International Journal of Mineral Processing 97 (2010) 26–30

Contents lists available at ScienceDirect

International Journal of Mineral Processing

j ourna l homepage: www.e lsev ie r.com/ locate / i jm inpro

Dependency of liquid overflow rate upon humidity of a pneumatic foam

Ryan Shaw a, Geoffrey M. Evans a, Paul Stevenson b,⁎a Centre for Advanced Particle Processing, University of Newcastle, Callaghan, NSW 2308, Australiab Department of Chemical and Materials Engineering, University of Auckland, 20 Symonds Street, Auckland 1142, New Zealand

⁎ Corresponding author.E-mail address: p.stevenson@auckland.ac.nz (P. Stev

0301-7516/$ – see front matter © 2010 Elsevier B.V. Adoi:10.1016/j.minpro.2010.07.005

a b s t r a c t

a r t i c l e i n f o

Article history:Received 8 December 2009Received in revised form 23 June 2010Accepted 10 July 2010Available online 15 July 2010

Keywords:FlotationHumidityFroth

In this work, the liquid flowrate in an overflowing pneumatic foam is studied as a function of environmentalhumidity. It is found that ingeneral, liquidflowrate is greaterwhen the relativehumidity of the air above the foamis high. This observation is attributed to bubble coalescence on the surface of the foam. Because the evaporationrate is diminishedwhen the relative humidity is high, less bubbles collapse on the surface of the foam. Because noexperiments have been conducted on mineralised flotation froths, a dependency on the operation of frothflotation devices on environmental humidity cannot be directly asserted; although the effect is so important indemineralised foams that it should not be overlooked in the context of flotation. The authors know of nolaboratory or plant studies of froth flotation that have controlled or measured the environmental humidity.

enson).

ll rights reserved.

© 2010 Elsevier B.V. All rights reserved.

1. Introduction

Froth stability is well-known to influence the performance offlotation devices. In a model that attempted to describe manyelements of the physics that underpin froth flotation, Neethling andCilliers (2003) gave a method for predicting the rate at which bubblesburst at the surface of a flotation froth. Extending this treatment,Neethling and Cilliers (2008) give a deterministic method predictingbubble coalescence by writing a force balance on the films andassuming that there is a critical pressure above which films fail.However, this model was not experimentally proven in a quantitativemanner. Ireland (2009) has asserted that the coalescence process isprobabilistic, and that non-equilibrium rupture mechanisms areresponsible for coalescence in pneumatic foam. Crucially, for thepurposes of the current work, the effects of humidity uponcoalescence rate at the surface of the foam were not considered byNeethling and Cilliers (2003, 2008).

Because foam stability is so important to the performance offlotation, a device that is based upon the Bikerman (1938, 1973) testof “foamability” has been patented by Triffett and Cilliers (2004). Theyplace a transparent vertical tube into the pulp, and foam rises up thecolumn. By tracking the column height as a function of time, as well asthe steady height that the foam can attain in the column, knowledgeof intrinsic foam stability is obtained. The connection betweenflotation performance and froth stability, as well as how measuringthe evolution of froth height can illuminate these considerations hasbeen extensively studied by Barbian et al. (2003, 2005, 2006) and

Aktas et al. (2008) who concluded that “flotation performance can beattributed to changes in froth stability”.

However, Li et al. (2010) have shown that the rate at whichpneumatic foam rises in a column, as well as the steady height that itattains, is strongly dependent upon the humidity gradient within thefreeboard (i.e. the space between surface of the foam and the top ofthe column). If the humidity gradient is large (so as to promoteevaporation from the surface of the foam) then the stability of thefoam is low and vice versa.

The effect of temperature upon flotation performance has beenwell-studied. Lazarov et al. (1994) found that flotation kinetics areenhanced with respect to smaller particles with increasing temper-ature, but there was no apparent dependency with respect to largerparticles. They measured the temperature dependency of dynamiccontact angle and surface tension, and attributed the improvedflotation kinetics to differences in the times for three-phase contactline expansion and induction.

