Role of Acoustic Fields in Promoting the Gas-Solid Contact in a Fluidized Bed of Fine Particles †

The present work is a review presenting the main results obtained by our research group in the field of soundassisted fluidization of fine particles. Our aim is to highlight the role of acoustic fields in enhancing the gas-solid contact efficiency, with specific attention to the phenomenological mechanism upon which this technique is based. In particular, the first section presents the characterization of the fluidization behaviour of four different nanopowders in terms of pressure drops, bed expansion, and minimum fluidization velocity as affected by acoustic fields of different intensity and frequency. The fluidization of binary mixtures comprising two powders is also investigated under the application of different acoustic fields and varying the amounts of the two powders. The second section focuses on the study of the mixing process between two different nanopowders both from a “global/macroscopic” and “local/microscopic” point of view and highlighting the effect of mixture composition, primary particles density and sound intensity. The last section presents a promising application of sound-assisted fluidization, i.e. CO2 capture by adsorption on a fine activated carbon, pointing out the effect of CO2 partial pressure, superficial gas velocity, sound intensity and frequency on the adsorption efficiency.


Introduction
Ultraf ine powders, in particular nanoparticles (< 100 nm), have received increased attention in recent years.Indeed, due to their unique properties arising from their very small primary particle size and very large surface area per unit mass, nanoparticles provide higher contact and reaction efficiencies than traditional materials (Hakim et al., 2005), thus finding application in different industrial sectors such as in the manufacture of cosmetics, foods, plastics, catalysts, energetic materials, biomaterials, and in micro-electromechanical systems (MEMS) (Nakamura and Watano, 2008).Within this framework, it has become increasingly important to understand how these nanoparticles can be handled and processed in large quantities (i.e.mixing, transporting, coating).However, before the processing of such materials can take place, the nanoparticles have to be well dispersed.In this respect, gas fluidization is one of the most effective available tech-niques in ensuring continuous powder handling and good mixing, chiefly because of the large gas-solid contact area (Zhu et al., 2005;Lepek et al., 2010a).Nevertheless, on the basis of their primary particle size and material density, nanosized powders fall under the Geldart group C (< 30 μm) classification, which means that their fluidization is expected to be particularly difficult (i.e.characterized by plug formation, channelling and agglomeration) because of cohesive forces (such as van der Waals, electrostatic-and moisture-induced surface tension forces) existing between particles and which become more and more prominent as the particle size decreases (King et al., 2008;Scheffe et al., 2009;Wang et al., 2002;Gündoğdu and Tüzün, 2006).Despite their Geldart classification, growing experimental evidence proves that nanoparticles can be smoothly fluidized for an extended range of gas velocities, thus implying that primary particle size and density are not representative parameters for predicting their fluidization behaviour.Indeed, due to the interparticle forces mentioned above, nanoparticles are always found to be in the form of large-sized porous aggregates (King et al., 2008;Scheffe et al., 2009;Saleh et al., 2006), rather than as individual nanosized particles when packed together in a gaseous medium.In other words, nanoparticles actually fluidize in the form of nanoparticle aggre-gates, and their properties (size/density) have a significant effect on the fluidization nature.Several experimental campaigns (Wang et al., 2002) show that highly porous nanoparticle aggregates can exhibit two distinct fluidization behaviours: APF (agglomerate particulate fluidization) and ABF (agglomerate bubbling fluidization).The former is characterized by very large bed expansion, smooth fluidization and very low minimum fluidization velocity; the latter, instead, shows little bed expansions and bubbling.
All things considered, even though the fluidization of such materials is made possible by agglomeration, this phenomenon on the other hand also represents a strong limitation to the exploitation of their full potential because of the undesired decrease in specific surface area.Accordingly, it is always preferable to contain the formation of aggregates as much as possible, i.e. the aggregate size should be relatively small to enable proper exploitation of the potential of nanoparticles.In other words, the achievement of a smooth fluidization regime is closely related to an efficient break-up of the large aggregates yielded by cohesive forces.To this aim and to overcome these inter-particle forces and achieve a smooth fluidization regime, externally assisted fluidization can be used, thus involving the application of additional forces generated, for instance, by acoustic (Ammendola and Chirone, 2010;Raganati et al., 2011a;Raganati et al., 2011b), electric (Lepek et al., 2010b;Valverde et al., 2009) and magnetic (Zeng et al., 2008;Yu et al., 2005) fields or mechanical vibrations (Nam et al., 2004;Yang et al., 2009a) to excite the fluidized bed, thus avoiding channelling and enhancing the dynamics of the powder.Among all these available techniques, sound-assisted fluidization has been indicated as one of the best technological options to smoothly fluidize fine and ultra-fine powders: under the influence of appropriate acoustic fields, channelling and/ or slugging tends to disappear, the bed expands uniformly and the minimum fluidization velocity is distinctly reduced (Ammendola and Chirone, 2010;Raganati et al., 2011a;Raganati et al., 2011b).Besides this, it is worth noting that this technique also holds advantages from a practical point of view: i) it is not intrusive, since neither additional equipment nor materials must be inserted in the bed; ii) the powders to be employed do not need to have any special property; iii) as widely reported in literature (Russo et al., 1995), the application of acoustic fields is capable of reducing the elutriation of fine particles from a fluidized bed, preventing problems related to downstream carry-over of fine particles such as clogging of valves; iv) last but not least, this technique is extremely economical and user-friendly, since the extra equipment (signal generator, audio amplifier loudspeaker and oscilloscope) required is very easily available on the market.
This review is not intended to be comprehensive, as it presents the main results obtained by our research group in the field of the sound-assisted fluidization of fine particles.Our focus here is on highlighting the role of acoustic fields in enhancing the gas-solid contact, with particular attention to the phenomenological mechanism at the basis of this technique.
In particular, the first part of section 2 presents a fundamental study describing the fluidization behaviour of four different nanopowders, Al 2 O 3 , Fe 2 O 3 , CuO and ZrO 2 , as affected by the application of acoustic fields.The effect of sound frequency (50-300 Hz) and sound intensity (SPL) (125-150 dB) on the fluidization quality in terms of pressure drops and bed expansion across the bed and minimum fluidization velocity is reported.The fluidization of binary mixtures of the two powders (Al 2 O 3 , Fe 2 O 3 ) is also investigated under the application of different acoustic fields (130-135 dB, 120 Hz) and varying the amounts of the two powders from 7 %wt to 90 %wt of Fe 2 O 3 .Then, aiming to explain sound-assisted fluidization from a phenomenological point of view, the second part of section 2 focuses on the mixing process between two different nanopowders.Scanning electron microscopy with X-ray microanalysis (SEM/EDS) of the captured samples lets us obtain key information about the dynamics of the mixing process both from a "global/macroscopic" and the "local/microscopic" point of view: the time dependence of the mixing degree, its asymptotic value and the mixing characteristic time are evaluated.The effect of mixture composition, primary particle density and SPL is also studied.
The positive and promising results obtained from this fundamental and phenomenological study form the basis for a second research activity aimed to apply and exploit this experimented capability of the acoustic fields in enhancing the fluidization quality and consequently the gas-solid contact in an actual gas-solid process such as adsorption on fine solid sorbents.Within the framework of the pressing environmental concerns about the escalating level of atmospheric carbon dioxide (Stone et al., 2009;Markewitz et al., 2012;Orr, 2009), CO 2 sequestration from flue gas streams by adsorption is recognized as being one of the most promising capture techniques (D'Alessandro et al., 2010).However, for adsorption to be used in large-scale applications, two main aspects need to be addressed.The first is the choice of the adsorbent materials (Drage et al., 2012).To date, great attention is focused on the development of highly specific adsorbent materials, namely materials with great affinity towards CO 2 molecules (Choi et al., 2009).Sure enough, the scientific community is moving in this direction, putting the emphasis on the manufacture of designed nanomaterials in which a molecular level of control can be achieved as a means of tailoring their CO 2 capture performance (Wang et al., 2009;Dawson et al., 2013;Baxter et al., 2009).However, although certain attributes of solid sorbents prove to be promising, they must still be integrated into a viable process (D'Alessandro et al., 2010) which includes the implementation of equipment that can take full advantage of the sorbent properties and maximize the separation efficiency (Sjostrom et al, 2011;Yang et al., 2009b).The choice of the reactor configuration is, therefore, the second crucial point to be considered.In this regard, fluidized beds can be a promising solution since they are characterized by a very efficient gas-solid contact and intense solid mixing, thus increasing the efficiency of reactor operation in terms of extremely high mass and heat transfer coefficients (Yang et al., 2009b).So, we proposed sound-assisted fluidization as a viable technological solution which would be capable of matching these two critical demands, i.e. fine solid sorbents and fluidized beds (Raganati et al., 2014a;Raganati et al., 2014b;Valverde et al., 2013).
To this end, section 3 reports the results obtained from CO 2 adsorption tests carried out on fine activated carbon particles (0.39 μm) in a sound-assisted fluidized bed.All the adsorption tests are performed at ambient temperature and pressure in a laboratory-scale reactor under ordinary conditions and under the effect of acoustic fields of different intensities (125-140 dB) and frequencies (20-300 Hz).In particular, the effectiveness of CO 2 adsorption is assessed in terms of the moles of CO 2 adsorbed per unit mass of adsorbent, breakthrough time and fraction of bed utilized at breakpoint.Then, the effect of sound frequency and SPL, the superficial gas velocity (1, 1.5, 2 cm/s) and CO 2 concentration in the feed stream (5, 10 and 15 %vol. in N 2 ) is studied.

