welcome it’s it’s my great pride and pleasure to to welcome you all here today Olympus we believe we’re leading the way in best practice application of this great technology portable XRF and I think this this workshops a manifestation of that huge thanks to all the speakers who are friends and who’ve agreed to present to you all today you’re going to hear some fantastic stories and really learn something I believe the other thing I would just say today we’re not talking about portable XRD today we’re only talking about portable XRF I think perhaps that we might like to consider doing a portable XRD workshop next year so if anyone’s interested in hearing more about portable XRD then grab one of us during the day and we can we can perhaps help you with that so first speaker is Dennis Dennis ionic get it out of the way Dennis Arnie he’s got over 30 years experience as a geologist specializing in geochemistry in a wide range of environments in both the minerals and petroleum indices as an academic and in geological surveys in recent years he’s been involved as a consultant in the development of exploration programs for precious and base metals exploration in Australia North America South America and Africa specifically this is included contributing to exploration targeting reviews and understanding gold and base metals mineral systems the design management interpretation of surficial geochemical surveys the interpretation of litho geochemical data preparation of qa/qc protocols reviews of geochemical data quality and contributions to ni 43-101 and joc Baumann technical reports a special area of interest is the incorporation of field portable analytical instrument instrumentation such as portable XRF inter exploration programs he’s published extensively in the areas of applied geochemistry economic geology and thermal history analysis for tectonic and petroleum exploration applications so I was asked to speak specifically about some things to think about if you’re looking to report portable Aref data publicly under international reporting codes so I’m only going to talk about two codes I’m gonna talk about 43-101 a little bit because I’m sure you’re all familiar with 43-101 those of you that report within the Canadian jurisdictions I’m gonna spend a bit more time talking about the Australian core code Joerg 2012 edition because I think it’s a little more specific and a bit more informative about some of the things you should be thinking about if you’re going to be reporting publicly portable XRF data so I’ll jump right in to jour to begin with York has three guiding principles in terms of public reporting of exploration and resource estimates and/or reserve calculations one is transparency that you need to be quite transparent in what you’ve done and and try to be non ambiguous about what you’re reporting materiality and we’ll talk a little bit about that that has in my mind it has two definitions one is is the information material to your company we wouldn’t expect Barrack to report Portable XRF data but if you’re a junior mining company on the venture exchange you’ve only got one project and you’re reporting portable XRF data it is quite material to to your company and finally this is probably the most important one is competence in the material that you’re presenting to the public so look I’ll just I’ll just reiterate materiality I won’t read all those words it’s too early in the morning to to hit you with a lot of text the key points are isn’t material to your company what you’re reporting the portable XRF data I’m going to show you an example where it was and have you provided sufficient information to allow a reader to make an informed decision about the data that you presented and and really the detail that you go into depends on how important that portable XRF data is to the overall report if it’s only a passing mention there’s no point in going into a lot of detail on calibrations and and all the analytical procedures that you followed so I’ve removed the name of the company and end the project to protect the guilty in this instance this is a press release that came out in November of 2013 on the Australian Stock Exchange and the company was reporting some drilling results as well as some aircraft results and you can see they got some quite

quite encouraging results but what really attracted people’s eyes was this reporting of gold data from portable XRF no really the only thing they said about the XRF work that they did was that they used in it–on and and then basically they disowned everything that they present so I should point out in November of 2013 you could sneak under the jerk 2012 guidelines for the for the up until the end of 2013 or December 2013 you could report under the previous jerk Islands that the jor 2012 code kicked in in December of 2013 so they were just sneaking in under the revisions to the code so this is a whoops moment down it may be hard to see the the printing here here is start off in November 2013 we’re going through – – January 2014 share price chart this is obviously a penny dreadfuls stock but you can see significant percentage movement in the stock here’s the the trades down on the bottom row and in millions of transactions at this point they released their portable XRF results on November 13th the the share price was already starting to ramp up a bit which often happens with your juniors and then it took a significant jump on release of those results these results for this particular good junior company were quite material to the share price and you can see they’ve they’ve reported copper silver and gold values well when people in Australia saw that who were familiar with portable XRF they jumped on top of it right away and there were complaints made to the Australian Securities Commission that that wasn’t a good thing to do oh and I should