Choung et al. (2004) investigated the equilibrium of a non-overflowing pneumatic foam made from recycle process water froman oil sand flotation plant. They found that equilibrium height (andtherefore froth stability) decreases markedly with increasing tem-perature. They attributed this observation to the formation of aninsoluble precipitate of surfactant at high temperature, therebyreducing concentration. Whilst the explanations of their observationsgiven by Choung et al. (2004) and Lazarov et al. (1994) may well bevalid, we will investigate whether the rate of evaporation from thefree surface of the foam can have a fundamental effect on thehydrodynamic state.

The liquid flux in an overflowing pneumatic foam is, inter alia,governed by the rate at which bubbles burst on the surface (Stevenson,2007). If the bubbles on the foam surface are less stable, then the liquidflux is diminished as is the flux of gas–liquid interface; this must surely

27R. Shaw et al. / International Journal of Mineral Processing 97 (2010) 26–30

have an important impact upon flotation performance. In addition,increased temperature decreases liquid viscosity and this increasesliquid drainage rate in the foam, thus decreasing liquid flux. Bhatta-charya and Pascoe (2005) performed an extensive review upon theeffect of temperature upon the flotation of coal, and noted that viscositydecreased with increasing temperature.

The increasing temperature has another important effect that hasbeen overlooked: At constant humidity ratio (i.e. value of absolutehumidity), not only does increasing temperature increase theenthalpy of the liquid (thereby promoting evaporation), but it alsodecreases the relative humidity of the air (thereby increasing thedriving force for evaporative mass transfer). Relative humidity isdefined as the quotient of the observed partial pressure of watervapour and the saturated partial pressure of water vapour at the sametemperature and pressure. For example, by reference to a psychro-metric chart for the air–water system at one standard atmosphere wefind that at an absolute humidity of 0.006 kg/kg bone dry air therelative humidity is 78.5% at 10 °C (dry bulb) but only 13.0% at 40 °C.(In fact these calculations were performed using CYTSoft Psychro-metric Calculator 1.0 software.) Thus on a cool night, the air isrelatively close to being saturated with water and consequently thedriving force of evaporation is low. However on a hot day, the relativehumidity at the same value of absolute humidity becomes relativelylow, and this increases the driving force for mass transfer andtherefore evaporation rate.

In this study, we investigate how the liquid overflow rate (i.e.liquid flux) in a gas–liquid foam varies with the humidity at the top ofthe column. We will argue that, by changing the humidity of the air atthe top surface of the foam, whilst keeping the temperature constant,under some circumstances the liquid flux in the column can bemanipulated. This is because the evaporation rate from the surface ofthe foam, and therefore the stability, is dependent upon relativehumidity. We have not studied mineralised froths, and this is thesubject of current investigation. Of course, if particle coverage of thesurface is complete then there will be no evaporation at all. This workbuilds on that of Li et al. (2010), who showed that the behaviour of anon-overflowing pneumatic foam was dependent upon humidity

Fig. 1. Schematic diagram of

gradient in the freeboard, by investigating the behaviour of ademineralised overflowing foam as a function of air humidity. Thisin turn, will provide a basis for future investigations into whether thehydrodynamic condition of mineralised flotation froth is dependentupon environmental humidity.