Experimental apparatus, materials and procedures
All the fluidization tests were performed at ambient temperature and pressure in a laboratory-scale soundassisted apparatus consisting of a fluidization column (40 mm ID and 500 mm high) made of Plexiglas.It is equipped with a porous plate gas distributor, a 300-mm-high wind-box filled by Pyrex rings to ensure an even distribution of gas flow, a pressure transducer installed at 5 mm above the gas distributor to measure the pressure drop across the bed, a sound wave guide at the top of the freeboard, a sound-generation system and a data acquisition system.This experimental set-up was also designed according to the Helmholtz resonator, i.e. one of the most-used engineering noise control methods, in order to reduce the sound insulation even for high intensity acoustic fields.
Nanopowders of Al 2 O 3 , Fe 2 O 3 , ZrO 2 and CuO with primary particle average sizes lower than 50 nm and densities of about 4000, 4500, 6000 and 6300 kg/m 3 , respectively, were used.
More detailed information about both experimental apparatus and materials can be found elsewhere (Ammendola et al., 2011a).
Fluidization of single powders: Aeration tests were carried out with an initial bed of 48, 33, 35 and 75 g for Al 2 O 3 , Fe 2 O 3 , CuO and ZrO 2 , respectively, corresponding to a bed height of about 15 cm for all the four powders.Dimensionless pressure drop (ΔP/ΔP 0 ) and bed expansion curves (H/H 0 ) were obtained both with and without the application of acoustic fields of different intensities (125-150 dB) and frequencies (50-300 Hz), ΔP being the actual pressure drop across the bed, ΔP 0 the pressure drop equal to the buoyant weight of particles per unit area of bed, H the actual bed height and H 0 the initial bed height under fixed bed conditions Fluidization of binary mixtures: the fluidization tests of binary mixtures were carried out using Al 2 O 3 and Fe 2 O 3 nanopowders.On the basis of the experiments carried out on the two powders alone, two different sound intensities (130/135 dB) at a fixed frequency (120 Hz) were applied.In particular, tests were performed by loading 30 g of the binary mixtures and increasing the amount of Fe 2 O 3 from 7 to 90 %wt, so that the effect of the relative amount of the two powders on the fluidization quality could be highlighted.Then, the effect of the initial loading order of the two powders inside the column was also investigated.
The experimental pressure drops and bed expansion curves of the single powders were elaborated and three parameters evaluated as an index of the fluidization quality: 1) the maximum value of dimensionless pressure drop (ΔP max /ΔP 0 ); 2) the maximum value of dimensionless bed expansion (H max /H 0 ); 3) the minimum fluidization velocity (u mf ).
Mixing: Mixing tests on Al 2 O 3 and Fe 2 O 3 nanopowders were performed under application of a fixed acoustic field (130 dB-120 Hz).Experiments were carried with an initial bed height of about 15 cm, corresponding to a bed of 30 g.The superficial gas velocity was fixed at about 0.45 cm/s, enough to fluidize the materials under application of the above acoustic field and to assure a ratio between the superficial gas velocity and the minimum fluidization velocity of about 7 for Al 2 O 3 , namely the powder characterized by the worse fluidization quality.The effect of the relative amounts of the two powders on the mixing quality was investigated by performing three tests, noted as A, B and C, with a Fe 2 O 3 amount of 17, 50 and 77 wt%, respectively.For all experimental conditions, the starting point was obtained by loading the Al 2 O 3 nanopowder and then the Fe 2 O 3 one.Each test was carried out for about 120 min.
Two mixing tests on ZrO 2 and CuO were also per-formed in order to study the effect of nanoparticle density.
The relative amounts of the two powders was fixed at 50 wt% of ZrO 2 and two different acoustic fields were applied (130/140 dB-120 Hz).In particular, the starting point was obtained by feeding 35 g of the ZrO 2 nanopowder and then 35 g of the CuO nanopowder, in order to have the same bed height used in the tests performed on Fe 2 O 3 and Al 2 O 3 (15 cm).The superficial gas velocity was fixed at about 1.5 cm/s, enough to fluidize the materials under application of the above-mentioned acoustic fields and so that the ratio between the superficial gas velocity and the minimum fluidization velocity was about 7 for ZrO 2 , that is the powder characterized by the worse fluidization quality.Each test was carried out for about 120 min.
Two possibilities of mixing two powders, both at a global scale (i.e.mixing between aggregates formed by only one powder) and at a local scale (i.e.mixing inside the aggregates, leading to the formation of hybrid aggregates formed by both powders) have been investigated.In this regard, the study was carried out by means of both a visual observation of the bed and SEM/EDS analysis.The visual observation of the bed gave some preliminary rough information on the uniformity of the mixing and the mixing characteristic times on the global scale.During experiments, samples of the fluidized materials were taken at different times by means of a non-destructive sampling procedure (Ammendola et al., 2011b) and then analysed by SEM/EDS analysis to determine the chemical composition of aggregates.In particular, a probe made of a silicon tube linked to the adhesive sample disc used for SEM analysis was carefully inserted from the top of the reactor so that the fluidized materials present in the upper part of the bed stuck to it.On this basis, the time dependence of the mixing degree, its asymptotic value and the mixing characteristic time on the local scale were evaluated.

