Proefschrift
ter verkrijging van de graad van doctor aan de Technische Universiteit Delft;
op gezag van de Rector Magnificus Prof. ir. K.C.A.M. Luyben; voorzitter van het College voor Promoties,
in het openbaar te verdedigen op vrijdag 15 maart 2013 om 12:30 uur
door
Ren´
e Martijn WAGTERVELD
Elektrotechnisch ingenieur geboren te IJsselmuiden
Prof. dr. E. Vlieg, Radboud Universiteit Nijmegen Prof. dr. ir. A. B. de Haan, Technische Universiteit Eindhoven Dr. ir. M. M. Mayer, EasyMeasure b.v.
Prof. dr. P. D. E. M. Verhaert, Technische Universiteit Delft, reservelid
ISBN: 978-94-6108-411-8
Martijn Wagterveld, 2013
Effect of Ultrasound on Calcium Carbonate Crystallization
PhD thesis Delft University of Technology, Delft, The Netherlands
been a serious limitation. Scaling causes flux decline, membrane degradation, loss of production and elevated operating costs. Prediction of scale formation tendency, together with implementing suitable scale prevention measures, is therefore essential for optimal operation. In this work a novel concept is proposed for the prediction of scale formation tendency. By enhancing the crystallization (kinetics) locally and monitoring the process itself, scaling can be predicted accurately before it occurs in the bulk solution. If successful, this will result in a typical proactive sensor because precautions can be made (based on the measurement - prediction) before damage has occurred (as in most if not all existing scaling monitoring devices). The better scaling risk assessment improves chemical dosage (preventing overdosing) and prevents the necessity of cleaning or membrane replacement. Ultrasound is selected as possible method for crystallization enhancement. Consequently, the topic of this thesis is the effect of ultrasound on crystallization of calcium carbonate.
After the introduction in chapter 1, the first part of the thesis considers the effect of ultrasound on CaCO3 crystallization from solutions without additives (chapter 2 - chapter 4). Chapter 2 describes the effect of ultrasound on seeded calcite (poly-morph of CaCO3) crystallization. The crystallization rate of calcite was enhanced by 46% through ultrasonic irradiation (42,150 Hz, 17 W dm−3). It was shown that this effect was related to the alteration of the seed crystals’ habit and size. During ultra-sonic irradiation disruption of conglomerates and erosion of single crystals occurred,
that breakage of accelerated crystals by interparticle collisions is unrealistic because of their small sizes and low velocities.
Chapter 4 delineates the way ultrasound exerts its effect by applying ultrasound in different treatment periods (time windows). Applying ultrasound during the first 10 minutes of the experiment did not result in any significant effect which rules out an influence on primary nucleation. The application of ultrasound starting later in the experiment enhanced precipitation of CaCO3. The proposed dominant mecha-nism responsible is deaggregation during the early growth phase (nuclei to crystals conversion regime). This effect is attributed to shear induced by micromixing and / or shear / stress induced by (supersonic) shockwaves, as a result of cavitation. Scan-ning electron microscope analysis shows that ultrasound increases the total number of particles that has, in addition, a more uniform size distribution compared with the untreated experiment. Consequently the available surface area for growth is higher resulting in a higher volumetric precipitation rate.
The effect of ultrasound in solutions without additives has now been determined. In practical applications, water contains more components than calcium and carbon-ate. For instance in reverse osmosis membrane processes antiscalants are added to the water to prevent scaling. Moreover, other ions present, such as magnesium and sulfate, influence the crystallization of CaCO3. Organic substances known as humic substances or humic acids, can also reduce the crystallization rate of CaCO3.
Chapter 5 covers the effect of ultrasound on the growth of calcite in the presence antiscalant nitrilotris(methylene phosphonic acid) (NTMP). During seeded growth experiments and in the presence of NTMP, ultrasound induced an approximately twofold increase in volumetric growth rate. In addition, after inhibition by NTMP ultrasound restored the growth more quickly. The results could be explained in part by the physical effect of ultrasound that causes breakage and attrition of poisoned crystals, which resulted in an increase in fresh surface area. Mass spectroscopy
anal-foreign ions, were based on reverse osmosis drinking water concentrate. The largest effect was found for solutions with antiscalant NTMP and subscribed to degradation of NTMP (see chapter 5). The presence of magnesium in addition to NTMP extended the inhibition time but ultrasound resulted in a similar effect as for the solution with NTMP only. For antiscalant HEDP the experiments were inconclusive. A smaller, but significant effect was seen for the experiments with foreign ions sulfate and magnesium (as a reduction in induction time). The experiments with humic substances did not result in any measurable effect on free-drift precipitation of CaCO3.
Chapter 7 discusses the effect of sulfate (and to a lesser extent magnesium) on the precipitation of CaCO3 in artificial drinking water reverse osmosis concentrate (without ultrasound). CaCO3 formation slows down with increasing sulfate concen-tration and the preferential polymorph shifts from vaterite to aragonite. With this polymorphic change, a new combined habit is observed where (presumably) aragonite spikes grow on top of vaterite (“morning star” habit).
Chapter 8 describes an extension to the software Minteq to systems in which the total amount of salt is a function of time. The model-based estimation of solid CaCO3 corresponds well with the experimental observations in the early growth stage. The estimate of CO2 exchange is evaluated by determining the rate-limiting step. For the sulfate experiments (without precipitation) the diffusion constant and equilibrium partial CO2pressure correspond well with the expected values.
Electric fields might influence the crystallization of CaCO3by affecting, for instance, the energetics of nucleation or the kinetics of either nucleation or growth. Chapter 9 discusses the application of an electric field as actuation method in CaCO3 crystal-lization. Two new measurement configurations were developed to increase the electric field strength compared to the methods described in literature. The first, the modi-fied sitting drop method, had low reproducibility and was therefore abandoned. The second method, the glass plate sandwich, was miniaturized using microfluidics. As for a possible effect of electric field strength on CaCO3crystallization, these
microflu-het meest voor in dergelijke aanslag. Scaling veroorzaakt significante beperkingen, met name in water productie faciliteiten die gebruik maken van omgekeerde osmose (RO) membranen. Het leidt tot vermindering van de vloeistof doorstroom, kwali-teit van membranen en capacikwali-teit van waterproductie. Doordat scaling moeilijk te voorspellen en te verwijderen is, heeft het hogere (onderhouds)kosten tot gevolg. Het waarnemen en voorspellen van de neiging tot aangroei, het risico op scaling, is dus noodzakelijk voor optimale bedrijfsvoering. In dit proefschrift wordt een innovatief concept gepresenteerd om risico op scaling te voorspellen. Door lokaal (de kinetiek van) de beginstappen van scaling te versnellen, i.e. de kiemvorming en kristalgroei (kristallisatie) en dus het scaling proces zelf te bestuderen, kan mogelijke vorming van scaling accuraat worden voorspeld voordat het plaats heeft in de bulk. Als dit suc-cesvol is, dan zal dit resulteren in een proactieve sensor omdat voorzorgsmaatregelen (op basis van de meting-voorspelling) genomen kunnen worden voordat er schade is ontstaan. De betere beoordeling van het risico op scaling zorgt voor een verbeterde afstemming van de dosering van chemicali¨en. Hierdoor wordt overdosering voorko-men en is het schoonmaken en vervangen van membranen niet langer noodzakelijk. Toepassing van ultrageluid is geselecteerd als mogelijke methode om de kristallisa-tie te versnellen. Daardoor is het effect van ultrageluid op kristallisakristallisa-tie van calcium carbonaat het hoofdonderwerp van dit proefschrift.
In hoofdstuk 3 wordt hoge snelheid fotografie gebruikt om de interactie tussen caviterende bellen en calciet entkristallen te bestuderen, zoals beschreven in hoofd-stuk 2. Clusters van imploderende bellen, ontstaan uit de initi¨ele vorming en implosie van bellen, veroorzaken attritie, disruptie van aggregaten en deagglomeratie van kris-tallen. Stromen van cavitatie zorgden alleen voor deagglomeratie. Het in twee delen breken van kristallen werd ook waargenomen. Cavitatie op of vlakbij het kristalop-pervlak zorgde voor hevige versnelling van deeltjes, maar het werd aangetoond dat opbreken van kristallen door botsingen zeer onwaarschijnlijk is door de kleine afmeting en lage snelheid van de kristallen.