mention they also reported some elk crop data as well again here’s the goal values you can see how people would get excited about that as well as the copper iron here’s the interesting column in my mind they had a fair bit of zinc as well and the thing about zinc is there’s an overlap between one of the excitation shells with gold and in fact you can see where we have the highest sink there we’re getting the highest gold values as well and if I lauded it up and there’s a there’s a decent correlation there’s only five data points but hey I’m a chief chemist so that doesn’t bother me too much so here comes the retraction of the gold and silver values didn’t seem to bother the market because the share price kept increasing maybe there was a delay effect in there maybe people thought well there probably is going to be pretty good even if they’re not precise the numbers that they put out from the XRF and here’s where the company actually released the lab results and here’s what the lab results look like for the drill holes that they reported on and you can see coppers way off and in my mind that sounds forgivable because copper is generally quite robust on portable XRF but you can see they they’ve obviously presented uncalibrated copper data silver is actually not that bad but it’s it’s not precise and it’s certainly not accurate compared to the two lab results and no gold at all so really what they were picking up is the interference from the zinc in the samples giving them a false Gold anomaly and now resulted in a 20% drop in their share price so you can see depending on the company involved presenting these results publicly can be quite material to the company’s share price I’ve put put up the definition of a competent person under jour 2012 this is a another bee in my bonnet because it is subtly different from the definition of a qualified person under ni 43-101 what I’ll show you a bit later the key thing here well just draw your attention five years relevant experience specifically in the style of mineralization or type of deposit upon which you’re reporting and that’s a little bit different from what 43-101 has but also they specify that you have to have confidence in the activity upon which you’re reporting so if you’re reporting portable XRF data you should be competent in the use of the equipment and we saw from the previous example that the company there called a very competent person in that instance was obviously not competent or somebody forgot to to sign the release before the press release one out of transparency just to develop a little bit more on that thing you do need to present as much information about what’s gone on in the work that you’re reporting and in and you have to be careful not to omit important information that’s relevant to the to the public release so for example even something like are you presenting data from a calibrated or an

uncalibrated instrument is very important for people to know that previous example I showed you there was no indication of a whether they’d done any calibration but it was clear when they when they got the lab results well maybe not so clear we’ll talk about that in terms of sampling issues with portable XRF but it’s unlikely that the machine was calibrated so jorik 2012 if any of you have seen a 2012 jorik release instead of having a full technical report to support a resource estimation or any sort of publicly reported data they have something which is called table 1 which is a very efficient and still quite detailed but as an efficient way to get all the relevant information associated with the release of the data out with the release and these are attached to all press releases now and they do specify the use of handheld XRF instrumentation so you need to talk about the nature and the quality of the sampling that’s occurred and that applies of course to any NGO chemical or assay data that you’re presenting and then they infer there any geophysical tools spectrometers handheld XRF instruments you need to be presenting the parameters used during the analysis and I give a few examples I think I would I would add a few more on to that as well just for the sake of thoroughness so jorik does actually mention NEX RF there’s no mention of x RF in 43-101 so some of the other things that York says and this applies to any data that you’re going to present him there’s no reason why wouldn’t apply to portable x RF data you know as are the samples appropriate and how was it prepared before it was introduced to the machine is it representative of the material that’s being sampled is it representative of the in-ground material or the sample size is appropriate and have you done QC run QC samples and more importantly are you getting acceptable levels of accuracy and precision and this is what I like about the juror code is it doesn’t just tell you that you should be doing QA QC it’s telling you that you know you need to have an acceptable level of accuracy in precision so you have to then think about what those levels should be unfortunately people seldom report these levels of accuracy and precision I think people were more comfortable doing QA QC protocols and they are in terms of interpreting QA QC data ni 43-101 in spite of its reputation internationally as being a very prescribed and detailed code look fact a lot of Australians are a little bit scared of doing 43-101 reports but really it covers off some of the things covered in George but it’s a little more vague but certainly how to talk sound about sample preparation and quality control and relevant information about a sane and analytical procedures Quality Assurance actions and whether you have adequate confidence in the data and then finally to provide an opinion about whether you think the data is any good or not so not as much detail I was in jorik but the onus is still on the