Fluidization of single powders
The fluidization quality of Al 2 O 3 , Fe 2 O 3 , CuO and ZrO 2 nanopowders is very poor without the application of acoustic fields, as clearly confirmed by the obtained fluidization and expansion curves (Fig. 1).
In particular, incrementing the gas flow rate causes the bed to be lifted at first by a slug which then collapses, giving rise to a structure which-even though displaying a certain expansion ratio-is characterized by channels, an uneven surface and a substantial lack of particle motion.Therefore, the application of an acoustic field is investigated to achieve a proper fluidization regime, as clearly shown in Fig. 2, reporting the dimensionless pressure drop and expansion curves obtained at 140 dB and 120 Hz.Indeed, the dimensionless pressure drop always reaches the asymptotic value of 1, thus indicating that the particle bed is completely fluidized, and also the expansion ratio is enhanced.
Fe 2 O 3 , on the contrary, is characterized by a good fluidization behaviour also under ordinary conditions as shown by both pressure drop and bed expansion curves (Fig. 1).In particular, a quite stable fluidization regime was achieved also in this case after the formation and break-up of a plug.However, as for the other three powders, the application of a suitable acoustic field can strongly improve the fluidization quality (Fig. 2).
In order to point out the most effective ranges of sound intensities and frequencies needed to obtain a good fluidization regime, ΔP max /ΔP 0 and H max /H 0 values were plotted as functions of the SPL at fixed frequency and vice versa for all the powders, as reported in Fig. 3a and b.The analysis of these curves suggests that, at a fixed SPL, the ranges of frequencies 100-125 Hz and 90-120 Hz were able to stabilize an optimum fluidization condition for Al 2 O 3 and CuO, respectively, whereas the best frequencies fall in the range 80-120 Hz for Fe 2 O 3 and ZrO 2 .On the other hand, at fixed frequency, Al 2 O 3 , CuO and ZrO 2 need acoustic fields of a SPL higher than 135 dB to obtain a good fluidization quality, whereas for Fe 2 O 3 , intensities higher than 125 dB are enough.
The pressure drop curves were elaborated in order to evaluate the minimum fluidization velocity for all the powders.The results obtained show that the denser powders (CuO and ZrO 2 ) are characterized by minimum fluidization velocities one order of magnitude higher than the lighter ones (Al 2 O 3 and Fe 2 O 3 ).
The trends obtained for the minimum fluidization velocities reported in Fig. 4 are consistent with the above-mentioned evidence concerning the influence of sound on the fluidization quality.In particular, at fixed frequency, an increase of the SPL results in a decrease of u mf , whereas at fixed intensity, an optimum range of frequencies exists for all the powders according to what was stated above.The beneficial effect generally shown by the SPL can be explained considering that an increase of the sound intensity implies an intensification of the energy introduced inside the bed, i.e. the external force exercised by the acoustic field on the aggregates is amplified.Therefore, large aggregates are more easily broken into smaller ones, thus consequently determining a reasonable decrease of u mf .On the other hand, sound frequency has a not monotone effect on the fluidization quality.This is due to the ability of sound to penetrate the bed as well as to promote aggregate reduction down to a scale that also depends on powder structure.In particular, the application of the acoustic field induces a relative motion between larger and smaller aggregates, thus leading to the break-up of the large aggregates originally present in the bed.For frequencies higher than about 125 Hz, the acoustic field is not able to properly propagate inside the bed, whereas for frequencies lower than about 80 Hz, the relative motion between smaller and larger sub-aggregates is practically absent.Between these values, there is a range of optimal frequencies able to maximize aggregate break-up.
All these considerations on the different fluidization behaviours were explained and clarified by referring to the different fluidized particle aggregates' properties, size and density, produced by the application of sound (Ammendola et al., 2011a).The results obtained show that: (i) the particle aggregate's size is far higher than the primary particle one (hundreds of microns), thus confirming that fine particle fluidization actually occurs in the form of aggregates; (ii) varying the sound parameters, the aggregate size and densities have opposite trends; in particular, at fixed frequency, aggregate size decreases at increasing SPL, as observed for u mf , whereas aggregate density increases; at fixed SPL, the ranges of frequencies which ensure the lowest values of u mf give the lowest values of aggregate size and the maximum values of the aggregate density.This behaviour can be related to the fact that the more efficient break-up mechanism produced by higher values of sound intensity implies a reasonable reduction of aggregate size and in turn, of u mf along with a greater bed compactness; in other words, higher sound intensities yield smaller as well as denser aggregates.
As regards the influence of frequency, on the one hand, optimal frequencies are able to promote the break-up of particle aggregates as well, thus determining lower values of aggregate size and consequently of u mf , whereas on the other hand, they promote a greater bed compactness.

