "-100") are used for control standards. The number of control standards should reflect the size of the analytical batch used by the laboratory. For some projects, it may be preferable to randomly insert control standards. This requires detailed record keeping and careful organization. Bags labeled with these numbers are filled with 5-10 grams of one of the control standards and the sample tag is inserted in the bag. Approximately 150-200 grams is suitable to be submitted if gold fire assays are requested. Care must be taken to ensure that control standards are not contaminated when handled and, if necessary, packets of the control standards should be prepared by suppliers or in a laboratory environment. Record which control standard was put in each bag in the sample log or sample cards. Control standards are inserted in numerical order with the samples prior to shipping. Ensure that the laboratory analyses the samples in numerical order. In some situations control standards may be inserted after samples have been pulverized. Generally the QAQC manager will supervise insertion of control standards and the description of the procedures should be included in quality control reports. The control standards are used on a rotational basis, i.e. the same control standard is not inserted at "1025" as at "1050". Sample Log Complete the internal sample log that records a minimum of the following information (table 2): the sample number; drill hole number; meterage/footage; sample shipment number or batch number; the date samples were shipped; the date results were reported and/or laboratory certificate number; sample numbers assigned for blanks; sample numbers assigned for control standards and which control standard was inserted; sample numbers assigned for drill core duplicates, if used.
Table 2: Example sample log recording control sample insertion batch numbers and dates shipped and validated (from Beaudry, 2003).
Sample Number 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 Hole-ID From To Shipment Shipped Date Recd/Cert # Oct.12 A990343
DH-06 DH-06 DH-06 DH-06 DH-06 DH-06 DH-06 DH-06 DH-06 DH-06
100 101 103.4 104.87 105.5 BLANK 107 108 110 113
101 103.4 104.87 105.5 107 108 110 111 113.5
33BH
Oct.3
12
QAQC Manual - 2007

1011 1012 1013 1014 1015 1016 1017 DH-06 DH-06 DH-06 DH-06 DH-06 DH-06 DH-06 113.5 114 115.23 115.9 Control 116.8 117.33 114 115.23 115.9 116.8 STD05 BM&S 117.33 118.75
Blanks Preparation of Blanks "Blank" material is submitted with samples to the laboratory to monitor contamination caused when crushing or pulverizing equipment is not cleaned properly after mineralized samples are processed, or due to dust. Suitable material consists of an unmineralized rock type (barren drill core is best) ideally where the metal content is below detection limit, which is commonly the case with assays of base metals but with ICP-ES or ICP-MS this is rarely the case. The rock selected for the blank should as hard or harder that the regular samples so that the preparation equipment is thoroughly scoured. Laboratories are also expected to analyze barren quartz chips or silica sand that is used to clean sample preparation equipment. Laboratories will routinely include analytical blanks in sample batches. The "blanks" described in this section are submitted without the knowledge of the laboratory and are designed to monitor contamination throughout both sample preparation and analysis. The ideal blank cannot be distinguished visually from regular samples, something that rarely happens in practices. The mass of the blank material used as a sample should be comparable to the average mass of unknown samples. Materials Required Unmineralized quartzite or sandstone (for example) 20-litre pails Procedure A geologist locates a source of suitable material (typically from a quarry or unmineralized drill core). Determine how much material to collect by (a) dividing the total of number of samples by the frequency of blank insertion to determine the total number of blanks for the program and (b) multiplying by 2 kg (the approximate weight of material submitted). Store the blank material in pails so that it is ready for routine core sampling. Record a description of the material and its origin. Submit five 200-500 gm sub-samples of the "blank" material to the primary laboratory to confirm low metal values.
13
QAQC Manual - 2007 Blank material is not processed in advance of its insertion into sample batches. Individual pieces of rock should be no larger than 5 inches by 5 inches so that they are small enough to pass the hopper of the crusher. Insertion of Blanks A blank sample is inserted routinely into sample batches. When the sample is analyzed, the reported analytical values should be near the detection limit of the method. If the reported values are higher than expected, contamination of the samples during crushing, pulverizing or analysis may be indicated. Sample preparation procedures would be reviewed to isolate the cause of contamination and corrective action taken.
Materials Required Sample bags Sample tags "Blank" material Procedure Two to four sample tags in each group of 100 tags are reserved for blanks. Reserved numbers usually end in "-10" and "-60" or, better still, two numbers chosen at random from each series of 100 sample numbers. The insertion of blanks randomly is a more robust test of the laboratory but requires careful training and record keeping. Prior to moving samples to a sample preparation facility, coarse blank samples are added as follows: - a) label a plastic sample bag with the sample tag ending in the number "-10" or "-60" or the pre-selected random number. - b) insert the sample tag in the sample bag. - c) add an amount of "coarse blank" material to the bag that is similar to that submitted for samples. Sort all samples into consecutive numerical order. Submit the "blank" sample to the sample preparation facility with the samples.
Note: Plastic bags can be filled with blank material in advance, then labeled and the sample tag added when preparing samples for shipment. Where possible, barren drill core is submitted so that the laboratory cannot recognize the "blank" material.
Validation of Control Materials and Laboratories All control materials selected for the project must be validated with the primary and secondary laboratories selected for the project. The objective of the validation is to demonstrate that both labs can analyse the standards to an acceptable level of accuracy and precision. If using project standard in many cases both laboratories were included in the
14
QAQC Manual - 2007 round robin survey. However if more than a year has passed since the round robin was undertaken the primary and secondary laboratories need to be re-validated. The digestion and analytical method must be the same that will be used for the project. Validating Standards Submit at least 6 aliquots of each standard to both laboratories and compare the results with the original round robin if a project standard or to the published statistics if the standards were purchased. The results should be within the tolerances of the standard for the labs to be approved for the project. In some cases the standards themselves are found to be a problem with the chosen digestion and analytical method. In such cases a different analytical method or different standards may be necessary. Validating Blanks Blanks also need to be validated. It is surprising how common it is to discover well into a sampling program that the blanks are not really blank contain trace amounts of pay elements that occasionally attain warning levels. The method of validating blanks entails selecting 10 blank samples manifesting a range of visual features, describing the samples in detail and comparing the assay results of the blanks with the description of the samples. If anomalously high values of pay metals occur in some samples and these values can be related to some observable feature (e.g. hairline sulfide fractures or presence of veining, etc.) then it may be possible to establish a blank selection protocol that will reject such samples. On the other if all blanks return with below detection limit then no special selection criteria is necessary.
SELECTION OF LABORATORIES A laboratory is selected on the basis of logistics, price, quality, and availability of services. A laboratory visit is highly recommended, unless impossible, to assess the laboratory's production and analytical facilities. Refer to appendix 2 for a list of details to record during a laboratory visit. A laboratory visit may also include submission of a series of control standards to test the laboratory's performance. Canadian publicly listed companies (called issuers) are required to follow NI-43-101 guidelines published by the Canadian Securities Administrators and adopted by most if not all Canadian provincial securities commissions. It is recommended (but not required) that accredited laboratories be used but the type of accreditation has not been specified. Joint venture partners and certain regulators may also have specific requirements. Each project shall have a primary laboratory where all routine analyses will be performed and a secondary laboratory to carry out check assays and to undertake sieve tests if these are required.
Laboratory Communication and Service Contract It is preferable to negotiate a contract with a commercial laboratory that clearly defines the services required, reporting formats, quality control parameters, pricing, turnaround time and penalties in the case turnaround times are not respected.
QAQC Manual - 2007 Laboratories should be asked to report on the following: Analyses of the second quartz chip sample (i.e. cleaner) to be passed through the pulverizer equipment prior to each batch of samples being prepared. Some labs do not routinely analyze the pulverizer blanks. Laboratory pulp replicate results. Blank and control standard results. Lab blanks are often solution blanks. Results of coarse crush particle size analysis performed by the laboratory to monitor sample preparation quality. Results of participation in round robins.
The contract should also specify the number of significant digits to be reported. The rule is that one additional digit should be reported. For example, if results are significant to two decimal places, three digits should be reported to the right of the decimal place. Although this last digit is not significant, its presence will prevent rounding at the next higher digit. This will avoid rounding errors and reduce a certain type of defect in the QAQC database and will contribute to lower overall variance. Refer to appendix 3 for a summary of recommended guidelines for preparing analytical contracts.
SETTING LABORATORY PERFORMANCE SPECIFICATIONS Laboratories have tolerances for their control materials that are generally equal to the precision stated in the advertized lab method codes. The fundamental objective of laboratory performance monitoring and of batch validation is to verify that the laboratory is actually respecting its own tolerances. The problem is that the standards used may not be matched, tolerance-wise, to the precision of the analytical method. A standard may have poor homogeneity such that two standards deviations is much larger than the stated precision of the method. As a result you will often not know if the lab is performing as well as it could with such a standard. Conversely, the standard may be much more precise than the lab method. Two standard deviations will be much larger than the methods precision leading to frequent and unjustified failures.
Standards A pulp standard is defined by 2 numbers, its mean or accepted value and its standard deviation which defines it level of homogeneity. Certified reference materials (CRMs) are often defined by the dispersion of the accepted value, also called the 95% confidence interval of the mean. This last value is typicically very small because it is meant to provide the possible range of values of the true mean of the standard, not its actual behaviour in routine analysis. In any event CRMs are typically much more precise than routine laboratory analytical methods and can always be used to monitor assays. During routine analysis of pulp standards the analytical results will vary so that apprimately 95% of the values will range between the interval defined by the mean and two standard deviations (figure 2).
16
QAQC Manual - 2007