In hoofdstuk 4 wordt ultrageluid in verschillende periodes toegepast (tijdvensters) tijdens vrije val precipitatie (ongecontroleerde pH) van CaCO3in een oplossing zonder additieven. Tijdens de eerste 10 minuten van het experiment resulteerde ultrageluid in geen enkel meetbaar effect. Een effect van ultrageluid op primaire kiemvorming is daardoor onwaarschijnlijk. Toepassing van ultrageluid later in het experiment re-sulteerde wel in versnelde precipitatie van CaCO3. Het veronderstelde mechanisme is deaggregatie tijdens de eerste groeifase, waar kiemen uitgroeien tot volwaardige kristallen. Dit effect wordt toegeschreven aan schuifspanningen ontstaan door micro-mixing en interne mechanische spanning veroorzaakt door (supersonische) schokgol-ven tijdens cavitatie. Elektronenmicroscopie laat zien dat het totaal aantal deeltjes groter is wanneer ultrageluid wordt toegepast en dat deze deeltjes meer uniform zijn verdeeld in vergelijking met het onbehandelde experiment. Dit resulteert eveneens in toename van het totale kristaloppervlak, wat leidt tot een grotere volumetrische kristalgroei.
Het effect van ultrageluid in oplossingen van CaCO3 zonder additieven is nu on-derzocht. Water zal in de praktijk veel meer componenten bevatten dan calcium en carbonaat alleen. In het geval van omgekeerde osmose worden bijvoorbeeld antisca-lants toegevoegd aan het water om scaling te voorkomen. Daarnaast kunnen andere
luid voor opbreken en vergruizen van NTMP vergiftigde kristallen, wat zorgt voor vers kristaloppervlak voor groei. Massaspectroscopie van met ultrageluid behandeld NTMP laat zien dat dit molecuul kan worden afgebroken. Het afbreken is mogelijk veroorzaakt door radicalen gevormd tijdens cavitatie of door pyrolyse door cavitatie. Hoofdstuk 6 behandelt het effect van ultrageluid op vrije val precipitatie van cal-cium carbonaat in aanwezigheid van antiscalants NTMP en HEDP, humuszuren, en de ionen magnesium en sulfaat. De toegepaste concentraties van CaCO3, magnesium en sulfaat zijn gebaseerd op praktijkwaarden van omgekeerde osmose concentraat (drink-water productie). Het grootste effect werd waargenomen voor NTMP (zonder andere toevoegingen) en toegeschreven aan het afbreken van het molecuul (zie hoofdstuk 5). De aanwezigheid van magnesium in deze oplossing (met NTMP) verlengde de inhibi-tietijd, maar ultrageluid zorgde voor een vergelijkbaar effect als in de oplossing met NTMP (zonder magnesium). De resultaten van de experimenten met HEDP waren niet afdoende. Een klein, maar significant effect werd waargenomen in de experimen-ten met sulfaat en magnesium (als een verkorte inductietijd). De experimenexperimen-ten met humuszuren resulteerden niet in een meetbaar effect.
In hoofdstuk 7 wordt het effect van sulfaat (en in mindere mate magnesium) on-derzocht op precipitatie van CaCO3 in omgekeerde osmose drinkwater concentraat (zonder ultrageluid). De vorming van CaCO3 gaat langzamer als de concentratie sulfaat wordt verhoogd. De voorkeurs-polymorf verschuift hierbij van vateriet naar aragoniet. Met deze verandering werd een nieuwe gecombineerde kristalvorm waarge-nomen waarbij (vermoedelijk) stekels van aragoniet bovenop vateriet kristallen groeien (“morgenster” kristalvorm).
Hoofdstuk 8 beschrijft een uitbreiding van de software Minteq voor systemen waarbij de totale hoeveelheid opgelost zout een functie van tijd is. De schatting van de hoeveelheid vaste stof CaCO3berekend met het model, komt goed overeen met de experimentele schatting in de vroege groeifase. De schatting van CO2uitwisseling is ge¨evalueerd door de limiterende stap te bepalen. Voor de experimenten met sulfaat
geminiaturiseerd. Deze tweede methode heeft nog niet afdoende resultaten opgeleverd om te kunnen concluderen dat het elektrisch veld een effect op CaCO3 heeft. Voor een succesvol vervolg is een aanpassing van de detectie techniek of het microfluidisch ontwerp noodzakelijk.
Als laatste wordt in hoofdstuk 10 de implicaties en perspectieven van de opgedane kennis vertaald naar de toepassing van een sensor voor de voorspelling van het risico op scaling.
Samenvatting ix
1. Introduction 1
1.1. Motivation . . . 2
1.1.1. Introduction . . . 2
1.1.2. Driving force for crystallization/scaling . . . 3
1.1.3. Scaling mitigation . . . 5
1.1.4. Scaling tendency estimation models . . . 5
1.1.5. Current (scaling) monitoring methods . . . 6
1.1.6. Proposed alternative scaling monitoring concept . . . 7
1.1.7. Ultrasound . . . 8
1.2. Research objectives . . . 10
1.3. Thesis outline . . . 11
2. Seeded calcite sonocrystallization 15 2.1. Introduction . . . 16
2.2. Experimental procedure . . . 17
2.3. Results and discussion . . . 20
3.3.4. Seed acceleration by bubble expansion and collapse . . . 37
3.3.5. Effect of cavitation on crystal habit . . . 41
3.4. Conclusion . . . 46
4. Effect of ultrasonic treatment on early growth during CaCO3 precipitation 51 4.1. Introduction . . . 52 4.2. Experimental Section . . . 53 4.2.1. Chemicals . . . 53 4.2.2. Experimental setup . . . 53 4.2.3. Experimental procedures . . . 55 4.3. Results . . . 56
4.3.1. Ultrasound during entire experiment (0 - 4500 s) . . . 57
4.3.2. Ultrasound during primary nucleation (0 - 600 s) . . . 62
4.3.3. Ultrasound during early growth (600 - 1200 s) . . . 62
4.3.4. Ultrasound during crystal outgrowth (1800 - 2400 s) . . . 63
4.4. Discussion . . . 64
4.5. Conclusion . . . 67
5. Ultrasonic reactivation of phosphonate (NTMP) poisoned calcite during crystal growth 73 5.1. Introduction . . . 74
5.2. Experimental . . . 76
5.2.1. Chemicals . . . 76
5.2.2. Calcite seed crystals . . . 76
5.2.3. Experimental set-up . . . 77
calants and ultrasound. 91 6.1. Introduction . . . 92 6.2. Experimental . . . 94 6.2.1. Chemicals . . . 94 6.2.2. Experimental setup . . . 95 6.2.3. Experimental procedures . . . 96
6.3. Results and discussion . . . 99
6.3.1. Antiscalants: NTMP, HEDP . . . 99
6.3.2. Foreign ions: Sulfate and magnesium . . . 108
6.3.3. Humic substances . . . 110
6.4. Conclusion . . . 112
7. Polymorphic change from vaterite to aragonite under influence of sulfate: the “morning star” habit 119 7.1. Introduction . . . 120
7.2. Experimental . . . 121
7.2.1. Chemicals . . . 121
7.2.2. Experimental setup . . . 121
7.2.3. Experimental procedures . . . 121
7.3. Results and discussion . . . 122
7.4. Conclusion . . . 131
8. Modeling the carbonate system 137 8.1. Introduction . . . 138
8.2. Newton-Raphson . . . 138
8.3. Calcium carbonate precipitation and pH . . . 142
8.3.1. Modeling solid CaCO3 formation from pH: Newton-Raphson . 142 8.3.2. Calcium carbonate solid formation from measured pH . . . 144
8.4. CO2 exchange and pH . . . 147
9.3.1. Modified sitting drop . . . 161
9.3.2. Microfluidics . . . 162
9.4. Conclusion . . . 168
10. Discussion and perspectives 173 10.1. External energy sources influencing crystallization . . . 174
10.1.1. Ultrasound . . . 174
10.1.2. Electromagnetic fields . . . 175
10.2. Scaling . . . 176
10.3. Device for monitoring scaling risk . . . 177
10.3.1. Miniaturization . . . 178
10.3.2. Dynamic model . . . 178
10.4. Additional process optimization . . . 179
A. Experiments sonoluminescence reactor 181 A.1. Experimental . . . 182
A.1.1. Chemicals . . . 182
A.1.2. Experimental setup . . . 182
A.1.3. Solution preparation . . . 183
A.1.4. Experimental procedures . . . 183
A.2. Results . . . 185 B. Supporting information for: Effect of US on early growth 187 C. Matlab-code: Solid CaCO3 as function of pH 191
Acknowledgements 195
the efficiency. That scaling is very persistent and hard to remove becomes certainly evident when cleaning the bathroom. The use of an acidic cleaning product is usually required.