Fluidization of binary mixtures
Fig. 5 reports the dimensionless pressure drop and bed expansion curves obtained for the different binary mixtures of Al 2 O 3 and Fe 2 O 3 under the effect of different acoustic fields.In particular, on the basis of the results obtained for the fluidization of single powders, two different sound intensities, namely 130 and 135 dB, were adopted at a fixed sound frequency of 120 Hz.Indeed, under these operating conditions the two nanopowders showed different behaviours: both a poor and a good fluidization quality was achieved for Al 2 O 3 and Fe 2 O 3 , respectively.The pressure drop and bed expansion curves obtained during the fluidization of single powders have also been reported for comparison.Obviously, mixtures with a high amount of Al 2 O 3 or Fe 2 O 3 behave like the powders alone, while mixtures with an intermediate composition have an intermediate behaviour in terms of both pressure drop and bed expansion.The addition of Fe 2 O 3 , the most fluidizable powder, to Al 2 O 3 generally improves the fluidization quality of the mixture in terms of higher values of both pressure drop and bed expansion.In particular, the maximum value of dimensionless pressure drop, ΔP max /ΔP 0 , reached at high superficial gas velocity, approaches unity (i.e. the limit value achievable in a condition of ideal fluidization) as the Fe 2 O 3 amount increases.
In particular, the beneficial effect deriving from an increasing amount of Fe 2 O 3 becomes relevant from a weight composition of Fe 2 O 3 of 1/3 %wt.Moreover, this beneficial effect is clearly stronger at a lower SPL (130 dB), where the Al 2 O 3 nanopowder alone showed a very poor fluidization quality.This evidence is likely due to the better fluidization quality of Al 2 O 3 at a higher SPL (135 dB).The initial loading order of the two powders inside the column has a negligible effect.The experimental values of u mf were plotted as functions of the weight composition of the mixture in Fig. 6, where the values of u mf obtained for the single Al

