ultrasonic dispersion of nanoparticles for environmental

Nanotoxicology, 2010; Early Online, 1–19

Ultrasonic dispersion of nanoparticles for environmental, health andsafety assessment – issues and recommendations

JULIAN S. TAUROZZI1, VINCENT A. HACKLEY1, & MARK R. WIESNER2

1Material Measurement Laboratory, National Institute of Standards and Technology, Gaithersburg, Maryland, and2Department of Civil and Environmental Engineering, Duke University, Durham, North Carolina, USA

(Received 9 July 2010; accepted 28 September 2010)

AbstractStudies designed to investigate the environmental or biological interactions of nanoscale materials frequently rely on the use ofultrasound (sonication) to prepare test suspensions. However, the inconsistent application of ultrasonic treatment acrosslaboratories, and the lack of process standardization can lead to significant variability in suspension characteristics. At present,there is widespread recognition that sonication must be applied judiciously and reported in a consistent manner that isquantifiable and reproducible; current reporting practices generally lack these attributes. The objectives of the present workwere to: (i) Survey potential sonication effects that can alter the physicochemical or biological properties of dispersednanomaterials (within the context of toxicity testing) and discuss methods to mitigate these effects, (ii) propose a method forstandardizing the measurement of sonication power, and (iii) offer a set of reporting guidelines to facilitate the reproducibilityof studies involving engineered nanoparticle suspensions obtained via sonication.

Keywords: Environmental assessment, toxicology, nanoparticle, nanomaterial, ultrasonics, sonication, dispersion, suspension

Introduction

Nanotechnology, and more specifically the emer-gence of engineered nanoscale particles (ENPs),shows promising potential for the development ofadvanced materials and devices in the fields of med-icine, biotechnology, energy, and environmentalremediation, among others (Zhang 2003; Salata2004; Raimondi et al. 2005). Similarly, ENPs aremigrating into a wide range of consumer products,many of which are now in the global market place(Project on Emerging Nanotechnologies 2010). How-ever, some of the properties that make ENPs attractivemay also pose environmental and health hazards(Colvin 2003; Gwinn and Vallyathan 2006;Kreyling et al. 2006; Dobrovolskaia and McNeil2007; Kahru and Dubourguier 2010). As both publicand private funding for application-oriented nanoma-terial research continues to grow and nanoparticle-based consumer products reach the market, there isan increasing demand for a more comprehensivescientific understanding of the interactions of ENPs

in biological and environmental systems (Maynard2006; Barnard 2006).Moreover, the assessment of potential risks associ-

ated with ENPs requires the evaluation of their behav-ior in environmentally and biologically relevantmatrices (e.g., cell media, serum, whole blood, stan-dardized seawater). When ENPs are not already in afully dispersed form, the test suspensions are typicallyobtained by dispersing dry powders or stock suspen-sions of the source material into the desired testmedium prior to in vitro or in vivo evaluation(Murdock et al. 2008; Jiang et al. 2009; Labilleet al. 2009). In these situations, the use of ultrasonicenergy, a process commonly referred to as ‘soni-cation’, is arguably the most widely used procedure.Sonication refers to the application of sound energy

at frequencies largely inaudible to the human ear(higher than » 20 kHz), in order to facilitate thedisruption of particle agglomerates through a processknown as cavitation. Ultrasound disruption is moreenergy efficient and can achieve a higher degree ofpowder fragmentation, at constant specific energy,

Correspondence: Dr. Vincent A. Hackley, National Institute of Standards and Technology, 100 Bureau Dr. Mail Stop 8520, Gaithersburg, 20899, MD, USA.E-mail: [email protected]

ISSN 1743-5390 print/ISSN 1743-5404 online � 2010 Informa UK, Ltd.DOI: 10.3109/17435390.2010.528846

Nan

otox

icol

ogy

Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

NIS

T N

atio

nal I

nstit

iute

s of

Sta

ndar

ds &

Tec

hnol

ogy

on 1

2/03

/10

For

pers

onal

use

onl

y.

than other conventional dispersion techniques (Parket al. 1993; Hielscher 2005; Mandzy et al. 2005).Sonication is a convenient, relatively inexpensive toolthat is simple to operate and maintain. Consequently,sonication has found extensive application in toxico-logical and environmental studies, where it is used tobreak down powders or re-disperse stock suspensionsto their nanoscale constituents and allow for theevaluation of the properties and behavior of the dis-persed nanoparticles in relevant systems.However, the inconsistent application of ultrasonic

treatment across laboratories and between individualoperators, combined with the lack of accepted processstandardization, has likely contributed to variability inobserved results. Furthermore, the complex physicaland chemical phenomena that occur during sonica-tion – and which may significantly alter the dispersedmaterial’s properties – are frequently underappreci-ated, perhaps resulting in the use of suspensions thatmay not meet the intended experimental design, orpotentially leading to artifacts that could compromiseconclusions as to how the observed ENP’s ecologicalor toxicological behavior relates to the material’sinherent properties. It is the authors’ contentionthat the inconsistent application and reporting ofsonication is a potential leading contributor to vari-ability in test results in which sonication is utilized;this conclusion has been echoed by the results ofrecent interlaboratory studies (Roebben et al. 2010).Moreover, while not the main focus of this work, it

is important to note that many of the sonication-induced physicochemical effects and recommenda-tions addressed herein are equally relevant to otherENP research areas where sonication is utilized, inparticular those involving the fate and transport ofnanomaterials in environmental systems.The objectives of the present work were to: (i)

Highlight salient aspects of the ultrasonic disruptionprocess and survey potential sonication-inducedeffects that may alter the physical, chemical or bio-logical properties and behavior of aqueous nanopar-ticle suspensions (within the context of toxicologicaltesting), (ii) propose a method for standardizing themeasurement of sonication power, and (iii) offer a setof reporting guidelines to facilitate the reproducibilityof studies involving engineered nanoparticle suspen-sions obtained via sonication.The authors aimed not only to indentify and discuss

key sonication issues relevant to ENP dispersion, butalso to address them by offering practical recommen-dations. For this reason, some discussions will followa traditional literature review format, while others aremore prescriptive in style. The authors also recognizethat additional sample-specific preparation steps areoften required to achieve the desired state of

dispersion, but their consideration is beyond thescope of the present work.

Terminology

We define at the outset several critical terms, whichare used throughout the text, in order to avoid ambi-guity, as these terms can have different meanings indifferent scientific and technical circles. For the mostpart, we follow definitions relevant to nanotechnologyas set forth in the standard E2456 (2006) by ASTMInternational (formally the American Society forTesting and Materials). Additional guidance isderived from recommendations of the InternationalUnion of Pure and Applied Chemistry (IUPAC2009).For the purposes of this publication, a nanoparticle

is defined as a sub-classification of ultrafine particlethat is characterized by dimensions in the nanoscale(i.e., between 1 and 100 nm) in at least two dimen-sions, with the recognition that the more rigorousdefinition of nanoparticle also requires that theseparticles exhibit novel size-dependent properties.A primary particle is the smallest discrete identifiableentity associated with a particle system; in this con-text, larger particle structures (e.g., aggregates andagglomerates) may comprise primary particles. Anaggregate is a discrete assemblage of primary particlesstrongly bonded together (i.e., fused, sintered, ormetallically bonded), which are not easily brokenapart. While the distinction is often blurred, aggre-gates might also be differentiated from agglomerates,which are assemblages of particles (including primaryparticles and/or smaller aggregates) held together byrelatively weak forces (e.g., Van der Waals, capillary,or electrostatic), that may break apart into smallerparticles upon processing (e.g., using sonication orhigh intensity mixing). Commonly, agglomeratesexhibit the same overall specific surface area (SSA)of the constituent particles, while aggregates have aSSA less than that of the constituent primary particles.In a related fashion, agglomeration and aggregationrefer, respectively, to the process by which agglom-erates or aggregates form and grow. It should be notedthat, although here we define them as distinct entities,the terms aggregate and agglomerate are often usedinterchangeably to denote particle assemblies. Ourusage is consistent with previous recommendationspublished in this journal (Oberdörster et al. 2007),but is by no means universally adopted.Additionally, in the present work, a suspension is

defined as a liquid (typically aqueous) in which par-ticles are dispersed. The term dispersion is used in thepresent context to denote the process of creating aliquid in which discrete particles are homogeneously

2 J. S. Taurozzi et al.

Nan

otox

icol

ogy

Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

NIS

T N

atio

nal I

nstit

iute

s of

Sta

ndar

ds &

Tec

hnol

ogy

on 1

2/03

/10

For

pers

onal

use

onl

y.

distributed throughout a continuous fluid phase (e.g.,to obtain a suspension), and implies the intention tobreak down agglomerates into their principal compo-nent particles. Dispersion is also used as an adjective incombination with terms relating to this process orrelating to the state of the resulting product.

Sonication basics and equipment

This brief overview highlights the basic aspects ofultrasonic generation and propagation relevant toparticle dispersion, and serves as background forthe ensuing discussions; while it is not a comprehen-sive review of the topic, readers can refer to the citedreferences for a more in-depth treatment of the sub-ject. During the process of ultrasonic disruption,sound waves propagate through the liquid mediumin alternating high and low pressure cycles at fre-quencies typically in the 20–40 kHz range. Duringthe low-pressure cycle (rarefaction), microscopicvapor bubbles are formed in a process known ascavitation (Figure 1). The bubbles then collapse dur-ing the high pressure cycle (compression) producing alocalized shock wave that releases tremendousmechanical and thermal energy.Ultrasonic waves can be generated in a liquid

suspension either by immersing an ultrasound probe(transducer horn) into the suspension (direct sonica-tion), or by introducing the sample container into aliquid that is propagating ultrasonic waves (indirectsonication). This is shown schematically in Figure 2.In a sonication bath or a so-called cup horn soni-

cator (indirect sonication), the sound waves musttravel through both the bath or cup liquid (typicallywater) and the wall of the sample container beforereaching the suspension. In direct sonication, theprobe is in contact with the suspension, reducingthe physical barriers to wave propagation and there-fore delivering a higher effective energy output intothe suspension. Bath sonicators typically operate at

much lower energy levels than are attainable using aprobe or cup horn. In the case of bath sonicators, thetransducer element is directly attached to the outsidesurface of the metal tank, and the ultrasonic waves aretransmitted directly to the tank surface and then intothe bath liquid. In a cup horn, the radiating surface ofthe horn is inverted and sealed into the bottom of a(typically) transparent plastic cup into which thesample container is immersed.At ultrasound frequencies, the alternating forma-

tion, growth and subsequent shock waves producedby the collapse of bubbles result in extremely largelocalized temperatures up to 10,000 K, rapid temper-ature changes (> 105 K s�1), pressure bursts on theorder of several MPa, and liquid jet streams withspeeds reaching 400 km � h-1 (Mason 1989; Hielscher2005). Such massive, local energy output is the basisof the disruptive effect of sonication. It must be notedthat these considerable pressure and temperaturedifferentials are a result of the cavitation processand occur at the local interface of the explodingbubbles; consequently, these effects are inherent tothe sonication process and will occur whether thesonicated container is cooled or not, or the treatmentperformed using a bath, cup horn, or a probesonicator.In all cases, the ultrasonic device transforms elec-

trical power into vibrational energy by means of apiezoelectric transducer that changes its dimensionsin response to an applied AC electric field. In directsonication, the transducer horn serves to transmit andfocus the ultrasonic waves into the targeted liquidsample. For direct sonication, the following equationrelates acoustic vibrational energy to probe andmedium parameters (Contamine et al. 1995):

P = cA f a12

)2 2ρ π( ( )2 1

Bubble

Lowpressure

cycle

Highpressurecycle

Implosionshock waves

Agglomeratefracture points

Figure 1. Schematic illustration of ultrasonic wave-inducedcavitation and agglomerate fracture.

