- Email: [email protected]
water ratio and reagents’ concentration on size distribution of manganese carbonate nanoparticles synthesized by microemulsion mediated route
Accepted Manuscript Title: Effect Of Surfactant/Water Ratio And Reagents’ Concentration On Size Distribution Of Manganese Carbonate Nanoparticles Synthesized By Microemulsion Mediated Route Author: Giuseppe Granata Francesca Pagnanelli Daisuke Nishio-Hamane Takehiko Sasaki PII: DOI: Reference:
S0169-4332(15)00126-9 http://dx.doi.org/doi:10.1016/j.apsusc.2015.01.101 APSUSC 29538
To appear in:
APSUSC
Received date: Revised date: Accepted date:
6-10-2014 11-1-2015 16-1-2015
Please cite this article as: G. Granata, F. Pagnanelli, D. Nishio-Hamane, T. Sasaki, Effect Of Surfactant/Water Ratio And Reagents’ Concentration On Size Distribution Of Manganese Carbonate Nanoparticles Synthesized By Microemulsion Mediated Route, Applied Surface Science (2015), http://dx.doi.org/10.1016/j.apsusc.2015.01.101 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
EFFECT OF SURFACTANT/WATER RATIO AND REAGENTS’ CONCENTRATION ON SIZE DISTRIBUTION OF MANGANESE CARBONATE NANOPARTICLES SYNTHESIZED BY MICROEMULSION MEDIATED ROUTE
ip t
Giuseppe Granataa,c, Francesca Pagnanelli*a, Daisuke Nishio-Hamaneb, Takehiko Sasakic
Department of Chemistry, Sapienza University of Rome, P.le Aldo Moro 5, 00185 Rome, Italy
b
Institute of Solid State Physics, The University of Tokyo, 5-1-5 Kashiwanoha, Kashiwa, Chiba,
us
cr
a
277-8581, Japan
Department of Complexity Science and Engineering, Graduate School of Frontier Sciences, The
an
c
M
University of Tokyo, 5-1-5 Kashiwanoha, Kashiwa, Chiba, 277-8561, Japan
Corresponding author email: [email protected]
d
Abstract
te
In this work nanoparticles of manganese carbonate were produced by microemulsion-mediated
Ac ce p
route at room temperature, without any post thermal treatment. All produced samples were characterized by XRD and by TEM and obtained images were analyzed in order to evaluate particle size distribution, mean size and polydispersity (variance). The influence of water-surfactant molar ratio and concentration of reagents were investigated in the range 5-7.5 and 0.25-1.0 M, respectively, according to factorial design. Significant effects on particle mean size and polydispersity were assessed by statistical analysis. Results showed that by increasing the watersurfactant molar ratio from 5 to 7.5, the average particle size increased from less than 10 nm to around 100 nm, and the standard deviation increased from less than 5 nm to 35 nm. Statistical analysis put in evidence that water-surfactant molar ratio has significant positive effect on both mean and variance of particle size.
Page 1 of 31
Concentration of reactants, in the investigated range, did not influence mean size of particles, while significant changes of variance were observed: passing from 0.25 to 1 M concentration, variances of particle size increased for w=5 and for w=6.25, while decreased for w=7.5.
ip t
Keywords: nanoparticles; manganese carbonate; microemulsion; particle size distribution; image
Introduction
us
1.
cr
analysis.
Manganese oxides have been extensively studied in the last years because of the large
an
number of manganese oxidation states, which allows to have different manganese oxides with wide range of applications, from catalysis [1, 2] to batteries [3]. In many cases manganese oxides
M
have been prepared by calcinating a precursor such as manganese carbonate [4-6] Manganese carbonate can be produced by different methods but the solvo-hydrothermal routes [7-10], and among them the microemulsion-mediated synthesis, seem to be the most
d
interesting due to their easiness and scalability.
te
Referring to the definition of Danielsson and Lindman, a “microemulsion” is a system of water,
Ac ce p
oil and amphiphile which is a single optically isotropic and thermodynamically stable liquid solution [11]. Water-in-oil microemulsions with spherical shape are the most used for the synthesis of nanoparticles and they are particularly interesting because the reverses micelles can be used as “nano-reactors”, where the precipitation of nanoparticles takes place [12]. Formation of nanoparticles in the micelles occurs according to a mechanism involving coalescence of droplets, exchange of reagents and nucleation of particles. Further mechanisms have been also assumed to explain experimentally observed particle dimensions larger than those predictable by the previous scheme. In particular the autocatalysis occurs when larger particles grow faster due to increased surface area [13] and the Ostwald Ripening takes place when the dissolution of small crystals or sol particles and the re-deposition of the dissolved species on the surface of larger crystals lead to a further growth of the large particles at the expense of smaller ones (as a result larger particles grow as smaller ones disappear) [14].
Page 2 of 31
Synthesis of nanoparticles is strictly related to microemulsion characteristics such as interfacial fluidity of surfactant layer (controlling intermicellar exchange) and charge density of surfactant polar heads (stabilizing newly formed nanoparticles and limiting their aggregation). In literature there are different exhaustive works, which well explain the base theory of reverse micelles and the
ip t
influence of operating parameters on micelle size, shape and flexibility [15-18]. Operating factors which mainly affect the characteristics of microemulsions are: type of surfactant and solvent,
cr
water/surfactant molar ratio (w) and concentration of reagents. In particular, by fixing surfactant
us
and solvent type and by increasing w, an increase of micelles size and then of nanoparticles size (both as primary particles and their aggregates emerging from coagulation phenomena) has been
an
observed [15].