qualified person to make sure the data are good and then finally what what do you have to be in Canada to be a kewpie well you have to have that five years of experience doing something in the industry and this is this is where we have a difference with jorik I believe for quite a significant difference although it’s really just a subtle difference in the wording and then experience relevant to the subject matter so if you’re reporting portable XRF you should have some experience in portable XRF notice it’s decoupled from the amount of time so as long as you’ve been doing geology for five years and you have a bit of experience with portable R X or F you’re you you would be considered a kewpie under 43-101 so let’s talk about something a little bit because of course it doesn’t matter how well we calibrate the instrument or how long we run the samples or how much qa/qc we do if we don’t have a good sample we’re just wasting our time so I like to go through just a few examples about sampling then I’ll talk a little bit about instrumentation and qa/qc and then I think I’m gonna be out of time so I don’t expect you to see what the numbers are here but this is a typical assembling sampling chart of sample mass on the y-axis versus maximum grain size on the x-axis and you can see sampling Theory allows us to predict the size of sample that we would expect to have the lines here represent different types of grades for different styles of mineralization uh and this has been around for a while and so as geologists we know if we’re looking at something that has a nuggety distribution in drill core or outcrop we need to have a large

sample mass so really we can’t forget that just because we’re using a portable XRF and I think we often do and as I said if you have a poor sample then everything afterwards is is a wasted effort so that begs the question well how how big could a portable XRF sample be so I’ve done some back of the envelope calculations as to what kind of sample mass we might be analyzing when we do a portable XRF analysis and this is really either the critical depth of penetration the the effective depth into a sample that you’re going to be able to obtain the this is dependent on the atomic mass so this is going to vary depending on what elements you’re looking at and also the density of the material so I’ve picked a number for purposes of calibration which is probably a bit on the conservative side something like phosphate would probably be less than that but some of your heavier elements like gold it’s probably one or two orders of magnitude too low but you can see if you do a spawn analysis you’re basically looking at a square centimeter you’re really sampling less than 0.1 of a gram so it’s a very small sample and that’s important to remember so the way a lot of people get around this well let’s do let’s do a whole bunch of spawn analysis and take an average so if you do 10 well we’re up to about 0.15 grams in this case or we can do a drag of the analyzer across the core and take a continuous read so maybe getting up to about one and a half grams or you could take a fit what I call a fill at sample which is to use a essentially an angle grinder and capture the material that comes off it and and collect that and analyze that so you’ve got you’ve got grain size reduction and your sample mass is getting up so almost 30 grams in that case or you couldn’t even just use the cuttings from having split the core in half off a diamond saw in which case we’re getting up to appreciable types of sample masses and also we’re dealing with fine grain material but the bottom line is particularly for a low-grade deposit and and sampling theory dictates that the lower the grade the more mass we need to get a representative sample that just using spawn analysis alone is going to be it’s going to be very challenging to get a representative sample and just to put this into perspective if you’re doing a ICP analysis these days they typically the labs will use half a gram so and that’s after homogenization so really we’re about in here somewhere for a typical analysis but if you’re looking at something like soils or perhaps brought chips from a percussion drilling program or an RC drilling program the question you have to ask yourself is it even worth going through calibration or if you’re just looking for relative differences in your data could you get by without doing a calibration or even any sample prep so I’ll show you a result from a big project we did in 2011 but we analyzed something like fifteen thousand soil samples from the Yukon oh we didn’t have time nor the tolerance from from our employer to hold the samples up and do sample prep on tot on site so all we did was dry the samples and we analyzed them straight through the bag and it was quick and easy to do and then they went off they were going off to the lab anyway for analysis so we were going to put a lot of effort in the field to to getting the XRF analysis what we wanted was an immediate some immediate information to tell us how we were going with the analysis and if you’re doing that and if you’re if you’re going to be shooting samples food through the bag you need to have a little bit of an understanding of what’s sort of an attenuation you’re going to get from from the material so here’s an example in the blue this is a standard or as 45c analyzed directly and here are the results analyzed through the sample bags that we were collecting the samples in and then going to be shooting the the samples through and you can see we’ve got a bit of a attenuation there so you need to be aware of that but again it’s not a problem because we’re looking at relative difference from sample to sample at this point we’re not interested in in the absolute values so I’ve just picked picked one example out of a paper that we put together on this work that was published in geochemistry exploration environment and analysis on the left-hand side these are the icp-ms data that came back the sample was sieve 2 – 105 microns and then analysis was done on a 30 gram sample so quite an in my mind quite a good sample you know on your right hand side or the portable XRF data and we’re looking at arsenic in both cases and I challenge you and I know it’s early in the morning but I challenge you to really spot the