Probe Bath Cup

Figure 2. Schematic illustration of direct (left) and indirect (middleand right) sonication configurations as described in the text.

Dispersion of nanoparticles 3

Nan

otox

icol

ogy

Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

NIS

T N

atio

nal I

nstit

iute

s of

Sta

ndar

ds &

Tec

hnol

ogy

on 1

2/03

/10

For

pers

onal

use

onl

y.

where P is the acoustic power (W) of the ultrasoundsource, a is the ‘emission area’ (m2), which is thesurface area of the emitting ultrasound source, r is theliquid density (kg m-3), A is the amplitude (m) ofoscillation of the ultrasound probe, c is the speed ofthe acoustic wave in the liquid medium (m s-1) and f isthe vibration frequency (Hz).

The amplitude of oscillation per unit time deter-mines the pressure difference between cavitationalrarefaction and compression cycles. Larger oscillationamplitudes yield larger high-low pressure gradientsand consequently greater energy outputs. The probe’soscillation amplitude is in turn determined by theamount of energy transferred by the instrument’sgenerator to the ultrasonicator probe, which can beregulated using the sonicator power settings.

Conventional direct sonicators operate by self-adjusting their power consumption from the electricalsource in order to maintain the vibration frequency ofthe transducer (e.g., quartz crystal) at a constant value(usually 20 kHz). As it oscillates, the sonicator probe –connected to the vibrating transducer – will experi-ence a resistance from the sonicated medium that willbe transmitted back to the vibrating element anddetected by the internal control unit of the instru-ment. The instrument’s control unit will in turnadjust the power consumption of the instrument tomaintain a constant vibration frequency. Highly vis-cous media will exert a greater resistance to theoscillating probe, and will therefore require greaterpower consumption to maintain a constant oscillationfrequency. For example, to produce an oscillationwith an amplitude of 3 mm at 20 kHz in deionizedwater, a sonicator may consume ‘N’ watts, while toproduce the same amplitude at 20 kHz in oil, it willconsume considerably more power.

The oscillation frequency value is usually fixed for agiven instrument and cannot be changed. Changingthe instrument power setting translates to a change inthe vibrating amplitude – not the frequency – of theprobe; that is, increasing the power setting results inlarger probe oscillation amplitudes. To maintain aconstant oscillation frequency, larger oscillationamplitudes, as well as higher medium resistances,

require greater power consumptions from the electricsource.A sonicator’s stated maximum power (e.g., 600 W)

refers to the maximum theoretical power it would beable to consume, should it require so; it does not reflectthe actual amount of ultrasonic energy delivered tothe suspension. That is, for the same frequency andamplitude (and thus the same sonicator setting), thesonicator would need to consume more power to treata high viscosity suspension than it would for a lowviscosity suspension, and if so, it could consume up toa maximum of 600 W. It is therefore erroneous toassume that selecting the highest setting value willresult in the delivery of the instrument’s nominalmaximum power to the sonicated suspension. Forlow viscosity media, even at the highest setting value,the delivered power will be a significantly lower frac-tion of the instrument’s maximum power rating.Similarly, the power value (measured in W) usually

shown on the instrument display reflects the powerthat the instrument is consuming from the electricalsource to produce the desired oscillation amplitude(from the selected setting value) in the sonicatedmedium. The consumed (i.e., displayed) power, how-ever, does not necessarily reflect the power that isactually delivered to the sonicated suspension, which isaffected by the probe’s oscillation amplitude in themedium (see Equation 1). Herein lies the deficiencywith reporting either the manufacturer’s stated max-imum power, the adjustable output power readingprovided by the instrument, or the chosen settinglevel. The flowchart in Figure 3 illustrates the energytransformations that occur in a conventional probesonicator.The net fragmentation effect from applying ultra-

sonic energy to a suspension is dependent on the totalamount of energy delivered to the sonicated medium.However, not all of the produced cavitational energyis effectively utilized in disrupting particle clusters.The delivered energy is consumed or dissipated byseveral mechanisms, including thermal losses, ultra-sonic degassing, and chemical reactions, such asthe formation of radical species. Only a portion ofthe delivered energy is actually used in breaking

Electrical, mechanical andthermal losses

Coupling, friction,thermal losses

SuspensionSonication

probeDisplayed

power

Generator&

Transducer

Electricsourcepower

Figure 3. Flowchart illustrating the energy transformation process in a conventional probe (direct) sonicator.


Nan

otox

icol

ogy

Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

NIS

T N

atio

nal I

nstit

iute

s of

Sta

ndar

ds &

Tec

hnol

ogy

on 1

2/03

/10

For

pers

onal

use

onl

y.

particle-particle bonds to produce smaller particleaggregates, agglomerates, and primary particles.Moreover, an excessive energy input can potentiallyresult in agglomerate formation or re-agglomerationof previously fragmented clusters, as well as induce avariety of physicochemical alterations on the materi-al’s surface or to the constituents of the suspendingmedium. Such sonication-specific side-effects areaddressed in the following sections, but first weexamine the particle disruption process as it is theprimary objective of ultrasonic treatment.

Sonication effects on particle disruption andaggregation

The properties associated with ENPs, and conse-quently their distinct biological and environmentalinteractions, are strongly dependent on the particlesize of the processed material. For instance, the size ofthe particles in suspension dictates – along with theparticles’ surface chemistry – their biodistribution,elimination, cellular uptake, toxicity, and environ-mental transport behavior (Rejman et al. 2004;Lanone and Boczkowski 2006; Pan et al. 2007;Powers et al. 2007; Li and Huang 2008; Sonavaneet al. 2008). Additionally, the agglomeration state

determines whether the particles will settle out orremain in suspension, which directly impacts assayefficacy and delivered dose. Numerous other exam-ples of size-dependent behavior can be found in therecent literature.Dry powders consist of particles that are bound

together into macroscopic structures by both weakphysical Van der Waals forces and stronger chemicalbonds including particle fusion (Zachariah andCarrier 1999; Mandzy et al. 2005; Jiang et al.2009). Commonly, these powders contain micro-meter – and submicrometer – scale aggregates, whichare in turn held together by physical forces to con-stitute larger agglomerates. This is depicted schemat-ically in Figure 4. For powders consisting of nanoscaleparticles and aggregates, and therefore having sub-stantial interfacial contact areas per unit volume, thenominally weaker Van der Waals forces can beextremely large, requiring the use of techniquessuch as sonication to effectively break down or disruptpowder agglomerates. In contrast, micrometer sizeprimary particles can often be dispersed with moder-ate mixing or stirring.Many of the commercially available metal oxide

ENP source powders (e.g., silica, titania and alumina)are synthesized by high temperature vapor phase

Ave

rage

par

ticle

siz

eA

vera

ge p

artic

le s

ize

Sonication energy(power, time, energy density)

Sonication energy(power, time, energy density)

Aggregateprimary particles held

together by sinter necksor strong bonds

Primary particle

Sinter neck

Asymptotictrend

Delivered sonication energy

Agglomerateaggregates held together

by Van der Waals orother weak forces

Smalleragglomerates &

aggregates

Dispersedaggregates and/orprimary particles

Sonicationinduced

aggregation(coagulation)

No furtherdisruption

A

B

Peaking trend

A

B

Figure 4. Schematic depiction of particle structures referred to in the text, and illustration of the typical effects of sonication on particle size as afunction of delivered sonication energy: (A) Asymptotic behavior (solid line) versus (B) peaking behavior (dashed line).


Nan

otox

icol

ogy

Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

NIS

T N

atio

nal I

nstit

iute

s of

Sta

ndar

ds &

Tec

hnol

ogy

on 1

2/03

/10

For

pers

onal

use

onl

y.

processes (Kraft 2005; Isfort and Rochnia 2009). Insuch processes, chemical precursors react at hightemperatures to produce small ‘hot’ nanoscale dro-plets of the desired material. As these hot dropletspass through the reaction zone, they collide and maycoalesce fully or partially, depending on the localtemperature. Full coalescence leads to a larger pri-mary particle, while partial coalescence leads to theformation of an aggregate of fused primary particles(bound via sinter necks), which later become ‘solidbridges’ upon particle solidification (Figure 4). Par-ticle coalescence and growth continue until the reac-tants are consumed or the process temperature isquenched (Singhal et al. 1999; Zachariah and Carrier1999; Teleki et al. 2008; Isfort and Rochnia 2009).Metal oxide and other ceramic powders frequently

show this configuration, wherein the ‘primaryparticle’ represents the solidified globules that con-stitute powder aggregates (frequently chain-like struc-tures). Yet, it must be noted that the solid bridges thathold the primary particles together in the dry powderaggregates are often too strong to be broken viasonication or other moderate dispersion processes.Therefore, in some cases, it can be unrealistic toexpect sonicated powders to break down to theirreported, nominal primary particle size. As a result,powders will be effectively de-agglomerated to aggre-gates of several primary particles (i.e., ‘primaryaggregates’), rather than isolated primary particles.These aggregates should be thought of as the effectiveprimary size for the ENP. Although the underlyingformation processes in solution phase synthesis (e.g.,via sol gel or precipitation) differ from the vapor phaseprocess, the resulting structures often express similarphysical properties to the above described structures(i.e., particle fusion, presence of aggregates andagglomerates), though typically to a lesser extent(Brinker and Scherer 1990).During sonication, shock waves from cavitational