Tojo et al. [14] experimentally validated a Montecarlo simulation (including particle-particle
M
coagulation) showing that diameter of particles increased by increasing the water-surfactant molar ratio because the size of microemulsion droplets acts as a physical constrain to the maximum particle size.
d
For a given microemulsion system, an increase of reagents concentration can determine an
te
increase of particle size ([15] and references within), even though some works reported a
Ac ce p
contrasting decrease of particles size by increasing reagents concentration [19-20]. Kumar et al. (2004) [21] developed a population balance model considering autocatalysis of large particles and they found out that the final size of nanoparticles has a no monotonous trend with respect to the concentration of reagents. In particular, at low concentrations, few numbers of nuclei are formed and an increase of reagents concentration results into growth thereby increasing the particle size. When the reagents concentration reaches the critical nucleation value, particle size will be the maximum and a further increase of concentration results into an increase of nucleation events in the micelles, leading to a decrease of particle size. This generic scheme can be a valuable explanation for the contradictory results about the effect of reagents concentration. Nevertheless, this scheme is valid for systems where agglomeration is negligible, i.e. for high film rigidity.
Page 3 of 31
Tojo et al. (1997) developed a Montecarlo simulation including agglomeration of particles and they showed the effect of the concentration of reagents on particle size distribution for different film rigidities [14]. Simulations results showed that higher concentrations give rise to a population of larger nanoparticles. Moreover, by looking at particle size distribution, for high concentrations of
ip t
reactants, systems with high film rigidity presented a bimodal distribution due to a limited intermicellar exchange of particles. Bimodal distribution was not present for low concentrated
cr
microemulsions, where nucleation and growth occurred in the same time scale. Bimodal
us
distribution disappeared also for high concentrations of reagents, where the film rigidity diminished. As for the effect of reagents concentration on particle polydispersity, some works reported a
an
decrease of polydispersity of particle size by increasing the concentration of reagents [22-24], whilst other works simulated an increase of polydispersity at high concentration values [14].
M
Pagnanelli et al. [25] demonstrated for the synthesis of MnCO3 nanoparticles in microemulsion that an increase of reagents concentration can result into different effects. For rigid systems with reduced intermicellar exchange (obtained by using solvents as hexane and isooctane, which can
d
interact tidily with surfactant tails making micelles more rigid), an increase of concentration acts as
te
a further obstacle to the exchange of reagents, thus determining a decrease of mean size and
Ac ce p
unchanged polydispersity. On the other hand, for fluid systems (obtained by increasing cosurfactant dosage) an increase of concentration did not alter mean size, but increased polydispersity.
Therefore, in spite of the advantages given by the microemulsion mediated routes, experimental results reported in literature seem to be strongly affected by the parameters influencing the characteristics of microemulsion (i.e. water-surfactant molar ratio, concentration of reactants etc.). A general framework for the understanding of the effect of single factors and their interaction should be then developed in order to overcome apparent contradictions in the experimental findings. Specifically concerning the synthesis of MnCO3 nanoparticles by microemulsion, few works can be found in the literature. In particular, Aragòn et al. (2007) produced submicron particles of MnCO3 with an average size of 200 nm by using inverse micelles made up of water, cyclohexane
Page 4 of 31
(proportion 1:4) and 0.2 M sodium bis(ethylhexyl) sulfosuccinate (Aerosol OT, AOT) [26]. Wu et al. (2006) used a CTAB microemulsion method followed by thermal treatment to synthesize MnCO3 nanocrystals performing also a shape control by modification of operating variables such as concentration of reagents and water-surfactant molar ratio [27]. Pagnanelli et al. [25] investigated
ip t
the synthesis of MnCO3 nanoparticles by CTAB microemulsion addressing the effects of reaction time (30-60 min), type of solvent (cyclohexane, isooctane and hexane), cosurfactant dosage (1.5-
cr
4.5 g) and reagents concentration (1.0-0.25 mol/L) by characterization with Scanning Electron
us
Microscopy. Previous results showed that size can be significantly increased by increasing reaction time and cosurfactant dosage. Nature of solvent also significantly affected particles size, whilst, as
an
above mentioned, the concentration of reagents had opposite effects depending on the system nature.
M
In this work we continued investigating the synthesis of MnCO3 nanoparticles by addressing different aspects such as the influence of post thermal treatment, water-surfactant molar ratio and reagents concentration in other different conditions, by using X-Ray, SEM and Transmission
Material and methods
te
2.
d
Electron Microscopy for the characterization of obtained nanoparticles.
Ac ce p
All chemicals used for this work (MnCl2, (NH4)2CO3, cyclohexane, n-pentanol, ethanol) were analytical grade reagents purchased from Kanto Chemicals Co. CTAB (99+%) was purchased from ACROS ORGANICS. A quaternary water in oil microemulsion of cyclohexane (20 g) + npenthanol (1.5 g) - CTAB (2g) – Water (MnCl2 and (NH4)2CO3 solutions) was employed as reaction medium. Pentanol was added as co-surfactant in order to increase the micelle permeability [28]. As shown in figure 1, MnCl2 and (NH4)2CO3 solutions were added dropwise in their respective organic phases until they became optically transparent. Then, after the mixing of the two initial microemulsions, reaction (1) occurred in MnCl2 + (NH4)2CO3 → MnCO3 + 2NH4Cl
(1)
Reaction was carried out at 30 °C, under stirring, for 2 hours. As showed in table 1 water-surfactant molar ratio (w) was investigated in the range 5-7.5 by keeping constant the amount of surfactant (2 g) and by modifying the amounts of aqueous
Page 5 of 31
solutions in the microemulsions: 0.49, 0.62 and 0.74 g of aqueous solutions were added in their respective systems in order to reach respectively w=5, w=6.25 and w=7.5. Concentration of reactants (MnCl2 and (NH4)2CO3) was investigated in the range 0.25-1.0 M. After precipitation, the insoluble product MnCO3 (KsMnCO32.24 x 10-11) was collected by
ip t
centrifugation and it washed twice by ethanol and then twice by distilled water. When the thermal treatment was tested, it was carried out in a teflon-lined autoclave reactor at 120°C for 4 hours
cr
after mixing. Size and morphology of the as synthesized particles were estimated by Scanning
us
Electron Microscope (JEOL JSM-5510LV) and by Transmission Electron Microscope (JEOL JEM2100 in ISSP, Univ. Tokyo) operating at 200 kV. Image J software was used for image analysis in
an
order to determine the size distribution of produced particles. For each sample, particle size distribution was evaluated by analyzing four TEM images resulting in a number of analyzed
M
particles included between 550 and 1150. Samples for microscopic observations were prepared by adding acetone and by dispersing the particles in an ultrasonic bath. After ultrasonication, a drop of the suspension was put on a copper grid and dried at air before microscope observation. Samples
d
were also analyzed by MacScience X-ray diffractometer M03X (XRD), with Cu Kα radiation (λ =
Ac ce p
of products.