70
60
Mean
50
Frequency
1StandardDeviation
40
Circa68%ofDistribution
30
20
2StandardDeviations
10
Circa95%ofDistribution
0
2.87 2.58 2.29 1.99 1.70 1.41 1.12 0.83 0.54 0.25 0.04 0.34 0.63 0.92 1.21 1.50 1.79 2.08 2.37 2.67 2.96
Data
Figure 2: Frequency distribution of a sample from a normal population.
Nena-B, Cu assay by OG46 method

Cu-OG46 Result 7 Mean +2Std -2Std +3Std -3Std
6.5
5.5 Cu_pct Cu_pct
4.5
3.5
3
74 75 75 77 77 79 81 73 72 83 86 86 89 87 88 74 75 77 79 90 90 81 72 73 86 90 92 84 89 88 70 70 70 70 70 70 70 70 70 70 70 70 70 70 70 70 70 70 70 70 70 70 70 70 70 70 70 70 70 70 70 92
SequentialOrder of Analysis
Figure 3: Example of a copper control chart for a pulp standard.
SDO Number
Conventional usage for control charts plot the control lines for the mean or accepted value for the standard along with the lines for 2 and 3 standard deviations (figure 3). Since the standard is normally distributed the results should plot alternately on either side of the mean, 95% of the results should plot in the region between the 2 Std control lines, about 4% of the results should plot between 2 and 3 Std and fewer than 1% should plot beyond the 3 Std control lines. Distributions that do not conform to these predictions are often the result of out of control situations that lead to batch failures.
17
QAQC Manual - 2007 Recommended tolerances are such that 1 sample can plot betyond the 2 Std control line from time to time but that consecutive samples cannot or a failure is generated. Moreover any sample that plots beyond the 3 Std control line is also considered a failure. Since laboratory precision tolerances are generally specified in percentage of the accepted value it is necessary to first convert the standard deviation to relative standard deviation as follows: RSD% = Std/Mean*100 (1)
Twice the RSD% for the standard should be less than or equal to the tolerance of the analytical method. Experience shows that a 2RSD% that is much smaller than the lab tolerance leads to excessive frequency of failures and can require relaxing the tolerance (i.e. increase the value of RSD%). On the other hand if 2RSD% is much greater than the lab tolerance the standard cannot be properly used to monitor laboratory performance. Unfornately it is not possible to reduce the RSD% of a standard since its value (and by extension the value of Std) is a fundamental minimum. A frequently asked question is whether we should just set the value of 2RSD% equal to the lab tolerance (at least for values of 2RSD% that are smaller than the lab tolerance) and use that number. The reality is that labs often perform much better than advertized and that using a smaller value results in better overall analyses. Gratuitously reducing the inherent homogeneity of a standard is, in a sense, leaving money on the table. In practice however values of RSD% much lower 1.5% do lead to many unnecessary failures for tolerances of 5% (and RSD% much lower than 3.5% for tolerances of 10%) and is an irritant to laboratory management. Besides the lab contract only requires the laboratory to deliver its advertized precision and anything better is in a sense a bonus they are not required to deliver. A good practicioner will monitor the evolution of the control chart over time and know the fundamental limits of the standard and the lab tolerances in making decisions about what constitutes a failure. The important thing is to document all decisions. In addition to results plotting outside of permitted regions, failures can be caused by other conditions. During the time frame of a sampling program sudden shifts (which may be detectable only after several batches have been received and approved) in either the average or the dispersion of the standard can occur that may not represent a failure as defined above but clearly indicate something happened in the laboratory. As soon as such a condition is encountered the lab should be contacted to ascertain what happened. Over time it may be apparent that the average value of a standard may change very gradually, possibly in response to settling and seggregation, or to oxidation of sulfides. Such changes should be monitored and at some point is may become necessary to re-homogenize the standard, particularly if it is stored in bulk, or it may need to be submitted to a new round robin to determine its accepted value. Laboratories include reference samples as part of their own QAQC measures. These should, and often are, reported to the client on the final certificates. The QAQC manager should plot these standards on control charts using the certified values provided by the laboratory and the published tolerances for the analytical method. One final comment about standards and tolerances. Gold standards often have very poor precision, especially for standards near the lower cutoff grade. The value of 2RSD% may be considerably higher than the published tolerance of the method. The standard may still be used to monitor gross errors but the QAQC manager should highlight the limitations of the standard in the QAQC report.
18
QAQC Manual - 2007 Blanks Blanks are samples with very low concentration of the elements or compounds we wish to control and as mentioned previously are used to detect intersample contamination caused by poor cleaning of the sample preparation equipment. The blank used on a project will be of comparable hardness to the unknown samples and will contain values for elements which near to or below the lower detection limit for the analytical method. Blank control charts are constructed like standard control charts by plotting the value of the element on the Y axis against some sequential measure such as order or time of reception of results on the Z axis (figure 4). Because they are so low grade blanks have no meaningfull measure of average or dispersion and hence have no determinable lower or upper control limits. A single Warning Level is determined based on the nature of the mineralization and an estimate of a cutoff grade. It is up to the QAQC manager in discussion with the project manager and laboratory personnel to determine the defect threshold concentrations for the blanks. This may vary according to project, mineralization type, and the average grades encountered but should always be less than 1/20th of the anticipated cut-off grade. For example, if the estimated cutoff grade is 4% Zn, the warning level would be placed somewhere below 1,000 ppm. For another ore type with a cutoff of 0.3% Cu the warning level could be placed at 150ppm. The main point to remember is that the warning level is not cast in stone and can be modified if conditions require it. However, like standard RSD% the lower the value for the warning level and the cleaner the samples. Discuss your on-going results with the lab if you feel low level contamination is endemic.
Crushing and Pulverization Samples are crushed and pulverized to reduce grain size so as reduce digestion time and to ensure that sub-samples are representative. Variance of sub-sampling is a complex function that contains several factors including the amount of material and the size of the sub-sample (Gy, 1982 and figure 5). Laboratories will typically quote >70 or 75% passing 10mesh (2mm) and will charge more for a better specification such as >90% passing 10mesh. If the coarse reject is to be split before pulverization in a LM1 or LM2 it is best in most cases to require a more expensive crushing specification. If a LM5 is used then presumably the whole sample will be used, in which case the lower specification is acceptable.
19
QAQC Manual - 2007

