Process engineers are often chal-lenged with the proper selec-tion of mixing technology for agiven operation. While there isa wealth of published informa-

tion on blending, solids suspensionand gas dispersion, there is anotherclass of mixing that is not well under-stood: the introduction of dry materialor powders into water or other liquids.The frequent result is an incorrect ap-plication of mixing equipment.

Stored dry powdered materials,such as chemical reagents, rheologymodifiers, fillers, abrasives or coloringagents, usually consist of aggregatesand agglomerates, which must be dis-persed in the liquid to produce “indi-vidual” entities. How effectively thisis done has significant implications forthe overall efficiency of a processingoperation. For instance, incompletepowder wetting and dispersion can re-sult in poor yields, unplanned filtra-tion expenses, batch variations andexcessive batch times. Tank cleanli-ness and cross contamination betweenbatches can also be issues. A poorlydesigned mixing system represents aprocess bottleneck that can bring anoperation to a halt.

Many such operating problemscan be attributed to the conven-tional approach of using in-tankmixers, typically turbine or rotor-stator versions. The alternatives ofselecting special mixer designs, ormixers that consume high horse-power per unit volume, lead to un-necessarily large capital expendi-tures and high operating costs. Inmost cases, single-pass in-line mix-ing is the better choice.

ChallengesThe process engineer is fortunate ifthe solid material to be added is wellbehaved; that is, if the powders wetout and disperse readily. Such casesoften can, in fact, be handled by in-tank mixing.

In practice, however, well-behavedsolids represent only a small percent-age of the applications in the processindustries. The engineer is more likelyto be confronted with difficult solids, afew examples being high-molecular-weight hydrocolloids, cellulose gums,starches, and proteins. Other exam-ples are named below, in the contextof the particular problems they cause.

When trying to mix difficult solidsinto liquids, two problems are particu-larly common:• The powders float, resisting all ef-

forts to “wet” them out, or• They lump together, forming impen-

etrable sticky masses or “fisheyes”(lumps of product with a hydratedskin but a non-hydrated, dry core)

When powders float, it is usually dueto surface tension effects. As a clumpof powder is introduced to the liquid,surface tension forces can prevent theliquid from penetrating the clump.Under inadequate mixing, this unwet-ted clump can float on the liquid andhave limited contact with it. Whensuch material is added to the surfaceof an agitated tank, a floating mass, or“raft” forms (Figure 1) leading tobuildup of solids on the wall andmixer shaft. These types of powdersare often referred to as hydrophobic.Fumed silica, carbon black, cocoa pow-ders and organic pigments are com-mon examples.

The second, fish eye, problem ariseswith powders that are instead hy-drophilic. When an agglomerate ofhydrophilic powder comes into contactwith water, the particles on the sur-face of the agglomerate become hy-drated and swell quickly, crosslinkingto form a tough, relatively imperme-able gel layer. The particles inside ofthis outer layer cannot be hydratedbecause they are shielded from thewater. At best, the result is small,transparent fisheyes (Figures 2 and3), but more frequently it consists oflumps of various sizes.

Good examples of hydrophilic solidsare the many powders classified asgums. These may be derived from aplant origin, such as locust bean orguar. Or, they can be either microbialin nature, such as xanthan, or syn-thetic, such as carboxymethylcelluloseor other cellulose derivatives. Anotherclass of hydrophilic powders that aredifficult to wet and disperse consists ofacrylic acid polymers (carbomers).

The conventional approachAs noted above, in-tank mixing, theoption with which most processors ap-proach the solids-into-liquids task, issuitable if powders wet easily and donot tend to form lumps. In these cases,the engineer need only concern himselfor herself with providing enough agita-tion to suspend the solids, so that theywill dissolve in the liquid, react with itor remain suspended, as the intendedcase may be. But, this approach leavesa lot to be desired if the powders aredifficult, as defined earlier.

For one thing, the surface turbu-lence generated with conventional

Engineering Practice

The Better Way to Mix Solids Into Liquids

The instinctive choice is in-tank mixing.But with problem powders, in-line mixing

gives far better resultsStephen Russell-HillQuadro Engineering Inc.

mixers is ordinarily not sufficient todraw in floating powders that haveformed rafts. What’s more, the pres-ence of baffles leads to in additionalsolids buildup at the wall. Long batchtimes, powder buildup and subsequentwaste during tank cleanup are the typ-ical consequences. Cross-contamina-tion of product can also be an issue ifmultiple formulations are processedthrough the same batch tank.

Some well-known strategies can beemployed to help the mixing processalong. For example, the baffles can becut back or eliminated altogether, topromote swirl or vortexing. This tac-tic can be very effective at drawing infloating powders. However, signifi-cant air entrainment normally occursas well, which requires the plant toadd a sometimes lengthy deaerationstep to the process.