te
1.5406 Ǻ) operating at 40 kV and 20 mA emission condition in order to determine the composition
Analysis of Particle Size
Particles were manually measured one by one by using first a graphic software (Adobe Illustrator) to draw the particles contours and to remove the original image from the background and then by using a software to analyze the resulting black/white image (ImageJ) by measuring the size of each particles whose contours had been preliminary drawn. In order to make sure the image analysis was representative, the first step in the image analysis was the determination of the minimum number of nanoparticles that should be counted in order to have representative sample statistics (i.e. mean size and standard deviation). This was done empirically by estimating mean and standard deviation for samples of nanoparticle with increasing numerosity and determining the minimum number of nanoparticles for which no significant variation was observed for these statistics.
Page 6 of 31
3.
Results and Discussion
3.1
Morphology and particle size characterization by image analysis Sample produced in condition 0 of table 1 (w=20 and C=0.5 M) denoted the formation of cubic
ip t
nanoparticles with mean size of 300 nm (Figure 2a). After thermal treatment (at 120°C for 4 h) shape remained unaltered (Figure 2b), while a clear increase of average particle size occurred
cr
(average size of 500 nm). The understanding of the mechanism which determined the increase of
us
particle size goes beyond the purpose of this paper but it might be due to coalescence and/or Ostwald Ripening since the final particle size was about twice than without thermal treatment
an
(suggesting two different phases might have come together) and because it was clearly temperature dependent, as showed by the pioneering works done by Lifshitz and Slyozov [29].
M
Aiming to produce nanoparticles smaller than 100 nm, further tests were performed without any thermal treatments as done elsewhere [27]. In addition, reduced w values were chosen due to the known effect of this parameter on particlesaverage size.
d
Experimental conditions for further tests were arranged according to a factorial design with two
te
factors (water-surfactant molar ratio and concentration of reagents) as listed in Table 1.
Ac ce p
As shown by TEM images and particle size distribution elaborated from them (figures 3-8), water-surfactant molar ratio (w) exhibited a relevant influence on particle size and polydispersity (evaluated as the standard deviation around the mean size) and also on particles shape. Regarding the histograms in figure 3-8, it should be noted that whereas particles were grouped as >x nm, it was because they were too few to be included to their real size group and only by grouping them together the histogram bar could be observed. Of course there are larger particles but in spite of being rather visible just because larger, when counted among the other particles they were just few (way less than 10) out 500-1150 counted particles. That is why just an observation without analysis can lead to mistakes. For w = 5 (CMn=0.25 M) particles had mostly a globular shape (fig. 3) with mean size 6.5 nm and with narrow dispersion (standard deviation 1.4 nm). By increasing water/surfactant molar ratio up to 6.25 (CMn=0.25 M)(fig. 4), particles become mainly rhombohedral (and also cubic) and their
Page 7 of 31
size and dispersion increased (mean size 71 nm and standard deviation 24 nm). When working at w = 7.5 (CMn=0.25 M) cubic nanoparticles were obtained (fig. 5) with increased particles size (mean size 89 nm) and dispersion (37 nm) with respect to lower levels of w. TEM images clearly showed the isotropic growth of particles according to spherical and cubic
ip t
shapes, even though literature works also reported anisotropic growth in presence of cationic surfactants, due to assembling phenomena on negatively charged surface of forming particles [18,
cr
30, 31].
us
As for the concentration of reagents, for the different levels of w, very similar results were obtained (Figures 6-8) in terms of shape, mean size and dispersion. Significance of observed
an
differences among mean size and dispersion for some conditions has been statistically evaluated in the following analysis (paragraph 3.3). X-ray characterization
M
3.2
Characterization by X-Ray diffraction was reported in Figures 9 for different w levels at CMn=1M. All reflection peaks can be indexed as a rhomb-centered hexagonal (rch) phase (space
d
group R3c(167)) of MnCO3 with lattice constants α=4.790 Ǻ and c=15.69 Ǻ (JCPDS card No 44-
te
1472). However, the XRD pattern of the sample obtained at w=5, the lowest water-surfactant molar
Ac ce p
ratio (Figure 9c), it also showed (peaks marked with *) the presence of a cubic structure of Mn2O3 (JPCDS No. 01-071-0636). Since all reactions were carried out by working under nitrogen atmosphere, it can be supposed that the reduced size of nanoparticles obtained in such condition and thus its high surface area, favored the oxidation of surface portions. In addition it should be mentioned that, for the lowest surfactant molar ratio, a modest amount of product was obtained giving a low signal to noise ratio of the XRD spectra (figure 9c). Average crystallite size (τ) was determined from full width at half maxima (FWHM) of peak 113 and by using the Scherer equation [32]: (2) where k is a dimensionless shape factor (0.9), λ is the X-ray wavelength (0.154 nm), β is the line broadening at half the maximum intensity, and θ is the Bragg angle. Samples obtained for w= 5, w=6.25 and w=7.5 (CMn=1 M) showed an average crystallite size respectively of 20, 35 and 45 nm. Page 8 of 31
These values are in agreement with mean size values obtained by TEM images that for the same conditions are 10, 66 and 89 nm, respectively. In particular, for the sample obtained at w=5, by considering the standard deviation due to sample polydispersity (2.6 nm), a 99% confidence interval of 10±7.8 is obtained, meaning that we have the 99% of probability that the true value of
ip t
the mean size of the population is included in this range. This consideration puts in evidence that similar predictions were obtained from image analysis and XRD, but also show the importance of
cr
evaluating sample polydispersity for nanoparticles population. Image analysis by TEM can be used
us
for such aim conversely to XRD spectra which, according to the elaboration proposed here, can only estimate the crystallite size. Resuming, the comparison between mean size by TEM and
an
crystallite size by XRD denoted that the first method give similar (w=5) or larger (w=6.25 and 7.5) values than the second. Slight underestimates for w=6.25 and w=7.5 could be due to
M
polycrystallinity of nanoparticles in which different crystal cells aggregate giving larger visible
3.3 Statistical analysis
d
particles in TEM images.