0.7
0.6
0.5
0.4
Au_gpt Au_gpt
0.3
0.2
Warning Line
0.1
-0.1
Figure 4: Example of Blank Control Chart.
Sieve tests or particle size analyses (PSA) may be undertaken on both reject samples and on pulps but rejects generally show more problems with the specifications and unless an LM5 pulverizer is used, should be mandatory for advanced projects. If on the other an LM5 is being used then sieve tests should be mandated for pulps. Coarse rejects and pulps returned from the primary laboratory are used as the source of material for the tests. A random selection of samples is retrieved; about 5% initially until about 30 specimens are available and afterwards about 1% on a continuous basis until the end of the program. A higher frequency may be used if these are concerns about laboratory performance. The main thing is to realize that you must wait for samples to come back from the primary lab in order to send out the samples for testing in the secondary lab. In some cases this could be a few months, which can lead to many poor results initially if the lab is not up to standard. In any event the laboratory contract should include specifications for crushing and pulverization. Results are plotted on control charts just like for standards and blanks and a control line agreed to in the contract is plotted (figure 6). Any significant pattern of low performance should be immediately brought to the attention of the laboratory. Remember that poor PSA results do not lead to batch failure but do indicate potential performance problems in the laboratory. In the Six Sigma approach PSA is an X that can lead to poor Y performance and warrants careful monitoring even if it is not necessarily easy to control. It is up to the QAQC manager to stay on top of the issue and to signal any problems as soon as they arise.
71 36 71 36 71 37 71 38 71 42 71 43 71 47 71 48 70 65 71 50 70 66 70 66 70 67 70 68 70 69 70 70 70 76 70 72 70 73 70 75 70 77 70 78 70 80 70 83 70 85 70 84 70 91 70 91 70 88 70 94 70 94
SequentialOrder
SDO Number
20
QAQC Manual - 2007
Starter sample
Coarse crush sub-sampling Pulp sub-sampling
Figure 5: Example of general safety chart providing guidelines for sample grain size reduction and acceptable sub-sampling masses (from ALX Chemex, 2000).
100.0%
90.0%
% Passing Mesh Size %passing 10mesh size
80.0%
70.0%
60.0%
50.0%
40.0%
30.0%
5N 04 OR 5N 0 4 04 OR 6N 04 04 OR 6N 05 04 OR 7N 0 5 04 OR 7N 05 04 OR 8N 0 5 04 OR 8N 05 04 OR 9 04 NO 05 9N R 0 04 OR 5 9N 05 O B 05 R 0 0N 5B 05 OR 2N 0 5 05 OR 4N 05 05 OR 4N 0 5 05 OR 4N 05 05 OR 5N 05 05 OR 6N 0 5 05 OR 6N 05 05 OR 7N 05 05 OR 05 7NO 05 8N R 05 OR 05 8N 05 06 OR M 0N 05 06 OR M 1N 05 06 OR M 1N 05 06 OR M 4N 0 5 06 OR M 5N 05 O M R 06 0 6N 5M 06 OR 06 6NO 05 7N R 06 OR 05 7N 0 5 06 OR M 9N 05 06 OR M 9N 05 06 OR M 9N 0 5 06 OR M 9N 05 07 OR M 0N 05 07 OR M 1N 05 07 OR M 2N 05 07 OR M 2N 0 O 5M R 05 M
SequentialOrder
Figure 6: Example of Sieve test control chart.
QUALITY ASSURANCE PROGRAMS The objective of quality assurance on sampling programs is to provide some way of ensuring the results of the analyses meet the minimal requirements of the project. Projects that are at an early stage do not warrant very complex procedures and call for fewer controls. Conversely, projects at an advanced discovery stage where the assays will be used for estimating resources, require much more stringent procedures and more numerous controls to satisfy the requirements of the project.
04
21
QAQC Manual - 2007 Project Stages Aligned with the progessive requirements of projects it is recommended that a staged approach to QAQC be adopted (table 3). Stage I corresponds to early stage projects concerned mainly with drilling geophysical or geochemical anomalies. The essential objective of the QAQC at this stage is to ensure the batch-to-batch (relative) bias is constant and that absolute accuracy at anomalous to near cut-off grades is measured and acceptable. This can be done with as few as 1 standard and 1 blank for 50 unknown samples. A check assay program should be implemented for additional confidence. Geochemical multi-element ICP analysis may be used to estimate the metal content of drill core as a less expensive alternative to assaying. Significant mineralization will automatically be assayed as described for Stage II and III projects, as will any indication of gold, silver or platinum/palladium.
Table 3: Recommended control sample instertion frequency for stages I, II and III projects.
Stage I Standards Blanks 2% 2% Stage II 2% 2% 5% of mineralized samples >= 5% Optional >= 5% 5% or 1 in 20 5% or 1 in 20 Stage III
Pulp Check Assays Coarse Crush and/or Pulp Replicates Core Duplicates Reject or Pulp PSA
5% Optional Optional Optional
5 to 10% of mineralized samples >= 5% initially until 55 samples collected, afterwards 1% Optional, 2-5% initially until 55 samples collected, afterwards 1% >= 5% initially until about 30 samples show PSA is aceptable, afterwards 1%
Stage II is a transitional stage that is initiated upon first discovery of potentially economic mineralization. It is directed towards additional QC checks, as well as the beginning of the characterization of a potential ore body. It is acknowledged there will have limited sample availability, at least initially. The discovery of potential economic mineralization automatically implies that assay methods are to be used for all samples within a mineralized interval in place of geochemical analyses more typically used in Stage I. Geochemical multi-element analyses can be used to "characterize" mineralization for elements other than base and precious metals with the objective of identifying deleterious components. Geochemical and assay analyses should not however be mixed in calculating the composite grade. QC procedures involve the submission of control standards and blanks on a more frequent basis than Stage I. Replicate samples (pulp and coarse crush) are incorporated in the QC procedure. Drill core duplicates are optional. Characterization studies for deleterious elements, optical predictive metallurgical characterization studies, and geotechnical data collection are initiated. Microprobe work can compliment this type of study to estimate liberation parameters and to locate host minerals of deleterious elements. Stage III applies to projects that have advanced to resource delineation and definition. Thorough QAQC procedures are designed and implemented for the specific project. Control standards (multiple control samples specific to the style, type and grade of mineralization) and blanks are submitted with each sample batch. QC data are verified upon receipt. Pulp cross-checks and coarse crush and/or pulp replicates, depending on the sample preparation
QAQC Manual - 2007 flowsheet being used, are submitted, and drill core duplicates remain optional. Coarse crush particle size analyses are initiated. Crush and pulverization sizes are optimized. Chemical analysis must be compared with metallurgical evaluations that may be run on individual holes, bulk samples composed of combined holes and/or large scale bulk samples. Analysis of deleterious elements is systematic so results can be incorporated in bulk composites. At this stage the objective is to characterize the deposit on the basis of geology (geometry, host rock composition, etc.), chemistry (distribution of grade, density and deleterious elements) and metallurgy (variations in grain size, anticipated recovery and therefore NSR variation). This leads to a high confidence and a comprehensive understanding of the deposit prior to large capital expenditures. The approach needs to reflect the fact that mineral deposits are not generally homogeneous. This manual is meant to provide practical advice and general procedures that are valid for most situations. Exceptions may occur with gold or other precious metals or if a project has unusual sampling and analytical requirements. In such cases it is highly recommended to consult with sampling specialists who can design or adapt special sampling and sample processing protocols to the particular needs of the orebody.
Batch Assembly The QAQC manager is responsible for the assembly of batches for shipment to the primary and secondary laboratories. Insertion of control materials is done at pre-determined frequencies dependent on the project stage. If insertion is done manually the QAQC manager will clearly indicate in the sample ticket booklets which sample numbers correspond to which controls whether a standard, a blank or a core duplicate. If the project is using drill core logging software the control samples may be issued automatically. It is recommended that the batch size be adjusted to the number of client samples in a digestion rack. Although not always possible, having all the samples in a batch fit in one rack ensures that the quality of the results of your unknown samples are closely predicted by the results of the control samples. When controls and unknowns are in different racks this predictive relation is not as strong. Ask the laboratory how many client-samples they use in a single rack. The number will depend on the type of digestion and type of instrumentation. Chain of Custody and Sample Integrity It is essential that proof be obtained that the samples were not tampered with between the time of sampling and arrival at the laboratory. Depending on the project this may be a trivial issue or a very complicated one. All movement of samples from the drill to the lab must be audited and the chain of custody documented and wherever there is a possibility of tampering procedures must be implemented to ideally prevent, or at least detect any tampering. The chain of custody must be documented before the start of the project and any change to the protocol, even if only temporary, must be recorded. The best way to prevent tampering is to have the samples transported by only the most reliable employees or by reputable transportation services. If samples must be transported in unsecured fashion (eg. on a public bus) use of locked strong boxes is strongly recommended
QAQC Manual - 2007 (figure 7). Reliable transportation can reduce but not prevent tampering. It is possible however to detect the occurrence by using plastic seals. Although not completely tamperproof the security seal when combined with the use of opaque pouches provides a reliable method to transport samples.
Figure 7: Methods for preventing (left) and detecting (detecting) tampering of samples (photo from Smee and Associates, 2002). The laboratory contract should require laboratory personnel to undertake a sample integrity audit of all sample batches on arrival at the laboratory. Any missing samples or samples that have been damaged (e.g perforated bags) should be noted and communicated. This information should be recorded in the table of failures.
Sample Preparation Flowsheet The QAQC manager should prepare in discussion with the project manager and the primary laboratory a sample preparation flowsheet for each project (figure 8). The flowsheet can be prepared using MS Powerpoint or Open Office Impress or using specialty software like Viseo. The flowsheet represents the process map of the sample processing procedures and should be included in the QAQC report.
24
QAQC Manual - 2007