A cutback or elimination of bafflesalso leads to increased radial flow pat-terns in the vessel, which can hinderblending and solids-suspension appli-cations that require good top-to-bot-tom turnover. And, the hydraulic loadacting upon the mixer shaft and bear-ings can increase significantly, lead-ing to premature component failure.

If powder dispersion is the mainchallenge, it is common to use high-shear rotor-stator mixers, sometimescalled homogenizers, which produce agreater shearing action with a rela-tively large invested horsepower perunit volume within the tank. Buteven though high-shear mixers, withhigh tip speed and tight clearance ro-tors and stators, are capable of highdegrees of shearing action, there is aninherent problem with these deviceswhen used in vessels: the maximumshear is found in the immediate area

of the rotor-stator, whereas the localshear rates are lower throughout therest of the vessel, effectively ap-proaching zero towards the wall. Theresult is a distribution of shear in thevessel, with the particles not beingtreated equally. Some particles aresubjected to intense shear, some seevery little shear, and the majority seean average shear.

Because of this shear distribution inthe tank and the unavoidably some-what random arrival of lumps to thehigh-shear rotor-stator, it is not possi-ble to accurately control the amount ofshear applied. As a result, it is diffi-cult to reproduce the desired processresult from one batch to the next.These mixers also tend to generatesignificant radial movement in thevessel, which tends to cause vortexingand air entrainment, the conse-quences of which have already beenhighlighted.

To overcome the limitations withhigh-shear rotor-stator devices, mostprocessors tend to run the mixer untilthey are satisfied with the dispersionor have eliminated most of the lumps.Apart from the obvious cost of longbatch times, this overprocessing canpotentially damage the end product ifit is either shear- or heat-sensitive.

A classic case is the overprocessingof some acrylic-acid-polymer solutionsthat are commonly employed to in-crease viscosity in liquids. The drymaterial is extremely hydrophilic, andfisheyes form immediately in contactwith water. As the lumps form, thebatch is run longer and longer in aneffort to eliminate them. However,the cross-linking polymers becomeshear-sensitive as they hydrate. As aresult, the overprocessing reduces theeffective viscosity. This, in turn, in-spires the processor to compensatewith the addition of yet more of the

Supplytank

Water mann

Finishedproduct

FIGURE 1. If not blended into liquidsproperly, hydrophobic powders maywell float on the liquid surface as a ‘raft’

FIGURES 2 and 3. As for powders that are hydrophilic, a characteristic problem withthem is the formation of lumps of product with a hydrated skin but a non-hydrated drycore. The particles on the inside become shielded by that skin, and thus they cannotinteract as desired with the liquid

FIGURE 4. In this typical, batch in-line arrangement, the powder is introduced into theliquid at a point in the piping. The same solid-feeding principle can be applied duringcontinuous processing

polymer, creating still more lumps. Atthe end of the batch, a significantamount of dry material must be re-moved by filtration, and thus wasted.

Another remedy consists of sprin-kling the powder onto the moving liq-uid surface slowly, thus encouragingthe particles to wet and hydrate “indi-vidually,” lessening the formation offisheyes. For most processors how-ever, this approach is manual, slowand tedious. For example, three ormore hours can be required to sprin-kle, agitate, disperse and then hydratea simple 250-gal batch of a carbomeringredient for a hair-styling gel.

Depending upon the process and thenature of the powder ingredients, fish-eye formation can be prevented bypre-wetting the powder in a non-aque-ous solution, such as a glycol. Oncedispersed, the solution is then blendedinto the aqueous phase to completethe hydration. This can be an effec-tive approach; but it, too, representsan additional process step, extraequipment and added material costs.

In summary, high-shear rotor-sta-tor mixers are designed to be shearproducers, not flow producers. Ifblending and solids suspension areadditional processing requirements inthe vessel, then this style of mixerrepresents a poor selection, as do theconventional in-tank mixers.

A better choiceThe issues summarized above can beaddressed by in-line processing. Itconsists simply of adding the powderphase into the liquid phase at somepoint in the piping, not in the processvessel itself. Single-pass powder-ad-dition device can be located in apipeline between two vessels (Figure4), in the water supply line leading toa batch tank, or at some point in apipeline that is part of a continuousprocess. As an alternative to single-pass addition, the powder can be in-troduced into a recycle or recirculationloop that includes a tank.

