Hi Brenton,

I wouldn't be too worried about that. Yes, it can be a problem, but it doesn't have to be.

Here is what I would do, and I have a fair bit of experience of hand-picking zircons: I would start by making an inventory of all types of zircon I have in a sample. That may take more time than the actual hand-picking and you'll become more experienced with time. Try to consider every aspect, such as shape, colour, inclusions, state of the edges, internal zonations, etc etc, everything you can think of and observe. Take abundant notes and as many photos as needed. Hopefully, after a fair bit of observation (it can take days), you'll have a few groups of zircons that exhibit the same characteristics. Well, that's almost job done, all you'll have to do after that is to pick the same number of zircons from each group. Place them in separate vials and analyze them separately. If you have grains that you can't place in a group, treat them as a separate group, such as "other".

This is all a bit simplistic, but that's all I can do remotely. Ideally, you'd be trained by someone with plenty of experience. One final piece of advise: make yourself very comfortable, as you'll spend long hours at the microscope and don't want to be hurting.

Now, if you want to count the numbers of zircons in each group, you'll have to make sure you've picked them all, which may take indeed some time.

Good luck,

Paul.

Dear Mr. Fairey,

at the beginning of your work you should precisely know what your study is all about. In your case, where source rock analysis and provenance analysis are paramount, the concern voiced in your question should be downgraded. It is to make a realistic family tree of mineral processing tailor-made and suitable for your study, this may involve grain size classification and separation. As you are going to hand-pick your grains I would suggest picking by morphology, color under polychromatic light and light having filtered out a special spectrum (e.g. UV lamp). After this first step you will have done most of that what you can achieve, as you have decided to take that approach. Be focused on your aims in terms of provenance analysis, have an overview of what you might expect and do not put the cart before the horse. This is important.

Good luck

Harald G. Dill

Hi Brenton,

You already have a few answers to your question. Here's another point to consider. You haven't mentioned how you have arrived at your final mineral separate. If you have used a magnetic separation step, it is best not to take this above 1.0A, as this could introduce (further) bias to your sample. If you have used heavy liquids you may have quite a pure mineral separate, consisting mostly of zircon. If this is that case, the simplest thing to do is to use a new plastic (disposable) transfer pipette and select several 100 zircons. You can pipette directly onto sticky tape to make a grain mount, so not only will this save you a lot of time picking, but it is more likely to give you a truly representative population. Other comments you have had in response to your question are also important (i.e. make sure you have studied the different morphology types etc). If there is not a unimodal size distribution, it is important to make separate grain mounts with grains of approximately equal size. Everyone has different techniques, but if you would like to see how I do this, you are welcome to visit me in Trinity College Dublin (Department of Geology).

Best of luck with your work,

Quentin.

Hi Brenton,

I use basically the same technique Quentin mentioned using pipette - you can either drop the zircons onto the tape or align them into rows for better orientation in your mount after polishing. Surely you should make sure you do not prefer any grains over the others when you do the alignment!

Some of the others use as few separation steps as possible - I have heard of separation technique where you actually load your crushed sample (without concentrating your heavies by any means) right into the heavy liquid (DIM) closed in a large centrifuge vial and let it centrifuge for some time - then you could be quite sure you got all your zircon populations of all sizes into the heavy fraction without much bias.

Even better seems to be to analyse the zircons from a polished block/thin-section. The automated software can help you finding all the zircon grains in such mount and by analysing all of them you avoid any extra bias besides those introduced by nature itself.

We performed quite a few experiments on this and concluded that the natural variations are usually in the same order of magnitude as the biases caused by the separation/picking technique. I do not myself sort out individual zircon population based on shape, color etc as this is not representative of the sample at all - if you have 0.5% population in your sample and get only 100 concordant analyses then you have obviously osmething like 50:50 chance that you do not detect this population at all. On the other hand there are specific applications where you want to search for certain grains (clear grains in sample with metamict population; euhedral grains to find out the magmatic age of volcanic rock that has plenty of inherited grains etc...).