collapse are principally responsible for powder break-age. The particles, in fact, serve as nuclei to initiate thecavitation process, which partly explains why sonica-tion is so effective (Suslick 1988). When ultrasoundenergy is applied to powder clusters in suspension,fragmentation can occur either via erosion or fracture.Erosion or ‘chipping’ refers to the detachment ofparticles from the surface of the parent agglomerates,while fracture or ‘splitting’ occurs when agglomeratespartition into smaller agglomerates or aggregates dueto the propagation of cracks initiated at surface defects(Figure 1). These fragmentation effects may occursimultaneously or in isolation, depending on the pow-der, its environment and the energy levels involved.For a given material in powder form, depending on

the physicochemical properties of the solid material

and method of manufacture, the occurrence of sur-face defects in the powder clusters, the material’ssurface toughness, and the bonding nature (e.g.,neck diameter or interfacial contact area) of aggre-gates and agglomerates, there is a system-specificacoustic energy threshold that must be crossed inorder to achieve agglomerate and aggregate breakage.It is thus worth noting that two powders of the samematerial, but obtained following different syntheticroutes, may show significantly different breakagebehavior. Material properties also affect the rate atwhich fragmentation occurs, and, consequently, theamount of time it takes at a given ultrasonic outputpower and for a given amount of material to reduce allof the powder to the smallest achievable dimensions.Thus, in principle, the cavitation process can effec-

tively break down powder agglomerates. Yet, undercertain conditions, the applied acoustic energy canconversely induce particle agglomeration or even leadto the formation of aggregates. Coagulation in ultra-sonic fields can occur from enhanced particle-particle interactions due to the increased collisionfrequency as well as the favorable reduction in freeenergy that accompanies the resulting reduction in theliquid-solid interface. As noted previously in discuss-ing the basics of ultrasonic disruption, cavitation givesrise to extreme local pressure and temperature gra-dients, as well as shock waves and jet streams of theorder of hundreds of meters per second. Under suchconditions, sonicated particles in the neighborhood ofa collapsing bubble collide with each other whilesimultaneously experiencing localized intense heatingand subsequent cooling cycles. Depending on theeffective energy delivered to the particles and thethermal properties of the material and medium, theseconcomitant effects can lead to re-agglomeration oreven thermally induced inter-particle fusion; that is,sonication-induced aggregate formation.Published studies have demonstrated the existence

of this phenomenon, which is often observed as anincrease in particle size beyond a given sonicationenergy threshold (Aoki et al. 1987; Vasylkiv and Sakka2001). In other cases, an asymptotic behavior isobserved, wherein sonicated powders break downto particles of a given size value, beyond which nofurther significant size decrease is obtained even athigher ultrasound energy inputs (Bihari et al. 2008;Murdock et al. 2008). The latter behavior furtherconfirms that, for some powders, primary particlescan be held together by bonds that cannot be brokenby sonication.Interestingly, both asymptotic and coagulation

behaviors (see Figure 4) as a function of differentsonication parameters have been observed for thesame source material (e.g., TiO2, Mandzy et al.


Nan

otox

icol

ogy

Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

NIS

T N

atio

nal I

nstit

iute

s of

Sta

ndar

ds &

Tec

hnol

ogy

on 1

2/03

/10

For

pers

onal

use

onl

y.

2005; Jiang et al. 2009), illustrating the system-specific dependence (i.e., source powder, dispersingmedium, ultrasound input) of powder sonicationprocedures. Asymptotic and peaking (coagulation)behaviors upon sonication have been reported for awide variety of engineered nanomaterials. The fol-lowing are some examples, where the respectiveasymptotic or peaking energy thresholds are alsoindicated in each case: TiO2 (rutile) – 60 s at 7 W(Bihari et al. 2008); TiO2 (anatase and rutile) – (100–300) s at 750 W (Jiang et al. 2009); ZrO2 – 100 s at160 W (Vasylkiv and Sakka 2001); Au – 60 min at40W � cm-2 (Radziuk et al. 2010); Ag – 10 s at (35–40)W (Murdock et al. 2008). It must be noted that inthese cases the reported delivered energies do notconform to the guidelines specified herein and there-fore may not accurately reflect the actual energydensities to which the ENPs were subjected.Furthermore, while the disruptive effect of sonica-

tion is in most cases instrumental to the achievementof nanoscale particles, it can also result in eitherdesired or undesired physical changes that may alterthe inherent properties of the dispersed material.A mainstream example of this phenomenon is thesonication induced scission of carbon nanotubes(Forrest and Alexander 2007; Huang et al. 2009;Lucas et al. 2009).While it may not be possible to completely elimi-

nate all of the aforementioned phenomena, it is crit-ical to optmimize the sonication procedure in order tominimize the undesirable effects, while attaining thedesired level of particle disruption. The recom-mended optimization steps are detailed in a subse-quent section.

Chemical effects of sonication

As with size, surface chemistry – and consequentlysurface charge – plays a critical role in governing theenvironmental and biological interactions of ENPs.Surface chemistry, whether it is that of the pristinematerial or includes the effects of a surface coating orcapping agent, largely determines the adsorptionbehavior at the solid-solution interface, the organic/inorganic phase partitioning, colloidal stability inaqueous media, reactivity with other medium com-ponents, and affinity towards cell membrane walls(Guzman et al. 2006; Li and Huang 2008; Jiang et al.2009; Kittler et al. 2010; Walczyk et al. 2010), toname just a few examples. Surface chemistry is thus akey parameter that affects the fate, transport, bioavail-ability, and bioaccumulation behavior of ENPs. It istherefore critical to consider the potential for changesin surface chemistry in ENP suspensions arising fromthe application of ultrasonic energy.

The extreme localized temperatures and pressuresgenerated by the cavitation process can yield highlyreactive species within a sonicated medium, in aprocess known as ‘sonic activation’. As an indicationof the importance, variety, and complexity of sonica-tion induced physicochemical effects in liquid media,in recent years, sonochemistry has emerged as a distinctdiscipline devoted to the study of a wide range ofsonication induced chemical reactions in both homo-geneous and heterogeneous solid and liquid systems(Mason 1989; Mason and Peters 2003).An example of sonic activation of particular rele-

vance to environmental and toxicological studies isthat of water sonolysis. Sonicated water partly dis-sociates into hydrogen and hydroxide radicals, withconcentration and lifetime dependent on the sonica-tion conditions (e.g., sonication time and energydensity) (Makino et al. 1983; Yanagida et al.1999), that subsequently recombine to form hydrogenperoxide (Brown and Goodman 1965; Mason andPeters 2003). Any materials or chemical species pres-ent in sonicated water will thus be exposed to low (andvariable) concentrations of highly reactive species,and as a result may undergo oxidative and otherchemical transformations.Uncoated, sonicated ENPs may undergo changes

in their surface chemistry due to the formation ofsonolysis radicals. As previously mentioned, the soni-cation of water can yield reactive radicals that caninitiate a variety of chemical reactions. The generatedhydrogen and hydroxyl radicals, and peroxide mole-cules, can interact with the material’s surface, chang-ing its oxidation state or inducing the hydroxylation ofthe materials’ surface, potentially altering the materi-al’s hydrophilicity and stability, as well as its chemicalcompatibility with other medium components.Chemisorbed hydrogen radicals generated duringsonication could potentially lead to the irreversibledisintegration of sonicated carbon nanomaterials,such as fullerenes and nanotubes, via ‘hydrogen-induced exfoliation’ (Berber and Tomanek 2009).Other changes in the material’s surface chemistry,

such as oxidation or hydroxylation, may result inan increased aqueous solubility and result in anenhanced leaching of ionic or soluble species intothe medium. The potentiality of sonication-enhanced leaching should be given special consider-ation when testing the biological interactions of ENPswith known toxicity in their molecular or ionic form; acase in point is nanoscale silver (Wijnhoven et al.2009; Elzey and Grassian 2010), in which the oxi-dized Ag+ form is well documented to exhibit biocidalproperties.Moreover, if the dispersion medium is not just

water, but incorporates other chemical species, the


Nan

otox

icol

ogy

Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

NIS

T N

atio

nal I

nstit

iute

s of

Sta

ndar

ds &

Tec

hnol

ogy

on 1

2/03

/10

For

pers

onal

use

onl

y.

extent and degree of formation of radicals from bothwater and other medium components, and the com-plexity of their interactions with the dispersedmaterialcan be further magnified. Such potentiality for soni-cation induced changes in the surface chemistry, andin particular its oxidation state or the hydroxylation ofthe material’s surface, should be given special con-sideration in light of recent findings that attributeENP’s toxic effects to oxidative stress and reactiveoxygen species (ROS)-mediated events, among othermechanisms (Unfried et al. 2007; Wang et al.2009a; Kim et al. 2010).While radical induced changes in the pristine mate-

rial’s surface might be restricted to the period ofsonication treatment and be potentially reversible –

although to the best of our knowledge this hasyet to be determined – it is unquestionable thatsonication-specific, radical-mediated changes havebeen observed as a result of sonication, as discussedbelow, and can significantly alter the suspensionmedium properties or the dispersed material’s surfacechemistry, and consequently influence its environ-mental and biological behavior.If organic molecules are present when sonicating

ENP suspensions, either as components of the dis-persing medium or as a surface coating on the ENPitself, they can be irreversibly degraded via directsonolysis from the dissipated cavitational energy(Williams 1983; Wang et al. 2009c). Furthermore,sonication-induced active radicals, such as those pro-duced from water sonolysis, have been shown tocleave polymers creating smaller oligomers and toactivate polymerization reactions (so called sonopo-lymerization), producing new polymeric species andirreversibly modifying the organic molecules presentin the suspension (Brown and Goodman 1965;Mason 1989). Sonication can also induce covalentbonding of organic functionalities to the ENP surface(Liu et al. 2005). The irreversible degradationand/or transformation of organic molecules due tosonication is a potentiality that should be consideredwhen preparing ENP suspensions for environmentalor biological testing.As a result of sonolysis, the sonicated medium may

not retain its native properties. Degraded mediumcomponents, such as denatured proteins, will likelyinteract differently with the ENPs or the intendedassay, than they would in their pristine, non-sonicated form. Sugars and proteins are two majorbiological and environmental media components.Basedow and Ebert studied the molecular degrada-tion of dextran as a function of exposure timeunder 20 kHz, low intensity (10 W cm-2) sonication(Basedow and Ebert 1979), while in a laterwork Lorimer et al. (1995) extended the study to

account for the effect of sonication intensity, temper-ature, and sugar concentration. In all cases, a signif-icant decrease in the sugars’ molecular weight atincreasing sonication times and powers was observed.Protein and enzyme sonolytical degradation has

