Tailings flocculation can be a problem hard to settle

An important parameter in any thickening application is the rate of liquor upflow (often referred to as the nominal thickener rise rate or velocity), determined by dividing the volumetric flow to the overflow by the available horizontal cross-sectional settling area.

Solids in a tailings slurry are roughly categorised by size and typical settling rates, as shown right. The coarsest phases will usually be well in excess of 40 µm with particle densities >2.5 g cm-3, and will typically have settling rates >10 m h-1 – in the majority of applications, these particles do not require flocculation. In terms of settling, silt can be taken as other non-clay main phases (e.g. oxides, sulphides) that may be in tailings. Some silt solids will have appreciable natural settling rates, but most will require aggregation for effective thickening.

While almost all clays have a ‘true’ or fully dispersed size at the low end of the range shown (i.e. in the colloidal range), they are rarely highly dispersed in tailings – if they were, they’d then require prohibitively high dosages and solids dilution. Clay type influences flocculation primarily through their propensity for dispersion and changes in shape (i.e. clays with a high aspect ratio will lead to more porous aggregate structures).

The colloidal range may also contain other phases that are of very high surface area and are not amenable to flocculation. Precipitates formed in the processing of effluent streams (e.g. hydroxides, hydrotalcites) are often colloidal and low density. Such solids coagulate in high ionic strength liquors, and these structures may then be flocculated. However, the flocculated aggregates will have very low density, with low settling rates even at large aggregate sizes that can form only at high dilution.

Thickeners over 80 m in diameter still exist, operating at throughputs that produce nominal rise rates <0.3 m h^-1. Under such conditions even fine clay tailings in moderate ionic strength liquors may settle without additional treatment. Such designs are now the exception, as the need for high throughputs was mostly met with the development of synthetic flocculants. Flocculated aggregates up to several millimetres in size can form from particles in the micron range, thereby raising settling rates by several orders of magnitude.

The problem with flocculation testwork

Small-scale batch testing of flocculants is common in process development and optimisation, but frequently gives results that are either irreproducible or inconsistent with full-scale performance. Many small-scale procedures originate from “jar tests” long applied in wastewater studies. However, these tests are designed for much lower solids concentration applications, primarily with coagulants, hence residence times under shear are longer and the sensitivity to shear is far less than in tailings applications. Coagulant solution properties are also easier to control than those of high molecular weight flocculants. Such issues are further heightened in the context of high solids/high dosage tailings applications. Even when cylinder tests are done carefully, it’s frequently under limited conditions that either miss or mask key aspects of the actual potential performance. CSIRO has developed tools to quantify flocculant performance under tightly controlled conditions that offers much greater insights into their response to key factors. Learn more …

Feedwell flocculation is not a simple process

While feedwells are currently expected to both dissipate the feed stream’s momentum and also provide conditions that are conducive to flocculation with high molecular weight polymers, it’s important to note the latter role is a comparatively recent addition, with thickeners and feedwells existing long before the advent of commercial polymer flocculants. Early thickeners had large diameters at low liquor rise rates, and so dissipating momentum in a feedwell was then enough to clarify suspension by utilising the natural setting of the solids; where slow settling fines were present, inorganic coagulants or natural products like starch

Why does it matter?

In many applications, the operational response to poor feedwell flocculation is to simply add more flocculant, with the added expense from higher dosages of a consumable often considered in the past to be an annoying but minor issue in the greater scheme of things. While that has changed in more recent times as sites have sought to minimise all possible costs, the potential impacts of excessive flocculant dosages downstream of the feedwell can actually be more of a concern. A poorly optimised feedwell can lead to dosages more than double that otherwise required, and this can favour bed networking that increases rake torque and limits the underflow densities achieved. Optimising feedwell flocculation can lead to incremental improvements in underflow density, which can be crucial in paste disposal applications and in counter-current decantation washing trains.

Even when flocculant dosages, underflow density or overflow clarity aren’t major operational concerns, optimised feedwell flocculation offers better unit stability and the potential for more responsive control, particularly in relation to shifts in mineralogy or feed flows. Long-term stability of thickener operation is highly desirable at many sites, where they’d much rather not have to think about tailings thickening at all. However, the need to increase throughputs well above design levels will almost inevitably arise in a site’s life, and feedwell optimisation can lead to a step-change in throughput without having to build additional thickeners.