te
From the preliminary analysis of the method, it emerged that representative values of both
Ac ce p
mean size and standard deviation required at least the counting of 500 nanoparticles. Then counting more than this number ensures the choice of figures does not affect the final estimates. After this preliminary empirical estimate of minimum number of nanoparticles in the counting, the error variability of data (SSE) was supposed to be due to both the execution of image analysis (SSIMAG) and random experimental errors (SSPE). SSE=SSIMAG+SSPE
Different replicates (three) of the same experiments were then performed (to estimate variability due to random error) and the same images were analyzed independently by the same operator (to estimate variability due to image analysis execution). Statistical analysis of variance denoted that variability due to execution of image analysis is not statistically significant with respect to variability due to random errors.
Page 9 of 31
Experimental tests were performed according to a full factorial design with two factors, A=w and B=CMn (the molar concentration of the manganese source), the first at three levels (5, 6.25 and 7.5) and the second at two levels (0.25 and 1.0 M). Different statistics were evaluated for particles populations obtained for each treatment (i.e. different operating conditions used in factorial design).
ip t
In particular mean (μ) and standard deviation (S) of particles size (xi) were evaluated.
These statistics were reported for each treatment in Figure 10, from where it can be noted that
cr
w is the most influencing factor for all statistics, meaning that, even for a small increase of w in the
us
investigated range, an increase of both mean (Fig. 10a) and standard deviation (Fig. 10b) was observed. Thus, by increasing water-surfactant molar ratio, both size and heterogeneity of particle
an
population tend to increase. On the other hand, by increasing the concentration of reagents, a less relevant effect on particle mean size and standard deviation was observed.
M
These qualitative observations about mean values were confirmed by ANOVA [33]. Estimate of variance component due to error (MSE) was obtained by independent replicated tests and it takes into account the variability associated to random experimental errors, the
d
variability due to image choice for statistic determination, and the variability due to image analysis
te
uncertainties. In this context eight TEM images from two replicates (four images for each replicate)
Ac ce p
of the same treatments were used to estimate variability due to random errors. ANOVA put in evidence that only A factor (w ratio) has a significant effect on mean size, whilst B factor (concentration of reagents) and AB interaction are not significant (99% confidence level). Estimates of the effects were evaluated considering the higher and lower levels for each factor (w = 5 and 7.5 and CMn= 0.25 and 1 M) and all effects were reported in figure 11 along with the confidence bands ±tα/2,n-1√(MSE/n), where α=0.01 and n=8 (dashed lines). Estimates of the effects which are larger than the confidence bands are those statistically significant (99% confidence level). Standard deviations in the different operating conditions were compared by considering the corresponding variances (the square of standard deviations) and performing statistical tests. In particular, F tests were carried out by comparing variances at fixed values of reagent concentration
Page 10 of 31
for various w and at fixed w for the two levels of reagent concentration. These statistical tests (99% confidence level) denoted that: -
For both concentration levels, dispersion of particle size for w=5 is significantly lower than at w=6.25 and w=7.5, while variances are the same for these high levels of w. Then, an
ip t
increase of w from 5 to 6.25 determined an increase of polydispersity which remained unaltered passing from w=6.25 to w=7.5.
As for the effect of the concentration of reagents, different results were obtained depending
cr
-
us
on w level. In particular, for w=5 and w=6.25 a significant increase of variance was observed by increasing the concentration of reagents, whilst for w=7.5 by increasing the
an
concentration a statistically significant decrease of variance was observed. As for the effect of water/surfactant molar ratio on nanoparticle mean size, it can be explained
M
by considering that this factor determines an increase of micelle size, an then an increase of nanoparticle size growing inside them due to the templating properties and physical constrains offered by micelles during nanoparticle growth.
d
As for the effect of w on polydispersity (variances of particle size), it was not specifically
te
addressed in previous literature: it could be argued that as micelles are larger they can contain a
Ac ce p
wide range of nanoparticles below their size being a physical constrain for nanoparticles during their growth.
As for reagent concentration, the experimental results reported in this work comprehended a test condition already tested (w=6.25) [25] plus other two levels (w=5 and w=7.25). The intermediate level already studied presented the same behavior found in the past: the increase of reagent concentration in the range 0.25-1M did not alter the average size of particles but a significant increase of variance was observed. The reproducibility of this result obtained in different experimental campaigns with different image characterization (TEM instead of SEM) gave strength to such finding. For the other two levels of w, the effect of reagent concentration on particle size remained the same, i.e. by increasing the concentration in the investigated range no significant effect on mean size was observed. Nevertheless different effects were found for variances: positive effect for w=5 and negative effect for w=7.25.
