Dewatering: Thickening, Filtering, CCD, Water Treatment & Tailings Disposal

Dewatering: Thickening, Filtering, CCD, Water Treatment & Tailings Disposal

  • To participate in the 911Metallurgist Forums, be sure to JOINLOGIN
  • Use Add New Topic to ask a New Question/Discussion about Thickening, Filtering or Tailings and Water.
  • OR Select a Topic that Interests you.
  • Use Add Reply = to Reply/Participate in a Topic/Discussion (most frequent).
    Using Add Reply allows you to Attach Images or PDF files and provide a more complete input.
  • Use Add Comment = to comment on someone else’s Reply in an already active Topic/Discussion.

Surface complexation modelling (14 replies)

U
Unterstarm
8 years ago
Unterstarm 8 years ago

I'm interested to learn if anyone has applied surface complexation models simulating the adsorption of metal species onto mineral surfaces, within a mine drainage project.

Also if anyone has performed synchrotron-based experiments to resolve surface complexes on samples from mine-related projects then I would be very keen to hear from them.

Paul Morrow
8 years ago
Paul Morrow 8 years ago

Sure, the GWB software, e.g. SpecE8 and react, offers multiple possible models [form which the modeler must make an appropriate choice for her/his problem] for addressing sorption onto mineral layers, and this is possible also with PHREEQC. Both these packages also allow the modeler to write custom code, for example looking at sorption onto Mn oxides, not just Fe. The problem is not so much the availability of a calculation tool; it is the characterization of the materials about which one is actually concerned, specifically the actual scale of the waste-management units that exist at mine sites.

An example of a more sophisticated approach [using the code UNSATCHEM] , and the site-specific characterization needed to execute the evaluation, involving variably saturated reactive transport of As and both arsenate and arsenite sorption on carbonate-hosted ores, can be found on the two part sequence by Decker et al (2006) in Vadose Zone Journal, vol 5, p 419-429 and 430-444.

To my knowledge, which may be woefully deficient, application of synchrotron-based experiments remains a research-level undertaking. Although invaluable to our growing understanding of the underlying mechanisms of surface chemistry, the scale of measurement is very hard to apply to problems where there may be hundreds of millions of tonnes (or more) of mine waste that are distributed heterogeneously over hundreds (or more) of hectares, the waste accumulating over, often, many decades. Decker, for instance, undertook synchrotron-based measurements of arsenic distributions in his samples as part of his dissertation research, but found that he needed a different approach altogether to describing the observed behaviors of even relatively small (compared to waste rock or tailing accumulations) heap-leach facilities. Perhaps people of your generation will find ways to implement the emerging science in robust, mine-scale models.

(unknown)
8 years ago
(unknown) 8 years ago

I've used a number of these types of codes to calculate metal sorption in mine water environments, as have many others. My latest project has involved groundwater following adit closures using GWB....but I like PHREEQC and MINTEQA2 just fine. And I've used synchrotron XANES methods to study biologically reduced and sorbed Se in low concentrations in mine waste. Have a look at the papers by Brandy Stewart, Scott Fendorf, and others from Stanford on U and As respectively, for methods and ideas.

U
Unterstarm
8 years ago
Unterstarm 8 years ago

Thanks to you both for your detailed responses. I have been using PHREEQC as part of my research, but haven't used GWB before. It is interesting that you mentioned Mn oxides as I have been studying surface complexation of Pb and Zn onto Mn oxides, and relating these to an area of the UK impacted by historic mining. Here I have observed the distribution of Pb and Zn to be strongly associated with Mn oxides, and compared natural samples with synthetic lab analogues, using EXAFS spectroscopy. My aim is to apply the derived surface complexation models to model the impacted area and assess the metal-host stability.

Natural heterogeneity of the waste that you mention is important, and the limited scale of measurement (relative to a mine-scale!) will always be relevant, and an inherent challenge with this approach. However, given the limitations and inadequacies of sequential extraction methods, synchrotron-based methods could present an advancement, or complementary technique.

I am aiming to gauge from the geochemistry community how widespread the use of SCMs are, and how they are applied in actual projects, rather than demos. I'd like to know what the assumptions are for their use, their limitations, and what evidence is used to support/justify their inclusion. Also, what sorts of checks are made to validate the initial models and calibrate them if required? Finally, if anyone would be willing to share non-confidential PHREEQC input files involving applied SCMs that would be great. 