Mechanism of nanopowder mixing
The efficiency of the mixing process between different nanopowders during aeration promoted by the application of a suitable acoustic field was verified.In particular, mixing occurs within different time periods depending on whether the phenomenon is observed from a macroscopic or microscopic point of view.The visual observation of the bed, also recorded by a video camera, gave some preliminary qualitative information on the uniformity of mixing and the mixing characteristic time from a macroscopic point of view.Fig. 7 refers to the B test: the starting point was obtained by feeding 15 g of the Al 2 O 3 nanopowder (white powder) and then 15 g of the Fe 2 O 3 nanopowder, which forms a red layer on the top of the bed (Fig. 7a).Fig. 7b refers to 1 min bed aeration with a nitrogen flow rate of u = 0.45 cm/s without the application of sound.Fig. 7c and d is relative to bed aeration assisted by the application of sound (SPL = 130 dB, f = 120 Hz) for 10 s and 2 min, respectively.Analysis of the figures shows that without the application of acoustic fields, no mixing occurs (Fig. 7b).Channelling is present, namely the formation of one or more channels through the bed can be noticed.The application of the acoustic field results in a relatively large bed expansion and solid mixing (Fig. 7c).After a few minutes, the entire bed turned brown and appeared to be well-mixed (Fig. 7d).On a macroscopic scale, the visual observation of the bed highlighted the effectiveness of sound application in promoting the fluidization and the global mixing of nanopowders within few minutes.Similar results have been obtained for A and C tests.
In order to obtain in-depth information, samples of the fluidized materials were taken from the upper part of the bed at different times during these experiments by means of the ad-hoc non-destructive sampling procedure described in the experimental section.The different samples were analysed by SEM/EDS analysis in order to determine the shape and the chemical composition of the aggregates.On the basis of the initial bed composition of the three tests, the Al weight composition corresponds to 79, 43 and 19 % for the tests A, B and C, respectively.These values represent the theoretical limit corresponding to a complete mixing both at the global scale (average composition of the bed) and at the local scale (average composition of aggregates).The EDS analysis performed on entire areas of bed samples pointed out that the Al weight composition was close to the theoretical one already after 1 min of sound-assisted aeration for all tests.These results confirm the information obtained by visual observation of the bed, i.e. at the global scale, the mixing of the bed really does occur within a very short time.
On the other hand, the EDS analysis carried out on the single aggregates pointed out that the microscopic mixing is characterized by dynamics developing over longer times.Fig. 8 reports the SEM images and EDS analysis of nanoparticle aggregates taken during A, B and C tests at 14, 4 and 1 min, respectively, after the application of sound.
The analysis of Fig. 7 shows that all the aggregates are formed of both alumina and iron oxide, in accordance with a mixing process also occurring at the local scale.Their compositions vary over a wide range.Aggregates richer in both alumina and iron oxide were detected, but their shapes and sizes appear to be similar to each other; therefore, the simple morphological analysis is useless to distinguish the different composition of the aggregates.
Fig. 9 reports the cumulative distribution of aggregate composition at different times obtained for A, B and C tests.N denotes the number of aggregates with an Al weight composition lower than the x-axis value, and N t represents the total number of aggregates.
The curve named "initial composition of the bed sur- face" is the first of the experimental curves, namely the one corresponding to the time t = 0; whereas the curve named "limit value" represents the theoretical limit curve, achievable in the case of uniform mixing within all aggregates.The analysis of these curves suggests that the chemical composition of the nanoparticle aggregates actually changes as time goes on.In particular, a general approach of the curves towards the limit curve with increasing time can be observed, even though this evolution is slower in the case of mixtures richer in Al 2 O 3 (Fig. 9a).This evidence is absolutely consistent with the theory of the nanoparticle aggregate undergoing an actual dynamic evolution (due to a continuous break-up and re-aggregation mechanism) during the sound-assisted fluidization process.As a matter of fact, all aggregates are mixed even after 1 min of sound-assisted fluidization and their composition varies over a wide range.In order to obtain more direct information, these data were elaborated to obtain the time-dependence of the aggregate's mixing degree M(t) for A, B and C tests (Fig. 10a), where M at a fixed time t was defined as the ratio between the number of aggregates whose Al composition differs from the theo-retical one less than 10 % and the total number of aggregates analysed at time t.Each data series was fitted with an exponential rise-to-maximum law M(t) = a(1 -e -t/b ).The analysis of these curves suggests that the mixing quality in terms of both characteristic time and maximum mixing degree is strongly affected by the relative amounts of the two powders; indeed, increasing the amount of Fe 2 O 3 from 17 to 77 wt%, the asymptotic value of M(t) increases from about 50 to 100 % and, at the same time, the characteristic time of the process decreases from 87 to 10 min.In other words, less time is needed to accomplish higher values of the mixing degree as the Fe 2 O 3 mixture content is increased.The explanation of this behaviour is likely to be found in the tight link existing between the mixing effectiveness and the nature of the two powders, namely their intrinsic nature and fluidizability.Indeed, an efficient mixing can only be achieved by means of an efficient break-up and re-aggregation mechanism promoted by the acoustic field.In view of that, the increasing Fe 2 O 3 amount, namely the powder that shows the better fluidization behaviour, promotes an evident improvement of the above-mentioned mechanism, thus resulting in both the increase of the asymptotic value of the mixing degree and the decrease of the time needed by the process to reach the steady stage.This is due to the fact that mixtures with higher contents of Fe 2 O 3 are logically characterized by aggregates with higher amounts of Fe 2 O 3 that are much more fluidizable than the Al 2 O 3 ones; in other words, aggregates with higher a concentration of Fe 2 O 3 are characterized by a much more prominent tendency to undergo a dynamic evolution (break-up and re-aggregation) during the fluidization process, thus making the acoustic field more effective.In other words, the break-up of the Al 2 O 3 nanoparticles, characterized, as said before, by a worse fluidization quality, is the limiting stage of the mixing process.
In the light of what has been said so far and in order to point out how the mixing quality is affected by the sound intensity, results obtained in another experimental campaign (Ammendola and Chirone, 2010) were considered and compared.Al 2 O 3 and CuO mixing tests were performed under application of an acoustic field of 140 dB and 120 Hz, at a superficial gas velocity of 0.45 cm/s and with a CuO weight composition of 14 %, corresponding to an Al weight percentage of 80 %, that is very close to the one used for the A test (79 %).Considering that in both the tests one material, Al 2 O 3 , is the same from both a qualitative and quantitative point of view, as well as the superficial gas velocity, a comparison between the obtained results is definitely reasonable; however, possible differences arising from the different nature of the second powder have to be considered.In Fig. 10b, the mixing degree curves of the two tests are shown.The analysis of this diagram shows that, using a more fluidizable powder, Fe 2 O 3 instead of CuO, the second material Al 2 O 3 being the same, similar results in terms of both maximum mixing degree (50 and 53 %, respectively) and characteristic time (87 and 80 min, respectively), can be obtained with the application of an acoustic field of much lower intensity, 130 dB instead of 140 dB, respectively.In other words, the lower CuO fluidizability is balanced by the application of a more intense sound level, thus determining more energy introduced inside the bed, in turn resulting in the enhancement of the external force yielded by the acoustic field on the aggregates.Accordingly, the sound capability to penetrate the bed and promote the break-up of the nanoparticle aggregates, namely the mechanism behind the mixing process between the two powders at a microscopic level, is enhanced.
In order to point out the influence of the nanoparticle's density on the mixing quality, tests were performed using two denser powders, ZrO 2 and CuO (about 6000 instead of 4000 kg/m 3 of Fe 2 O 3 and Al 2 O 3 ), with a relative amount of the two powders of 50 wt%, under the application of the same acoustic field used for the Fe 2 O 3 and Al 2 O 3 mixing tests (130 dB-120 Hz).Also in this case, the bed samples have been analysed by SEM/EDS analysis in order to determine the shape and chemical composition of the aggregates.Fig. 11a reports the cumulative distribution of the aggregates' composition at different times, whereas Fig. 11b reports the time-dependence of the mixing degree along with the curve obtained for the Fe 2 O 3 and Al 2 O 3 B mixing test, reported for comparison.
The analysis of this diagram shows that with the sound, the bed height and the relative amount of the two powders being the same, the ZrO 2 and CuO mixing quality is lower than the Fe 2 O 3 and Al 2 O 3 one, in terms of both the mixing degree asymptotic value (44 % instead of 71 %) and characteristic time (64 instead of 38 min).This evidence can likely be attributed to the higher density of ZrO 2 and CuO nanoparticles, thus resulting in denser aggregates which break with more difficulty during the dynamic evolution.In other words, the ability of the sound to penetrate the bed is reasonably hindered by the higher density of the bed.

Experimental apparatus, materials and procedures
An activated carbon DARCO FGD (Norit) was used as the adsorbent material.According to its cumulative size distribution (Raganati et al., 2014b), it is characterized by a Sauter mean diameter of 0.39 μm, i.e. it belongs to the group C of Geldart's classification.It is also characterized by a large surface area (1060 m 2 /g) and broad pore size distribution.More detailed information can be found elsewhere (Raganati et al., 2014b).
Preliminary fluid-dynamic characterization: the activated carbon was previously characterized to assess its fluidization quality both under ordinary and soundassisted conditions .All the fluidization tests were performed at ambient temperature and pressure in the same experimental apparatus described above.
Adsorption tests: All adsorption tests were carried out at ambient temperature and pressure.The sorbent material was treated prior to each adsorption test by heating the powder up to 140 °C in order to remove any trace of moisture.In a typical experiment, the sorbent (110 g) is loaded in the column in order to obtain a bed height of 15 cm.Then, in a pre-conditioning step of about 10 min, N 2 is introduced into the column in order to stabilize a fluidization regime at fixed operating conditions in terms of superficial gas velocity and sound parameters.This is followed by the adsorption step in which a CO 2 /N 2 gas mixture at a fixed CO 2 concentration is fed through the column.N 2 and CO 2 flow rates were set by means of mass flow controllers (Bronkhorst) and were subsequently mixed before entering the bed.The CO 2 concentration in the column of effluent gas is continuously monitored as a function of time (breakthrough curve) by an ABB infrared gas analyser (AO2020) until the gas composition approaches the inlet gas composition value (99 %), i.e. until bed saturation is reached.CO 2 concentration profiles were obtained as a function of time t, which was counted from the time the gas mixture takes to flow from the fluidized bed to the analyser.This transit time was previously measured for each gas flow rate by flowing the gas mixture through the empty bed (about 90 s).Each adsorption test was performed both under ordinary and sound-assisted conditions.In particular, the effect of sound parameters (SPL and frequency), fluidization velocity and CO 2 partial pressure on adsorption efficiency was investigated.