been broadly studied as well. Wang et al. (2009c)examined the sonication-induced damage to bovineserum albumin (BSA) both in the presence andabsence of metal complexes. In all cases, even atsonication powers as low as 20 W, ultrasonic irradi-ation was shown to damage the BSA structure, theextent of which increased with increasing sonicationtime. Coakley et al. (1973) measured the sonolyticalinactivation of alcohol dehydrogenase and lysozymeexposed to a 20 kHz ultrasonic field as a function ofsonication time (at 90 W), and measured a 70%inactivation of alcohol dehydrogenase and lysozymeafter 10 and 25 min of sonication, respectively.A more comprehensive review and discussions onsonication induced alterations of relevant biologicalcompounds (e.g., enzymes, DNA, biological mem-branes) can be found in the works of Williams (1983)and Riesz and Kondo (1992).Alterations in the molecular structure of common

organic compounds (e.g., sugars, proteins) as a resultof sonication may elicit specific biological responses inthe animal model or assay upon inoculation with thesonicated suspension, thereby yielding false positivesor negatives that could incorrectly be attributed to theENPs, rather than the degraded medium. To accountfor this possibility, we recommend using a superna-tant control in which the medium, in the absence ofENPs, is subjected to the sonication procedure andtested in parallel with the ENP suspensions.Furthermore, sonication-induced transformations

in the medium could potentially result in a loss ofthe medium’s stabilizing properties, therefore reduc-ing its ability to maintain the ENPs in suspended non-agglomerated form. For example, species-specificalbumin homologs, such as BSA, have been utilizedas stabilizing components in dispersion mediaintended for toxicity testing of ENPs (Bihari et al.2008; Porter et al. 2008; Kittler et al. 2010;Tantra et al. 2010), while humic acid is commonlyused as a medium component and organic stabilizeradded to ENP suspensions to simulate environmentalmatrices (Chen and Elimelech 2007; Pallem et al.2009; Saleh, Pfefferle and Elimelech 2010). Typi-cally, the albumin or humic acid is present whenthe suspension is sonicated. Naddeo et al. (2007)demonstrated that humic acid is degraded (eithervia oxidative mechanisms or physical aggregation)when subject to 20 kHz sonication, the extent ofwhich was found to increase with sonication intensityand organic concentration. Preliminary studies in our


Nan

otox

icol

ogy

Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

NIS

T N

atio

nal I

nstit

iute

s of

Sta

ndar

ds &

Tec

hnol

ogy

on 1

2/03

/10

For

pers

onal

use

onl

y.

own laboratory on the stabilization effect of BSA onsonicated TiO2 powders have shown a measurableloss of nanoparticle stability against agglomerationwhen BSA is sonicated together with the sourcepowder, as opposed to when BSA is added post-sonication. In the latter case, the particles did notexhibit agglomeration over the same time frame.These results will be published as part of a separateresearch article currently in preparation.Moreover, the ultrasonic degradation of organic

molecules has the potential to irreversibly modifythe chemical structure of ENP functional surfacecoatings (e.g., polyethylene glycol [PEG], and poly-vinylpyrrolidone [PVP]), or even a complete loss ofthe coating itself. Kawasaki et al. (2007) studied theultrasonic degradation of PEG as a function of son-ication time at 28 kHz and 20 W, over a wide range ofmolecular weights. In all cases, radical mediatedsonolysis caused C-O bond scission and moleculebreakage. Taghizadeh and Bahadori (2009) measuredthe degradation kinetics of PVP when subject to24 kHz 100 W sonication, revealing a dependenceof the degradation rate on molecular weight andpolymer concentration. Such potential sonication-specific alterations of organic surface coatings mayadversely impact the ENP’s capacity to remain insuspension, as particle coatings often provide protec-tive electrostatic or steric stabilization. Ultrasonicdegradation of organic surface coatings could alsomodify the ENP’s interactions with other mediumcomponents, interfering with the ENP’s transport,adsorption, or toxicological behavior.Additionally, ENPs may act as sonolysis catalysts,

synergistically enhancing the aforementioned effectswhen sonicated in the presence of organics.Wang et al demonstrated the sonocatalytic effect ofnano-sized silica and titanium dioxide towards thedegradation of BSA, showing an enhanced degrada-tion of BSA when the ENPs were present in thesonicated solution, with respect to ENP-free controls(Wang et al. 2005, 2009b).For these reasons, the chemical complexity of the

sonicated suspension and the particle coating shouldbe reduced to the greatest possible extent. Wheneverpossible, organic compounds should be spared fromsonication and the formation of radicals minimized.A proposed approach for the preparation of ENPsuspensions is, when applicable, to first sonicate thepristine, uncoated powder in de-ionized water (orother appropriate pH adjusted aqueous solution)and then mix the sonicated suspension into thedesired, non-sonicated, dispersion medium, or addthe stabilizing agent and other organic additives to thesuspension after sonication. This may require signifi-cant additional effort to properly formulate and

disperse the ENPs, but the extra effort will result inless perturbation of the chemical species present. Theformation of radicals can also be mitigated by mini-mizing the sonication power, again prompting theneed for optimizing the dispersion procedure to min-imize undesired effects while obtaining the desireddispersion properties. Furthermore, nitrogen or car-bon dioxide bubbling, or the addition of dry ice orother radical scavengers, can quench free radicalsduring the sonication process and minimize theirpotential impact (Kondo et al. 1986; Honda et al.2002).

Measuring and reporting delivered power

A critical parameter for the reproducible preparationof sonicated ENP suspensions is the delivered soni-cation power. As previously mentioned, only a por-tion of the total output power is effectively consumedin particle disruption.The energy loss and consequently the efficiency of

the energy transformation (or the ‘energy yield’)from the electrical source to the acoustic powereffectively received by the sonicated suspension isheavily instrument dependent (on the instrument’sparticular driving circuitry, self-adjusting feedbackloops, and its various electrical components). Fur-thermore, the displayed power omits thermal andmechanical energy losses, which are also instrumentspecific, that occur as the output power is transferred‘downstream’ through the sonication probe and intothe suspension (Figure 3).Therefore, simply reporting the sonicator’s power

setting value (e.g., ‘setting 5’ or ‘60%’) or the dis-played power does not provide an accurate indica-tion of the effective acoustic power delivered to thesonicated suspension, and are as such reportingparameters that do not allow for transferable orreproducible results. Two different instrumentsoperating at a ‘level 5’, or a ‘60%’ setting can deliversignificantly different effective acoustic powers to thesame suspension, as they may translate into differentoscillation amplitudes for different instruments.Alternatively, instrument ‘A’ may consume – anddisplay – 50 W from the electrical source to producea probe oscillation amplitude of 3 mm, while instru-ment ‘B’ may produce a probe oscillation amplitudeof 2 mm by consuming 50 W, thereby deliveringdifferent powers to a suspension even when showingthe same displayed power. Published works havediscussed and demonstrated that the power readingdisplayed by the instrument bears poor correlationwith the measured power received by the sonicatedsuspension (Contamine et al. 1995; Kimura et al.1996; Raso et al. 1999). This issue is broadly known


Nan

otox

icol

ogy

Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

NIS

T N

atio

nal I

nstit

iute

s of

Sta

ndar

ds &

Tec

hnol

ogy

on 1

2/03

/10

For

pers

onal

use

onl

y.

in the sonochemistry field as ‘the reproducibilityproblem’ (Mason 1991; Contamine et al. 1995;Kimura et al. 1996).While there is yet no available standardized proce-

dure for reporting the effective acoustic energy deliv-ered to liquid suspensions, we hereby propose thecalorimetric method as a simple procedure toestablish a standardized, instrument-independent,transferable and reproducible method for the mea-surement and reporting of sonication power. It isour belief that by implementing such an approach,variabilities derived from the inconsistent applica-tion and reporting of ultrasonic disruption will bereduced.

The calorimetric method

We propose a simple and easily transferable set ofexperimental guidelines for the determination of calo-rimetric curves. The calorimetric method has beenproposed to standardize sonochemistry studies andused to calculate the amount of acoustic energy deliv-ered to a liquid medium subject to direct sonication(Kimura et al. 1996; Zhu et al. 2009). Calorimetry is arelatively simple, rapid, and inexpensive procedurethat allows for the direct measurement of the acousticenergy delivered to a sonicated liquid (Mason 1991).The method is based on the measurement of thetemperature increase in a liquid medium over time,as a bulk effect of the cavitational process induced in aliquid by an ultrasound probe.At a given sonicator setting, the temperature

increase in the liquid is recorded over time and theeffective delivered power can then be determinedusing the following equation:

P =dTdt

MCp , ( )2

where P is the delivered acoustic power (W), T and tare temperature (K) and time (s), respectively, Cp isthe specific heat of the liquid (J g-1 K-1) and M is themass of liquid (g).

By obtaining calorimetric curves following a stan-dardized procedure, users can report and reproducethe power level applied to a sonicated suspension in amanner that is easily transferrable and instrumentindependent.

The following protocol is proposed for the deter-mination of direct sonication calorimetric curves(Taurozzi et al. 2010a):

(1) Fill a 600 mL cylindrical borosilicate beakerwith 500 mL of de-ionized water (resistivity>18 MW � cm);

(2) Determine the mass of the liquid using a top load-ing balance. First tare the empty 600 mL beaker;

(3) Immerse the sonicator probe (horn) approxi-mately 2.5 cm (1 inch) below the liquid surface;

(4) Immerse a temperature probe connected to atemperature meter and data logger (e.g., anExtech HD 2001 temperature meter and data-logger coupled to a Type K immersion temper-ature probe). The temperature probe tip shouldbe about 1 cm from the sonicator probe (seeFigure 5);

(5) Select a sonicator output power setting and,operating in continuous mode, record the watertemperature increase for the initial 5 min. Dur-ing sonication, ensure that the beaker does notshift position, especially when operating at highpower settings; this can be accomplished forinstance, using a clamp attached to a ring stand;

(6) Using the recorded values, create a temperaturevs. time curve and obtain the best linear fit forthe curve using least squares regression.

Figure 5. Set-up for the measurement of calorimetric curves. Insetshows sonicator tip and temperature probe immersion depths.Sound protection enclosure was omitted for clarity of the image.

1. Certain trade names and company products are mentioned in the text oridentified in illustrations in order to specify adequately the experimentalprocedure and equipment used. In no case does such identification implyrecommendation or endorsement by National Institute of Standards andTechnology, nor does it imply that the products are necessarily the bestavailable for the purpose.


Nan

otox

icol

ogy

Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

NIS

T N

atio

nal I

nstit

iute

s of

Sta

ndar

ds &

Tec

hnol

ogy

on 1

2/03

/10

For

pers

onal

use

onl

y.

With the obtained slope, calculate the delivered powerusing Equation 2. The procedure should be repeatedfor all desired power settings, relating each settinglevel to the measured power. The power obtained viacalorimetry is the value we recommend for reportingpower used in sonication-based dispersion proce-dures. By doing so, laboratories using different com-mercial ultrasonic disruptors should be able toreproduce a similar delivered power by selecting theappropriate calibrated setting. It is important to notethat the intended use of calorimetry curves is not tomeasure the actual fraction of power utilized forpowder disruption under specific dispersion condi-tions, but simply to establish a standardized proce-dure to relate instrument setting levels to measurablepowers under a fixed set of controlled conditions, andthus allow for the reporting and transference of son-ication power levels between laboratories and users.