EL Morro Project Sample Preparation Process January 2002 5-7 kg Core Sample
Dry Sample at < = 700 C
Screen Test #10 Mesh (2.00mm) 1/20 Samples
Single Stage Jaw Crusher >95% #10 Mesh (< = 2.00 mm)
Quality Control
Quality Control
Clean with compressed air every sample & coarse quartz every 10 samples
Jones Style Riffler
Quality
Control
1.0 kg Coarse Crush (Labelled & stored for future compositing)
1.0 kg Coarse Crush
1 sample in 20 Duplicate 1.0 kg Coarse Crush Label Sample #A
2.0 - 5.0 kg Coarse Reject (Labelled & stored)
LM-2 Pulveriser >95% -150 mesh (~0.1006 mm) Screen Test #150 Mesh (~0.1006 mm) 1/20 Samples
Quality Control
Clean with quartz wash after every sample
Quality Control
Pulp Duplicate Upon Request Sample # DUP
750 grams Pulp Labelled & Stored
Quality Control Data Report
250 grams Pulp to Analytical Laboratory
ANALYTICAL LABORATORY
Figure 8: Example of sample preparation flowsheet.
The sample preparation flowsheet should be used for training purposes and exposed prominently in the samples preparation facilities. Any discussion about sampling procedures should refer to the flowsheet. Keep in mind that sample processing requirements may change during the evolution of a project from Stage I through to Stage III and the flowsheet must reflect the reality of the actual process. Hence flowsheets should be dated and versioned. Similar flowsheets or process maps can be prepared for other procedures in the project. Use brainstorming techniques to map out processes but be sure to validate your process maps before incorporating the flowsheets as standard operating procedures (SOPs).
25
QAQC Manual - 2007 Insertion of Control Materials Insert control standards and blanks as per the ratio determined for the project stage. Pack samples tightly into pails, boxes or sacs in numerical order. numbered according to the order that they were used. The containers are
Complete the Analytical Requisition Form to indicate which samples are included in the shipment, when samples were shipped, the total number of samples in the shipment and the analytical procedures. The laboratory should be advised when samples with high sulphide content or other special characteristics are in the shipment. Enclose a copy of the Analytical Requisition Form in the first container. A second copy of the form is retained in the files. This form may be submitted in person with the samples where high security is an issue. Fax or e-mail a copy of the Analytical Requisition Form to the laboratory, where possible.
Logic Table of Failures and Table of Failures The Logic Table of Failures is a tool that formalizes the logical criteria for approving or rejection a batch (table 4). A table should be built for each project on the basis of the essential pay elements and any critical deleterious elements that may be expected. The Logic Table of Failures is best constructed using brainstroming techniques in a team composed of the principal profesionnals involved on the project. Along with the Logic Table of Failures, prepared before the start of the sampling program, a Table of Failures is maintained and regularly updated to record any failures noted during routine batch validation (table 5). This table records in a compact fashion the identity of the batch, the control material that failed, the nature of the failure and any follow up actions taken to correct the failure. In some cases failures may not lead to batch re-assay. However it is up to the QAQC manager to decide if and what actions need to be taken to ensure the quality of analytical results are acceptable for the project. The Table of Failures can be included in regular reporting documents along with control charts in order for management to quickly identify any developing quality issues with the project.
QUALITY CONTROL Quality control measures are all actions undertaken to prevent out-of-control conditions and to detect and correct such conditions when they do occur in order to prevent the errors from being incorporated into the resource database.
26
QAQC Manual - 2007

Table 4: Example of Logic Table of Failures for a copper-gold porphyry project.
If one Cu standard fails between 2 and 3 SD, and no other failure occurs in the batch, the standard is accepted. If adjacent Cu standards fail between 2 and 3 SD in a single batch, the standards are classified as failures. If both a Cu standard and a Field Blank fail in a single batch, both are classified as failures. If a Cu standard fails beyond 3 SD, the standard is classified as a failure. If a Au standard fails between 2 and 3 SD and the Cu standard passes, the standard is accepted. If a Au standard fails beyond 3 SD, the standard is classified as a failure (unless close to the detection limit). If both Cu and Au fail beyond 2 SD, the standard is classified as a failure. If a Field Blank fails in Cu in a minor way, the analytical batch is examined for other QC failures in the same batch. If there are no other failures the laboratory is informed of potential contamination but no other action is necessary. If a Field Blank shows a significant failure in Cu, and other QC samples also fail in the same batch, the most likely cause is sample mis-ordering and appropriate action is taken to find the extent of the mis-ordering. If a Field Blank alone shows a significant failure in Cu, the surrounding batch is classified as a failure. No failures will be attributed to Mo, as the concentrations of the standards are too close to the detection limit to be useful.
Table 5: Example of Table of Failures for a copper-gold porphyry project.