The term scaling, also known as inorganic or precipitation fouling, comprises the formation of hard mineral deposits. Typical examples of equipment that suffer from scaling are process (fig. 1.1b) or membrane equipment (fig. 1.1c) such as boiler sys-tems, cooling towers, mining process water, oil wells and membrane water purification systems. Besides calcium carbonate, common scaling minerals are calcium sulfate (CaSO4· H2O), barium sulfate (BaSO4), strontium sulphate (SrSO4), silicates (forms
of silicon oxide, SiO2), calcium phosphate (Ca3PO4), ferric- and aluminum hydroxides (Fe(OH)3 and Al(OH)3).
Integrated water supply systems are increasingly utilizing non-traditional water sources such as seawater, excessively hard or brackish groundwater, poorer quality surface waters, and wastewater. These sources commonly require treatment with
(a)
Domestic: Washing machine heating element [1]
(b)
Industrial: Scaling in pipe [2]
(c)
Industrial: Scaling on mem-brane [3]
increase of the relative dissolved salt concentration [5] which will lead to a driving force for scaling. In particular for brackish water RO, scaling is the main limiting factor (due to low biofouling risk) [5]. Chapters 6 and 7 discuss such brackish groundwater with a concentration factor of 5.
Scaling causes flux decline, membrane degradation, loss of production and elevated operating costs. The operating costs of a large brackish RO plant producing 700,000 m3per day have been estimated in 2006 to be$20,000,000 per year (Yun et al. [6]), of which 40% can be ascribed to prevention of scaling, so$8,000,000 per year. Membrane replacement takes half of this sum, the other half is due to chemicals added. Improving the operating costs for scale prevention with only 5% will already save$400,000 per year.
Given the damage-related costs mentioned above, process optimization demands scaling prevention based on accurate prediction. In general, scaling leads to blockage of water flow (or heat flux) through pipes and membranes causing:
Higher energy consumption Lower production capacity Higher maintenance
Shorter lifetime of the process installation
1.1.2. Driving force for crystallization/scaling
The process of scaling can only commence when the condition of supersaturation is achieved. The fundamental dimensionless driving force for crystallization (start of the scaling process) as function of supersaturation is defined as follows, where S (-), ∆µ (J mol−1), R (J mol−1 K−1), T (K) and v (-), are the supersaturation ratio, change in chemical potential, gas constant, absolute temperature, and number of different ion species in the formula unit (v = 2 for CaCO3), respectively [7]:
∆µ
Figure 1.2.:
Schematic representation of the metastable zone. The thick line represents Ksp.
Be-low this line the solution is undersaturated and stable, implying Ca2+ and CO2–3 remain
permanently in solution. Far above the line the solution is supersaturated and crystal-lization takes place. Between the dotted and thick line the solution is supersaturated but metastable due to kinetics. The arrow indicates the effect of antiscalants, shifting the metastable limit further from the solubility line.
The driving force is positive when S > 1, or supersaturated. The supersaturation ratio is then defined in terms of activities, where IAP is the ion activity product and Kspthe thermodynamic equilibrium solubility product:
S = IAP Ksp
v1
(1.2) A region exists, just above saturation, where kinetics are so slow that the system is considered to be metastable. This region is also referred to as metastable zone and is schematically depicted in fig. 1.2. The position of the “metastable limit” (dashed line), which together with the solubility defines the metastable zone, is not well defined and depends on the measurement method and application. Above the “metastable limit” crystallization is “spontaneous” and scaling tendency is high. Scale formation comprises complex phenomena involving both crystallization and transport mechanisms. In addition to supersaturation conditions, kinetics of crystallization should also be considered.
Nucleation (critical nucleus formation), followed by crystal growth can result in membrane surface (pore) blockage, the first form of scaling. The crystals might ag-glomerate and form layers which leads to cake formation, the second form of scaling. Whether supersaturation, through crystallization, leads to scaling, is ruled by var-ious operating conditions such as pH, temperature, operating pressure, permeation
The most common scale mitigating techniques can be grouped into three categories [8]:
Altering feed water characteristics, i.e. acidification
Optimization of operating parameters and system design i.e. limiting product recovery
Antiscalant addition
Application depends on the nature of the feed water, membrane compatibility with acid or scale inhibitor and cost. In practice the use of antiscalants is almost inevitable in reverse osmosis membrane systems. Antiscalants inhibit growth of crystals and thereby increase the apparent metastability (fig. 1.2). The economic benefit of the use (costs) of antiscalants eventually arises from higher product recovery [9–12]. Currently antiscalants are highly overdosed since the scaling tendency is difficult to assess [13, 14]. Besides the high cost, due to stricter environmental regulations, a reduction of antiscalant dosage is necessary. Moreover with the use of antiscalants the risk of (bio)fouling is 4-10x higher [5], all implying an improvement of scaling tendency determination [4]. With better scaling risk assessment, processes can operate closer to the metastable limit, implying a higher efficiency at lower costs.
1.1.4. Scaling tendency estimation models
Most research on scale formation and control focuses on synthetic (or analytical) solutions. Although this helps in gaining insight in individual components, it fails to mimic the chemistry of water of more complex composition. In an attempt to account for the complexity of water chemistry empirical models have been developed. All practiced scale prediction models use the theoretical concept of saturation [15] and do not include, for instance, that antiscalants effectiveness is not equal for all scalants [5].
Langelier saturation index (LSI) was the first introduced scaling index and is based on the comparison measured pH and an theoretical “pH”, based on water composition
1.1.5. Current (scaling) monitoring methods
The dynamic behavior of the water chemistry forces an on-line scaling tendency es-timation. Successful application of scaling tendency models just discussed would comprise the online measurement of every component in the system, which is very impractical. An alternative would be to measure a dominant component, for instance the “consumption rate” of antiscalant. Since the concentration of antiscalant is very low, additional labeling is necessary for accurate measurement, as is applied in the Trasar technology [16]. In this technology the antiscalant is labeled with fluorescent dye and measured with optical means. However, besides being more expensive, the variation in antiscalant is not a direct measure of scaling tendency. The dosage of antiscalant is better controlled, but is still based on (static) scaling tendency models. Since scaling cannot be precluded, scaling monitors have been applied. To date, scaling monitoring methods rely on the detection of scaling (and not on scaling ten-dency) and the following methods have been researched:
Permeate flux decline [17]
Ultrasonic time-domain reflectometry analysis [18] Surface acoustic wave [19]
Quartz crystal [20] Visual observation [21]
Electrical impedance spectroscopy analysis [22]
All methods (except for the permeate flux decline according to the authors) share the same drawback: They monitor scaling, but not the risk of scaling. In other words, scaling already takes place and mitigation such as cleaning or membrane replacement is still a necessity. The monitoring method based on permeate flux decline takes place
Figure 1.3.:
Artist impression of the proposed concept. Part of the (supersaturated) bulk solution is directed to the measurement device, either flow through or batch sampled. An actuator enhances the crystallization kinetics and a sensor detects the course of crystallization. Suggested actuation is by application of ultrasound or a large electric field.
in a bypass stream and is said to monitor scaling locally before it occurs in the main process [17]. However, flux decline can have many other causes [23] which may lead to erroneous scaling risk assessment.
1.1.6. Proposed alternative scaling monitoring concept
Current scaling prediction is based on empirical models which are not very accurate. Scaling measurement is based on the actual scaling, which makes the response inher-ently too late. The scaling prediction and scaling measurement can be improved by measuring the scaling process and detect scaling before it occurs in the bulk. Pro-posed concept: By enhancing the crystallization (kinetics or thermodynamics) locally and monitoring the process itself, scaling can be predicted accurately before it occurs in the bulk solution. If successful, this will result in a typical proactive sensor because precautions can be made (based on the measurement - prediction) before damage has occurred. The better scaling risk assessment improves chemical dosage (preventing overdosing) and prevents the necessity of cleaning or membrane replacement.