The robustness of the method was tested underdifferent scenarios, including variations in tempera-ture probe location, water volume, and the use ofstirring. While variations in temperature probe loca-tion and stirring only resulted in minor variations ofthe measured power (< 5%), it should be emphasizedthat using a different water volume, a container of adifferent material or a different type of horn tip thanthose recommended here can result in changes in themeasured temperature versus time slope for a givenpower setting due to variations in the rate of heattransference between the sonicated water and theenvironment. If conditions other than those recom-mended here are used to measure calorimetric curves,

these modifications to the protocol would need to bereported.Figure 6 shows calorimetric curves and calculated

delivered power at different sonicator settings,obtained using the above method for a Branson450 analog probe sonicator with a standard 1/2-inchprobe and tip.

Additional considerations

As noted previously, sonication is a highly systemspecific dispersion procedure, involving a variety ofcomplex concomitant physicochemical interactions inboth the sonicated ENP and the suspension medium.Therefore, for a given system, optimal sonicationconditions can only be determined by consideringthe effects of a variety of sonication parameters onthe dispersion state of the suspension under a broadrange of relevant conditions. As follows, a discussionof the effects of these parameters is offered along withrecommendations for implementation of ultrasonicdisruption as a preparatory tool. Two parameters,namely powder properties and sonication power,are not discussed here as they have already beenaddressed in previous sections.

Temperature

During sonication, the extreme local heating cyclesthat take place at the micro-scale bubble interface dueto cavitation will result in bulk heating of the liquidover time. Excessive bulk heating can cause

1.9

Setting 1, 8.85 W

Setting 3, 32.71 W

Setting 6, 60.50 W

Setting 8, 104.1 W

Setting 10, 134.7 W

1.8

1.7

1.6

1.5

1.4

Nor

mal

ized

tem

pera

ture

(T

/T0)

1.3

1.2

1.1

10 50 100 150

Time (s)

200 250 300

Figure 6. Calorimetric curves, linear fits, and corresponding calculated power values for different operational settings, obtained using aBranson 450 analog sonicator. For ease of illustration, the y-axis shows the recorded temperatures normalized to the initial temperature (T0) ineach case.


Nan

otox

icol

ogy

Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

NIS

T N

atio

nal I

nstit

iute

s of

Sta

ndar

ds &

Tec

hnol

ogy

on 1

2/03

/10

For

pers

onal

use

onl

y.

evaporative loss of liquid, resulting in changes in thesample volume after sonication. Moreover, high sus-pension temperatures may lead to degradation of theENP or other medium components if proper precau-tions are not employed.A common approach to minimize temperature

driven side effects is to keep the sample from expe-riencing substantial high-temperature excursions byimmersing the suspension container in a cooling bath.The container should be immersed to a level roughlyequal to that of the contained suspension.An ice-water bath is often sufficient to keep suspen-

sions from overheating. If, however, ice water is notsufficient to cool the sample under intense sonication(e.g., the sample volume is too low, the sonicationtime is too long, or high sonication powers are used)an ice-salt bath could be considered as an option.Working with containers made of materials with

high thermal conductivities ensures a rapid release ofthe generated heat from the suspension into thesurrounding cooling bath. With regard to thermalproperties, the following container materials workbest, in order of decreasing thermal conductivity:aluminum, stainless steel, glass, and plastic (e.g.,polyethylene). When selecting the container, consid-eration must also be given to the chemical compat-ibility between the container material and thesuspension components, especially under intenselocal heating conditions; for instance, aluminum isincompatible with acidic suspensions, and glass isincompatible with alkaline solutions.Additionally, an aluminum foil or thin thermoplas-

tic film with an opening just large enough for theprobe to enter untouched, will reduce the evaporativeloss of liquid content, especially when sonicating involatile solvents or for substantial durations. Coolingthe sample will also reduce evaporative loss. Thiscover will also minimize the potential release of aero-sols generated by the sonication process.

Sonication time and operation mode

The total amount of energy (E) delivered to a sus-pension not only depends on the applied power (P)but also on the total amount of time (t) that thesuspension is subject to the ultrasonic treatment:

E = P t× ( )3

Consequently, two suspensions treated at the samepower for different times can show significantly dif-ferent dispersion states.

Ultrasonic disruptors typically can operate in eithercontinuous or pulsed mode. In pulsed mode, ultra-sonic intervals are alternated with static (sonication

off) intervals. The duration of on and off intervals canbe regulated. Operating in pulsed mode retards therate of temperature increase in the medium resultingfrom the cavitation process, minimizing unwantedside effects and allowing better temperature controlthan with continuous mode operation.

Sample volume and concentration

While sonication power and time describe the amountof energy delivered to the suspension, the work doneon samples of different volumes and particle con-centrations can differ. At constant volume, higherparticle concentrations result in an increased particlecollision frequency. In principle, an increased colli-sion frequency can enhance particle breakage due toan increase in particle-particle impact events. How-ever, if sufficient local activation and sintering ener-gies are achieved, increased collision frequencies canalso induce agglomerate or aggregate formation asparticles collide and coalesce (Figure 4). The effect ofconcentration is thus dependent on both the energydelivered into the suspension and the physiochemicalproperties of the suspension.Particle concentration can also alter the acoustic

properties of the suspension, affecting the medium’sviscosity and acoustic conductivity and thus theeffectively delivered power. However, at the concen-trations commonly used for environmental andtoxicological evaluations, typically in the 50–2000 mg � mL-1 range, a significant impact on themedium’s acoustic properties is not anticipated.Where particle concentration may alter acoustic prop-erties is when working with concentrated stock sus-pensions prior to dilution.The effect of suspension volume (at constant par-

ticle concentration) is often measured as energy density(W s mL-1). This magnitude expresses the amount ofdelivered energy per unit of suspension volume. Inprinciple, at equal power and particulate concentra-tion, higher energy densities (i.e., smaller suspensionvolumes) will result in a higher disruptive effect. Itshould be emphasized that when working with smallvolumes the temperature of the suspension will rise atfaster rates and more intense cooling conditions maybe required in this case.

Sonicator probe, container geometry, and tip immersion

The sonicator probe is the acoustic element thatconducts the acoustic energy from the transducerinto the sonicated suspension. The amount of acous-tic energy transferred to the suspension will heavilydepend on the shape of the sonicator probe and itsimmersion depth in the suspension.


Nan

otox

icol

ogy

Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

NIS

T N

atio

nal I

nstit

iute

s of

Sta

ndar

ds &

Tec

hnol

ogy

on 1

2/03

/10

For

pers

onal

use

onl

y.

The vibrational amplitude of the horn at a givenpower is dependent on the horn taper (Berlan andMason 1992). Tapered probes act as vibration ampli-tude magnifiers, and consequently allow for higherdelivered powers (see Equation 1). For example,stepped horn probes allow for up to 16-fold poweramplifications, while exponentially tapered probescan yield power densities on the order of hundredsof W � cm-2 (Mason 1989).At equal sonicator power settings, microtip probes

vibrate with larger amplitudes than conventional flat tiphorns. However, microtips are lessmechanically robustand often limited in terms of the maximum powersetting at which they can be used. The manufacturer ofthe sonicator will typically specify a maximum powersetting (based on a percentage of the maximum poweroutput setting) for use with microtips.The way in which the sonic energy is distributed

within the suspension is also heavily influenced by thecontainer geometry. For example, in one publishedstudy (Pugin 1987) acoustic energy distribution pro-files measured in round bottom flasks subject toindirect sonication significantly differed from thoseof flat bottom (e.g., Erlenmeyer) flasks subject tosimilar sonication conditions. Conical bottom, flatbottom, and round bottom flasks will also showdifferent energy maxima and minima distributionprofiles, which will in turn vary significantly for dif-ferent probe tapers and probe tip immersion depths.When possible, using the smallest diameter vessel

that allows for the probe to be inserted withouttouching the container walls is probably the bestapproach. Using smaller container diameters raisesthe height of the liquid and maximizes the liquid-probe surface area exposed to the acoustic waves, aswell as the container wall surface/volume ratio fordissipation of heat by the cooling bath. Cylindricalshaped beakers work well, especially for smallvolumes.Probe immersion depths between 2 and 5 cm are

generally recommended when operating with stan-dard horns having a flat radiating surface (tip) or withmicrotips. Probes should be placed no closer thanabout 1 cm from the bottom of the sample container,and contact between the probe and the container wallsshould always be avoided.

Aerosoling and foaming

If the sonicator probe is not sufficiently immersed inthe suspension it can give rise to surface agitationresulting in nebulization (i.e., the formation andrelease of aerosols). This could pose a risk if ENPsor other potentially harmful medium components arereleased in this manner.

Aerosoling may be indicated by fluctuations in theaudible sound pitch during operation, fluctuatingpower readings, or by the appearance of a fine sprayin the vicinity of the probe. If aerosoling is observed, itcan usually be eliminated by lowering the sonicationprobe deeper into the suspension.Additionally, if surfactants are present, the suspen-

sion could generate a foam during sonication, and thepresence of a stable foam in contact with the hornsurface will interfere with the delivery of ultrasonicenergy to the suspension in a self-limiting process.Pulse mode operation with long off periods will helpavoid foaming in samples subject to this effect. Theappearance of foaming may require that operation isceased until the foam dissipates. Anti-foaming agentscan be effective, but would probably not be compat-ible with toxicological studies.