I can imagine that reading all those theoretical/experimental papers makes a lot of confusion so you would better be off in the lab actually doing the thing rather than spending the precious time on listening the "good advices" ;-)

Hi Brenton, 

I share the different points of views of the precedent comments. It is true that you have to distinguish very carefully all types of zircons in your sample. You can take a look on the works on zircon morphology (Pupin and others...) because it cans be sometime not so evident under binoculars, and you can have a little idea about the original petrologies of zircons (althought this has to be used carefully). I did not know the technic with pipette, but it sounds to me a good option to not waste to much time. But depending of the technic of analyses, I would go for around 250 analyses. 

Good luck, 

Jérémie

I personally like the UV lamp approach. That helps me pick any type zircon I come across while picking without getting a certain number of each population, which would be a bias in itself if I am using the strength of the peaks to determine provenance.

The previous suggestions are all great!

Cheers

Vali

This paper may be of your interest.....

Thank you all for your very helpful contributions. Jiri, I was actually having a read G3, 2012 paper. It serves to show how debate is on going in terms of how many zircons one should analyse. Quentin, thank you very much for the invitation! Unfortunately the lab time has been booked and I have little time to do anything else but prepare my samples! For the UV, how is the apparatus set up? I am hoping you will say that its just a run-of-the-mill black light!

The one I have used was a handheld UV light that you could switch between long wave and short wave, nothing special as far as I know. I found one of the web that may work and I may actually buy myself. http://www.hometrainingtools.com/shortwave-ultraviolet-lamp/p/RM-SWLAMP/

The one I have used was about 12in long and skinny, but I am pretty sure it is not going to matter. One crucial part is that you will have to sit in a dark room to pick zircons with UV light. Hope this helps. Good luck!

Hi Brenton,

Just to throw an extra "de-biasing" thought into the mix. In general I agree with the others in that you should definitely think about what you are doing, but not get too stressed about it.

Meticulous 'pre pick' characterisation as suggested is important, and may help you find each age peak with good clarity. But if you are doing detrital and concerned about relative probabilities of populations then this may introduce a personal bias of course, resulting in all peaks being of a similar height. Relative probabilities of peaks can be very important if you are trying to determine  how similar different units are, or if they are equally distal/proximal to certain sources. These are of course produced by the "natural sample biases" (mentioned by Richard Albert) that are inherent in the sourcing of sediments. If it is important to reduce yours down to just these natural biases and produce the most accurate relative probabilities then the only way to do that is to throw a lot of zircon on the tape and hit a lot of grains - often far more than you really need (statistically) to find those smaller populations, as you have to overcome your analytical bias too i.e. hitting the grains you liked the look of. The only trip up of throwing lots on without looking is there might be a really fine grain size population you just polish away without realising, hopefully not!

(that was a bit longer than expected, on to the proper bit)

If you have to hand pick due to a poor separate for some reason, i.e. you've got

Hi Brenton,

as you see, there are many different ways to pick out a statistically valid number of zircon(types).Therefore I don't want to add that much.

The most important thing is to recognise the zircons that are not "typical", e.g. metamict grains of various colour (white, grey, orange..., most of them can only be found using reflected light, as they are not transparent) or uncommon morphotypes (e.g. soccer ball zircons from high grade metamorphic rocks etc.). Some grains do also have different optical characteristics resulting from tiny pits on their surfaces that originate from their sedimentary transport. Thus, they seem to be darker and so on. You just have to see them, but not to search them (--> statistics).

As you can see, there are many things to be aware of. But I am sure that you will pick mounts that are close to the (naturaly biassed) average of your samples when you get used to it.

Maybe you could pick a second mount of your first sample at the end of your work and compare both of them (maybe just optically). Than you may can estimate the quality of your first samples.

The hand picking of the detrital zircons for U/Pb dating is related  to their grain size. If the grains are medium  ( from  0.315 to 1.00 mm ), they can be hand picked under the binocular microscope. If they are  coarse to very coarse ( from 1.5 to more than 5 mm), they can hand picked under the Binocular loupe  based on zircon external morphological features. If the grains a very fine ( less than 0.315 mm), it would not be easy to hand pick. You can use density differentiation to separate the grains. 