Figure 1.3 provides an artist impression of the proposed concept. Part of the bulk solution is directed to the measurement device, either flow through or batch sampled. An actuator enhances the crystallization kinetics and a sensor detects the course of crystallization. The ideal actuator should enhance crystallization in a generic way and should be additional to the conditions present, e.g. keep the bulk conditions constant.
ment or conductivity detection but both parameters might not give enough sensitivity when complex solutions are applied. An optical measurement seems to be the best candidate as long as the bulk solution is relatively transparent, otherwise an acousti-cal measurement is an option. With optiacousti-cal detection one can apply light scattering / transmission, laser diffraction, optical attenuation for particle detection, or even infrared or Raman detection for identification of newly formed material.
In this work the focus will be on CaCO3 crystallization, with application of ul-trasound and pH measurements, sometimes assisted with optical scattering measure-ments.
1.1.7. Ultrasound
Ideally, enhancing the precipitation kinetics is done by adding external energy to the system. Applying ultrasound might be a good candidate as such an external source. Ultrasound can help controlling the course of precipitation and crystallization processes and is also referred to as sonocrystallization [27]. The positive effects seen in sonocrystallization are usually ascribed to cavitation that appears in high power ultrasound.
Cavitation is the interaction of (acoustic) pressure waves with cavities (microbub-bles), caused by the rupture of the fluid in the negative pressure cycle. Microscopic bubbles oscillate or grow and collapse under the varying pressure field inside the treated liquid. Several effects (can) occur during this process; the formation of radi-cals, generation of shockwaves and microjets, local hotspots of high pressure (up to 200 MPa or 2000 bar) and temperature (up to 6000 K), micromixing, macromixing and rise of bulk temperature [28]. These effects by cavitation might influence the thermodynamics (local hotspots of high pressure and temperature) or kinetics of the crystallization / scaling process, and thereby shifting, or reducing, the metastable
tial (or transient) cavitation [29]. Non-inertial cavitation is the stable oscillation of a bubble. Surface oscillations can cause interesting bubble geometries [30], as shown in fig. 1.4. Mixing is enhanced around these bubbles but they do not exhibit the other effects found during cavitation. Inertial cavitation is the growth and collapse of bubbles, as depicted in fig. 1.5. In the collapse phase a local hotspot of high pressure and temperature is created. Shockwaves are generated, and for asymmetric collapse, microjets are created. The fast expansion and collapse and presence of shockwaves lead to enhanced mixing. Shockwaves and microjets might cause surface erosion, like pitting or attrition, or even breakage. Furthermore particle (de)agglomeration and (de)aggregation have been observed previously [27]. During application of ultra-sound the bulk temperature will rise. Controlling the bulk temperature is therefore a necessity.
Figure 1.4.:
Examples of non-inertial cavitation e.g. ultrasound causing bubble oscillations. A) Surface oscillation causing transitions between circular and square bubble geometry projection. Time between consecutive images is 11 µs, frequency of ultrasound 42.2 kHz. B) Surface oscillation causing transitions between circular and hexagonal bubble geometry projection. Time between consecutive images is 22 µs, frequency of ultrasound is 20.7 kHz
Figure 1.5.:
Example of inertial cavitation e.g. ultrasound causing bubble growth and collapse. Time between consecutive images is 8 µs, frequency of ultrasound 42.2 kHz. The series shows exactly one period. The sine wave shows a typical sequence of inertial cavitation. In a negative pressure cycle a cavity is formed. It effectively grows faster in size in consecutive negative pressure cycles than it shrinks in positive cycles. At a certain critical size the bubble completely collapses and a local hotspot is created.
1.2. Research objectives
By enhancing the crystallization (kinetics) locally and monitoring the process itself, scaling can be predicted accurately before it occurs in the bulk solution. Ultrasound is selected as principle method for (possible) crystallization enhancement. Application of an electric field is also briefly addressed. Consequently, the main research objectives are:
Characterize the effect of ultrasound on crystallization of CaCO3 in:
– Solutions of CaCO3 without additives. – Solutions of CaCO3 with antiscalant added.
– Solutions with (variations on) components and concentrations based on drinking water RO concentrate.
Clarify the underlying mechanisms in CaCO3 sonocrystallization.
Explore experimental methods to investigate the effect of an electric field on CaCO3 crystallization.
water concentrate is investigated by using synthetic solutions. Additionally, the spe-ciation is modeled, specifically in case of time varying pH. Finally the application of an alternative method, electric field, is briefly addressed.
Chapter 2 describes the effect of ultrasound on seeded calcite (polymorph of CaCO3) crystallization. In a (seeded) constant composition experiment, the con-ditions during crystal growth, such as temperature, solution composition and thus pH, were kept constant. With this methodology, application of ultrasound should, if effective, directly lead to an affected (volumetric) crystallization rate. With scanning electron microscopy the size and habit of the crystals with and without treatment were investigated, as well as the particle size distributions.
In chapter 3 the interaction of acoustic cavitation with suspended calcite crystals, as discussed in chapter 2, is visualized using high speed imaging. Possible breakage by high velocity interparticle collisions is investigated and particle acceleration and deceleration by cavity expansion and collapse is modeled. The cavitation phenomena responsible for disruption of agglomerates and aggregates are discussed, as well as the effect of cavitation on crystal habit (using SEM analysis).
Chapter 4 delineates the way ultrasound exerts its effect by applying ultrasound in different treatment periods (time windows). For that, three stages in precipitation are distinguished: The first is primary nucleation, either homogeneous or heteroge-neous. The second is “early growth”, when secondary nucleation can take place and crystals grow to detectable size. The last is “late growth” during which formed crys-tals continue to grow and the supersaturation is reducing to saturation. Calcium carbonate formation, using the free-drift method, is monitored by three independent parameters: pH, light scattering, and scanning electron microscopy (SEM).
Chapter 5 covers the effect of ultrasound on the growth of calcite in the pres-ence antiscalant nitrilotris(methylene phosphonic acid) (NTMP). The calcite crystal growth was measured using the constant composition method at various NTMP con-centrations with and without ultrasonic irradiation. Mass spectrometry is used to detect break down products of NTMP.
this software has been applied to determine the initial speciation. In this chapter the model is applied to systems in which the total amount of salt is a function of time, e.g. due to CaCO3 deposition or gaseous CO2exchange with the environment.
Chapter 9 discusses the application of an electric field as actuation method in CaCO3 crystallization. Two experimental methods are explored, a variation of the sitting drop technique, and the glass plate sandwich in combination with microfluidics. Finally, Chapter 10 discusses the implications of the gained knowledge in this work with respect to a sensor for scaling tendency prediction, and perspectives are outlined.
References
[1] Unknown. Limescale is caused by Hard Water. 2012. url: http://www.plumber24hours. co.uk/plumber-blog/ (cit. on p. 2).
[2] Unknown. Pipe full of scale formed by Calcium Carbonate and other salts solved in the water. 2012. url: http://www.merusonline.com/in- general/water- containing-calcium-carbonate (cit. on p. 2).
[3] Unknown. Mixed sulphate scaling rapidly causes deterioration in RO membrane per-meability. 2012. url: http://www.environmentalthinking.com/et/environmental_ thinking / Reducing - the - environmental - impact - of - membrane - cleaning / 1 / 200336175636 (cit. on p. 2).
[4] S. J. Khan, D. Murchland, M. Rhodes, and T. D. Waite. “Management of concentrated waste streams from high-pressure membrane water treatment systems”. In: Crit. Rev. Environ. Sci. Technol. 39.5 (2009), pp. 367–415 (cit. on pp. 3, 5).
[5] L. F. Greenlee, D. F. Lawler, B. D. Freeman, B. Marrot, and P. Moulin. “Reverse osmosis desalination: Water sources, technology, and today’s challenges”. In: Water Res. 43.9 (2009), pp. 2317–2348 (cit. on pp. 3, 5).
[6] T. I. Yun, C. J. Gabelich, M. R. Cox, A. A. Mofidi, and R. Lesan. “Reducing costs for large-scale desalting plants using large-diameter, reverse osmosis membranes”. In: Desalination 189.1 (2006), pp. 141–154 (cit. on p. 3).
p. 5).
[11] J. Guo and S. J. Severtson. “Inhibition of Calcium Carbonate Nucleation with Amino-phosphonates at High Temperature, pH and Ionic Strength”. In: Ind. Eng. Chem. Res. 43 (2004), pp. 5411–5417 (cit. on p. 5).