2. Experimental method

A schematic representation of the experimental apparatus isshown in Fig. 1. The arrangement was similar to that employed byStevenson and Stevanov (2004). A Perspex column of internaldimensions 70 mm(d)×80 mm(w)×1150 mm(h) was utilised, theheight is that between the top of the column and the pulp/frothinterface which was maintained at a constant level by manipulatingthe rate of the peristaltic pump that feeds the bottom of the column.The system was isolated from the ambient environment by attachingan enclosed lid to the launder vessel so that the humidity of the airnext to the top surface of the foam can be manipulated, and thehumidity and temperature inside the closed launder was monitoredthrough the use of a digital thermometer/hygrometer (Lutron HT-3009). Although an opening remains in this box to collect theoverflowing foam, the separation from the external environment wasmaintained through the creation of a positive pressure environmentby flowing conditioned air through a rotameter followed by a separatevalve and out through this hose. Solutions of sodium dodecylsulphate(SDS) at concentrations of 8.33 mM (2.4 g/l), 4.17 mM (1.2 g/l) and2.09 mM (0.6 g/l) were prepared in batches of 20 l for eachexperiment; the critical micelle concentration of SDS is approximately2.4 g/l. Air was supplied at superficial velocities of approximately 9,12, 15, 18, 21 and 27 mm/s.

Humidity conditioning of the supplied environmental air wasachieved by the use of a drierite cell packed with Na2SO4(s) for dry air,and by bubbling through 2 sequential containers of saturated NaCl(aq).A constant flowrate of conditioned air to the top of the column wasmaintained. Prior to foam generation, the apparatus' internalenvironment was conditioned for 10 min with the appropriate airfor that test. The temperature of each experiment was also monitored

experimental apparatus.

Fig. 3. The data of Fig. 2 plotted on logarithmic–linear axes to highlight the lack ofdependency upon relative humidity at the low flowrates.

28 R. Shaw et al. / International Journal of Mineral Processing 97 (2010) 26–30

and found to be between 18 and 23 °C. Pneumatic foams were createdthrough the volumetric introduction of air at a measured flow rate,through a sintered glass frit. The foam was allowed to reach steadystate by allowing the generation of an overflowing foam to occur for30 min. Steady state was inferred if, after 30 min, 2×5min collectionswere observed to produce results which were in agreement. Inaddition, the system was monitored for maintenance of the relativehumidity and temperature, as well as air supply rate. The samples ofoverflowing foamwere collected for a period of 5 min each of 10 timesper experiment, and the flowrate averaged to give the superficialvelocity of the overflowing liquid. After each sample was collected,the foamwas returned to the sample reservoir to ensure that the totalconcentration of surfactant was not significantly changed during eachtest.

3. Results

3.1. Effect of humidity on liquid rate at relatively high surfactantconcentration

Fig. 2 shows liquid overflow rate (i.e. superficial velocity) as afunction of gas superficial velocity for three different relativehumidities (10, 40 and 80%) for a foam stabilised by SDS atapproximately the critical micelle concentration (8.33 mM). Aswould be expected, liquid superficial velocity increases monotonicallywith gas superficial velocity. At low gas superficial velocities very littledependence of liquid flowrate upon humidity is observed. Indeed,when the same data is plotted in Fig. 3 upon logarithmic–linear axesso that dependency at low rates can be ascertained, it is seen thatthere is no variation beyond the range of experimental readings; theerror bars signify the range of ten readings taken for each experiment.

However, at higher gas rates (N20 mm/s), and consequentlyhigher foam liquid fraction, a significant dependency of liquidflowrate upon humidity is observed. Liquid flowrate is greater whenthe air above the foam surface is of relatively high relative humidity(80%) than when the humidity is low (10%). We hypothesise that thisis because the bubble coalescence at the surface of the foam is greaterwhen the air above it is drier, and this is because evaporation ratesfrom the surface are greater.

3.2. Effect of humidity on liquid rate at moderate surfactantconcentration

For a foam stabilised by 4.17 mM SDS (i.e. approximately one-halfof the critical micelle concentration) a similar dependency to thatobserved at the highest concentration of liquid flowrate upon gas rateand humidity is seen (Fig. 4). Only at the highest gas superficial

Fig. 2. Superficial liquid velocity versus superficial liquid velocity for a foam stabilisedby 8.33 mM SDS. Error bars represent the range of observed experimental results. Eachdata point is the mean of 10 values. The legend denotes the relative humidity.

velocity (27 mm s−1) is a humidity dependence observed, althoughthis is still significant.