Tip maintenance

As the source of ultrasonic waves, all sonicator probesare subject to progressive erosion due to the intensecavitation that occurs in the immediate vicinity of theprobe tip. Tip erosion is therefore an unavoidable sideeffect in direct sonication. There are two effects of tiperosion that are relevant to the present discussion.During erosion, microscopic tip residues (typicallytitanium metal) are released from the tip into thesonicated suspension, introducing an impurity andpotentially contaminating the suspension. Operation-ally, tip erosion also results in a reduced energyoutput. The extent of potential contamination fromtip erosion can be evaluated by performing a controlexperiment, or by measuring the amount of metalreleased into the sonicated medium (e.g., via diges-tion followed by ICP-MS analysis) under relevantsonication conditions. Erosion contamination canbe minimized by adequate and frequent tip mainte-nance procedures, as explained below, or by usingindirect sonication.The extent and occurrence of tip erosion depends

on the intensity, duration, and frequency of sonica-tion cycles, as well as the chemical properties andconcentration of the sonicated suspensions. If used ona daily basis, tips should be inspected weekly for signsof erosion; tip erosion can be recognized by theappearance of a grayish matting, as opposed to theshinny lustrous appearance of a non-eroded tip (seeFigure 7).A mildly eroded (i.e., matted) tip or horn surface

can be reconstituted (dressed) by buffing with a veryfine grit carbide paper or emery cloth (refer to the tipor sonicator manufacturer for the correct grit size; donot use conventional sand paper) (Berliner 2010).Substantially eroded tips or horn surfaces (i.e., with


Nan

otox

icol

ogy

Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

NIS

T N

atio

nal I

nstit

iute

s of

Sta

ndar

ds &

Tec

hnol

ogy

on 1

2/03

/10

For

pers

onal

use

onl

y.

visible pitting) should be replaced, as these cannot bereconstituted. Standard 1/2-inch diameter horns typ-ically have a removable threaded tip, but larger probesare likely to consist of a solid horn and thus greatercare may be required to dress the horn’s active surfaceif it becomes matted.It is also essential to ensure that removable probe

tips are tightly coupled to the horn, and do not loosenas a result of prolonged operation; a loose tip willresult in a reduced acoustic power conversion andenergy yield.

Medium properties

As previously noted, the ultrasonic energy delivered toa suspension is partly attenuated and dissipated by thesuspending medium. The amount of energy lost to themedium at a given acoustic frequency is governed byits physicochemical properties and chemicalcomposition.Generally, the attenuation of acoustic energy in a

given medium can be estimated using the followingequation (Brown and Goodman 1965):

I = I exx

02 4− α ( )

where I0 is the acoustic intensity (power/area) at thesource, Ix is the intensity at a distance x from thesource and a is the absorption coefficient.

The magnitude of the absorption coefficient, a, isdependent on the medium’s physicochemical prop-erties and composition, as well as the acoustic fre-quency. In principle, higher medium viscosities andacoustic frequencies will result in larger a values. Fora given liquid, a can be estimated by using thefollowing equation (Brown and Goodman 1965):

αμρ

∝ 23

5f 2

( )

where m is the dynamic viscosity of the liquid (kg m-1

s-1), f is the acoustic frequency (s-1) and r is theliquid’s density (kg m-3).

As shown by Equation 5, acoustic attenuationincreases with frequency for a given medium. Forthis reason, to achieve equivalent energy densities,higher power inputs are required when operating athigher frequencies. For example, to achieve an acous-tic intensity of 20 W � cm-2 at a distance of 10 cm,would require approximately 54% greater sourceintensity at 10 MHz compared with that required at10 kHz (Mason 1989).Additionally, higher viscosities will dampen the

cavitation process, requiring higher power inputs toachieve dispersion. The medium’s density and acous-tic wave speed (c) will also impact on the amount ofvibrational energy that is transformed into acousticenergy (see Equation 1).The medium’s viscosity, density, acoustic wave

speed, and chemical composition are all importantparameters that impact on the amount of deliveredenergy that is effectively utilized to disrupt powderclusters. For media relevant to environmental andbiological testing, the salt content is likely to haveonly a very small impact on the acoustic propertiesrelative to pure water, as the presence of physiologicalsalt levels will slightly alter the density and viscosity ofthe solution. Overall, at ultrasound frequencies, theseeffects are probably insignificant compared to otherfactors discussed previously.Yet, as noted previously, the medium’s chemical

composition is of particular relevance in terms of theoccurrence of potential side reactions during sonica-tion (e.g., radical activation, sonopolymerization) thatmay alter the physicochemical properties of the son-icated medium and/or the sonicated nanomaterial.

Optimizing sonication for dispersion of ENPs

As discussed in preceding sections, the treatmentconditions required to achieve complete ENP disper-sion are highly system specific. It is therefore advi-sable to conduct dispersion optimization studies (e.g.,measure particle size as a function of sonicationparameters) for the test system (Bihari et al. 2008).While it would be unrealistic and impractical to offer a

Figure 7. Active surface of removable 1/2-inch flat tips, ranging from pristine (left) to matted (center) to eroded with extensive pitting (right).


Nan

otox

icol

ogy

Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

NIS

T N

atio

nal I

nstit

iute

s of

Sta

ndar

ds &

Tec

hnol

ogy

on 1

2/03

/10

For

pers

onal

use

onl

y.

unique set of sonication conditions for all possiblesystems, the ultimate goal of the optimization proce-dure should be to achieve the desired degree ofparticle dispersion with the least possible energy input,in order to minimize unwanted side effects.The process to determine such conditions is based

on a trial and error approach. The guidelines offeredbelow are intended as tools for the interpretation ofthe observed behavior at a given set of sonicationconditions, and to determine the appropriate param-eter modifications to be made in response.As noted previously, the relationship between dis-

persion efficiency and the different sonication para-meters is not linear. The convolution of deliveredenergy input and suspension properties can resultin either a net agglomerate breakage or cluster re-formation effect.For this reason, the optimization process should

scan parameter values covering a range both aboveand below a starting point. The first trial set ofsonication conditions for an untreated powder insuspension should be intended to ensure that disper-sion can in fact be achieved. Being the first trial andtherefore likely not the optimal set of conditions, itshould be designed to confirm that under reasonablyattainable sonication conditions, the powder in sus-pension can be effectively dispersed. The startingpoint of the optimization should therefore provide ahigh energy input to the suspension. This is achievedby selecting a relatively large sonication output powerlevel, a prolonged sonication time, and a high energydensity (small suspension volume).Table I offers a general guideline of starting points

for power, time, and volume selection, which mightbe utilized to investigate new ENPs or media/ENPcombinations (Hielscher 2005).The choice for the starting powder concentration is

subject to the desired application. It is also possible tostart with concentrations higher than the target level,and then if needed dilute the processed suspensionpost-sonication.The effect of different processing parameters on the

dispersion state of the suspension can be analyzedindependently by varying one parameter while keep-ing the others constant. For each parameter, theeffectiveness of the applied sonication conditions

can be evaluated by measuring the particle size dis-tribution (PSD) of the sonicated suspension. In prin-ciple, a reduction in the average particle size andpolydispersity (breadth of the distribution) are indic-ative of more effective dispersion. The parameteroptimization sequence is explained as follows.Once a starting set of values for time, concentra-

tion, and volume has been selected, and keeping allother processing parameters constant, sonicationpower is varied. Sonication powers above and belowthe starting point should be tested. As a generalguideline, the effect of power can be scanned inincrements of 10 W both above and below the startingsonication power. Using the power value that yieldsthe smallest measured size and lowest polydispersity,then sonication time should be scanned (e.g., inincrements of 30 s) above and below the startingpoint. Operation in pulsed mode is recommendedfor sonication times greater than 1 min, particularlywhen using small volumes (i.e., below 50 mL).For the optimized time, power, and selected vol-

ume, the effect of particle concentration both aboveand below the concentration starting point and downto the minimum acceptable concentration for thedesired application, may be evaluated if desired. Ifthe optimal concentration is above the desired valuefor the particular application, the suspension’s PSDshould be measured following dilution to ensure therehas been no change.The above described approach for optimizing the

sonication conditions for dispersion is both simpleand practical, but the authors recognize that theremay be other factors to consider or other optimizationapproaches that work equally well; it should be con-sidered as a recommendation not a prescription forsuccess.

Reporting guidelines

In light of the relevance of the above discussed para-meters and their significance in dictating the state andquality of suspensions obtained via sonication, wepropose that the following parameters be reportedin published work involving the use of ultrasonicenergy to disperse ENPs prior to environmental,exposure, or biological testing (Table II). It shouldbe noted that many of these parameters or materialcharacteristics would likely be reported anyhow, orcertainly should be according to several recent pub-lications and recommendations (Murdock et al. 2008;Warheit 2008; Minimum Information for Nanoma-terial Characterization [MinChar] Initiative 2009;Boverhof and David 2010).The lengthiness of this list may preclude full dis-

closure within a typical experimental section in a

Table I. Guidelines for sonication starting points.

Energy density(W � s/mL)

Samplevolume (mL) Power (W) Time (s)

<100 10 50 <20

100–500 10 50 20–100

>500 10 50 >100


Nan

otox

icol

ogy

Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

NIS

T N

atio

nal I

nstit

iute

s of

Sta

ndar

ds &

Tec

hnol

ogy

on 1

2/03

/10

For

pers

onal

use

onl

y.

Tab

leII.Listof

parametersan

dmaterialch

aracteristicsreco

mmen

dedforrepo

rtingpu

rposes.

Son

ication

ENPor

source

material

Testmed

ium

Final

suspen

sion

Typ

eof

sonicatorused

(bath,

prob

eim

mersion

,inverted

cup)

Description

ofpo

wde

ror

ENPas

received

(totheextent

relevant

andkn

own)

Com

position

includ

ingionicstreng

thor

molar

saltco

ncen

trations

Order

inwhich

compo

nentsweread

ded

andwhe

ther

orno

tthey

werepresen

tdu

ring

sonication

Son

icator

make,

mod

el,

ratedpo

wer

output,freq

uenc

y,prob

etype

/diameter

(ifused

)

For

commercial

ENPs,

prod

ucttrad

ena

me

and/or

prod

uctnu

mbe

rtogether

with

batch/lotnu

mbe

rifavailable

For

commercial

med

ia,source,prod

uct

andba

tchnu

mbe

r(s),as

applicab

lePost-treatm

entstorageco

nditions

priorto

testing(e.g.,co

ntaine

rmaterial,tempe

rature,

expo

sure

tolig

ht,expo

sure

toox

ygen

,time)

Sam

pleco

ntaine

rvo

lumean

dtype

Primarypa

rticle

size

andmetho

dof

measuremen

t(e.g.,TEM

)InitialpH

(preferablymeasured,

othe

rwise

nominal

buffer

value)

Whe

ther

endo

toxinwas

evalua

tedor

proc

edures

appliedto

avoid/remov

een

dotoxins

(relevan

tifsuspen

sion

isto

beused

for

toxicity

testing)

Tip/probe

immersion

depth(ifused

)Gross

particle

morph

olog

y(e.g.,sphe

roidal,

tubu

lar,

fibrou

s,platelet,aggregate)

Dispe

rsingagen

t(s)

andco

ncen

tration(s),

ifused

Con

dition

sof

filtration

and/or

centrifugation

used

(whe

reap

prop

riate)

Ope

ration

mod

e(con

tinu

ousor

pulsed

),ifpu

lsed

indicate

pulsedu

ration

Spe

cificsurfacearea

(for

powde

rs)an

dmetho

dof

measuremen

tSou

rcean

dqu

alityof

water

iftest

med

ium

was

prep

ared

ordilutedon

-site

Post-treatm

entmassco

ncen

trationof

particle

phase

Total

cumulativetreatm

enttime

Iden

tity

ofco

atings

orsurfacefunc

tion

ality

ifkn

own

Any

relevant

details

relating

tostorage,

dilution

,or

prep

arationof

med

ium

Final

pH

Pow

erinpu

tto

sample(asmeasuredfor

prob

esonicators:See

abov

ean

dalso

Tau

rozziet

al.(201

0a)