[12] L. F. Greenlee, F. Testa, D. F. Lawler, B. D. Freeman, and P. Moulin. “The effect of antiscalant addition on calcium carbonate precipitation for a simplified synthetic brackish water reverse osmosis concentrate”. In: Water Res. 44.9 (2010), pp. 2957– 2969 (cit. on p. 5).
[13] R. A. Dawe and Y. Zhang. “Kinetics of calcium carbonate scaling using observations from glass micromodels”. In: J. Petrol. Sci. Eng. 18.3-4 (1997), pp. 179–187 (cit. on p. 5).
[14] T. Chen, A. Neville, and M. Yuan. “Calcium carbonate scale formation–assessing the initial stages of precipitation and deposition”. In: J. Petrol. Sci. Eng. 46.3 (2005), pp. 185–194 (cit. on p. 5).
[15] V. A. Prisyazhniuk. “Prognosticating scale-forming properties of water”. In: Appl. Therm. Eng. 27.8 (2007), pp. 1637–1641 (cit. on p. 5).
[16] E. H. K. Zeiher, B. Ho, and K. D. Williams. “Novel antiscalant dosing control”. In: Desalination 157.1 (2003), pp. 209–216 (cit. on p. 6).
[17] C. A. C Van de Lisdonk, B. M. Rietman, S. G. J. Heijman, G. R. Sterk, and J. C. Schippers. “Prediction of supersaturation and monitoring of scaling in reverse osmosis and nanofiltration membrane systems”. In: Desalination 138.1 (2001), pp. 259–270 (cit. on pp. 6, 7).
[18] X. Lu, E. Kujundzic, G. Mizrahi, J. Wang, K. Cobry, M. Peterson, J. Gilron, and A. R. Greenberg. “Ultrasonic sensor control of flow reversal in ro desalination part 1: Mitigation of calcium sulfate scaling”. In: J. Membr. Sci. 419–420 (2012), pp. 20–32 (cit. on p. 6).
[19] M. L¨offelmann and A. Mersmann. “How to measure supersaturation?” In: Chem. Eng. Sci. 57.20 (2002), pp. 4301–4310 (cit. on p. 6).
[20] O. J. Joung, Y. H. Kim, and K. Fukui. “Determination of metastable zone width in cooling crystallization with a quartz crystal sensor”. In: Sens. Act. B. 105.2 (2005), pp. 464–472 (cit. on p. 6).
[21] M. Uchymiak, A. Rahardianto, E. Lyster, J. Glater, and Y. Cohen. “A novel RO ex situ scale observation detector (EXSOD) for mineral scale characterization and early detection”. In: J.Membr. Sci. 291.1 (2007), pp. 86–95 (cit. on p. 6).
[26] Z. Hammadi and S. Veesler. “New approaches on crystallization under electric fields”. In: Prog. Biophys. Mol. Biol. 101.1 (2009), pp. 38–44 (cit. on p. 8).
[27] M. L. de Castro and F. Priego-Capote. “Ultrasound-assisted crystallization (sonocrys-tallization)”. In: Ultrason. Sonochem. 14 (2007), pp. 717–724 (cit. on pp. 8, 9). [28] Y. Shah, A. Pandit, and V. Moholkar. Cavitation reaction engineering. Plenum Pub
Corp, 1999 (cit. on pp. 8, 9).
[29] T. G. Leighton. “What is ultrasound?” In: Prog. Biophys. Mol. Biol. 93.1-3 (2007), pp. 3–83 (cit. on p. 9).
[30] M. Versluis, S. M. van der Meer, D. Lohse, P. Palanchon, D. Goertz, C. T. Chin, and N. de Jong. “Microbubble surface modes [ultrasound contrast agents]”. In: Ultrason. Symp., 2004 IEEE. Vol. 1. IEEE. 2004, pp. 207–209 (cit. on p. 9).
abstract The seeded sonocrystallization of calcite was investigated by measuring the vol-umetric crystallization rate at constant composition conditions. The crystallization rate of calcite was enhanced by 46% through ultrasonic irradiation (42,150 Hz, 17 W dm−3) of a supersaturated crystal suspension. It was shown that this effect was related to the alteration of the seed crystals’ habit and size. During ultrasonic irradiation disruption of conglomerates and erosion of single crystals occurred, accompanied by the production of many fines. The increased surface area available for crystal growth resulted in the observed crystallization rate enhancement.
This chapter has been published as: L. Boels, R.M. Wagterveld, M.M. Mayer, G.J. Witkamp. Seeded calcite sonocrystallization. Journal of Crystal Growth 312 (2010) 961-966
A few papers mention a decrease of the metastable zone width [8, 10, 13, 14, 18] as well as an enhanced solubility of a sparingly soluble salt [20]. Furthermore, effects are seen on the mean particle size [1, 2, 5–7, 9, 11–13, 16, 21], the particle size distribution and crystal morphology [1, 2, 5–7, 10–14, 16].
Acoustic streaming and cavitation are the most important phenomena that occur when ultrasound is passed through a liquid-solid system. These phenomena can pro-duce a series of unique chemical and physical effects. In acoustic streaming, acoustic waves produce a stirring effect. In the vicinity of the crystal surfaces, this stirring effect causes a reduction of the diffusion layers’ thickness, thereby increasing the liquid-solid mass transport [23]. In addition, ultrasound produces small imploding cavities that generate high-energy shock waves which impinge on the particle sur-face. This can create high velocity interparticle collisions that can alter the particle morphology and size dramatically. It was reported that these interparticle collisions occur with such a great force that even metal particles tend to melt together [24]. Microjet formation in the vicinity of the particle surface is another well-established mechanism for accounting the effect of cavitation in solid-liquid systems. However, in literature it is mentioned that this mechanism can only occur if the surface is several times larger than the resonant bubble size [25].
Calcium carbonate is one of the most abundant minerals, and the problem of its scaling propensity is encountered in many industrial water treatment processes. Al-though the topic of sonocrystallization has received much attention in the last decade, the effect of ultrasonic irradiation on the precipitation of calcium carbonate has hardly been investigated and therefore the literature on this topic is sparse. Dalas [15] re-ported a retarded crystal growth in the presence of ultrasonic irradiation. No effect of this irradiation on the nature, morphology or the size of the formed calcium carbonate crystals was observed. In the study conducted by Nishida [19], it was observed that the spontaneous precipitation of calcium carbonate was enhanced in the presence of
between vaterite and calcite particles.
Considering the physical effects that occur due to cavitation, it is reasonable to expect that the habit and size of calcium carbonate crystals can be altered through ultrasonic irradiation. The aim of the present work is to investigate whether or not this occurs. For this purpose the volumetric crystal growth rate (i.e. crystallization rate) of calcite seed crystals during ultrasonic irradiation was measured using the proven constant composition methodology [27, 28]. If the habit and/or size of the seed crystals can be indeed affected, the volumetric crystal growth rate should also change. Therefore, the crystallization rate after the irradiation period was also measured.
2.2. Experimental procedure
Only analytical grade reagents, high quality water (MilliQ Reagent Water System, resistivity >18 MΩ cm) and grade A glassware were used throughout the experi-ments. Calcite seed crystals were prepared, using the method described by Reddy and Nancollas [29], by slowly adding 2 dm3 of a 0.20 M calcium chloride solution
to 2 dm3 0.20 M sodium carbonate solution at 25◦C. The freshly precipitated seed
crystals were aged overnight in mother liquor and were subsequently washed with MilliQ water each day for 1 week. Afterwards, the washed seed crystals were aged for 3 weeks before filtering. The dried crystals were characterized by scanning electron microscopy (Jeol JSM-6480LV), nitrogen adsorption (Micromeritics Tristar 3000) and ATR-FT-IR spectroscopy (Shimadzu 4800). The specific surface area of the seed crys-tals was found to be 0.17 m2g−1as determined with a five-point BET method [30] on
3 replicate samples. The ATR-FT-IR spectra showed characteristic adsorptions for calcite at 1795, 1392, 871 and 711 cm−1. The characterization confirmed the crystals to be pure calcite and appeared as interpenetrated conglomerates.