Batch Number Nor-020203-01 B05001 BC302-46177.0 Nor-020203-01 B05002 BC302-46177.0 Nor-020203-04 B05003 BC302-46175.0 Nor-020205-03 B05004 BC302-464400 Date Feb-09-2002 Sample Number 306675 Sample Type Blank Pulp Std EM-1 Cause of Failure Cu value 0.007 is > 3S.D above expected value Ag = 0.63 ppm exp. value <50 ppm. 2ond sample from blank was 2.02ppm Ag. Subsequent Action Subsequent control standard & field blank were in spec for Cu Batch accepted. Cu, Mo, & Au values in Field Blank were in range. Re-run Ag on 306706 & 306708. Batch accepted. Conclusions ~
Feb-09-2002
306706
Field Blank
Feb-09-2002
305800
Std EM-2 Leached Zone Blank Pulp Std EM-1
Cu & Au values in field blank and Cu value 0.012 between blank pulp within specs. 2SD & 3SD Batch accepted. Cu value of 0.005 is > 3 Cu for adjacent standard & Au in both standard & blank pulp & field blank S.D. below expected within specs. Batch accepted. value Au = 0.02 ppm is > 3 S.D. below expected value Cu for adjacent standard & Au in both standard & blank pulp & field blank within specs. Batch accepted.
Feb-13-2002
305821
B05005 Nor-020218-01
Mar-08-2002
306040
Blank Pulp Std EM-1
Nor-020303-01 B05006 BC302-461400
Mar-09-2002
307677
Std EM-3 Enriched Zone
Cu for adjacent standard & Au in both Cu value of 1.147 is < 3 standard & blank pulp within specs. Blank pulp 0.03 in Cu. Possible S.D. >2 S.D. over contamination. Batch not accepted It expected value should be???
Integrity of Samples The first line of defence against biased results is the integrity of the samples as they move from the drill to the core logging and processing facility, to the laboratory and through sample
QAQC Manual - 2007 preparation. The QAQC manager must actively monitor all movement of samples and instruct all personnel to signal any problems or irregularities in any samples. Sample batches should be audited upon arrival at the laboratory. Any damaged or perforated sample bags should be identified and reported to the QAQC manager. Subsequent irregularities in assay results can be explained by problems during transportation but only if the damaged samples are identified.
Core Recovery Core recovery is a very important and often poorly controlled source of bias in assays. Because of measurement errors it is normal that recoveries are less than 100%. However some core loss may occur because of poor rock conditions or poor drilling technique. If the elemental concentration of the lossed interval is the same as that in the recovered portion of the sample then no bias will occur. Unfortunately this assumption is almost never justified. Core recoveries should be estimated by comparing the length of core samples against the drill meterage markers. These results should be plotted on Control Charts, one for each drill. A control line at 98% should be used as a lower recovery tolerance. Individual low recoveries are usually cause by cavities or soft and friable rock and generally do not warrant action. However any systematically low recoveries or sudden changes in the pattern of recoveries should trigger a review of drilling methods with the drilling company. In extreme case (e.g. lateritic deposits) it may be necessary to audit recoveries on a real-time continuous basis with company personnel at the drill site.
Chain of Custody The Chain of Custody operational plan must be regularly tested and any deviations from the protocol must be discussed and approved. Often problems arise because the original protocol cannot be implemented, perhaps temorarily, due to infrastructural breakdown (bridge washed away; scheduled flights cancelled, etc.). A back up plan is always advizeable. Any change in the transportation protocols should be clearly noted and discussed in relation to QAQC results. Validating Analytical Batches The validation of analytical batches is the last chance at preventing poor results from entering the resource database. Batches should be reviewed immediately upon reception and before the results are uploaded to the database or any interpretation of data is performed. This advice is particularly important for qualified persons (QP) working on projects funded by Junior mining companies or when the results will be issued as a news release. Failure to undertake batch validation before assay results are made public can lead to severe sanctions for the company and the QP, including temporary or permanent loss of his or her professional license. When the batch results arrive the first step is to identify the standards, blanks and duplicates and copy these results to appropriate tables for plotting of control charts. Once the control charts have been updated the QAQC manager applies the criteria of the Logic Table of Failures to judge whether any failures have occurred and if so, what actions must be taken.
QAQC Manual - 2007 Standard failures generally result in re-assaying of the batch in part or in whole from the left over pulp samples. Failures in blanks however generally require re-assaying from coarse reject material. Failure modes can be complex and must be anlyzed in detail for action is taken. Laboratories include reference standards in each batch of analyses. The laboratory contract should stipulate reporting requirements for the labs own standards and the QAQC manager should plot these standards on control charts using the labs own published method tolerance for control limits. Any standards that plot beyond the tolerances should be signaled immediately to the laboratory.
Quality Control Failures: Request for Repeat Analyses If results for blanks or control standards are outside predetermined acceptable limits, it is necessary to determine the appropriate action. The following items should be checked prior to contacting the laboratory: If the values for blanks are unacceptably high, determine if the previous samples are high grade as this may explain sample carry-over in preparation or analysis. Further discussion with the laboratory is required to isolate the cause; If all the values for a control standard are outside the acceptable limits check that the control standard was recorded correctly; To determine if a sample mix-up is likely, review the data for several samples before and after the questionable quality control sample to see if any results are similar to the expected values. If a sample mix-up is identified, request reanalysis of the suspect samples; If the value for a control standard is nominally outside the acceptable limits and all other quality control data for the sample batch are acceptable, no further action is necessary; If the values for several quality control standards are similarly biased high or low, even though close to acceptable values, repeat analyses are requested for all the samples (and control standards) bracketed by the questionable control standard results.
If it is determined that there is no immediate explanation for unacceptable quality control results, the following guidelines apply: If a sample mix-up is suspected or there is no explanation for unacceptable quality control results, request re-analysis of sample pulps that incorporate the suspect quality control sample(s). Repeat analyses should be requested for approximately 10 samples before and 10 samples after the suspect quality control sample(s). It is preferable (albeit not always practical) to request return of the pulps for the samples that are subsequently re-numbered and submitted with different quality control samples; If sample contamination is suspected based on poor results for the blank, request that a new pulp be prepared from the reject and re-analyzed. Sample crosscontamination is more likely at the pulverization stage than the crushing stage of sample preparation, but is also possible during analysis; When re-analyses are reported, compare these values against the original results. If results are similar no further action is required and it is acceptable to retain the original results in the database. The repeat analyses should be retained in a separate table. The Century Systems database will store all values but will automatically use the most recently entered value by default;
29
QAQC Manual - 2007 If repeat values are different than the original results, the laboratory must provide an explanation and state changes to their internal procedures that will prevent another occurrence. The original values are replaced with the corrected data; If quality control results are consistently unacceptable for a sample submission, the entire sample batch must be re-analyzed.
It is preferable to discuss the quality control requirements of a program with the laboratory providing services prior to the start of a program. Most major laboratories will provide repeat analyses, based on a quality control program, free of charge. A laboratory contract is a useful business tool which can define the conditions under which repeat analyses are free of charge or when charges may apply. Invoices for analytical services should not be paid until quality control data has been verified and any necessary repeat analyses completed. Record Keeping and Audit Trail It is very important that a clear audit trail be recorded that identifies which batches have failed and what actions were taken to correct the failures. During subsequent auditing of the resource database, the auditor will request see all QAQC reports and will want to review all failed batches and compare with the actual contents of the database. It is a good idea to maintain before and after control charts, one set including all the failed control results and a second with only the control samples for the batches captured in the database. Commercial drilling databases such as Acquire and DHLogger capture all data but allow manipulation of results from validaded batches only. Other advantages make these databases highly desirable for the management of advanced projects.
Verification or Check Assays Verification analyses should be carried out in the secondary laboratory. Pulps are generally preferred because of their lower inherent variability makes it easier to demonstrate bias. Check assays provide additional accuracy control on the primary laboratory and complement (but cannot replace) standards because check results cover the entire grade curve of the deposit whereas standards, although more precise, control only specific points on the curve. Approximately 5 to 10% of the sample pulps should be submitted for verification analysis at the beginning of a project for Stage III projects. Pulps should be returned and re-numbered. In general, 30 pairs should be accumulated prior to assessing the reproducibility of the results. If analytical reproducibility is determined to be adequate, the frequency of checks can be reduced but should continue through the end of the sampling program. For Stage II projects, less than 5% of mineralized samples can be submitted for verification analysis. Replicate analysis is generally only carried out for a Stage I project if significant mineralization has been intersected. It is very important to keep in mind that check assays only measure relative accuracy, i.e. the bias between the two laboratories. The absolute measure of accuracy can only be determined with standards. Therefore it is critical that the same standards used to monitor the primary laboratory be used in the secondary laboratory.
30
QAQC Manual - 2007 No blanks are generally included (although used blank pulps are sometimes used) and are not required if pulps are used. For check assays from coarse rejects you may want to include blanks. Batch results for check assays are validaded just like normal batches but the control samples are plotted on separate control charts. Once the batch has been validated the results are combined with the original analytical results from the primary laboratory and plotted on XY scatter plots (figure 9a). This plot should include a control line for equal values as a visual reference. Another excellent chart is the percent difference between the check and original assays [(check-original)/mean*100] plotted on the X axis vs the mean of the two assays on the Y axis (figure 9b). Samples that plot on the right of the zero difference line are higher in the check assay and conversely samples which plot on the left of the zero line are higher in the orginal assay. The big advantage of this chart is that one can quickly vizualize the bias for each part of the grade curve.
3 Central North South
X=Y
Cu_Orig_pct
2
0 0 1 2 3 4
Cu_New_pct
Original Assay > New Assay
3
A
Central North South
New Assay > Original Assay
Avg_Cu_pct
0 -60 -50 -40 -30 -20 -10 0 10 20 30 40 50 60
Diff_Cu_pct
Figure 9: Recommended charts for check assays. A) XY plot with X=Y control line; B) Percent difference between check and original assays vs mean of the two assays (it is recommended to avoid confusion that the original assay always be subtracted from the check or new assay).
Additional Control/Monitoring Measures The control measures presented above are sufficient to ensure assays are accurate and meet the requirements of the project. Other complementary measures can be undertaken, however, to improve the evidential basis of the quality interpretation. Duplicate assays of pulps, coarse rejects or of drill core provide, along with repeated standard measurements (and check assays) information to estimate the precision of the various sample preparation steps. In addition particle size analyses are useful to record and monitor the performance of the crushing and pulverization steps. Neither of these measures leads to batch failures but they can provide some proof that the sampling flowsheet is adequate for the project.
31
QAQC Manual - 2007 Duplicate Assays and Precision Duplicate assay results can be plotted on absolute difference vs mean of duplicates scatter plots (figure 10). These charts along with a review of sample descriptions can sometimes help to isolate difference ore types.
Figure 10: Example of absolute difference vs mean of duplicates plot, identifying possibly different ore type (from Smee and Associates, 2002) .
However the problem with this chart is that it is not possible to easily compare the significance of differences in absolute difference in terms of the implications on quality. But with enough pairs of duplicates it is possible to estimate the precision of the sub-sampling for any grade using the method developed by Howard and Thompson (1978) (figure 11). Unfortunately a minimum of 55 samples are required for the exercise but unless at least 100 pairs are available the results are not stable. The authors present in the same paper a different method when few pairs are available but it too suffers from large confidence intervals.
32
QAQC Manual - 2007
Figure 11: Example of Howard and Thompson Precision plot (from Smee and Associates, 2002).
Another method of estimating the variance of sub-sampling is to plot the percent absolute difference between the pairs vs the cummulative percentage of the pairs (figure 12) and report some baseline value such as the %difference for the 90th percentile. Unfortunately this graph does not show well the impact of low concentration on the cummulative %absolute difference which in figure 12 are mainly clustered on the left and right hand sides of the graph.
120 1.8 1.6 100 1.4 1.2 1 60 0.8 0.6 40 0.4 0.2 0 0 0 20 40 60 80 100 0.2
%Absolute difference Cu_pct
80
Average Cu_pct
%AbsDiff_Cu AvgCu
20
Cummulative %
Figure 12: Example of percent absolute difference between duplicate values vs the cummulative percentage. In this case 90% of the samples have a difference of 30% or less between the duplicate values (see blue line). Average Cu_pct values plotted in red for reference.
A log10 (or Box-Cox) transformed dataset could be interpreted using analysis of variance methods but because of non locatility of assays (i.e. different parts of the deposit may have different distributions) the estimates of the partitioning of variances can be inacurate. Nevertheless these methods are often used but are beyond the scope of this manual.
QAQC Manual - 2007
Particle Size Analysis A program of systematic sieve testing is recommended on all advanced projects. Sieve tests should be plotted as soon as they are received and any problems should be immediately communicated to the laboratory. In addition laboratories often undertake their own test sieving program as part of their own QAQC. If possible request this data from the laboratory and plot the results as a control chart. Remember however that the labs own QAQC data cannot replace the projects QAQC but are at best a complement to show the behaviour of the laboratorys own data compared to project controls.