differences between those two percentile gridded images so I’ll use percentile images because I’m not so interested in absolute amounts I’m really I’m looking for where the high arsenic is and we’re seeing it in the XRF data we saw that we saw it in the XRF data within a week of collecting the sample and we had confirmation several months later when we finally got the lab analysis back 2011 was a busy season up in the Yukon and turnaround times weren’t that great but the absolute values aren’t too bad either if you look at the percentile breaks on the lab data versus percentile breaks on the XRF data they’re quite they’re quite comparable and really the only differences at the top end the very the very highest or most anomalous samples are a little bit different now I’ll move on to drill core and sampling and sampling drill core is a little bit more challenging what we did was I did a study with the the help of Kiska metals in Vancouver and we didn’t multiple spot analysis both on the so on side of the drill core and on the rounded side of the drill core note this is a measure of accuracy so this is bias relative to the assays we had for that core interval so you can see that the biases are different depending which surface we analyzed which was interesting but by the time we got up to about 16 or 17 analysis we were getting a reasonably accurate result but we could get to the same point by doing multiple drag analysis in this case we did three dragon alysus across each of the surfaces and plotted up that bias and you can see that’s pretty good you know within about plus or minus 10% of the actual assailant and the other aspect is the precision and as a measure of the precision I’m using the coefficient of variation or the relative standard deviation and again didn’t matter how many spot analysis we did it was very difficult to get precise readings from the portable XRF but pretty good with the with the dragon alysus so my conclusion is I wouldn’t have a problem reporting this information that the the handheld instrument was calibrated we had QA QC program and play we’ve assessed both the accuracy and the precision of the data so as a kewpie I wouldn’t have a problem reporting this these data the publicly now if we look at something that’s a little more nuggety in its distribution so a similar style of information I forgot to mention the previous example was from a poor free copper occurrence and in this case we have some coarser grain chocopie right in the sample doing the same thing the the saw on side the rounded side and again we had to go out to about 20 plus analysis now to get an accurate result comparative to what the assay was and you know we’re not too far off but we’re starting to see a little bit of bias and the portable XRF data from the dragon alysus if we look at precision we just never got anywhere close with precision even taking 20 analysis we’re still up with a relative standard deviation or coefficient of variation of on the order of a hundred and seventy percent I don’t think anybody would consider that acceptable precision in their data we’re a little bit better with the Dragon alysus so my conclusion I’d be very careful about publicly reporting this and then finally a different style of mineralization and also a different style of diagram I’ve got the coefficient of variation plotted in blue and you can see in this case we’re dealing with high angle veins moly veins so the the drillings been at a high angle to those veins and we’re looking at at the surfaces that cut those veins so very difficult to use a surface to represent three-dimensional veins and drill core but if we look at the coefficient of variation you know we’re not too bad in terms of precision doing multiple drag analysis in this case three dragon alysus but the accuracy is just not any good we’re we’re a hundred percent below what we should be compared to the assays so there’s no way you’d report this but you have to do the work you have to figure this out first and each project is a little bit different I’m not a great fan of boilerplate qa/qc and sampling protocols with the XRF I think you’ve got to do the work you’ve got to figure out what your accuracy and precision are going to be and then from that you did you derive your sampling protocol and in some cases you’re just not going to get decent data so it shouldn’t be reported it’s just that simple I’ll touch on a few analytical aspects because I think other people today will speak on this in more detail than I will but it’s important to select the right instrument so you you want to make sure you’ve got the correct