Preliminary fluid-dynamic characterization
In Fig. 12a and b, the dimensionless pressure drops and bed expansion curves obtained under ordinary conditions are reported.
The fluidization quality under these conditions (i.e.without the application of any acoustic field) is particularly poor (channelling), as clearly confirmed by the fact that the asymptotic value reached by the pressure drops is lower than 1.Therefore, the application of sound is required to achieve a proper fluidization regime, which is closely related to an efficient break-up of the large aggregates produced by cohesive forces into smaller structures that are easier to fluidize.In particular, an in-depth study was carried out in order to evaluate the most effective acoustic conditions, namely whether it is possible or not to find optimal values of SPL and frequency.Fig. 13a and b report the effect of SPL (at fixed frequency, 80 Hz) and frequency (at fixed SPL, 140 dB) on the fluidization quality, respectively.
In the first place, the analysis of these curves confirms that under application of the acoustic field, more regular pressure drops and expansion curves were obtained.Then, as regards the role played by the SPL, it is clear from Fig. 13a that sound intensities higher than or equal to 125 dB are enough to obtain a good fluidization quality.In other words, 125 dB is a kind of threshold value for this activated carbon.Indeed, all the tests performed at a higher SPL (125, 135 and 140 dB) are characterized by quite similar pressure drops and expansion curves, which means that any additional increase of sound intensity does not succeed in further enhancing the fluidization quality, since the break-up mechanism is already efficient at 125 dB.On the contrary, the test performed at 120 dB is significantly worse in terms of both pressure drops and expansion ratio.As regards the sound frequency, the results reported in Fig. 13b show that it has a not monotone effect on the fluidization quality.Actually, it is possible to find an optimum range of frequency (50-120 Hz) to achieve the best fluidization quality.Either too low or too high frequencies which fall out of this range (20, 300 Hz) correspond to worse fluidization qualities.All these remarks regarding the effects of SPL and frequency can be even more clearly inferred from Fig. 14, which reports the minimum fluidization velocity, u mf , (evaluated from the pressure drop curves by means of a graphical method) as a function of SPL and frequency.Firstly, all the sound-assisted tests are characterized by a lower u mf with respect to the test performed under ordinary conditions, thus confirming the ability of the sound to enhance the fluidization quality.As for the SPL, u mf is sharply decreased passing from 120 to 125 dB and then it holds steady, regardless of the further increase of SPL.Provided the SPL is higher than 120 dB, a reasonably good fluidization quality can be attained.As for the frequency, the curve is characterized by a minimum value (in the range 50-120 Hz), corresponding to the best sound frequencies.All these results are in agreement with those obtained from the experimental campaign on the fluidiza-tion of fine powders presented in the previous section of this review.In order to highlight the most significant portion of the curve, namely the section before and soon after t b , the graph has been reported in logarithmic scale.In particu-  lar, a CO 2 inlet concentration of 10 % and a fluidization velocity of 1.5 cm/s were used in this case.These curves were worked out to evaluate: (i) the moles of CO 2 adsorbed per unit mass of adsorbent, n ads , calculated by integrating the breakthrough curves; (ii) the breakthrough time, t b , or breakpoint, which is the time it takes for CO 2 to reach the 5 % of the inlet concentration at the adsorption column outlet; (iii) the fraction of bed utilized at breakpoint (W), namely the ratio between the CO 2 adsorbed until the breakpoint and that adsorbed until saturation.

Adsorption tests 3.2.2.1 Effect of sound application
Analysis of the curves suggests that the application of sound greatly enhances the breakthrough time, which, as reported in Table 1, in sound-assisted tests (63 s) is more than four times the value obtained under ordinary conditions (12 s).The application of sound affects also the global adsorption capacity.Indeed, the total amount of CO 2 adsorbed until saturation, n ads , moves from 0.31 mol/ kg under ordinary conditions, to 0.37 mol/kg under sound-assisted conditions (Table 1).The fraction of bed utilized at breakpoint (W) is also greatly enhanced by sound, moving from values lower than 3 % in the tests performed under ordinary conditions up to values more than five times larger in the sound-assisted tests.
Finally, the application of sound greatly improves the kinetics of the entire process.Indeed, the application of acoustic fields makes it possible to speed up the adsorption process: under sound-assisted conditions, the time for CO 2 to approach the saturation value is significantly decreased (60 min for the sound-assisted test against 120 min for the test performed under ordinary conditions, Fig. 15), both the values of n ads and average rate of CO 2 adsorption being higher than those obtained under ordinary conditions.
It can be observed that soon after t = t b , the CO 2 concentration rises abruptly for the tests performed under ordinary conditions.The explanation of this evidence is likely to be found in the fluidization quality being extremely poor and unstable under ordinary conditions; indeed, unable to overcome the cohesiveness of the fine powder, most of the fluid only manages to flow across the bed by finding channels of minimum resistance.Spreading across the bed, these channels allow for a bypass of an appreciable volume of gas, thus hampering the quality of fluid-solid contact, which is the main factor ruling any adsorption process.Therefore, it is most likely that adsorption takes place mainly on those aggregates placed at the wall of the gas channels, whereas most of the adsorption surface is almost precluded by the fluid.Soon after the extremely quick saturation of the adsorption sites on those easily available aggregates, CO 2 is logically found to appear in the effluent gas.Therefore, the rather steep slope of the CO 2 concentration profile is due to the above-mentioned majority of fluid bypassing the bed without truly taking part in the adsorption process.After this sharp rise, a likewise abrupt decrease of slope can be observed in the breakthrough curve, which very slowly goes to saturation.The explanation of this behaviour is twofold: the channels are subjected to perturbation which brings fresh aggregates into contact with CO 2 , and the smaller portion of fluid actually permeating the bed.In other words, while most of the inflow CO 2 bypasses the bed, only a small fraction takes part in the adsorption, which is actually very slow, as clearly confirmed by the extremely slow breakthrough curve tail, which practically accounts for the whole adsorption process.In particular, the extreme slowness of the breakthrough tail is probably due to the action of two different aspects: the slow adsorption kinetics yielded by the poor fluidization quality and the fact that, despite a fixed flow of CO 2 being fed to the bed, only a limited fraction takes part in adsorption while the majority flows through the bed unaltered (i.e. the CO 2 flow actually undergoing adsorption is smaller than the nominal one because of the bypassing gas), thus reasonably slowing down the process.Obviously this slowness is also due to the fact that with time passing, the bed becomes more and more saturated with CO 2 , thus making further CO 2 adsorption increasingly difficult.In contrast, the sound-assisted tests are characterized by more regular breakthrough curves.Indeed, no abrupt change of slope can be noticed.The beneficial effect shown by the sound is probably due to the enhancement of the fluidization quality (which brings better gas-solid contact and mass transfer coefficients) with respect to the tests performed under ordinary conditions, namely without the aid of any external force.
In particular, the enhancement of the break-up mechanism and re-aggregation of fluidizing aggregates produced by the sound application implies the constant renewal of the surface exposed to the fluid.In other words, the continuous aggregate break-up and re-aggregation mechanism makes the surface of the activated carbon more readily available for the adsorption process.In order to verify these considerations, a further test was carried out.This test was started under ordinary conditions, and only at a time t = t*, corresponding to the above-mentioned change of slope typical of ordinary adsorption tests, the sound was switched on (Fig. 16).
Analysis of the obtained breakthrough curve clearly shows that for t < t*, the CO 2 concentration profile is about the same as that obtained under ordinary conditions (i.e. the bypassing gas makes the CO 2 concentration rise abruptly ).Then, at t = t*, the CO 2 concentration suddenly drops before rising again, but now following the typical trend of the sound-assisted tests.This behaviour confirms the ability of the sound to better exploit the adsorption capacity of the activated carbon.Indeed, as soon as the sound was switched on, that specific surface, precluded by the fluid under ordinary conditions, suddenly becomes available, causing the CO 2 concentration to drop because of the renewed activated carbon adsorption capacity.