Paul Morrow
8 years ago
Paul Morrow 8 years ago

I see where you are going, and I am delighted to see someone try to establish a mechanistic model, based on actual substrate characterization, for the Mn system.

A couple of questions. Might one expect that the surface complexation capacity of a system is related to the biogeochemistry of the site-specific conditions under which there is neo-formation of Mn oxides (and other possibilities)? For example, might one expect the surface chemistry of actual materials to be the same in Derbyshire as in say classic Irish-type deposits, or volcanogenic massive sulfides at Mount Isa in the Australia desert or Faro in the Canadian sub-arctic? Perhaps your research will illuminate exactly such questions.

I think the answers are quite important to how one views modeling. If one is using models to assist heuristically in your understanding of what you observe in terms of mineralogy and aqueous chemistry, that is one thing. If modeling is to be used on a strong predictive basis, then one needs to understand how the parametric characterization being used relates to the site-specific development of the MnOx surfaces.

I ask because I would say that most exercises, certainly in the consulting field, use the default structure and data of the available models (i.e. implementations of Dzombak and Morel), assume the characterization of HFO by Dzombak.

U
Unterstarm
8 years ago
Unterstarm 8 years ago

Thanks for the reply and the interesting points raised. Microbiology is definitively significant, and some of the oxides that I have sampled are most likely biogenic in origin. Characterizing the mineral products from the myriad biota is a challenge, as is assessing the consistency of the mineral products, and their relative surface complexation capacities.

Testing the universality and demonstrating applicability of data for a Mn oxide system in one setting to other settings is work that I am starting, but there are many aspects for research. It's certainly a frontier of geochemistry that has much to be explored and especially relevant since surface complexation reactions have strong controls in contaminant transport. Whether it is possible for the information to be used beyond (site-specific) heuristic purposes is a salient question that I'm exploring, and I'm hoping to be able to test my results and make comparisons at other sites in future.

I have used the Dzombak and Morel HFO model in the past, and this model seems well established. I'd like to know how discriminating people are in selecting this approach, and if any sensitivity analysis is performed.

Paul Morrow
8 years ago
Paul Morrow 8 years ago

Can really speak only for our approach to uncertainty analysis. We would typically examine the sensitivity to ranges of

  • pH
  • Mass of HFO
  • Solution chemistry, at least for the trace metals of potential concern
  • Competition.

In some cases, we have looked at impacts of uncertainty in specific surface area. I strongly prefer doing these as deterministic sensitivity analyses, selecting the ranges from the empirical distributions of such data as we have or enhancing the empirical ranges with some arbitrary factors ("suppose there were twice the observed Zn and also twice the Ca and Mg, balanced by SO4..."). We could explore why I prefer the deterministic approach either off-line or in a different thread, as it is not related specifically to surface complexation.

(unknown)
8 years ago
(unknown) 8 years ago

Two quick follow up responses. The input values for the substrates are critical, and not easily come by....particularly for the Mn system. Empirically developed data are therefore highly valuable, and in my most recent experience, most useful in regulatory frameworks. But, you raise interesting and challenging questions. Can you incorporate all of the relevant variables in those empirical frameworks?

Next, in your sensitivity work, you might want to consider co-variance of key factors in your deterministic approach....if you find that there are some that are interdependent. For example, Fe and pH, when your objective is to assess metal sorption. 

Paul Morrow
8 years ago
Paul Morrow 8 years ago

It is exactly the existence of covariance that leads me to be cautious about the use of "probabilistic" methods in dealing with natural waters. Well, that and a belief that natural waters do not develop their multivariate compositions as a result of random processes (except for the gross alpha and gross beta concentrations).

Empirical adequacy of a model does not, of course, require that all information being used has to have been derived from de novo measurements at a site. We are all quite prepared to believe that g is a constant with a value that is knowable to adequate precision for our purposes. However, empirical data tied to the site in question are not just useful, but should be seen as essential to all stakeholders, Proponents as well as Regulators. We may have substantial latitude in how we demonstrate the empirical connection, but surely there must be a connection.

Victor Bergman
8 years ago
Victor Bergman 8 years ago

Regarding your last paragraph on the use of the default Dzomback and Morel I am pretty sure you mean using the fit Kint included in that volume. It's a great work in that it internally consistent, as is the companion volume by Karamalidis and Dzomback on Gibbsite surface complexation. With no detriment intended anywhere, I feel that it is well established more so due to availability than providing a great match to real systems. That said, it’s where I start when I am looking for Kint. I generally agree with everything that Mark L. has posted on this topic, and also thank you for exploring the Mn system.