A double walled thermostatted glass reactor equipped with a floating magnetic stir-rer bar to minimize any grinding effects, a by-pass loop and an ultrasonic transducer were used (fig. 2.1). The ultrasonic transducer was part of a dedicated homebuilt
sys-Figure 2.1.:
Scheme of the constant-composition experimental set-up consisting of (1) a double walled glass reactor, (2) floating magnetic stirrer bar, (3) membrane pump, (4) ultrasonic trans-ducer, (5) free port for seed addition, (6) pH electrode and reference electrode, and (7) a temperature sensor.
tem that could be controlled precisely in terms of shape, frequency and amplitude of the alternating current for driving the transducer. For the control of pH, a combined Pt-ring pH electrode and a shielded Ag/AgCl reference electrode in combination with two coupled automatic burettes (Metrohm Titrino 785 and Dosimat 665) were used. In order to avoid any possibility of damaging the electrodes by the ultrasonic irradi-ation, both electrodes and a temperature sensor were positioned in a glass cell in the by-pass stream. A positive displacement membrane pump was used to pump the crys-tal suspension with 1.5 dm3min1through the by-pass. Metastable working solutions
were prepared by slow addition of 500 cm3 of a 4 mM calcium solution to 500 cm3
of a 4 mM bicarbonate solution in the reactor. Fresh titrant solutions were prepared every day as shown in table 2.1. Potassium chloride was added to all solutions to maintain the ionic strength constant at 0.1.
was adjusted to the desired value of 8.500 using a 0.05 M KOH solution. The stability of the working solution was verified by observing a constant pH for at least 45 min. After the addition of 0.250 g dry seed crystals, the solution pH started to decrease as a result of calcite precipitation. This triggered the automatic burettes to add concurrently equivalent amounts of calcium chloride and sodium carbonate solutions in order to achieve and maintain the pH at the target value ( ± 0.002 pH units). In this way, a constant degree of supersaturation was maintained throughout the growth experiment.
After 30 min of normal growth, the ultrasonic irradiation was initiated. The solu-tion was treated with ultrasound for 45 min at a frequency of 42,150 Hz and with an intensity of 17 W (real output power intensity pre-determined by measuring the adi-abatic temperature rise in time using a similar well isolated glass reactor filled with 1 dm3of water). The volume of added titrant solution was monitored over time and this
data represented the calcium carbonate precipitation rate. During the experiment, aliquots of solution were rapidly removed, filtered through a 0.2 µm filter and ana-lyzed for the calcium content with inductive-coupled plasma spectrometry (Optima 3000XL, Perkin-Elmer) to verify the constancy of the degree of supersaturation.
After the ultrasonic treatment the crystals were allowed to grow for another 75 min at the same supersaturation. All experiments were conducted at atmospheric pressure with ambient levels of CO2. By keeping the reactor lid closed and all ports sealed during the experiments, the exchange of atmospheric CO2 with the reactor solution was minimized. Seed addition and reactor sampling were performed as rapidly as possible. The temperature was maintained constant at 25 ±0.1 ◦C by circulating water from a thermo-bath through the jacket of the reactor. During the ultrasonic irradiation, the thermo-bath was set to cool in order to keep the reaction mixture temperature at 25 ±0.1◦C.
In order to investigate the effect of ultrasound on the crystal habit, 0.250 g dry seed crystals were added to a 1 dm3saturated calcium carbonate solution and the resulting
Figure 2.2.:
Titrated volume of Ca2+ vs. time of constant composition calcite growth experiments: three sets of control experiments and two sets with ultrasonic treatment (start after 30 min). Experimental conditions: S = 2.11, pH= 8.500, T = 25◦C.
suspension was stirred and treated with ultrasound in the same way as during the growth experiments. Throughout the experiment, samples of 15 cm3 of solution were
collected and filtered through a 0.2 µm filter. After drying, the samples were analyzed with SEM.
For comparison, the experiment was repeated with a stirred saturated suspension without applying ultrasound (control).Afterwards, both suspensions were transferred to a particle size and shape (video) analyzer equipped with a liquid flow cell (Eyetech, Ankersmid) in order to measure the number distributions of the particle size. Three replicate analyses were performed in which 5000 particles were analyzed to reach a confidence level above 99%.
2.3. Results and discussion
The driving force for crystallization can be expressed as ∆µ
RT = v ln S (2.1)
where R (J mol−1 K−1) is the gas constant, T (K) is the absolute temperature, ∆µ (J mol−1) is the change in chemical potential, and S (-) is the supersaturation. For CaCO3, S is best expressed in terms of the solubility product:
Control 03 1.70 × 10−6 0.1199 0.999995 Average 1.82 × 10−6 0.0388 0.999999 IAP Ksol 1v (2.2) where IAP (mol kg−1 ) is the ion activity product, Ksol (mol kg1) is the solubility
product, and v (-) is the number of ions in the formula unit. The supersaturation, S, with respect to calcite of the working solution used here was calculated to be 2.11 (software: Visual Minteq v2.53, model: Davies) and the change in chemical potential, ∆µ, of transfer from supersaturated solution to equilibrium, measured 1.85 kJ mol−1. If the surface area of the used seed crystals is known, the constant linear calcite growth rate, R (mol m−2 s−1), can be easily obtained from the initial slope of the titrant addition versus time curve. However, growth leads to an increase of crystal volume and an accompanying increase of surface area. This causes the slope of the titrant addition versus time curve to increase over time. The first time derivative of the added volume of Ca2+ titrant is denoted here as the (volumetric) crystallization rate (cm3s−1).
fig. 2.2 shows three datasets of constant composition calcite growth experiments (control), and two datasets of similar experiments but with 45 min of ultrasonic treatment (ultrasound). A cubic regression was used to fit the control experiments and their average. The linear growth rate can be derived from the first order constant [31, 32]. The first 15 min show an initial growth surge. This period is excluded from the fit. The growth rate was estimated to be 1.82 mol m−2 s−1 (6.76 10−11) m s−1)
(table 2.2). The results in fig. 2.2 and table 2.2 reveal that the experiments were highly reproducible. Furthermore, the measured growth rate corresponds well with values found in literature [33]. This suggests that the experimental set-up is suitable to study the effect of ultrasound on the crystallization of calcium carbonate.
After the ultrasonic irradiation is started, the slope of the ultrasound curve starts to increase faster compared to the slope of the control curves fig. 2.2. This effect is
Figure 2.3.:
Crystallization rate (volumetric) defined as the time derivative of the added volume of Ca2+ titrant (cm3 s−1
) (see fig. 2.2). A moving average (span = 15 of 448 data points) was used on all datasets.
more clearly seen in fig. 2.3 where the crystallization rate of the average values of control and ultrasound datasets are plotted as function of time. At the start of the ultrasonic treatment at t =30 min, the crystallization rate of the ultrasonic treatment shows a clear increase compared to the control curve.
After termination of the ultrasonic treatment (t = 75 min), when all conditions are exactly the same compared to the control experiments, the ratio of both crystallization rates remains the same. This is demonstrated in fig. 2.4 where the ratio between the ultrasound and control crystallization rates is shown. Before the ultrasonic treatment, this ratio is close to 1 during the first 30 min. Then, the figure shows a clear increase of precipitation rate during the treatment and approaches a final value of approximately 46% after the treatment.
After termination of the ultrasonic treatment, there is no decline in precipitation rate. This indicates that the enhanced precipitation rate is caused by permanent physical changes of the seed crystals, and not by a temporarily enhancement of the liquid-solid mass transport.
The retardation effect reported by Dalas [15] was not observed. The precipitation rate already increased during the ultrasonic treatment. It is possible that this sug-gested retardation effect was camouflaged because of the relatively large increase in crystal reactivity. The used calcite seeds and supersaturation levels in both studies are quite similar. However, the differences in the ultrasonic field applied and the way
Figure 2.4.:
Ratio of crystallization rates: ultrasound/control (-). The increase in crystallization rate caused by the ultrasonic irradiation is approx. 46%.
the pH was measured during the ultrasonic irradiation differed markedly. Therefore, the causes of the contradictory observations remain unclear.