CONCLUSIONS Delivering quality assays is the responsibility of the QAQC manager who is a Qualified Person as defined by Canadian Regulatory authorities. The objective of this manual is to provide the QAQC manager with a staged approach to QAQC in projects in order to adapt the procedures to the needs of projects. The reason being that early stage projects have simpler QAQC requirements than advanced projects where analytical results will be used in resource estimation. This manual presents concepts to help design quality assurace programs and details all the essential procedures required for effective monitoring and control of sampling, sample preparation and analysis to ensure the quality of results meet the minimum requirements of the project. By following these guidelines the QAQC manager will have the tools to detect and correct any failures that occur in routine analysis of drill core samples. The methods described here are primarily oriented towards diamond drilling but can easily be applied to other types of exploration sampling as the same principals apply.
34
QAQC Manual - 2007 REFERENCES
Beaudry, C., 2002. Quality in diamond drill assaying: reducing the variation. Six sigma Brown Belt project. Gy, P. M., 1982. Sampling of particulate materials, theory and practice. Amsterdam, 431 p. Elsevier,
Howarth, R.J. and Thompson, M., 1978. A new approach to the estimation of analytical precision: Jour Geochem. Explor., 9, 23-30. Smee and Associates Consulting Ltd, 2002. Quality Control in Mineral Exploration. Powerpoint presentation.
35
QAQC Manual - 2007 APPENDIX 1: GLOSSARY OF TERMS
ANALYSES A.A.S.: Atomic absorption spectroscopy is a single element, solution based technique. Geological samples must be dissolved prior to A.A.S. analysis. Sample solutions are compared against the quantity of light adsorbed for calibration solutions at a specific wavelength to determine elemental concentrations. Accuracy: The measure of how close the average value of a group of analyses is to the true value. Assays: Assays are distinguished from geochemical analysis by being more precise and designed for higher-grade material. In general, assays are assumed to represent "total" metal. Assays are used to quantify metal contents for ore reserve definition. The precision of assays is not dependent on the type of instrumentation used for the analysis. The precision of the method is dependent on the amount of sample used, the digestion (or fusion) technique used to dissolve the sample, dilution procedures, spectral interference and optimization of the instrumentation for specific grade ranges. Deleterious Elements: Some slements is in high enough concentration can cause problems in the smelter either through lower recoveries of pay metals or can environmental problems or even be health hazards. Smelters will reduce payments for concentrates with concentrations of certain elements beyond specified levels and hence these may have significant risk impact on project economics. Fire Assay: Fire assay is a traditional assay method for precious metals, specifically gold, silver, platinum and palladium. There are four steps consisting of fusion, cupellation, parting and weighing of the precious metal bead. A gravimetric finish includes weighing of the precious metal bead. To achieve lower detection limits and to determine platinum and palladium, the bead is dissolved and the solution analyzed by A.A.S. The traditional fire assay method includes collection of the precious metals by lead during the fusion. Geochemical Analyses: Geochemical analyses generally cost less than assays. The laboratory uses partial digestions and lower sample weights for geochemical analyses than for assays. In general, geochemical analyses are less precise than assays and have a lower upper detection limit. Geochemical analyses often provide multi-element data and compromises are made in order to report a wide range of elements. I.C.P.-O.E.S.: Inductively coupled plasma - optical emission spectroscopy is primarily used as a multi-element, solution based technique. Quantification is achieved with reference to multi-element aqueous solutions. Matrix effects can be more significant than for A.A.S. determinations but mathematical corrections can usually be applied to correct for interference and spectral overlap. I.C.P. has a greater dynamic range than A.A.S. meaning that a broader range of concentrations can be determined without requiring dilutions or calibration changes for higher concentration samples. I.C.P.-M.S.: Inductively coupled plasma - mass spectrometry is a multi-element, solution based technique. Mass spectra are measured, as opposed to emission spectra I.C.P.-O.E.S., which are less complicated so that there is less potential interference. I.C.P.-M.S. is primarily used for the determination of rare earth elements, low level elemental concentrations (in the ppb-range) and some isotopes.
QAQC Manual - 2007
QUALITY ASSURANCE Accepted Value: The accepted values refers to the mean value of analyses of a standard in a round robin or to the recommended value for a certified or purchased standard. Blanks: A material with negligible metal values that is used to monitor contamination during sample preparation or analysis. Certified Reference Materials ("CRM"): Certified reference materials ("CRM") are a special class of control samples that are high quality materials. CRM's have been subjected to rigorous international testing. CRM's are seeing wider use in exploration and development circles, however they are typically used in routine testing to develop analytical methods and calibrate equipment. Check Assays: Analyses undertaken in a secondary laboratory to provide a check against the primary laboratory. The principal objective of the check assays is to measure relative accuracy across the whole range of grades encountered in a deposit. Pulps standards provide tight control but only at certain specific concentrations. Between these concentrations the accuracy performance of the primary laboratory is interpolated. Coarse Crush Reject: In most circumstances, an entire drill core interval or rock sample is crushed. The crushed material is referred to as the "coarse crush". A sub-sample is usually removed for pulverizing. Any coarse crush material remaining is referred to as the coarse crush reject. The reject is stored for future reference and coarse crush particle size analyses. Control Chart: Plot of results of some measure along the Y axis generally against some chronological order of measurement along the X axis. Control lines for the mean and upper and lower tolerance limits are typically plotted for reference. Control Samples: Control samples are any material that is used to control or monitor the quality of the analytical results. This includes standards, blanks and duplicates. Drill Core Duplicate: The second half of the drill core may be submitted for preparation and analysis, which is referred to as a drill core, duplicate. This can give an estimate of the total variance of drill core sampling. Field Duplicates: A second sample collected at the same time, using the same sampling protocol as the primary sample, to measure sample representivity. Field duplicates are collected for rock samples, stream sediments, soils and other sample media. An analogous procedure for drill core programs is submission of the second half of the drill core for analysis. Logic Table of Failures: A table containing the formalized logic used to approve or fail analytical batches. This table is unique to each project. Lower cutoff: The lowest concentration of an element or compound in an orebody that can be mined at a profit. The definition can vary and is strictly only applicable to mining reserves or mining operations. However, the concept is very useful and is often applied in early stage projects, even if inacurately, to provide some economic baselines for comparisons. Particle Size Analysis (PSA): Measurement of the percentage of material passing a specified mesh size. Experimental work by Gy (1982) and Franois-Bongaron (1998a) show that this measure should be stated as the mesh size that is required for 95% of the sample to pass.
QAQC Manual - 2007 Normally however laboratories will quote or warrant somewhat less than 95% passing 2mm (10 mesh). For base metals this is usually not a problem and larger sieves can be used to obtain the parameters for experimental determination of sub-sampling sampling variance. For gold however PSAs often need to be monitored closely and special sampling procedures adopted if conventional crushing is not adequate. Pay Elements or Pay Metals: Those elements for any given deposit or project that are expected to concentrate and be sold to a smelter or refinery for a profit. Precision: Precision is related to thestandard deviation and is measure of the dispersion around the average value of a group of analyses. It is generally defined as twice the relative standard deviation (i.e. RSD% = standard deviation / average *100). Primary Laboratory: The primary or principal laboratory where samples are originally submitted for analysis. Primary Samples: The first sample collected at any stage of sample collection or sample preparation is arbitrarily referred to as the "primary" sample. Project Standards: Pulp standards prepared from project materials and submitted to a round robin survey to establish the accepted values of elements. Pulp: A finely-ground rock sample or fine-fraction of a soil or stream sediment that is usually only a portion of the original sample collected. In most cases, analysis is done on the pulp. Pulp Replicate: A sample pulp may be analyzed a second time to measure analytical precision. A sub-sample is removed from the original pulp for analysis that typically weighs 0.5 to 2.0 g for base metal determinations. Most commercial laboratories routinely perform pulp replicate determinations and these are also referred to as "laboratory duplicates". Alternatively, the pulp may be submitted to a secondary laboratory for analysis, which is referred to as a "cross-check" or check analysis. Approximately 10% of pulps will be submitted to the secondary laboratory on advanced projects. A comparison of check assays with the original assay results provides an additional measure of accuracy over the whole of the grade range of the deposit. QA (Quality Assurance): A management system that consists of all the planned procedures that must be followed in order to ensure measurements attain a certain level of quality or that a product or service will satisfy given needs. QC (Quality Control): Those actions that need to be executed to ensure that a given level of quality is attained in a management system. QAQC Manager: The QAQC manager is a Qualified Person and is responsible for the quality of analytical results and all other measurements on the exploration project.
Qualified Person or QP: The QP is a licensed professional with at least 5 years experience and some experience in the field of the project. Reject Replicate: A second split of the coarse reject may be submitted for pulverization and analysis, which is referred to as a reject replicate. The same preparation and analytical procedures are performed on the reject replicate as the primary sample.
38
QAQC Manual - 2007 Relative Standard Deviation (RSD%): Measure of dispersion of values around the mean for a normal or near- normal distribution. Derived dividing the standard deviation (see below) by the mean or accepted value and multiplying by 100. Also called the coefficient of variation. Secondary Laboratory: A laboratory selected for analysis of a selection of sample pulps in order to check the accuracy or bias of the primary laboratory's results. Selection Bias: Introducing bias in results by systematically selecting among a pair of samples (e.g. two core halves) the sample which is either better or less mineralized. The only way to prove there is selection bias is to emplement a systematic core duplicate sampling program. Standard Deviation (Std): Measure of dispersion of values around the mean for a normal or near- normal distribution. The value is located at the inflection points on either side of the curve of the frequency distribution. Table of Failures: Used in conjunction with the logic table of failures and is a table that lists all QC failures and actions undertaken to correct the failures. This table is inserted in the QAQC report. Unknown Samples: Those samples for which we seek to know the concentration of certain elements. The purpose of QAQC is to prove that analyses of unknown samples are accurate and precise. Variance: The square of the Standard Deviation. A measure of dispersion used in inferencial statistics. Variance is additive in a sequential process. Conversely the total variance of a system can be decomposed into its component parts. Foundation concept of analysis of variance or ANOVA. Warning Level: Concentration of pay metal above which the blank is suspected of habing been contaminated by the sample preparation equipment.
OTHER TERMS Geotechnical Data: Rock property data and physical property measurements such as rock quality description (RQD), fracture frequency, hardness (mineralization, gangue and host rock) and alteration on fracture surfaces. Predictive Metallurgical Analyses: Predictive metallurgical analyses allow for an early stage characterization of the metallurgy of significant mineralization with the objective of an early stage characterization of a potential ore body. Predictive metallurgy evolves into more advanced bulk and bench scale testing. Both stages rely on good sampling techniques, abundant sample availability and most importantly, a uniform drill density and spacing that is representative of the variable nature of most deposits. Samples are subjected to optical, chemical, microprobe and, with increased sample availability, bench testing. Optical analyses investigate the mineralogical features (minerals present, oxides vs. sulphides, etc.) and textures (grain size, grain to grain relationships, intergrowths, etc.) including the nature of the gangue (hardness, etc.) and host rock (dilution). Chemical analyses determine the presence of potentially deleterious elements. In combination with the optical investigations, the microprobe analyses determines the distribution of these elements in the minerals present and will give a good indication of where, if at all, the deleterious elements will report. Bench testing subjects a larger quantity of material to the milling and recovery process.
39
QAQC Manual - 2007
40
QAQC Manual - 2007
APPENDIX 2: LABORATORY AUDIT CHECK LIST
Sample Preparation Lab Name : Equipment required for audit: - Flashlight - Digital camera - Ruler/measuring tape - Protractor/compass - Notebook Building - Security - Safety - Location - Address - Access Manager - Name - Phone Sample Receiving and Sorting - Organization of work space - Work order book - LIMS? Sample Preparation - Space - Cleanliness/Dust level during visit Dust Control system - Description Drying Facility - Description of ovens, number - Heat source - Capacity - Description of drying procedures Crusher equipment - Types/Models and number - For each machine: - Condition - Sieve attachment present?
QAQC Manual - 2007
- Description of crushing procedures - Dust control - Type of reject bags Pulverizer equipment - Types/Models and number - For each machine: - Condition - Description of pulverizing procedures - Dust control - Type of pulp bag Cleaning material for crusher and pulverizer - Source - Scheduled usage Splitters - Types and number - For each splitter: - Overall length and width - Vane spacing - Angle of chute - Description of splitting procedure - Dust controls Air hoses and compressor - Air pressure
Preparation QC methods - Frequency of preparation duplicates - Crusher - Pulverizer - Frequency of PSA tests - Coarse rejects - Pulps - QC log book for preparation - Specifications for crusher - Specifications for pulverizer - Control charts plotted? Preparation blanks - Source - Frequency of analysis - Analysis of wash material?
Analytical Laboratory Lab Name :