a node for the project that you’re gonna be working on have you set it up correctly for the elements that you’re going to be analyzing and have you got suitable detection limits for the elements of interest so what what you often see and this is going back a few years to be to be fair to the manufacturers but you often see these limits of detection of the lod s in this case comparing the old pin detectors against the the drift detectors just to show what the limits of detection from the manufacturer are but these of course are based on a pure silica matrix so they don’t include a lot of the interferences that you’re going to see in real geological material but you can actually work out your own limits of detection from your own data there’s a few tricks to getting the data out of the machine you have to set it up properly to obtain this information but you can you can actually work these out you can calculate them these were done on some soil samples and then you can also determine what percentage of your samples are above that limit of detection and you can see if only four percent of your sample population is above the detection limit for tungsten well it’s obviously not a suitable detection limit or a suitable instrument if you’re going to be analyzing tungsten but for copper were pretty good lead and zinc were pretty good antimony not very good moly possibly okay that’s the first pass but something that’s not really matched to all these different elements now have to spend a minute explaining this this is out of the chimera program or study that was portable x-ray handheld instruments so we’ve got five instruments up at the top three different elements that we’re looking at in in certified reference materials different counting times now 20 40 60 80 100 120 the the study was done to see how the count times required to get reasonable data might vary from instrument to instrument and also from element to element and you can see there’s there’s a lot of scatter but at least here you know once we’re past about 40 seconds we’re getting good precision you can see the data starting to tighten up so basically precision is a function of how many counts you collect it’s a function of the number of data you have to play with and in this case reasonably accurate you may not even have to calibrate to use arsenic on this instrument this is good too it’s not accurate out of the box from the factory but that’s okay the data are precise we can apply calibration and we can use calcium with a lot of confidence this instrument the data are all over the place they’re neither accurate or precise really you can’t do anything with these data now accuracy you need to monitor monitor using appropriate certified reference materials you want them to be match somewhat as close as you can get anyway to the material that you’re gonna be analyzing here’s an example for that program I showed you previously from the Yukon we were running or as 42p as our routine standard and you can see we’ve got beautiful precision and great accuracy in fact there we didn’t even do a calibration for the arsenic this was these were the arsenic data straight out of the factory for this instrument so so really good gave us all the confidence in those arsenic values and then it was validated one once we got the lab data back as well and you need to monitor this on a regular basis just like a lab needs to monitor it’s it’s icp-ms just to make sure there’s no instrumental drift over the life of the project and you can see in this case there really isn’t and then if you do a calibration you need to be looking at the XRF data plotted against the certified reference material data and you want to see a good correlation close to one you want a y-intercept near zero and you want a slope close to one and then you know that these data or this instrument is calibrated and ready to analyze comprar so in summary really the I like the approach from work in terms of an underlying philosophy work for reporting any data including portable XRF data the information needs to be material you have to be transparent in what you’re doing and you need to be competent representative sampling is critical if we don’t have a good sample then there’s no point in worrying about the calibration or count times or running an extensive QA QC program because you’re just fooling yourself this if the sample is not representative the material you’re sampling really it’s all a bit of a waste of time you have to make sure you have a suitable instrument I’ve run into a few companies that are using X RFS built specifically for gold analysis

but these are built specifically for gold analysis of jewelry not for rock material so they don’t include any Corrections or any matrix effects for geological materials so it it was an inappropriate machine for what they were trying to do and I couldn’t understand why the lab data didn’t match what they were getting off their XRF instrument if you want to report absolute values you’re going to have to calibrate but if you if you only want to show a map of relative values you may not have to calibrate or even do a lot of sample preparation but again you have to do enough work to give yourself confidence that it’s all working you need to have a count time that gives you sufficient precision and again you need to check that using some certified reference materials and and certainly if you’re going to be reporting data you need to have a QA QC program in place so that’s all I have to say just one last word I would like to thank Olympus and inner vex Canada for jumping into the breach when the course was cancelled by P DAC I appreciate the opportunity to be here today and to speak and I’m happy to answer any questions if we have any time left what’s the effect of moisture yes moisture will have an effect in that that was something that was a study I think in phase 2 of the chimera project basically you’re changing the the composition of the sample so you will reduce your values but people have been analyzing quite wet samples but again they’re calibrating and accounting for that so they’re measuring the moisture of the sample and then they can make a correction back to a dry analysis but it does have an effect and that’s why the one thing we did on that UConn case study was we dried those samples in a drying tent for two or three days we had to do that anyway because we’re gonna ship them out and it’s no point in shipping water out of the Yukon so what we did was just dry them and just before we were ready to put them in the the poly weave bags to put them on the plane to go out they spent a couple minutes on the portable