The SEM pictures of the crystals at the end of the growth experiments (t = 150 min) showed no clear difference between the control and ultrasound experiments. Never-theless, the effect of ultrasonic irradiation on the crystals could easily be observed from the samples taken during the treatment period in a saturated solution (without growth). In figs. 2.5 and 2.6, SEM pictures of, respectively, untreated and treated seed crystals in a saturated solution are shown. These pictures are a cross-selection of 21 pictures taken from 7 samples. The appearance of some fines in fig. 2.5B can be explained in two ways. First, the introduction of the dry seed crystals probably led to some initial breeding. Second, the seeds may have experienced some attrition due to the pumping and stirring action in the set-up. Attrition may contribute to the increasing slope of the control growth curves. In Fig. 6, seed crystals treated with ultrasound are shown. It can be seen that in the first period the interpenetrated con-glomerates were disrupted. At the end of the ultrasound treatment, numerous fine particles and many damaged crystals were observed. Because supersaturation was absent in these experiments, the appearance of these fine particles can be subscribed to the attrition and breakage of the parent crystals during the ultrasonic treatment. Although the same mechanism is expected to have played a pivotal role in the actual growth experiments, it is not possible to exclude nucleation mechanisms driven by supersaturation.
Figure 2.5.:
SEM pictures of untreated seed crystals in a saturated solution: (A) starting seeds and (B) after 60 min stirring and pumping.
the ultrasonic irradiation caused a rapid increase of the total surface area available for growth. It is likely that the rapid healing of the fractured parent crystals and the outgrowth of fines after the ultrasonic irradiation period in the growth experiments camouflaged any differences in crystal habit at the end. Furthermore, the wide crystal size distribution of the seed crystals makes it difficult to distinguish grown particle fragments from original small seeds.
Besides calcite, other polymorphs of calcium carbonate are aragonite and vaterite. Although the preference for the calcite polymorph is thermodynamically favored under the experimental conditions of the bulk liquid (25◦C, 1 atm pressure) [29], the local
conditions in the vicinity of an imploding cavity are significantly different [23]. It is questionable, however, if the change in local conditions is maintained for a sufficient time to allow the growth of a different polymorph. In addition, the presence of calcite seeds kinetically dictates the growth of this polymorph. Also, ATR-FT-IR spectra of the grown seeds did not show any characteristic adsorptions other than those for calcite. Therefore, it is reasonable to expect that no other polymorphs grew besides calcite.
The particle size distributions by number of the treated and untreated seed crystals are shown in fig. 2.7. Note that the lower detection limit of the particle size measure-ment was 1.5 µm. Therefore, the large number of fines present in the treated samples according to the SEM pictures did not contribute to these distributions. Although the small seeds and seed fragments were not counted, there is a shift in the distributions. The arithmetic (number-length) mean of the number distribution (fig. 2.7A) shifts
Figure 2.6.:
SEM pictures of seed crystals treated with ultrasound in a saturated solution: (A) after 5 min and (B) after 45 min of ultrasonic irradiation.
from 10.3 to 8.9 µm. This indicates that there is a decrease in the number of larger particles (note the drop in the distribution between 10 and 20 µm) and the number of particles in the lower part of the distribution increases (1.5-3.2 µm). This shows that the seed crystals are fragmented into smaller parts. These observations are in line with the SEM analysis.
The SEM and particle size analysis showed that the physical effects that occurred during ultrasonic irradiation gave rise to a change in the habit and particle size of the calcite crystals. The resulting increment of the surface area is a major factor in the observed growth rate enhancement. In addition, it can be assumed that the density of active growth sites on the fractured crystal surfaces is much higher and facilitates a higher growth rate compared to the ripened smooth crystal surfaces. It is well known, however, that these fractured crystal surfaces rapidly heal and proceed to grow at a much lower rate [34]. This might explain why the ratio between the ultrasound and control crystallization rates in fig. 2.4 is still increasing right after termination of the ultrasonic treatment before it reaches a constant value.
2.4. Conclusion
The effect of ultrasound on the seeded calcite crystallization was investigated. A ded-icated experimental set-up based on the constant composition method gave highly reproducible results. It was shown that the volumetric crystallization rate was in-creased by 46% after 45 min of ultrasonic irradiation (17 W dm−3). Control
ex-Figure 2.7.:
Particle size distribution (equivalent area diameter) by number of the untreated and treated seed crystals.
periments without supersaturation showed that during the irradiation period, seed crystals were disrupted and many fines appeared. SEM and particle size analysis showed that it is possible to alter the habit and size of the calcite crystals through ultrasonic irradiation. The increased surface area available for crystal growth resulted in the observed crystallization rate enhancement.
Acknowledgements
This work was performed in the TTIW-cooperation framework of Wetsus, Centre of Excellence for Sustainable Water Technology (www.wetsus.nl). Wetsus is funded by the Dutch Ministry of Economic Affairs, the European Union Regional Development a Fund, the Province of Fryslˆan, the City of Leeuwarden and the EZ/Kompas program of the “Samenwerkingsverband Noord-Nederland”. The authors like to thank the participants of the research theme “Concentrates” and theme “Sensoring” for the discussions and their financial support.
References
[1] A. Abbas, M. Srour, P. Tang, H. Chiou, H. Chan, and J. A. Romagnoli. “Sonocrystalli-sation of sodium chloride particles for inhalation”. In: Chem. Eng. Sci. 62.9 (2007), pp. 2445–2453 (cit. on p. 16).
[5] H. Li, J. Wang, Y. Bao, Z. Guo, and M. Zhang. “Rapid sonocrystallization in the salting-out process”. In: J. Cryst. Growth 247.1 (2003), pp. 192–198 (cit. on p. 16). [6] H. Li, H. Li, Z. Guo, and Y. Liu. “The application of power ultrasound to reaction
crystallization”. In: Ultrason. Sonochem. 13.4 (2006), pp. 359–363 (cit. on p. 16). [7] M. Louhi-Kultanen, M. Karjalainen, J. Rantanen, M. Huhtanen, and J. Kallas.
“Crys-tallization of glycine with ultrasound”. In: Int. J. Pharm. 320.1 (2006), pp. 23–29 (cit. on p. 16).
[8] M. D. Luque de Castro and F. Priego-Capote. “Ultrasound-assisted crystallization (sonocrystallization)”. In: Ultrason. Sonochem. 14.6 (2007), pp. 717–724 (cit. on p. 16). [9] E. Miyasaka, S. Ebihara, and I. Hirasawa. “Investigation of primary nucleation phe-nomena of acetylsalicylic acid crystals induced by ultrasonic irradiation–ultrasonic energy needed to activate primary nucleation”. In: J. Cryst. Growth 295.1 (2006), pp. 97–101 (cit. on p. 16).
[10] A. Kordylla, S. Koch, F. Tumakaka, and G. Schembecker. “Towards an optimized crystallization with ultrasound: Effect of solvent properties and ultrasonic process parameters”. In: J. Cryst. Growth 310.18 (2008), pp. 4177–4184 (cit. on p. 16). [11] L. McCausland and P. Cains. “Ultrasound to make crystals.” In: Chem. Ind. 9 (2003),
pp. 15–16 (cit. on p. 16).
[12] G. Ruecroft, D. Hipkiss, T. Ly, N. Maxted, and P. W. Cains. “Sonocrystallization: the use of ultrasound for improved industrial crystallization”. In: Org. Proc. Res. Dev. 9.6 (2005), pp. 923–932 (cit. on p. 16).
[13] N. Amara, B. Ratsimba, A. Wilhelm, and H. Delmas. “Growth rate of potash alum crystals: comparison of silent and ultrasonic conditions”. In: Ultrason. Sonochem. 11.1 (2004), pp. 17–21 (cit. on p. 16).
[14] N. Amara, B. Ratsimba, A. M. Wilhelm, and H. Delmas. “Crystallization of potash alum: effect of power ultrasound”. In: Ultrason. Sonochem. 8.3 (2001), pp. 265–270 (cit. on p. 16).
[15] E. Dalas. “The effect of ultrasonic field on calcium carbonate scale formation”. In: J. Cryst. Growth 222.1 (2001), pp. 287–292 (cit. on pp. 16, 22).
[16] J. Dodds, F. Espitalier, O. Louisnard, R. Grossier, R. David, M. Hassoun, F. Baillon, C. Gatumel, and N. Lyczko. “The effect of ultrasound on crystallisation-precipitation processes: Some examples and a new segregation model”. In: Part. Part. Sys. Charact. 24.1 (2007), pp. 18–28 (cit. on p. 16).