Direct addition of powder to aprocess line has tangible advantagesover in-tank mixing:• No waiting for the wetting of float-

ing powders• No buildup on the tank wall, mixer

shaft, and baffles. This reduces

tank-cleaning time, lowers thechance for cross contamination, andvirtually eliminates powder loss

• As there is no tank vortexing, signif-icantly less air is entrained in theprocess fluid and, as a result, thereis no need to install expensivedeaerating devices or to wait fordeaeration to occur

A single-pass, in-line approach allowsfor not only the controlled introduc-tion of powders but also the controlledapplication of the shear necessary forthe specific behavior characteristics ofthe powders. The controlled applica-tion of shear to the process fluid yieldssome significant benefits:• There is repeatable processing from

batch to batch• Overprocessing to get rid of lumps is

eliminated; this benefit is especiallyvaluable when dealing with heat- orshear-sensitive products

• The applied energy is focused uponthe task at hand – generating uni-form, lump- and raft-free disper-sions. No energy is wasted on try-ing to recirculate multiple passes offluid through an in-tank dispersingmechanism (one application, for ex-ample, involving the dispersion of a2,000-gal batch of gum, required20–25-hp if carried out by an rotor-stator in-tank mixer, but only 7.5 hpif done via an in-line disperser)

• The performance of the dispersingdevice is independent of the charac-teristics of the dispersed product

• Batch times are reducedThe in-line approach is also attractivebecause a single in-line device canfeed a number of batch tanks. Thereis no need to provide a high-shear in-tank mixer for each process vessel.Furthermore, powders can be intro-duced and dispersed into the liquidduring the filling of the vessel, thusreducing batch times.

In-line alternativesA number of in-line technologies areavailable for processing solids into liq-uids. All can be used in a single-passprocessing approach. However, eachoption has its particular characteris-tics that tend to limit the concentra-tion or quality of dispersion that it canachieve with a given powder.

Regardless of equipment type, thepowder becomes incorporated into thefluid because the device generates avacuum. It is the rate at which thisvacuum builds that dictates the pow-der incorporation rate. Dispersion oc-curs in a localized region of shearing,or where large velocity differencesoccur over relatively short distances. Funnel-and-eductor arrange-ments: The simplest and oldest tool ofadding powders into liquids in-line is

Engineering Practice

Funnel

Air vent holes

Make-upwater in

Educatoroutlet

FIGURE 5 (above). The funnel-and-eductor system was the firstsystem employed for in-lineblending of powders into liquids

FIGURE 6 (left). An

improvementupon the funnel-

and-eductorsystem employs

this injector,which is fed withliquid under high

pressure

the powder funnel and eductor (Fig-ure 5). The vacuum created due to afluid flowing past an eductor drawsthe powder into the fluid.

This option features a simple con-struction with no moving parts, apartfrom the presence in many installa-tions of a powder valve that separatesthe funnel from the eductor. Theflowing powders begin to individual-ize as they are drawn down into theeductor.

However, powder incorporationrates are generally low with this op-tion, as are the shear rates. The dis-persing capability is limited, espe-cially if one is dealing withhydrophilic powders. And the perfor-mance of the equipment is very sensi-tive to process upsets in flow rate,temperature, and viscosity. Specialized eductors: More-sophis-ticated eductor systems have been de-veloped that improve upon the basiccapabilities of the funnel-and-eductorsystem described above. These de-vices incorporate a specially designedinjector that is fed with the processliquid under high-pressure (Figure 6).As the liquid is forced through an an-nular orifice at high velocity, a power-ful vacuum is created, and the liquidforms a spray or curtain. Meanwhile,in a separate stream, the solids aredrawn under vacuum to the eductor.The solids are introduced into thespray or curtain, which wets and dis-perses them.

Due to the higher feed pressuresand the nature of the eductor nozzledesign, greater concentrations ofsolids can be introduced, and the pow-der agglomerates encounter greater

shearing forces than they would in atypical funnel-and-eductor system.Drawbacks to this specialized-eductorapproach include the need for higher-pressure feed pumps, a risk of vapor-ization and cavitation when operatingwith hot fluids, and the fact that it canbe difficult to disperse high concentra-tions of difficult powders.Pump-style blenders: Later varia-tions of the powder funnel-eductor ap-proach employ some form of rotatingmechanical device to boost the abilityto incorporate and disperse the pow-der into the liquid. One of the mostcommon types is effectively a modifiedcentrifugal pump with a powder fun-nel attachment, often referred to as apowder blender. With this device, liq-uid is pumped into a verticallymounted pump chamber that incorpo-rates a centrifugal pump wheel, a per-forated screen or set of baffles at theperiphery of the impeller to help redi-rect flow and disperse powders, and apowder feed tube that keeps the pow-der agglomerates separate from theliquid phase until the last moment,when the impeller draws the powderand liquid phase together and dis-charges the blend through the screenor baffles. Vacuum is created by theaction of the fluid being dischargedfrom the unit. Some vendors offer two-stage models (with two pump im-pellers) that offer increased vacuumand, hence, a higher capacity forsolids addition.