Page 11 of 31
As for the lack of the effect of reagent concentration on particle mean size, by considering the non-monotonous trend predicted by simulations [21], it could be reasoned that, for this system, the investigated concentrations are near the critical concentration (maximum) of the curve “nanoparticle size” versus “reagent concentration” being the concentration of reagents scarcely
ip t
influencing nanoparticle mean size.
As for the opposite effect of reagent concentration on nanoparticle polydispersity, it is in
cr
agreement with previously reported experimental results showing both the increase and the
us
decrease of polydispersity by increasing the concentration of reagents [14, 22, 23, 24]. As in the case of mean size, it is evident the concentration of reagents can play a different role in
an
nanoparticle growth in microemulsions, thus requiring further experimentation in order to complete the mapping of experimental behavior for different operating conditions.
M
It should be noted that average size of nanoparticles ranged from few nanometers up to 100 nm. In particular, for w=7.5, only an average of 2-3% of particle population were larger than 150 nm, whilst for w=5 and w=6.25, even more than respectively 99% and about 80% of particles were
d
smaller than 100 nm. Thus, since the size of individual microemulsion droplets for CTAB systems
te
can range from 10 to 200 nm depending on system specific composition and operating conditions
Ac ce p
[34], only a very small percentage of particle population could be considered too large to be contained in the micelles. Therefore, it is likely that coalescence of primary nanoparticles occurred mostly inside the micelles, whilst a further stacking in rhombohedral structures might occurred in a second phase out of the micelle cores, resulting into the larger particles.
4
Conclusions
Nanoparticles of rhomb-centered hexagonal phase of manganese carbonate were obtained by a microemulsion-mediated route at room temperature without any post-thermal treatment. Operating conditions allowing to obtain the smaller particles with globular shape were w = 5 and CMn = 0.25 M. Experimental data and statistical analysis revealed that, in the investigated range, water-surfactant molar ratio had a significant positive effects on both particle size and particle dispersion, whilst the
Page 12 of 31
concentration of reagents did not show any significant effect on average size and opposite effects on variances (depending on w value), thus requiring a further experimentation in order to complete the mapping of experimental behavior for different compositions of the system.
ip t
Acknowledgement
This study was supported in part by KAKENHI 22110503 for Scientific-Research on Innovative
us
for the Promotion of Science and Technology (MEXT, Japan).
cr
Areas (Frontier Science of interactions between plasmas and nano-interfaces) and Strategic Funds
The use of the facilities of the Materials Design and Characterization Laboratory at the Institute for
an
Solid State Physics, University of Tokyo, is gratefully acknowledged.
[1]
M
References
M. Baldi, E. Finocchio, F. Milella, G. Busca, Evaluation of the mechanism of the oxy-
dehydrogenation of propane over manganese oxide, Appl. Catal. B 16 (1998) 43-51. E.R. Stobbe,
B.A. De Boer,
J.W. Geus, The reduction and oxidation behaviour of
d
[2]
J.K. Chang, W.C. Tsai, Material characterization and electrochemical performance of
Ac ce p
[3]
te
manganese oxides. Catal. Today 47 (1999) 161-167.
hydrous manganese oxide electrodes for use in electrochemical pseudocapacitors, J. Electrochem. Soc. 150, (2003) 1333–1338. [4]
S. Lei, K. Tang, Z. Fang, Q. Liu, Q. Zheng, Peparation of α-Mn2O3 and MnO from thermal
decomposition of MnCO3 and control of morphology, Mater. Lett. 60 (2006) 53-56. [5]
X. Shen, Z. Ji, H. Miao, J. Yang, K. Chen, Hydrothermal synthesis of MnCO3 nanorods and
their thermal transformation into Mn2O3 and Mn3O4nanorods with single crystalline structure, J. Alloy Compd. 509 (2011) 5672-5676. [6] L.X. Yang, Y.J. Zhu, H. Tong, W.W. Wang, Submicrocubes and highly oriented assemblies of MnCO3 synthesized by ultrasound agitation method and their thermal transformation to nanoporous Mn2O3, Ultrason. Sonochem. 14 (2007) 259–265.
Page 13 of 31
[7] H. Hu, J.Y. Xu, H. Yang, J. Liang, S. Yang, H. Wu, Morphology-controlled hydrothermal synthesis of MnCO3 hierarchical superstructures with Schiff base as stabilizer, Mater. Res. Bull. 46 (2011) 1908–1915. [8]
S. Lei, Z. Liang, L. Zhou, K. Tang, Synthesis and morphological control of MnCO3 and
ip t
Mn(OH)2 by a complex homogeneous precipitation method, Mater. Chem. Phys. 113 (2009) 445– 450.
M. Sanchez-Dominguez, K. Pemartin, M. Boutonnet, Preparation of inorganic nanoparticles
cr
[9]
us
in oil-in-water microemulsions: A soft and versatile approach, Curr. Opin. Colloid. In. 17 (2012) 297–305.
an
[10] X. Wang, Y. Li, Hydrothermal reduction route to Mn(OH)2 and MnCO3 nanocrystals, Mater. Chem. Phys. 82 (2003) 419–422.
M
[11] I. Danielsson, B. Lindman, The definition of Microemulsion, Colloids and Surface 3 (1981) 391-392.
[12] A. Vàzquez-Olmos, R. Redòn, A.L. Fernàndez-Osorio, J.M. Saniger, Room-temperature
d
synthesis of Mn3O4 nanorods, Appl. Phys. A 81 (2005) 1131–1134.
te
[13] M. Ethayaraja, K. Dutta, D. Muthukumaran, R. Bandyopadhyaya, Nanoparticle Formation in
Ac ce p
Water-in-Oil Microemulsions: Experiments, Mechanism and Monte Carlo Simulation, Langmuir 23 (2007) 3418-3423.