[17] Z. Guo, A. G. Jones, and N. Li. “The effect of ultrasound on the homogeneous nu-cleation of BaSO4 during reactive crystallization”. In: Chem. Eng. Sci. 61.5 (2006), pp. 1617–1626 (cit. on p. 16).
[22] A. Kordylla, T. Krawczyk, F. Tumakaka, and G. Schembecker. “Modeling ultrasound-induced nucleation during cooling crystallization”. In: Chem. Eng. Sci. 64.8 (2009), pp. 1635–1642 (cit. on p. 16).
[23] Y. T. Shah, A. B. Pandit, and V. S. Moholkar. Cavitation reaction engineering. Springer, 1999 (cit. on pp. 16, 24).
[24] S. J. Doktycz and K. S. Suslick. “Interparticle collisions driven by ultrasound”. In: Science 247.4946 (1990), pp. 1067–1069 (cit. on p. 16).
[25] K. Suslick. “Sonochemistry”. In: Science 247.4949 (1990), pp. 1439–1445 (cit. on p. 16). [26] S. S. Berdonosov, I. V. Znamenskaya, and I. V. Melikhov. “Mechanism of the vaterite-to-calcite phase transition under sonication”. In: Inorg. Mater. 41.12 (2005), pp. 1308– 1312 (cit. on p. 17).
[27] T. F. Kazmierczak, M. B. Tomson, and G. H. Nancollas. “Crystal growth of calcium carbonate. A controlled composition kinetic study”. In: J. Phys. Chem. 86.1 (1982), pp. 103–107 (cit. on p. 17).
[28] M. B. Tomson and G. H. Nancollas. “Mineralization kinetics: A constant composition approach”. In: Science 200.4345 (1978), pp. 1059–1060 (cit. on p. 17).
[29] M. M. Reddy and G. H. Nancollas. “The crystallization of calcium carbonate: I. Iso-topic exchange and kinetics”. In: J. Colloid Interface Sci. 36.2 (1971), pp. 166–172 (cit. on pp. 17, 24).
[30] S. Brunauer, P. H. Emmett, and E. Teller. “Adsorption of gases in multimolecular layers”. In: J. Am. Chem. Soc. 60.2 (1938), pp. 309–319 (cit. on p. 17).
[31] J. N`yvlt and M. Matuchov´a. “Determination of linear growth rates of crystals (II). The shape factors method”. In: Krist. Tech. 11.3 (1976), pp. 245–253 (cit. on p. 21). [32] M. Matuchov´a and J. N`yvlt. “Determination of linear growth rates of crystals (I).
Calculation of linear growth rates of individual crystal faces from overall rates”. In: Krist. Tech. 11.2 (1976), pp. 149–161 (cit. on p. 21).
[33] A. R. Hoch, M. M. Reddy, and G. R. Aiken. “Calcite crystal growth inhibition by humic substances with emphasis on hydrophobic acids from the Florida Everglades”. In: Geochim. Cosmochimi. Acta 64.1 (2000), pp. 61–72 (cit. on p. 21).
abstract The acoustic cavitation (42,080 Hz, 7.1 W cm−2 or 17 W) effects on suspended calcite crystals, sized between 5 and 50 µm, have been visualized for the first time using high speed photography. High speed recordings with a duration of 1 s containing up to 300,000 frames per second, revealed the effect of cluster and streamer cavitation on several calcite crystals. Cavitation clusters, evolved from cavitation inception and collapse, caused attrition, disruption of aggregates and deagglomeration, whereas streamer cavitation was observed to cause deagglomeration only. Cavitation on the surface gave the crystals momen-tum. However, it is shown that breakage of accelerated crystals by interparticle collisions is unrealistic because of their small sizes and low velocities. Crystals that were accelerated by bubble expansion, subsequently experienced a deceleration much stronger than expected from drag forces, upon bubble collapse. Experiments with pre-dried crystals seemed to support the current theory on bubble nucleation through the presence of pre-existing gas pockets. However, experiments with fully wetted crystals also showed the nucleation of bub-bles on the crystal surface. Although microjet impingement on the crystal surface could not be directly visualized with high speed photography, scanning electron microscopy (SEM) analysis of irradiated calcite seeds showed deep circular indentations. It was suggested that these indentations might be caused by shockwave induced jet impingement. Furthermore, the appearance of voluminous fragments with large planes of fracture indicated that acoustic cavitation can also cause the breakage of single crystal structures.
This chapter has been published as: R.M. Wagterveld, L. Boels, M.M. Mayer, G.J. Witkamp. Visual-ization of acoustic cavitation effects on suspended calcite crystals. Ultrasonics Sonochemistry 18 (2011) 216-225.
inertial cavitation (or transient cavitation), which is the high energetic fast growth and collapse of bubbles [1].
Besides the pressure constraint, inertial cavitation can only occur if the radius of the initial cavity lies between a lower and upper critical value for cavitation. With decreasing frequency of the applied ultrasound, this size range increases which results in more inertial cavitation [2].
In many practical systems the inception of cavitation is observed at much lower tensile stress than calculated using the classical cavitation nucleation theory. The generally accepted theory is that the cavitation inception is mediated by cavities already present in the liquid. The existence of these nucleation sites is usually sub-scribed to gas pockets in crevices on surfaces and particles. The gas pockets arise by the introduction of water in the container, or by introduction of dry particles into liquid. Some reported experiments, however, cannot be explained by this theory [3]. Mørch developed a slightly modified theory based on surface structure [4]. In this theory it is assumed that in case of concave surface structures, stress arises at the liquid solid interface. Mørch stated that in this situation, the liquid ruptures at lower tensile stress, without the need of gas pockets.
When ultrasound is passed through a liquid-solid system, bubble cavitation causes a series of unique physical phenomena that can affect the solid. Microjets and high energetic shockwaves are produced by inertial cavitation. Shockwaves are formed when cavities rapidly expand or collapse. Recently, it has been shown experimentally that spherical particles can be accelerated significantly by the shockwave that occurs during the explosive growth of a cavity on the particle surface (cavitation inception) [5, 6]. During the collapse of a cavity, high local temperatures and pressures arise which result in a pressure shockwave [7–12]. Shockwaves may cause mechanical damage to close objects and are known to cause material erosion [1]. This phenomenon is used in lithotripsy to fragment kidney stones [13–16]. The collapse of bubbles close to a large
processes are related to the process of crystal nucleation [18]. Here, the focus is on the physical effects that ultra sound has on existing crystals.
In this paper a distinction is made between agglomerates and aggregates using the following definitions. An agglomerate is an assemblage of particles which are loosely coherent, and an aggregate is an assemblage of particles rigidly joined together (ISO14887 [19]). Shockwaves can reduce agglomeration. Ultrasonic treatment of crystals might lead to erosion which can change the crystal habit [20, 21]. Recently, the authors reported that ultra sound can increase the volumetric crystal growth rate of calcite significantly [22] (chapter 2). It was shown that this effect was related to the alteration of the seed crystals’ habit and size. Seed crystals seemed to be subjected to attrition and breakage during ultrasonic irradiation. Although it was suggested that the rupture of agglomerates occurs in close proximity of a collapsing bubble [23], no direct visualization of this phenomenon has been reported yet. One attempt was made by Guo et al. who investigated the acoustic effects on large sugar crystals. It was suggested that both the collisions of crystals and vibration and implosion of cavitation bubbles lead to crystal deagglomeration and breakage [24]. However, the reported observations were recorded with a frame rate of only 4500 fps. This is far too low to capture the 20,000 Hz acoustic phenomena (4 cavitation cycles per frame) that actually caused the observed breakage and deagglomeration of the sugar crystals. Several in situ methods have been used to investigate the effect of cavitation on par-ticles or surfaces, like high speed (HS) imaging (in combination with schlieren/shadow imaging [7–10]) and optical beam deflection (OBD) measurements [9, 11]. Further-more, Scanning Electron Microscopy (SEM) analysis can be used to investigate the resulting particles or surfaces after ultrasonic treatment, to show attrition or breakage of solids, deagglomeration and pitting [25–27]. In high speed photography, most of the research involved large surfaces [28–34] and relatively large (metal) particles [5, 6, 24, 35]. Also, high speed photography was used to study bubble-bubble interaction [36], cavitation clusters [37, 38] and resonant bubbles [39–43].
The aim of the present work is to identify the cavitation phenomena that are re-sponsible for crystal deagglomeration, disruption of aggregates, attrition and breakage