I am currently working on detrital zircons ( trace element geochemistry and U/Pb dating): application  to mineral exploration and source rock location. I published a paper in 2012 (Resource Geology Journal), and new paper is going to be on line in this same journal in April 2015.

Dr. KANOUO

Dear Kanouo,

the typical zircon grain size as I experienced is ca. 100 micron (50-200 micron zircons make up 90% of my samples - both detrital and magmatic) so I can confirm that handpicking such grains is quite easy. Going down to 50 micron becomes more difficult I must admit.

There is considerable good advice above.  But any hand picking will introduce additional bias in any analysis.  If you want the analysis to be as un-biased as possible, make sure your mineral separates are clean, and this may mean refining the separation process.  Then mount everything, image the mount, and then do a random walk from zircon grain to zircon grain.  

Bear in mind that part of your approach should be directed to what you are tying to accomplish.   You may, for example, select euhedral grains because you are interested in the provenance of first-cycle grains only.  Or perhaps only the rounded grains that may be recycled are of interest.    As you note, the euhedral grains are relatively easy to identify, and but zircons get increasingly more difficult to identify if they are broken, frosted, and they have radiation damage.   As such, it is very easy to bias your data by just focussing on the euhedral grains.

It's so nice when people share analytical experience and advice about zircon separations that may save tons of time. I was always wondering whether heavy liquid separation or hand picking bias zircon populations in the study which is very important for a prov.analysis.

However, sometimes we need to bias our samples depending on our age target. So, the main question that you need to make is what your target of ages is? If you start entirely from scratch and there is no previous literature or you are not interested about a specific range of ages then you do not continue to read the below info. I will differ from the above suggestions by providing you few alternative ideas describing ways of focusing on specific ages instead of shooting to any zircon grain and run the destructive analysis of LA-ICPMS on your valuable zircon grains.

If there is an available lab separation facility or/and micro-Raman spectroscopy analysis then it would be a very good approach prior to CL and later geo- or thermo-chronology. Hand-picking is a first stage of filtering zircon grains by age (that is by color or shape). For example a white or yellow zircon is possibly younger than a black zircon. A micro-Raman will provide you with more information between the same color or the same shape as it will allow you to distinguish the distortion of the crystal lattice under higher definition than our eyes and subsequently to separate age ranges better than couple of identified colors or shapes.

Keep in mind that the uranium content always affect the amount of metamictization in specific time (distortion of crystal lattice, amount of alpha-particles, etc). However, a uniform uranium distribution among grains is not possible at all the times through detrital grains.

Selecting zircons of specific age range as micro-Raman may reveal the zircon lattice "health" and subsequently the zircon's possible range of age prior to any other geochronological dating. This may isolate the population of your interest (i.e. very old zircon grains will have 1006 or less of Raman wavenumber; I would avoid anything more than 1007 cm-1 to avoid young crystallization ages; see Marsellos & Garver, 2010 Amer.Mineralogist).

A good trick about using the non-expensive for analysis Micro-Raman is to holding all grains onto a double-sticky tape and map the zircon grains that belong to your age of interest (very old or very young). I think the number of zircon grains varies depending on the initial sample. Starting with 20 zircons and going up to 50 per sample, the number of peaks relative to your age-target of interest in your research will modify the total number of zircon grains which may go up to 120 or more grains. In one of my research, I had 4 grains out of approximately 80 grains of approximately 3.6-3.9 Ga. If this is your interest, then you will need an adequate number of zircon grains to build up a statistically significant peak in your age-histogram.

Good luck, and thank you for sharing your question.

Hi! According to my experience of performing detrital zircon geochronology, I never hand-picked the zircon grains. I mounted all the heaviest fraction of minerals and measured the zircon grains randomly according to their different shape, size, colour and CL response. The advantage of doing this is two-folds: 1) It saves time for hand-picking of ~200 grains per sample, 2) Biased is reduced as no picking is involved.