3.3. Effect of humidity on liquid rate at relatively low surfactantconcentration

Whereas for the higher concentrations (shown in Figs. 2 and 4)liquid flowrate dependency is observed at the highest gas superficialvelocities, when the concentration of SDS is reduced to 2.09 mM(0.6 g/l, approximately one-quarter of the critical micelle concentra-tion), another type of behaviour was observed. The data is depicted onlogarithmic–linear axes in Fig. 5). At the higher gas flowrates (i.e. thewetter foams) there is no dependency upon humidity above the rangeof experimental readings. However at lower gas rates (b15 mm/s),there is a systematic dependency of liquid flowrate upon relativehumidity, with the foams in more humid environments exhibitinggreater liquid rates. Thus, it is hypothesised that for this relatively lowsurfactant concentration, at low flowrates the rate of coalescence ofbubbles at the surface of the foam is increased, causing low liquidflowrate.

3.4. Dichotomy of liquid rate dependency upon concentration at lowhumidity

When the data, given above, for liquid flowrate in 10% relativehumidity air is replotted as series of constant concentration (seeFig. 6), they reveal a dichotomy that we cannot readily explain. Whenthe surfactant concentration is lower, the bubbles are less stable, andtherefore one would expect that bubble coalescence on the surface

Fig. 4. Superficial liquid velocity versus superficial liquid velocity for a foam stabilisedby 4.17 mM SDS.

Fig. 5. Superficial liquid velocity versus superficial liquid velocity, depicted onlogarithmic–linear axes, for a foam stabilised by 2.09 mM SDS.

29R. Shaw et al. / International Journal of Mineral Processing 97 (2010) 26–30

would be greater causing a reduction in liquid rate. This is preciselythat trend that is observed at lower gas superficial velocities.

At higher gas flux, the liquid flux is significantly higher for foamsstabilised at the lowest concentration. However, at lower rates of gasflux, the liquid flux is observed to be significantly lower at this lowconcentration. Such behaviour is only apparent at 10% relativehumidity. A possible explanation, albeit experimentally untested, forthis unexpected result may lie in the mechanics of the foam flowingfrom the top of the column and into the launder vessel. At higher gasflowrates, and therefore higher foam liquid fraction, the foam exhibitslower viscosity (i.e. it has higher flowability). It is possible that,although the bubbles on the free surface of the foam are experiencingcoalescence, the bulk of the foam, in fact, is convected to the laundervessel from the side because the viscosity is the foam is low.Stevenson (2007) showed that surface coalescence created a streamof “washwater” in the bulk of the foam, thereby making it wetter. It ispossible that, when the surfactant concentration and relativehumidity is low, the coalescence rate on the free surface is higher,and this creates a wetter foam that can more easily flow from the sideinto the launder vessel. This, in turn, would cause the liquid flowrateto be enhanced.

3.5. Inter-day reproducibility

When performing NMRI experiments on the addition of wash-water to columns of pneumatic foam, Stevenson et al. (2009) found itdifficult to obtain quantitative reproducibility vis a vis liquid fractionwithin the foam. Keeping all possible experimental parametersconstant, they found that the liquid fraction observed in the foam

Fig. 6. Superficial liquid velocity versus superficial liquid velocity at 10% relativehumidity.

on one day could be very different to that observed in nominally thesame experiment on the next day. They supposed that the lack ofreproducibility was caused by an inability to precisely control thebubble size distribution from one experiment to the next, but offeredno further explanation.

Due to the length of time that it took to conduct the currentexperiments, the experimental campaign described herein wasconducted over many days. A concern was that similar reproducibilityproblems to those experienced by Stevenson et al. (2009) maycompromise the possibility of comparison between data collected fordiffering conditions. Thus, exactly the same experiment was con-ducted on 6 separate days in order to assess inter-day reproducibilityof the liquid flowrate results. A foam stabilised by the intermediateSDS concentration (i.e. 4.17 mM) and sparged with air at a superficialvelocity of 12 mm/s was chosen. On each of the 6 days, the gas wasswitched on and, once the column had attained steady state, the liquidoverflow flowratewasmeasured by collecting the foam in a bucket for5 min, and this was done 10 times.