Effect of SPL and frequency
The effect of SPL and frequency on the CO 2 adsorption efficiency was evaluated by carrying out tests at fixed frequency (80 Hz) and different sound intensity (from 120 up to 140 dB) and at fixed SPL (140 dB) and varying the sound frequency (from 20 to 300 Hz), respectively.The comparison among all the tests performed in terms of moles of CO 2 adsorbed, t b and W are reported in Fig. 17a  and b.The data obtained under ordinary conditions were also reported for comparison.
The analysis of these results is quite clear: the SPL effect on the CO 2 adsorption process reflects what was observed in the fluidization tests.Indeed, the adsorption process undergoes a significant enhancement only when SPLs that are higher or equal to 125 dB are applied, which is perfectly consistent with the activated carbon fluid-dynamic behaviour obtained.
Indeed, 125 dB is a sort of threshold intensity beyond which any further increase of SPL is ineffective, and sure enough all the tests performed at a higher SPL are very similar in terms of breakthrough curve shape, moles of CO 2 adsorbed, t b and W. Whereas the behaviour observed at 120 dB is intermediate.
As well as for the SPL, the results obtained on the effect of sound frequency are also in perfect agreement with those obtained from the fluidization tests.Indeed, the best results in terms of CO 2 adsorption efficiency can be achieved when sound frequencies in the same optimum range (50-120 Hz) are applied.Indeed, the tests performed at intermediate frequencies (50, 80 and 120 Hz) are characterized by very similar behaviours (breakthrough curves, n ads , t b and W).Whereas the adsorption tests carried out at 20 and 300 Hz are significantly worse.
These results are additional proof of the tight link existing between the adsorption efficiency and the fluiddynamics of the system.

Effect of CO 2 partial pressure
The effect of the CO 2 partial pressure on the adsorption process was highlighted by performing tests at three different CO 2 inlet concentrations (5, 10 and 15 %vol. in N 2 ) for each investigated superficial gas velocity (1, 1.5 and 2 cm/s) and both under ordinary and sound-assisted conditions (125/140 dB-80 Hz).The overall results are reported in Table 1.As expected, the CO 2 capture capacity of the adsorbent (n ads ), at fixed superficial gas velocity and fluidization conditions (with or without the sound), is increased with CO 2 partial pressure.This trend is absolutely consistent from a thermodynamic point of view, since the CO 2 partial pressure represents the driving force of the adsorption process.
As clearly reported in Table 1, an increase of the CO 2 inlet concentration also results in a decrease of the breakthrough time, in spite of the increased adsorption capacity.This behaviour is probably due to the adsorption process becoming faster at a higher CO 2 inlet concentration; indeed, the higher the CO 2 inlet concentration, the more CO 2 molecules enter the bed per unit time, the quicker the bed saturation, all the other operating conditions being the same.The obtained experimental values of the moles of CO 2 adsorbed when the bed is completely saturated, namely the sorbent capacity at equilibrium conditions, were elaborated and fitted by the Langmuir equation (Raganati et al., 2014b).Fig. 18 reports the adsorption iso-therms calculated under ordinary and sound-assisted tests.
Analysis of the curves highlights the beneficial effect played by the application of the acoustic field on adsorption performance.Under sound-assisted fluidization, the adsorption isotherms move to more favourable adsorption conditions.

Effect of fluidization velocity
The results of the same tests were elaborated in order to point out the effect of the fluidization velocity on the adsorption process.In particular, the dependence of the breakthrough time on the contact time, defined as the ratio between the mass of adsorbent and the CO 2 volumetric flow, was highlighted.The curves obtained for different CO 2 inlet concentration are shown in Fig. 19a, b and c.As a matter of fact, the fluidization velocity is expected to affect the breakthrough time because the mere increase of the fluidization velocity results in a decrease of the contact time.However, the dependence of the breakthrough time on the contact time, i.e. the fluidization velocity, is as one could expect linear only for the tests performed under ordinary conditions.Whereas the breakthrough time is found to exponentially increase with the contact time, namely decreasing the fluidization velocity from 2 to 1 cm/s for the sound-assisted tests.This evidence is likely due to the role played by fluidization velocity in soundassisted tests.Indeed, under ordinary conditions, the system is quite insensible to changes of the fluidization  velocity, the fluidization quality always being very poor.Therefore, the observed linear increase of the breakthrough time with the decrease of the fluidization velocity is only due to the CO 2 taking more time to flow through the bed.On the other hand, in sound-assisted fluidization tests, changes of the fluidization velocity greatly affect the fluid dynamics of the system.In particular, the decrease of the fluidization velocity results in a more homogeneous fluidization regime, which is characterized by a lower bypass of gas through the bed compared with the tests performed at higher fluidization velocity.This is confirmed by the bed expansion curves reported in Fig. 13, which show a quite sharp change of slope at superficial gas velocities higher than 1 cm/s, thus confirming the occurrence of bubbles (i.e.bypass of gas).As a consequence, the breakthrough time is more than tripled, passing from 2 to 1 cm/s.Moreover, the fluidization velocity has a slight effect on the adsorption capacity of the activated carbon and the fraction of bed utilized until breakpoint, as clearly shown in Table 1.