As for as project work, I feel that SCM's are very valuable, and have been using them for Monitored Natural Attenuation - Inorganic remedial actions for some time now, a growing area of consulting. Again data quality, as you points out, is the simple stuff that affects a models ability to represent reality. Personally, I fell our next largest weakness is the difficulty of separating biotic from abiotic effects. In the big picture this detail is important, but to close a site, it may suffice to reducing biogeochemistry to a lumped kinetic term.

Another area of sensitivity and difficulty for field application is competitive complexation. Missing small concentrations of one competing species can have a bigger effect on SCM model fit than a mistaken assumption that all your iron is present as goethite, rather than HFO/ferrihydrite. Also, while you can find Kint for every nasty metal and metalloid out there on a broad variety of substrates, there is much less data on very common species that are known to compete for sites. You can find the sensitivity analysis I did on that topic here:http://www.ees.nmt.edu/outside/alumni/papers/2001d_miller_gp.pdf

O
Obersturmbann
8 years ago
Obersturmbann 8 years ago

We have derived generic SCM models for contaminant transport in mine waste using equilibrium leaching methods such as Pr EN 14429 or Pr EN 14997 and then fitting logKs using phreeqc COM in excel. In another project we combined BET surface area determinations, PZC titrations (to understand surface charge behavior) and selective extractions for HFO, with sorption isotherms and pH envelopes to refine the logKs for a "generic SCM" and then calibrated the parameter set for a fully coupled reactive transport model using column studies.

Unfortunately synchrotron based techniques tend to be seldom used in commercial projects due to the prohibitive cost and availability of beam time. A good understanding of the matrix mineralogy and potential sorption capacity and affinity can however be achieved through more conventional techniques (XRD, SEM, BET). Where polyatomic sorbates are of interest (oxyanions for example) ATR-FTIR can provide some surface sensitive information on the local bonding environment (i.e. bi-dentate v mono-dentate).

U
Unterstarm
8 years ago
Unterstarm 8 years ago

Thanks to all for their contributions (and encouragement) in this discussion so far – they have all been extremely thought provoking and helpful. Bridging the gap between theoretical understanding of surface complexation modeling and confident application is certainly a challenge. It's good to hear from practicing geochemists about some of the considerations that must be made in field applications of SCMs, and limitations to be aware of. Main key is demonstrating that there is sufficiency and quality of input data for both the natural system being modeled and for the SCM parameters. The role of competitive complexation of other species is also definitely critical and further supports the thermodynamic approach – plus the need for measuring a wide suite of analytes other than the ones specifically of concern.

One of the strands of evidence that I have in my work is from element mapping of the soils and sediment using SEM/EDX. Pb and Zn are observed to be invariably associated with Mn oxides. In addition, the EXAFS spectra that I have for both my synthetic and natural Mn oxide samples are consistent, and I’m going to be testing if the SCMs I have derived can account for the observed concentrations in pore water, and groundwater, or be combined with other modeling that incorporates factors such as dilution and equilibrium with other phases.

It’s good to hear that SCMs are being applied across industry (thanks for your link, and thanks for the insight into the use of SCMs for MNA, an aspect that I would like to look into further. Also, regarding As, my supervisor, Prof David Sherman has written several papers on surface complexation As on Fe oxides which you might be aware of – I can send links if not). Your work sounds really interesting, and I remember your poster at the mine water conference in Cardiff a while back. Perhaps you can send me some further details about the types of project you work on and the tests you do relating to SCMs.

It does seem to be the case that the applied users of SCMs are rarely able to make use of synchrotron facilities for various reasons, but it does seem like an area of geochemistry where collaboration would be very dynamic, rewarding and interesting!

Amar
8 years ago
Amar 8 years ago

Complex issue, touched on the subject in my thesis, see linkhttp://pure.ltu.se/portal/sv/publications/process-water-geochemistry-and-interactions-with-magnetite-at-the-kiirunavaara-iron-mine-northern-sweden%28571ce718-44ab-470d-ad4f-1c85ced55b58%29.html

U
Unterstarm
8 years ago
Unterstarm 8 years ago

I will take a good look at your thesis. Surface complexation reactions are undoubtedly important in process water chemistry.

Gruppen
8 years ago
Gruppen 8 years ago

Thanks for sharing your thesis, I will also take a look at it!

Please join and login to participate and leave a comment.