Particlesize

distribu

tion

intherelevant

suspen

ding

med

ium

andmetho

dof

measuremen

t

Whe

ther

coolingba

thwas

used

;ifno

ttempe

rature

rise

during

sonication

Massof

powde

ran

dtotalvo

lumeof

suspen

sion

sonicated


Nan

otox

icol

ogy

Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

NIS

T N

atio

nal I

nstit

iute

s of

Sta

ndar

ds &

Tec

hnol

ogy

on 1

2/03

/10

For

pers

onal

use

onl

y.

refereed journal, but could easily be accommodatedin the supplemental information supported by mostscientific journals. The recommendations may also beviewed as a checklist for identifying issues that mayneed to be addressed during the selection or devel-opment of a test material or dispersion procedure. Fora more detailed and formal guidance document onreporting, refer to Taurozzi et al. (2010b) or consultthe references cited above.Additional parameters and conditions may need to

be reported in order to fully describe the test sampleand experiment in their entirety. For commercialsources, steps should be taken, if possible, to ensurethat the as-received material reflects the man-ufacturer’s description – for instance, comparisonof powder characterization to nominal specificationsand typical values. Examples might include moisturecontent by TGA, primary particle or aggregate size byelectron microscopy, or purity determination usinganalytical methods. Under certain circumstances itmay be prudent to test for the presence of contami-nants arising from the dispersion treatment processitself, such as additional dissolved species (e.g., metalions, organic carbon). It may be equally important totest for the loss of species that should be present, suchas constituents of the original test medium. The fateof dispersing agents (e.g., surfactants, proteins,sugars) may be relevant for both environmental andbiological testing.Other material-specific parameters that are com-

monly mentioned as reportable include aspect ratio,crystalline form and content, amorphous content, andzeta potential from electrophoretic mobility or othermeasure (such as the isoelectric point or zero point ofcharge) related to the particle surface charge proper-ties; zeta potential is both material and mediumdependent, and typically varies with pH for materialsthat contain acidic or basic moieties, so the pH andionic strength of the sample should always accompanyzeta potential results. It may also be necessary, andcertainly prudent to demonstrate that the post-dispersion suspension is stable over the time periodof the experiment; for instance, particle size could bemonitored in situ (e.g., using dynamic light scatter-ing) over a period of time congruent with the testingprotocol. The potential list of reportable physico-chemical endpoints is nearly unlimited, so prioritiza-tion within the context of the intended use is critical.

Closing remarks

Sonication has proven to be a simple and energyefficient technique for the disruption of coarsepowders and, as such, has become a mainstreamtool for the preparation of ENP suspensions for

environmental and toxicological studies. While wedo not intend to discourage the use of ultrasonicdisruption to produce such suspensions, our goalhere is to promote a rational use of sonication, inorder to ensure that unintended side effects, such assonolysis, are minimized and that the results obtainedusing such suspensions are accurately interpreted.Additionally, by providing a set of guidelines forthe standardized measurement and reporting of rel-evant sonication parameters, we seek to improve thereproducibility of related studies and allow for acomprehensive peer review process, which may beotherwise hindered by insufficient informationregarding the preparation of the tested suspensions.We do not attempt to address the nearly infinite rangeof possible experimental scenarios, but instead offer astarting point towards the reproducible preparationand comprehensive evaluation of ENP suspensions.

Acknowledgements

The authors thank Fred Klaessig of Pennsylvania BioNano Systems, LLC, for helpful suggestions and forhis critical review of the draft manuscript. The authorswould also like to thank participants of the Interna-tional Alliance of NanoEHS Harmonization (IANH)for useful discussions, some of which helped motivatethe present work.

Declaration of interest: This work was funded inpart by a cooperative research agreement(70NANB9H9166) between NIST and the Centerfor the Environmental Implications of Nanotechnol-ogy (CEINT). The authors report no conflict ofinterest. The authors alone are responsible for thecontent and writing of the paper.

References

Aoki M, Ring TA, Haggerty JS. 1987. Analysis and modeling ofthe ultrasonic dispersion technique. Adv Ceram Mater2(3A):209–212.

ASTM International (formally the American Society for Testingand Materials). 2006. Standard terminology relating to nano-technology. E2456-06. West Conshohocken, PA: ASTMInternational.

Barnard AS. 2006. Nanohazards: Knowledge is our first defence.Nat Mater 5(4):245–248.

Basedow AM, Ebert KH. 1979. Effects of mechanical stress on thereactivity of polymers – activation of acid-hydrolysis of dextranby ultrasound. Polym Bull 1(4):299–306.

Berber S, Tomanek D. 2009. Hydrogen-induced disintegrationof fullerenes and nanotubes: An ab initio study. Phys Rev B80(7):075421–075425.

Berlan J, Mason TJ. 1992. Sonochemistry – from research labora-tories to industrial plants. Ultrasonics 30(4):203–212.

Berliner S. 2010. Changing materials with high-intensity sound.Accessed from the website: http://berliner-ultrasonics.org/index.html.


Nan

otox

icol

ogy

Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

NIS

T N

atio

nal I

nstit

iute

s of

Sta

ndar

ds &

Tec

hnol

ogy

on 1

2/03

/10

For

pers

onal

use

onl

y.

Bihari P, Vippola M, Schultes S, Praetner M, Khandoga AG,Reichel CA, et al. 2008. Optimized dispersion of nanoparticlesfor biological in vitro and in vivo studies. Part Fibre Toxicol 5:14.

Boverhof DR, David RM. 2010. Nanomaterial characterization:Considerations and needs for hazard assessment and safetyevaluation. Anal Bioanal Chem 396(3):953–961.

Brinker JC, Scherer GW. 1990. Sol-gel science: The physics andchemistry of sol-gel processing. New York: Academic Press.

Brown B, Goodman J. 1965. High intensity ultrasonics: Industrialapplications. Princeton, NJ: D. Van Nostrand Company.

Chen KL, Elimelech M. 2007. Influence of humic acid on theaggregation kinetics of fullerene (C-60) nanoparticles in mono-valent and divalent electrolyte solutions. J Colloid Interf Sci309(1):126–134.

Coakley WT, Brown RC, James CJ, Gould RK. 1973. Inactivationof enzymes by ultrasonic cavitation at 20 kHz. Arch BiochemBiophys 159(2):722–729.

Colvin VL. 2003. The potential environmental impact of engi-neered nanomaterials. Nat Biotechnol 21(10):1166–1170.

Contamine RF, Wilhelm AM, Berlan J, Delmas H. 1995. Powermeasurement in sonochemistry.UltrasonSonochem2(1):S43–47.

DobrovolskaiaMA,McNeil SE. 2007. Immunological properties ofengineered nanomaterials. Nat Nanotechnol 2(8):469–478.

Elzey S, Grassian VH. 2010. Agglomeration, isolation and disso-lution of commercially manufactured silver nanoparticles inaqueous environments. J Nanoparticle Res 12(5):1945–1958.

Forrest GA, Alexander AJ. 2007. A model for the dependence ofcarbon nanotube length on acid oxidation time. J Phys Chem C111(29):10792–10798.

Guzman KAD, Finnegan MP, Banfield JF. 2006. Influence ofsurface potential on aggregation and transport of titania nano-particles. Environ Sci Technol 40(24):7688–7693.

Gwinn MR, Vallyathan V. 2006. Nanoparticles: Health effects –

pros and cons. Environ Health Persp 114(12):1818–1825.Hielscher T. 2005. Ultrasonic production of nano-size dispersions

and emulsions. ENS’05, Paris, France.Honda H, Zhao QL, Kondo T. 2002. Effects of dissolved gases and

an echo contrast agent on apoptosis induced by ultrasound andits mechanism via the mitochondria-caspase pathway. Ultra-sound Med Biol 28(5):673–682.

Huang YY, Knowles TPJ, Terentjev EM. 2009. Strength of nano-tubes, filaments, and nanowires from sonication-inducedscission. Adv Mater 21(38–39):3945–3948.

Isfort CS, Rochnia M. 2009. Production and physico-chemicalcharacterisation of nanoparticles. Toxicol Lett 186(3):148–151.

International Union of Pure and Applied Chemistry (IUPAC).2009. Compendium of chemical terminology. Version 2009-09-07: Release 2.1.5.

Jiang JK, Oberdörster G, Biswas P. 2009. Characterization of size,surface charge, and agglomeration state of nanoparticle disper-sions for toxicological studies. J Nanopart Res 11(1):77–89.

Kahru A, Dubourguier HC. 2010. From ecotoxicology to nano-ecotoxicology. Toxicology 269(2–3):105–119.

Kawasaki H, Takeda Y, Arakawa R. 2007. Mass spectrometricanalysis for high molecular weight synthetic polymers usingultrasonic degradation and the mechanism of degradation.Anal Chem 79(11):4182–4187.

Kim KT, Klaine SJ, Cho J, Kim SH, Kim SD. 2010. Oxidativestress responses of Daphnia magna exposed to TiO2 nano-particles according to size fraction. Sci Total Environ408(10):2268–2272.

Kimura T, Sakamoto T, Leveque JM, Sohmiya H, Fujita M,Ikeda S, et al. 1996. Standardization of ultrasonic power forsonochemical reaction. Ultrason Sonochem 3(3):S157–161.

Kittler S, Greulich C, Gebauer JS, Diendorf J, Treuel L,Ruiz L, et al. 2010. The influence of proteins on the

dispersability and cell-biological activity of silver nanoparticles.J Mater Chem 20(3):512–518.

Kondo T, Kuwabara M, Sato F, Kano E. 1986. Influence ofdissolved gases on chemical and biological effects of ultrasound.Ultrasound Med Biol 12(2):151–155.

Kraft M. 2005. Modelling of particulate processes. KONA23:18–35.

Kreyling WG, Semmler-Behnke M, Moller W. 2006. Healthimplications of nanoparticles. J Nanopart Res 8(5):543–562.

Labille J, Masion A, Ziarelli F, Rose J, Brant J, Villieras F, et al.2009. Hydration and dispersion of C-60 in aqueous systems:The nature of water-fullerene interactions. Langmuir25(19):11232–11235.

Lanone S, Boczkowski J. 2006. Biomedical applications and poten-tial health risks of nanomaterials: Molecular mechanisms. CurrMol Med 6(6):651–663.