XRF before they went out the door and and that that workflow worked pretty well but I would always dry because trying to measure the moisture content to make the corrections is it’s quite difficult in the field it would if you have sufficient heat you just have to we didn’t actually test everything we we sort of got them as dry as we needed to have them before we shipped the material off we did in that case because we were having an occasional rain as well we air dried them for a few days and then before they were going to be measured we brought them into we took over the the core tent and which had a heater in it and we put them in the core tent and they sat there for at least 24 hours before they’re put across the machine but again remember you you’re only going to be measuring the X rays on the outer surface of the sample so if the core doesn’t need to be particularly dry just the outer rim of the sample I’d say going in at least two centimeters should be drive before you analyze so from your comments I’d assume then at the start of a project or start of a new activity win the project say drilling you’d be extremely cautious about reporting anything until you have sufficient time to build the experience necessary to study your results from the XRF analyses to those of the lab a short answer yes yeah I think like any geochemical technique I’m a great believer in doing orientation studies yet you have to figure out your material your project you have to do some cross validation against laboratory data and and here’s but here’s where you also optimize so you can work out maybe you don’t have to count for 120 seconds maybe 4 if you’re interested in copper maybe 40 seconds is gonna be good enough and if you’re going to be doing several thousand analysis that that 80 seconds that you’ve saved on each analysis is translates into a lot of time saved so that’s that’s where you that’s where you can optimize everything you can figure out what’s going on and then once you’ve done that process you’ll know you’re getting you’re getting reliable numbers yeah okay I might be confident to to report those publicly but if you haven’t done that work you’re taking a chance to release results I might wind up asking this question of a couple people today I’m a mining engineer so I I haven’t seen a lot of these test programs in the field and particularly for core but you’ve mentioned the dragon Alice’s as as a method of getting it may be a more accurate or precise or whatever

measurement and you I think set up to around a meter potentially for that for that interval for that length of drag for somebody who’s perhaps interested in what happens at the sub meter level the heterogeneity or the variability of the sub meter level do you see a lot of variation and in core at that sub meter level and it might preclude perhaps the drag test but I’m interested in that heterogeneity yeah that I mean that would be one of the things that you’d want to figure out in your orientation work so I mean that that comes back to how you’re sampling the core anyway for if ultimately you’re going to be doing lab analysis on it so that they is what what should your sample interval be for this project and for this drill core if your if you’re sending it off to an assay what sort of sample length would you be looking to take originally so you could you could break up your core into sample intervals and and in fact you could do this very quickly and and before you start to send stuff off to a laboratory you could take it in in short intervals maybe in fact those those examples were about 30 or 40 centimeters take 30 40 centimeters in your core break it up and then compare results and and see if you do have that heterogeneity yeah I guess the the issue that we find is that when you make a sample of say a meter or 2 meters or whatever and send it off you’ve already smooth out everything that you want to know about the sub meter level so it’s just in general do you see heterogeneity that gets smoothed out as a result of passing oh you would yes certainly I think for example the the course shell co-pirate example there was a lot of heterogeneity in in that certainly the Molly veins a great example of very heterogeneous distribution of mineralization so yeah I mean it’s a great technique to investigate that headed unity in the core to get an understanding of how a variable might be over that that standard sample with that you might be collecting for and and plus you’re getting an instant result to it it’s a it’s a good question and actually what it was it was part of my question and I’m realizing now how simplified my question is gonna be on this but but roughly rule of thumbs what kind of what kind of times let’s say we know we’ve established a meters are you know is appropriate but how long would you do a scan yeah well I guess that is again that hit that protocol you got to work out but look doing it by hand isn’t isn’t the greatest way to do it and and there are instruments and I think you’re going to hear about one today where this whole process can be automated and controlled and but you have to the the thing is you set up a countin on your instrument and you say we’re gonna do we’re gonna go over a meters so you might set your count time to forty or sixty seconds and then really the trick is okay I’ve got to get from one end of the core over one end of the meter to the other end of the meter at about one minute so getting that rate of drag is quite tricky we just did it by trial and error and sometimes we’d get to the end of the meter and we still had 30 seconds left to count so we had to scrub that analysis and go back and adjust I think automating the approach that’s using robotics would allow you to control that and I think that’s one of the major advantages of using an automated system for for scanning core okay thank you