QAQC Manual - 2007
Equipment required for audit: - Digital camera - Notebook - Pulp Standards to be submitted Building - Security - Safety - Location - Address - Access Manager - Name - Phone
Fire Assay Description of Facilities: Fusion Furnaces: Number: Size: Condition: Cupel Furnaces: Number: Size: Condition:
Source of litharge: Method of mixing: Method of order control: Method of inquart: Ventilation and Safety: Management of crucibles (reject levels): Wet Chemistry Distilled Water: Calibration/testing:
43
QAQC Manual - 2007
Hot Plates: Auto-pipettes: Make: Calibration: Glassware: Instruments: Makes and Models: Calibration Standards: Weighing Room: Balances (makes and models): Room isolated? Balance tables? Calibration: Frequency: Warehouse and Storage: Pulps: Rejects: Organization: Quality Control Methods: Standards: Frequency: Round Robin Limits: In house Standards: Preparation Duplicates: Frequency Analyzed: Pulp Duplicates: Frequency Analyzed: Blanks: Frequency Analyzed: Laboratory QC Manager: Usage of control charts: Reporting:
QAQC Manual - 2007
LIMS? Database? Fax/E-mail? Custom reporting formats? Personnel: Total: No trained technicians: No. shifts: Description of procedure used for samples:
45
QAQC Manual - 2007
APPENDIX 3: EXAMPLE LABORATORY CONTRACT
The objective of an analytical service agreement is to clearly articulate sample processing and analytical specifications and the required level of quality in analytical results in addition to price and turn around times. The agreement should include the following : Lab code for each analysis package. The maximum weight of samples before crushing overcharges are applied. The make and model of crushers to be used for the crushing of samples. Unless special sample preparation procedures are required all samples will be crushed to minimum 90% -10mesh. The coarse crush splitting procedure and type of splitter to be used. The size of coarse split. The make and model of pulverizers to be used for pulverization of samples. Unless special sample preparation procedures are required all samples will be pulverized to minimum 95% -150mesh. The pulverization procedure (i.e. time and additives). Frequency of laboratory preparation blanks. Size of aliquots for digestion and fire assays. Types and quantities of reagents used in digestion. Source and quantity of litharge and inquart used in fire assay fusion. Description of digestion and/or fusion procedures. Number of client samples and number and description of control samples in digestion and fusion rack. Make and model of analytical instruments. Source of calibration solutions. Frequency of solution blanks. Certified reference materials and in-house standards to be used with each lab package and the tolerances the lab uses for those standards. Units for each element in package and number of significant digits to be reported. The detection limits for each element. All analytical results for control samples should be reported in assay/analysis certificate and in digital copies of lab reports. The sample preparation and analytical specifications should be summarized in a flowsheet (one for each principal lab package) to be included as part of the laboratory service agreement. The logic table of failures should be discussed with the lab and the lab agree with the tolerances placed on project standards and blanks. Cost for each lab package and for other services. Maximum turn-around time and penalties for delays. Method of delivery of results. Instructions regarding return of pulps and rejects.
46