[14] C. Tojo, M.C. Blanco, F. Rivadulla, M.A. Lopez-Quintela, Kinetics of the Formation of Particles in Microemulsions, Langmuir 13 (1997) 1970-1977. [15] J. Eastoe, M.J. Hollamby, L. Hudson, Recent advances in nanoparticle synthesis withreversed micelles, Adv. Colloid Interfac. 128-130 (2006) 5-15. [16] A.K. Ganguli, A. Ganguly, S. Vaidya, Microemulsion-based synthesis of nanocrystalline materials, Chem. Soc. Rev. 39 (2010) 474-485. [17] M.A. Michaels, S. Sherwood, M. Kidwell, M.J. Allsbrook, S.A. Morrison, S.C. Rutan, E.E. Capenter, Quantitative model for prediction of hydrodynamic size of nonionic reverse micelles, J. Colloid Interf. Sci. 311 (2007) 70-76.
Page 14 of 31
[18] M.P. Pileni, Control of the size and shape of inorganic nanocrystals at various scales from nano to macrodomains, J. Phys. Chem. 111 (2007) 9019-9038. [19] M. Fernandez-Garcıa, X. Wang, C. Belver, A. Iglesias-Juez, J.C. Hanson, J.A. Rodriguez, Ca doping of nanosize Ce-Zr and Ce-Tb solid solutions: Structural and electronic effects, Chem. Mater.
ip t
17, 4181-4193.
[20] Z. Ye, M. Tan, G. Wang, J. Yuan, Preparation, characterization and application of fluorescent
cr
terbium complex-doped zirconia nanoparticles, J. Fluoresc. 15 (2005) 499-505
us
[21] R.A. Kumar, G. Hota, A. Mehra, K.C. Khilar, Modeling of Nanoparticles Formation by Mixing of Two Reactive Microemulsions, AIChE J. 50 (2004) 1556-1567.
micelles, Colloids Surf. A 19 (1996) 123-131.
an
[22] J. Eastoe, S. Stebbing, J. Dalton, R.K. Heenan, Preparation of colloidal cobalt using reversed
M
[23] I. Lisiecki, M.P. Pileni, Synthesis of well-defined and low size distribution cobalt nanocrystals: the limited influence of reverse micelles, Langmuir 19 (2003) 9486-9489. [24] M.P. Pileni, The role of soft colloidal templates in controlling the size and shape of inorganic
d
nanocrystals, Nat. Mater. 2 (2003) 145-150.
te
[25] F. Pagnanelli, G. Granata, E. Moscardini, L. Toro, Synthesis of MnCO3 nanoparticles by
Ac ce p
microemulsions: quantitative evaluation of the effect of operating conditions on particle size distribution, J. Nanopart. Res. 15 (2013) 1887-1898. [26] M.J. Aragón, C. Pérez-Vicente, J.L. Tirado, Submicronic particles of manganese carbonate prepared in reverse micelles: A new electrode material for lithium-ion batteries, Electrochem. Commun. 9 (2007) 1744-1748.
[27] X. Wu, M. Cao, H. Lu, X. He, C. Hu, Microemulsion-mediated solvothermal synthesis and morphological evolution of MnCO3 nanocrystals, J. Nanosci. Nanotechno. 6 (2006) 2123-2128. [28] G. Palazzo, F. Lopez, M. Giustini, G. Colafemmina, A. Ceglie, Role of the cosurfactant in the CTAB/water/n-pentanol/n-hexane
water-in-oil
microemulsion.
1.
Pentanol
effect
on
the
microstructure, J. Phys. Chem. B 107 (2003) 1924-1931. [29] I.M. Lifshitz, V.V. Slyozov, The kinetics of precipitation from supersaturated solid, J. Phys. Chem. Solids 19 (1961), 19-35.
Page 15 of 31
[30] Y. Liu, Z. Liu, G. Wang, Preparation of Mn3O4 nanowires by calcining the precursor powders synthesized in a novel inverse microemulsion, Appl. Phys. A-Mater. 76 (2003) 1117–1120. [31] S. Vaidya, P. Rastogi, S. Agarwal, S.K. Gupta, T. Ahmad, A.M. Antonelli, K.V. Ramanujachary, S.E. Lofland, A.K. Ganguli, Nanospheres, nanocubes, and nanorods of nickel
ip t
oxalate: Control of shape and size by surfactant and solvent, J. Phys. Chem. C 112 (2008)12610– 12615.
cr
[32] B.D. Cullity, Elements of X-ray diffraction. Addison-Wesley, Reading, 1978.
[34] G. Garg, P.A. Hassan, S.K. Kulshreshtha,
us
[33] D.C. Montgomery, Design and analysis of experiments, McGraw Hill, Milano, 2005. Dynamic light scattering studies of rod-like
Ac ce p
te
d
M
an
micelles in dilute and semidilute regime, Colloid Surf A 275 (2006)161–167
Page 16 of 31
CMn [M]
w
0
0.50
20.00
1
0.25
5.00
2
0.25
6.25
3
0.25
7.50
4
1.00
5.00
5
1.00
6.25
6
1.00
7.50
Ac ce p
te
d
M
an
us
cr
test
ip t
Table 1: investigated factors of full factorial experimentation
Page 17 of 31
Flowchart for the synthesis of nano-crystalline manganese carbonate
Figure 2
SEM images of MnCO3 particles obtained for w=20, 0.5 M at 30°C (a) and for w=20, 0.5 M at 90 °C for 4 hours (b)
Figure 3
Bright field TEM image of MnCO3 particles obtained for w=5, 0.25 M at 30 °C (a) and their size distribution (b).
Figure 4
Bright field TEM image of MnCO3 particles obtained for w=6.25, 0.25 M at 30 °C (a) and their size distribution (b).
Figure 5
Bright field TEM image of MnCO3 particles obtained for w=7. 5, 0.25 M at 30 °C (a) and their size distribution (b).
Figure 6
Bright field TEM image of MnCO3 particles obtained for w=5, 1 M at 30 °C (a) and their size distribution (b).