Li SD, Huang L. 2008. Pharmacokinetics and biodistribution ofnanoparticles. Mol Pharmaceut 5(4):496–504.

LiuMH, Yang YL, Zhu T, Liu ZF. 2005. Chemical modification ofsingle-walled carbon nanotubes with peroxytrifluoroacetic acid.Carbon 43(7):1470–1478.

Lorimer JP, Mason TJ, Cuthbert TC, Brookfield EA. 1995. Effectof ultrasound on the degradation of aqueous native dextran.Ultrason Sonochem 2(1):S55–57.

Lucas A, Zakri C, Maugey M, Pasquali M, van der Schoot P,Poulin P. 2009. Kinetics of nanotube and microfiber scissionunder sonication. J Phys Chem C 113(48):20599–20605.

Makino K, Mossoba MM, Riesz P. 1983. Chemical effects ofultrasound on aqueous-solutions – formation of hydroxyl radi-cals and hydrogen atoms. J Phys Chem-Us 87(8):1369–1377.

Mandzy N, Grulke E, Druffel T. 2005. Breakage ofTiO2 agglomerates in electrostatically stabilized aqueous disper-sions. Powder Technol 160(2):121–126.

Mason TJ. 1989. Sonochemistry: Theory, applications and uses ofultrasound in chemistry. Chichester, UK: Ellis Horwood.

Mason TJ. 1991. Practical sonochemistry: User’s guide to applica-tions in chemistry and chemical engineering. Chichester, UK:Ellis Horwood.

Mason TJ, Peters D. 2003. Practical sonochemistry: Power ultra-sound uses and applications. Chichester, UK: Ellis Horwood.

Maynard AD. 2006. Nanotechnology: A research strategy foraddressing risk, WoodrowWilson International Center for Scho-lars – project on emerging nanotechnologies.

Minimum Information for Nanomaterial Characterization (Min-Char) Initiative. 2009. Recommended minimum physical andchemical parameters for characterizing nanomaterials on toxi-cology studies. Accessed from the website: http://characteriza-tionmatters.org/parameters.

Murdock RC, Braydich-Stolle L, Schrand AM, Schlager JJ,Hussain SM. 2008. Characterization of nanomaterial dispersionin solution prior to In vitro exposure using dynamic lightscattering technique. Toxicol Sci 101(2):239–253.

Naddeo V, Belgiorno V, Napoli RMA. 2007. Behaviour of naturalorganic mater during ultrasonic irradiation. Desalination210(1–3):175–182.

Oberdörster G, Stone V, Donaldson K. 2007. Toxicology of nano-particles: A historical perspective. Nanotoxicology 1(1):2–25.

Pallem VL, Stretz HA,Wells MJM. 2009. Evaluating aggregation ofgold nanoparticles and humic substances using fluorescencespectroscopy. Environ Sci Technol 43(19):7531–7535.

Pan Y, Neuss S, Leifert A, Fischler M, Wen F, Simon U, et al.2007. Size-dependent cytotoxicity of gold nanoparticles. Small3(11):1941–1949.

Park BS, Smith DM, Thoma SG. 1993. Determination of agglo-merate strength distributions. 4. Analysis of multimodal particlesize distributions. Powder Technol 76(2):125–133.


Nan

otox

icol

ogy

Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

NIS

T N

atio

nal I

nstit

iute

s of

Sta

ndar

ds &

Tec

hnol

ogy

on 1

2/03

/10

For

pers

onal

use

onl

y.

Porter D, Sriram K, Wolfarth M, Jefferson A, Schwegler-Berry D,Andrew M, et al. 2008. A biocompatible medium for nanopar-ticle dispersion. Nanotoxicology 2(3):144–154.

Powers KW, Palazuelos M, Moudgil BM, Roberts SM. 2007.Characterization of the size, shape, and state of dispersion ofnanoparticles for toxicological studies. Nanotoxicology 1:42–51.

Project on Emerging Nanotechnologies. 2010. Woodrow WilsonInternational Center for Scholars and the PewCharitable Trusts.Accessed from the website: http://www.nanotechproject.org/inventories/consumer/.

Pugin B. 1987. Qualitative characterization of ultrasound reactorsfor heterogeneous sonochemistry. Ultrasonics 25(1):49–55.

Radziuk D, Grigoriev D, Zhang W, Su DS, Mohwald H,Shchukin D. 2010. Ultrasound-assisted fusion of preformedgold nanoparticles. J Phys Chem C 114(4):1835–1843.

Raimondi F, Scherer GG, Kotz R, Wokaun A. 2005. Nanoparticlesin energy technology: Examples from electrochemistry andcatalysis. Angew Chem Int Ed 44(15):2190–2209.

Raso J, Manas P, Pagan R, Sala FJ. 1999. Influence of differentfactors on the output power transferred into medium by ultra-sound. Ultrason Sonochem 5(4):157–162.

Rejman J, Oberle V, Zuhorn IS, Hoekstra D. 2004. Size-dependentinternalization of particles via the pathways of clathrin-andcaveolae-mediated endocytosis. Biochem J 377:159–169.

Riesz P, Kondo T. 1992. Free-radical formation induced by ultra-sound and its biological implications. Free Radical Bio Med13(3):247–270.

Roebben G, Ramirez-Garcia S, Hackley VA, Roesslein M,Klaessig F, Kestens V, et al. 2010. Interlaboratory reproducibil-ity of size and surface charge measurements on nanoparticlesprior to biological impact assessment. J Nanopart Res, in review.

Salata O. 2004. Applications of nanoparticles in biology and med-icine. J Nanobiotechnol 2(1):1–6.

Saleh NB, Pfefferle LD, Elimelech M. 2010. Influence of biomacro-molecules and humic acid on the aggregation kinetics of single-walled carbon nanotubes. Environ Sci Technol 44(7):2412–2418.

Singhal A, Skandan G, Wang A, Glumac N, Kear BH,Hunt RD. 1999. On nanoparticle aggregation during vaporphase synthesis. Nanostruct Mater 11(4):545–552.

Sonavane G, Tomoda K, Makino K. 2008. Biodistribution ofcolloidal gold nanoparticles after intravenous administration:Effect of particle size. Colloid Surf B 66(2):274–280.

Suslick KS. 1988. Ultrasound: Its chemical, physical and biologicaleffects. New York: VCH Publishers.

Taghizadeh MT, Bahadori A. 2009. Degradation kinetics of poly(vinyl-pyrrolidone) under ultrasonic irradiation. J Polym Res16(5):545–554.

Tantra R, Tompkins J, Quincey P. 2010. Characterisation of thede-agglomeration effects of bovine serum albumin on nanoparti-cles in aqueous suspension. Colloids Surf B-Biointerfaces75(1):275–281.

Taurozzi JS, Hackley VA, Wiesner MR. 2010a. CEINT/NISTProtocol for the preparation of nanoparticle dispersions frompowdered material using ultrasonic disruption.

Taurozzi JS, Hackley VA, Wiesner MR. 2010b. Reporting guide-lines for the preparation of nanoparticle dispersions from drymaterials.

Teleki A, Wengeler R, Wengeler L, Nirschl H, Pratsinis SE. 2008.Distinguishing between aggregates and agglomerates of flame-made TiO2 by high-pressure dispersion. Powder Technol181(3):292–300.

Unfried K, Albrecht C, Klotz LO, Von Mikecz A, Grether-Beck S,Schins RPF. 2007. Cellular responses to nanoparticles: Targetstructures and mechanisms. Nanotoxicology 1(1):52–71.

Vasylkiv O, Sakka Y. 2001. Synthesis and colloidal processing ofzirconia nanopowder. J Am Ceram Soc 84(11):2489–2494.

Walczyk D, Bombelli FB, Monopoli MP, Lynch I,Dawson KA. 2010. What the cell ‘sees’ in bionanoscience.J Am Chem Soc 132(16):5761–5768.

Wang F, Gao F, Lan MB, Yuan HH, Huang YP, Liu JW. 2009a.Oxidative stress contributes to silica nanoparticle-induced cyto-toxicity in human embryonic kidney cells. Toxicol in Vitro23(5):808–815.

Wang J, Ding N, Zhang ZH, Guo Y, Wang SX, Xu R, et al. 2009b.Investigation on damage of Bovine Serum Albumin (BSA)catalyzed by nano-sized Silicon dioxide (SiO2) under ultrasonicirradiation using spectral methods. Spectrosc Spect Anal29(4):1069–1073.

Wang J, Wang YF, Gao J, Hu P, Guan HY, Zhang LQ, et al. 2009c.Investigation on damage of BSA molecules under irradiation oflow frequency ultrasound in the presence of Fe-III-tartratecomplexes. Ultrason Sonochem 16(1):41–49.

Wang J, Wu J, Zhang ZH, Zhang XD, Wang L, Xu L, et al. 2005.Sonocatalytic damage of Bovine Serum Albumin (BSA) in thepresence of nanometer Titanium dioxide (TiO2) catalyst.Chinese Chem Lett 16(8):1105–1108.

Warheit DB. 2008. How meaningful are the results of nanotoxicitystudies in the absence of adequate material characterization?Toxicol Sci 101(2):183–185.

Wijnhoven SWP, Peijnenburg WJGM, Herberts CA, Hagens WI,Oomen AG, Heugens EHW, et al. 2009. Nano-silver – a reviewof available data and knowledge gaps in human and environ-mental risk assessment. Nanotoxicology 3(2):109–178.

Williams AR. 1983. Ultrasound: Biological effects and potentialhazards. New York: Academic Press.

Yanagida H, Masubuchi Y, Minagawa K, Ogata T, Takimoto J,Koyama K. 1999. A reaction kinetics model of water sono-lysis in the presence of a spin-trap. Ultrason Sonochem5(4):133–139.

Zachariah MR, Carrier MJ. 1999. Molecular dynamics computa-tion of gas-phase nanoparticle sintering: A comparison withphenomenological models. J Aerosol Sci 30(9):1139–1151.

Zhang WX. 2003. Nanoscale iron particles for environ-mental remediation: An overview. J Nanopart Res 5(3–4):323–332.

Zhu ZL, Minasny B, Field DJ. 2009. Adapting technology formeasuring soil aggregate dispersive energy using ultrasonic dis-persion. Biosystems Eng 104(2):258–265.

NOTICE OF CORRECTION

The Early Online version of this article published online ahead of print on 15 November 2010 contained an erroron page 3. Equation 1 should have read

P = cA f a12

)2 2ρ π( ( )2 1

This has been corrected for the current version.


Nan

otox

icol

ogy

Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

NIS

T N

atio

nal I

nstit

iute

s of

Sta

ndar

ds &

Tec

hnol

ogy

on 1

2/03

/10

For

pers

onal

use

onl

y.

ultrasonic dispersion of nanoparticles for environmental

Documents