The data are plotted in Fig. 7 with the height of the columndenoting the mean values of 10 readings of each experiment, with theerror bars showing the full range of all readings. It is apparent that ourexperimental design was able to give excellent quantitative inter-dayreproducibility.

The question therefore arises as to why Stevenson et al. (2009)were not able to gain similar reproducibility using similar apparatus?The answer to this is not knownwith certainty, although Stevenson etal. did not control, nor did theymeasure, the relative humidity in theirlaboratory.

4. Discussion

Although the experiments that have been described herein are allconducted upon demineralised foams, it is possible that the resultscould have an important bearing upon knowledge of flotation froths.To the authors' knowledge there have been no experimental or plantstudies of froth flotation that have reported the environmentalhumidity and yet, in this study of demineralised foam, it is shownthat humidity can have a great effect on liquid flowrate from apneumatic foam. If a similar trend is indeed apparent in flotationfroths then it is suggested that further studies of this process areconducted under conditions of controlled or, at the very least,measured humidity. We acknowledge that the presence of attachedparticles can act to stabilise bubbles at the free surface of the froth,and that there can be no evaporation (and therefore no dependencyon humidity) if the surface coverage of particles is total. However, thepotential dependency of the performance of flotation froths upon the

Fig. 7. Liquid overflow rate for the same experiment (12 mm/s air rate, 4.17 mM SDS)showing that there was inter-day quantitative reproducibility exhibited by ourexperimental method. The error bars show the range of ten readings taken for eachexperiment.

30 R. Shaw et al. / International Journal of Mineral Processing 97 (2010) 26–30

weather should not be generally overlooked. Li et al. (2010) havegiven metrological data for Cessnock (NSW) and Mount Isa (QLD)where significant flotation plants are located in Australia. The intra-year variation of relative humidity is apparent.

Theoretical studies of the hydrodynamic state of pneumaticdemineralised foam that have attempted to model the coalescenceof bubbles at the free surface of the foam (Neethling et al. (2005),Grassia et al. (2006) and Neethling and Cilliers (2008)) have notconsidered a humidity dependency at all. However, as noted by Li etal. (2010), we should be guided by the well-informed opinion ofCilliers (2009) who stated that “Developing and verifying models fordetermining from first principles the rates of surface lamellae failureand internal bubble coalescence remains one of the most challengingaspects of understanding in detail the froth behaviour.” We haveinferred that surface lamellae failure (i.e. bubble coalescence), insofaras it affects liquid overflow rate in a demineralised pneumatic foam, isdependent upon environmental humidity. Alas, the development of amechanistic model of surface coalescence is currently out of theauthors' reach. For example, factors that confound the estimation ofevaporation rate from the free surface of the foam include theobservation of Burnett and Himmelblau (1970) that adsorbedsurfactant molecules at the gas–liquid interface gas control the rateof inter-phase mass transfer.

5. Conclusions

1. Led by the observation that the behaviour of a non-overflowingpneumatic foam is dependent upon the gradient of relativehumidity within a column, experiments have been conducted onan overflowing system to ascertain whether the liquid flowrate of ademineralised pneumatic foam has dependency upon environ-mental humidity.

2. It is observed that relative humidity has an important affect uponliquid overflow rate at relatively high gas rates when highsurfactant concentration as used to stabilise the foam, and atrelatively low gas rates when a low concentration is used.

3. Unexpectedly, it was observed that, at a low relative humidity of10% and high gas rates, the liquid flux was greater for foamsstabilised by a lower concentration of surfactant. A possibleexplanation is that enhanced surface bubble coalescence acts tocause the foam beneath to be wetter, and this is transported to thelaunder vessel from the side at the exit of the column.