Figure 7
Bright field TEM image of MnCO3 particles obtained for w=6.25, 1 M at 30 °C (a) and their size distribution (b).
Figure 8
Bright field TEM image of MnCO3 particles obtained for w=7. 5, 1 M at 30 °C (a) and their size distribution (b).
Figure 9
X-ray diffraction patterns of nanocrystalline MnCO3 obtained for w=7.5 (a), w= 6.25 (b) and w=5 (c)
Figure 10
Mean size (a) and standard deviation (b) in the different treatments
Figure 11
Estimates of effect with confidence bands on mean size.
Ac ce p
te
d
M
an
us
cr
ip t
Figure 1
Page 18 of 31
Highlights MnCO3 nanoparticles were produced by inverse microemulsion at room temperature. Water-surfactant molar ratio and reactant concentrations were investigated.
Water-surfactant molar ratio has positive effect on both size and polydispersity.
Ac ce p
te
d
M
an
us
cr
Reactant concentration influences only polydispersity.
ip t
Effects on mean size and polydispersity were assessed by statistical analysis.
Page 19 of 31
Graphical Abstract (for review)
Ac
ce pt
ed
M
an
us
cr
ip t
Graphical abstract
Page 20 of 31
Figure
Ac
ce pt
ed
M
an
us
cr
ip t
Fig. 1
Page 21 of 31
Ac
ce pt
ed
M
an
us
cr
ip t
Fig. 2
Page 22 of 31
Ac
ce pt
ed
M
an
us
cr
ip t
Fig. 3
Page 23 of 31
90
d>130 nm
120
110
us
an
M
16
100
12
80
ed
14
70
60
6
50
8
40
10
30
d<30 nm
frequency (%)
ce pt
Ac
cr
ip t
Fig. 4
4
2
0
Page 24 of 31
Ac
ce pt
ed
M
an
us
cr
ip t
Fig. 5
Page 25 of 31
Ac
ce pt
ed
M
an
us
cr
ip t
Fig. 6
Page 26 of 31
Ac
ce pt
ed
M
an
us
cr
ip t
Fig. 7
Page 27 of 31
Ac
ce pt
ed
M
an
us
cr
ip t
Fig. 8
Page 28 of 31
Ac
ce pt
ed
M
an
us
cr
ip t
Fig. 9
Page 29 of 31
Fig 10
120
ip t
80 60
cr
CMn=0.25 M 40
CMn=1 M
us
mean size (nm)
100
20
4.5
5
5.5
6
6.5
7
7.5
40
ed
30
15 10 5
ce pt
25
Ac
standard deviation (nm)
35
20
8
M
w
an
0
CMn=0.25 M CMn=1 M
0
4.5
5
5.5
6
6.5
7
7.5
8
w
Page 30 of 31
Ac
ce pt
ed
M
an
us
cr
ip t
Fig. 11
Page 31 of 31
S0169-4332(15)00126-9 http://dx.doi.org/doi:10.1016/j.apsusc.2015.01.101 APSUSC 29538
To appear in:
APSUSC
Received date: Revised date: Accepted date:
6-10-2014 11-1-2015 16-1-2015
Please cite this article as: G. Granata, F. Pagnanelli, D. Nishio-Hamane, T. Sasaki, Effect Of Surfactant/Water Ratio And Reagents’ Concentration On Size Distribution Of Manganese Carbonate Nanoparticles Synthesized By Microemulsion Mediated Route, Applied Surface Science (2015), http://dx.doi.org/10.1016/j.apsusc.2015.01.101 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
EFFECT OF SURFACTANT/WATER RATIO AND REAGENTS’ CONCENTRATION ON SIZE DISTRIBUTION OF MANGANESE CARBONATE NANOPARTICLES SYNTHESIZED BY MICROEMULSION MEDIATED ROUTE
ip t
Giuseppe Granataa,c, Francesca Pagnanelli*a, Daisuke Nishio-Hamaneb, Takehiko Sasakic
Department of Chemistry, Sapienza University of Rome, P.le Aldo Moro 5, 00185 Rome, Italy
b
Institute of Solid State Physics, The University of Tokyo, 5-1-5 Kashiwanoha, Kashiwa, Chiba,
us
cr
a
277-8581, Japan
Department of Complexity Science and Engineering, Graduate School of Frontier Sciences, The
an
c
M
University of Tokyo, 5-1-5 Kashiwanoha, Kashiwa, Chiba, 277-8561, Japan
Corresponding author email: [email protected]
d
Abstract
te
In this work nanoparticles of manganese carbonate were produced by microemulsion-mediated
Ac ce p
route at room temperature, without any post thermal treatment. All produced samples were characterized by XRD and by TEM and obtained images were analyzed in order to evaluate particle size distribution, mean size and polydispersity (variance). The influence of water-surfactant molar ratio and concentration of reagents were investigated in the range 5-7.5 and 0.25-1.0 M, respectively, according to factorial design. Significant effects on particle mean size and polydispersity were assessed by statistical analysis. Results showed that by increasing the watersurfactant molar ratio from 5 to 7.5, the average particle size increased from less than 10 nm to around 100 nm, and the standard deviation increased from less than 5 nm to 35 nm. Statistical analysis put in evidence that water-surfactant molar ratio has significant positive effect on both mean and variance of particle size.
Page 1 of 31
Concentration of reactants, in the investigated range, did not influence mean size of particles, while significant changes of variance were observed: passing from 0.25 to 1 M concentration, variances of particle size increased for w=5 and for w=6.25, while decreased for w=7.5.
ip t
Keywords: nanoparticles; manganese carbonate; microemulsion; particle size distribution; image
Introduction
us
1.
cr
analysis.