Conclusions
In this work, a review of the main results obtained by our research group in the field of sound-assisted fluidization was performed.The fluidization behaviour of four different nanosized powders, Al 2 O 3 , Fe 2 O 3 , CuO and ZrO 2 , and of binary mixtures of Al 2 O 3 and Fe 2 O 3 affected by the application of acoustic fields of different intensities (125-150 dB) and frequencies (50-300 Hz) was studied.The results obtained show that the application of sound is necessary for all the powders to obtain a smooth and regular fluidization regime.This observed beneficial effect resulting from the application of sound is due to the fact that it promotes a continuous break-up and reaggregation mechanism of the fluidizing aggregates (produced by the cohesive forces) into smaller structures, which are easier to fluidize.In particular, the parameters of the applied acoustic field strongly affect the fluidization quality: increasing SPLs generally enhance the fluidization quality, whereas the sound frequency has a not monotone effect, it always being possible to find an optimum range giving the best fluidization quality.As regards the binary mixtures, the addition of Fe 2 O 3 , i.e. the powder characterized by the best fluidization quality, to the mixtures generally results in a better fluidization behaviour.
These considerations about the mechanism upon which the sound-assisted fluidization is based were verified by carrying out mixing tests of two different nanopowders.The results obtained show that the fluidizing aggregates actually undergo a dynamic evolution (break-up and re-aggregation) during the fluidization process.In particular, the mixing occurs in a different time period depending on whether the phenomenon is observed from a macroscopic or microscopic point of view.Visual observation shows that the bed was already homogeneously mixed after a few minutes.In contrast, the local mixing, namely the mixing occurring among sub-aggregates due to a continuous break-up and re-forming of single aggregates, leading to the formation of mixed ones, needs dynamics developing over longer time periods.The maximum value of the mixing degree and the time needed by the process to reach it are strongly affected by the relative amount of the two powders.Increasing the amount of Fe 2 O 3 from 17 to 77 wt%, the asymptotic value of the mixing degree is enhanced from about 50 % to 100 % and the characteristic time is decreased from about 87 to 10 min, thus confirming the tight link between the mixing effectiveness and the intrinsic fluidizability of the two powders.
On the basis of this experimented capability of the acoustic fields to enhance the fluidization quality and consequently the gas-solid contact, CO 2 adsorption on a fine activated carbon was performed in a sound-assisted fluidized bed, thus meeting the increasing environmental concerns and interest in the CCS technologies.CO 2 adsorption on a fine activated carbon was performed in a sound-assisted fluidized bed.The experimental results show that the acoustic field positively affects the adsorption efficiency of the powder in terms of a significantly higher breakthrough time, adsorption capacity, fraction of bed utilized until breakthrough and adsorption rate.This beneficial effect shown by the sound is due to enhancement of the fluidization quality (which brings better gas- solid contact and mass transfer coefficients) with respect to the tests performed under ordinary conditions.In particular, the above-mentioned enhancement of the break-up and re-aggregation mechanism of the fluidizing aggregates constantly renews the surface of activated carbon exposed to the fluid, thus making it more readily available to the adsorption process.

Fig. 4
Fig. 4 Minimum fluidization velocity (u mf ) as a function of (a) SPL at fixed frequency (120 Hz) and as a function of (b) frequency at fixed SPL (140 dB).

Fig. 9
Fig. 9 Cumulative distribution of aggregate composition at different times for Al 2 O 3 and Fe 2 O 3 mixing tests.(a) A mixing test; (b) B mixing test; and (c) C mixing test.

Fig. 11
Fig. 11 (a) Cumulative distribution of aggregates at different times for the ZrO 2 and CuO mixing test (130 dB-120 Hz); (b) time-dependence of aggregates' mixing degree for the ZrO 2 and CuO mixing test and for the Al 2 O 3 and Fe 2 O 3 B mixing test (130 dB-120 Hz).

Fig. 12
Fig. 12 (a) Dimensionless pressure drops and (b) bed expansion curves obtained in ordinary tests.

Fig. 15
Fig. 15 reports the typical breakthrough curves (i.e.C/ C 0 vs time, C and C 0 being the CO 2 concentration in the effluent and feed stream, respectively) obtained under ordinary and sound-assisted conditions (140 dB-80 Hz).In order to highlight the most significant portion of the curve, namely the section before and soon after t b , the graph has been reported in logarithmic scale.In particu-

Fig. 13
Fig. 13 (a) Effect of SPL on pressure drops and bed expansion curves at fixed frequency (80 Hz); (b) effect of frequency on pressure drops (c) and bed expansion curves at fixed SPL (140 dB).
2 O 3 and Fe 2 O 3 powder were been reported (0 %wt and 100 %wt of the Fe 2 O 3 amount, respectively) for comparison.The analysis of these curves suggests that the values of u mf of binary mixtures are intermediate between those of single nanopowders, and increasing amounts of Fe 2 O 3 result in lower values of u mf .This evidence is in agreement with the better fluidizability shown by Fe 2 O 3 and it is most likely related to a reduction of the aggregate size.In particular, increasing amounts of Fe 2 O 3 do not result in a linear decrease of u mf : this decrease is much more relevant for low amounts of Fe 2 O 3 ; for Fe 2 O 3 amounts higher than 1/3 %wt, an evident change of the slope of u mf vs Fe 2 O 3 curve was observed.It is most likely that mixtures with Fe 2 O 3 amounts of 1/3 %wt are already characterized by good fluidization behaviour, so that further increasing the Fe 2 O 3 amount can only result in smaller changes of u mf .Moreover, it can be noticed that the influence of SPL obtained for the powders alone is confirmed -in fact, and also in the case of binary mixtures, the u mf values are higher at 130 dB than at 135 dB.

Table 1
Results of CO 2 adsorption tests.