4. A study has been carried out to confirm that our results exhibitquantitative inter-day reproducibility. It is possible that previoussimilar studies of pneumatic foam lacked inter-day reproducibilityprecisely because environmental humidity was not controlled.

5. Although we have not conducted experiments on mineralisedflotation froths, the potential for environmental humidity to affectthe hydrodynamic condition of the flotation process should not beoverlooked. The authors know of no laboratory or plant studies offroth flotation that have controlled or reported the environmentalhumidity.

Acknowledgments

This work was funded under the Discovery Projects scheme of theAustralian Research Council (project no. DP0878979). Useful discus-sions with Xueliang (Bruce) Li are gratefully acknowledged.

References

Aktas, Z., Cilliers, J.J., Banford, A.W., 2008. Dynamic froth stability: particle size, airflowrate and conditioning time effects. Int. J. Min. Proc. 87, 65–71.

Barbian, N., Ventura-Medina, E., Cilliers, J.J., 2003. Dynamic froth stability in frothflotation. Min. Eng. 16, 1111–1116.

Barbian, N., Hadler, K., Ventura-Medina, E., Cilliers, J.J., 2005. The froth stability column:linking froth stability and flotation performance. Min. Eng. 18, 317–324.

Barbian, N., Hadler, K., Cilliers, J.J., 2006. The froth stability column: measuring frothstability at an industrial scale. Min. Eng. 19, 713–718.

Bhattacharya, S., Pascoe, H.D., 2005. Effect of temperature on coal flotationperformance: a review. Min. Proc. Extr. Metall. 26, 31–61.

Bikerman, J.J., 1938. The unit of foaminess. Trans. Faraday Soc. 34, 634–638.Bikerman, J.J., 1973. Foams. Springer-Verlag. pub.Burnett, J.C., Himmelblau, D.M., 1970. Effect of surface active agents on interphase mass

transfer. AIChE J. 16, 185–193.Choung, J., Walker, J., Xu, Z., Masliyah, J., 2004. Effect of temperature on the stability of

froth formed in the recycle process water of oil sands extraction. Can. J. Chem. Eng.82, 801–806.

Cilliers, J.J., 2009. Physics-based froth modelling: new developments and applications.Int. J. Comp. Fluid Dyn. 23, 147–153.

Grassia, P., Neethling, S.J., Cervantes, C., Lee, H.T., 2006. The growth, drainage andbursting of foams. Coll. Surf. A 274, 110–124.

Ireland, P.M., 2009. Coalescence in a steady-state rising foam. Chem. Eng. Sci. 64,4866–4874.

Lazarov, D., Alexandrova, L., Nishkov, I., 1994. Effect of temperature on the kinetics offroth flotation. Min. Eng. 7, 503–509.

Li, X., Shaw, R., Stevenson, P., 2010. Effect of humidity on dynamic foam stability. Int. J.Min. Proc. 94, 14–19.

Neethling, S.J., Cilliers, J.J., 2003. Modelling flotation froths. Int. J. Min. Proc. 72, 267–287.Neethling, S.J., Cilliers, J.J., 2008. Predicting air recovery in flotation cells. Min. Eng. 21,

937–943.Neethling, S.J., Lee, H.T., Grassia, P., 2005. The growth, drainage and breakdown of

foams. Coll. Surf. A 263, 184–196.Stevenson, P., 2007. The hydrodynamic theory of rising froth. Min. Eng. 20, 282–289.Stevenson, P., Stevanov, C., 2004. Effect of rheology and interfacial rigidity on liquid

recovery from rising froth. Ind. Eng. Chem. Res. 43, 6187–6194.Stevenson, P., Mantle, M.D., Tayler, A.B., Sederman, A.J., 2009. Quantitative NMRI

studies of transient washwater addition to rising foam. Chem. Eng. Sci. 64,1001–1008.

Triffett B.B., Cilliers J.J. 2004 Measuring Froth Stability, International Patent ApplicationNumber: PCT/AU2004/000331.