A valve on the powder funnel iso-lates the powder from the liquid priorto the blending operation. Some ven-dors recommend the further use ofthat valve to throttle the powder addi-

tion rate, in order to control the dis-charge consistency (for example, if theprocessor wants a low concentration).However, this strategy should be con-sidered with care, because partialvalve openings tend to result in abuildup of partially hydrated materialin the powder feed tube, and can alsolead to an increase in air entrainmentin the product and a decline in theperformance of the blender.

This equipment should always beused with a feed pump and liquid-flowcontrol. Depending upon the dis-charge conditions (line length, bends,elevation, other considerations) a sec-ondary discharge pump may be re-quired to deliver the product to down-stream processing.

In general, these pumpwheel-styledevices are suitable for relatively highcapacities and solids concentrations.On the other hand, they run into diffi-culties when dealing with gums andthickeners that are difficult to dis-perse; the shearing effect delivered bythe pump impeller and screen is gen-erally not sufficient to effectively dis-perse such powder agglomerates insingle-pass processing. Recirculationcan be attempted, but the changingphysical properties of the dispersedproduct (such as an increase in viscos-ity) can cause the performance of theunit to quickly drop off.

Similarly, these devices cannot beused with high-viscosity liquids(above 250 cP), and are susceptible toproblems, due to cavitation, when pro-cessing hot fluids. The possibilities ofcavitation in the reactor head andvapor wetting of the feed-tube andpowder-valve area can cause powderclumping and bridging.High-shear blenders: Many vendorsoffer significant improvements overthe original pumpwheel-style blender.Instead or relying upon a pump im-peller and screen/baffle plate as thedispersing mechanism, designs havebeen introduced that incorporatemore-aggressive dispersion tooling(Figures 7 and 8) with similarities tothe aforementioned high-shear rotor-stator mixers for in-tank usage. Witha design employing a vaned rotor anda slotted stator, or a toothed rotor-sta-tor design, the operating clearancesand tip speeds are such that the pow-

FIGURES 7 and 8. Whilehigh-shear mixing doesnot effectively mix difficultpowders with liquid duringin-tank mixing, the high-shear principle can beharnessed with goodresults for in-line blending

der is subjected to intense shearing asit is introduced into the liquid phaseand passes through the rotating ele-ment. The most challenging hy-drophilic gums and polymers canreadily be handled with these me-chanical devices.

In some designs, the establishmentof a “liquid ring” in the reactor headcontaining the rotating element,much like the configuration in a liq-uid-ring vacuum pump, generatesthe vacuum that is required to drawin powders. Other designs rely uponhigh-pressure flows past an eductorupstream of the dispersing mecha-nism as the motive force for powderincorporation.

Table 1 offers a summary compari-son of the technologies available todayfor in-line mixing of solids into liquids.The table should be used solely as ageneral guideline; and the numbersshown for capacity and solids concen-tration are for relative comparison

purposes only, and may vary depend-ing upon the physical properties andother characteristics of the powderbeing handled.

Most vendors of in-line devices arewilling to help the engineer match theright device to the task, through a va-riety of testing programs, onsitedemonstrations, monthly rentals andother means. For the engineer, themain goal is to end up with the bestdevice for the job at hand, and to aimto achieve the desired mixing result ina single pass. ■

Edited by Nicholas P. Chopey

Engineering Practice

AuthorStephen Russell-Hill isgeneral sales manager formixing equipment at QuadroEngineering Inc. (613 ColbyDrive, Waterloo, Ont. N2V1A1, Canada; www.quadro.com). Previously, he was gen-eral manager, Flow Div., forAlfa Laval Canada, and prod-uct manager and applicationengineering manager forProchem Mixing Equipment

Co. He has 20 years of experience in industrialequipment sales and application engineering,with a focus on mixing. He is a member ofAIChE, the Professional Engineers of Ontarioand the North American Mixing Forum. A reg-istered professional engineer in the Province ofOntario, he holds a bachelor of applied sciencedegree in chemical engineering from the Univer-sity of Waterloo.

TABLE 1. RELATIVE PERFORMANCE OF IN-LINE TECHNOLOGIESTechnology Relative Relative Typical Liquid Typical Powder

Capacity Shear Throughputs, Concentrations, lb/h wt. percent

Funnel-and-eductor Low Low 250 <10Specialized eductor Moderate Moderate 50,000 <25Pump-style blender High Moderate 100,000 <35High-shear blender Moderate High 50,000 <15

Reprinted from CHEMICAL ENGINEERING, November 2004, copyright 2004 by Access Intelligence with all rights reserved.Additional reprints may be ordered by calling Chemical Engineering Reprint Department (212) 621-4958. To subscribe to Chemical Engineering, call (212) 621-4656.