QAQC Protocols 2007

Uploaded by

Document Information

Original Description:

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

QAQC Protocols 2007

Uploaded by

Copyright:

Available Formats

DRILL CORE SAMPLING AND ANALYSIS PROTOCOLS (QAQC MANUAL)

Charles Beaudry, M.Sc., P.Geo., go 26 August, 2007

QAQC Manual - 2007

Charles Beaudry, M.Sc., P.geo., go

Charles Beaudry, M.Sc., P.geo., go

QAQC Manual - 2007

Charles Beaudry, M.Sc., P.geo., go

QAQC Manual - 2007 TABLE OF CONTENTS

Charles Beaudry, M.Sc., P.geo., go

QAQC Manual - 2007

Charles Beaudry, M.Sc., P.geo., go

QAQC Manual - 2007

Charles Beaudry, M.Sc., P.geo., go

QAQC Manual - 2007

Charles Beaudry, M.Sc., P.geo., go

Charles Beaudry, M.Sc., P.Geo., go

Split & sub-sample

Figure 1: Simplified flow chart of drill core sample processing process.

Charles Beaudry, M.Sc., P.Geo., go

Charles Beaudry, M.Sc., P.Geo., go

Charles Beaudry, M.Sc., P.Geo., go

Charles Beaudry, M.Sc., P.Geo., go

QAQC Manual - 2007

QAQC Manual - 2007

Charles Beaudry, M.Sc., P.Geo., go

Charles Beaudry, M.Sc., P.Geo., go

101 103.4 104.87 105.5 107 108 110 111 113.5

Charles Beaudry, M.Sc., P.Geo., go

QAQC Manual - 2007

Charles Beaudry, M.Sc., P.Geo., go

Charles Beaudry, M.Sc., P.Geo., go

Charles Beaudry, M.Sc., P.Geo., go

QAQC Manual - 2007

Figure 2: Frequency distribution of a sample from a normal population.

Nena-B, Cu assay by OG46 method

5.5 Cu_pct Cu_pct

Charles Beaudry, M.Sc., P.Geo., go

Charles Beaudry, M.Sc., P.Geo., go

Charles Beaudry, M.Sc., P.Geo., go

QAQC Manual - 2007

Figure 4: Example of Blank Control Chart.

Charles Beaudry, M.Sc., P.Geo., go

QAQC Manual - 2007

Coarse crush sub-sampling Pulp sub-sampling

% Passing Mesh Size %passing 10mesh size

Figure 6: Example of Sieve test control chart.

Charles Beaudry, M.Sc., P.Geo., go

5% Optional Optional Optional

Charles Beaudry, M.Sc., P.Geo., go

QAQC Manual - 2007

Screen Test #10 Mesh (2.00mm) 1/20 Samples

Jones Style Riffler

1.0 kg Coarse Crush (Labelled & stored for future compositing)

1.0 kg Coarse Crush

1 sample in 20 Duplicate 1.0 kg Coarse Crush Label Sample #A

2.0 - 5.0 kg Coarse Reject (Labelled & stored)

Clean with quartz wash after every sample

Pulp Duplicate Upon Request Sample # DUP

750 grams Pulp Labelled & Stored

Quality Control Data Report

250 grams Pulp to Analytical Laboratory

Figure 8: Example of sample preparation flowsheet.

Charles Beaudry, M.Sc., P.Geo., go