Manganese oxides have been extensively studied in the last years because of the large
an
number of manganese oxidation states, which allows to have different manganese oxides with wide range of applications, from catalysis [1, 2] to batteries [3]. In many cases manganese oxides
M
have been prepared by calcinating a precursor such as manganese carbonate [4-6] Manganese carbonate can be produced by different methods but the solvo-hydrothermal routes [7-10], and among them the microemulsion-mediated synthesis, seem to be the most
d
interesting due to their easiness and scalability.
te
Referring to the definition of Danielsson and Lindman, a “microemulsion” is a system of water,
Ac ce p
oil and amphiphile which is a single optically isotropic and thermodynamically stable liquid solution [11]. Water-in-oil microemulsions with spherical shape are the most used for the synthesis of nanoparticles and they are particularly interesting because the reverses micelles can be used as “nano-reactors”, where the precipitation of nanoparticles takes place [12]. Formation of nanoparticles in the micelles occurs according to a mechanism involving coalescence of droplets, exchange of reagents and nucleation of particles. Further mechanisms have been also assumed to explain experimentally observed particle dimensions larger than those predictable by the previous scheme. In particular the autocatalysis occurs when larger particles grow faster due to increased surface area [13] and the Ostwald Ripening takes place when the dissolution of small crystals or sol particles and the re-deposition of the dissolved species on the surface of larger crystals lead to a further growth of the large particles at the expense of smaller ones (as a result larger particles grow as smaller ones disappear) [14].
Page 2 of 31
Synthesis of nanoparticles is strictly related to microemulsion characteristics such as interfacial fluidity of surfactant layer (controlling intermicellar exchange) and charge density of surfactant polar heads (stabilizing newly formed nanoparticles and limiting their aggregation). In literature there are different exhaustive works, which well explain the base theory of reverse micelles and the
ip t
influence of operating parameters on micelle size, shape and flexibility [15-18]. Operating factors which mainly affect the characteristics of microemulsions are: type of surfactant and solvent,
cr
water/surfactant molar ratio (w) and concentration of reagents. In particular, by fixing surfactant
us
and solvent type and by increasing w, an increase of micelles size and then of nanoparticles size (both as primary particles and their aggregates emerging from coagulation phenomena) has been
an
observed [15].
Tojo et al. [14] experimentally validated a Montecarlo simulation (including particle-particle
M
coagulation) showing that diameter of particles increased by increasing the water-surfactant molar ratio because the size of microemulsion droplets acts as a physical constrain to the maximum particle size.
d
For a given microemulsion system, an increase of reagents concentration can determine an
te
increase of particle size ([15] and references within), even though some works reported a
Ac ce p
contrasting decrease of particles size by increasing reagents concentration [19-20]. Kumar et al. (2004) [21] developed a population balance model considering autocatalysis of large particles and they found out that the final size of nanoparticles has a no monotonous trend with respect to the concentration of reagents. In particular, at low concentrations, few numbers of nuclei are formed and an increase of reagents concentration results into growth thereby increasing the particle size. When the reagents concentration reaches the critical nucleation value, particle size will be the maximum and a further increase of concentration results into an increase of nucleation events in the micelles, leading to a decrease of particle size. This generic scheme can be a valuable explanation for the contradictory results about the effect of reagents concentration. Nevertheless, this scheme is valid for systems where agglomeration is negligible, i.e. for high film rigidity.
Page 3 of 31
Tojo et al. (1997) developed a Montecarlo simulation including agglomeration of particles and they showed the effect of the concentration of reagents on particle size distribution for different film rigidities [14]. Simulations results showed that higher concentrations give rise to a population of larger nanoparticles. Moreover, by looking at particle size distribution, for high concentrations of
ip t
reactants, systems with high film rigidity presented a bimodal distribution due to a limited intermicellar exchange of particles. Bimodal distribution was not present for low concentrated
cr
microemulsions, where nucleation and growth occurred in the same time scale. Bimodal
us
distribution disappeared also for high concentrations of reagents, where the film rigidity diminished. As for the effect of reagents concentration on particle polydispersity, some works reported a
an
decrease of polydispersity of particle size by increasing the concentration of reagents [22-24], whilst other works simulated an increase of polydispersity at high concentration values [14].
M
Pagnanelli et al. [25] demonstrated for the synthesis of MnCO3 nanoparticles in microemulsion that an increase of reagents concentration can result into different effects. For rigid systems with reduced intermicellar exchange (obtained by using solvents as hexane and isooctane, which can
d
interact tidily with surfactant tails making micelles more rigid), an increase of concentration acts as
te
a further obstacle to the exchange of reagents, thus determining a decrease of mean size and
Ac ce p
unchanged polydispersity. On the other hand, for fluid systems (obtained by increasing cosurfactant dosage) an increase of concentration did not alter mean size, but increased polydispersity.
Therefore, in spite of the advantages given by the microemulsion mediated routes, experimental results reported in literature seem to be strongly affected by the parameters influencing the characteristics of microemulsion (i.e. water-surfactant molar ratio, concentration of reactants etc.). A general framework for the understanding of the effect of single factors and their interaction should be then developed in order to overcome apparent contradictions in the experimental findings. Specifically concerning the synthesis of MnCO3 nanoparticles by microemulsion, few works can be found in the literature. In particular, Aragòn et al. (2007) produced submicron particles of MnCO3 with an average size of 200 nm by using inverse micelles made up of water, cyclohexane
Page 4 of 31
(proportion 1:4) and 0.2 M sodium bis(ethylhexyl) sulfosuccinate (Aerosol OT, AOT) [26]. Wu et al. (2006) used a CTAB microemulsion method followed by thermal treatment to synthesize MnCO3 nanocrystals performing also a shape control by modification of operating variables such as concentration of reagents and water-surfactant molar ratio [27]. Pagnanelli et al. [25] investigated
ip t
the synthesis of MnCO3 nanoparticles by CTAB microemulsion addressing the effects of reaction time (30-60 min), type of solvent (cyclohexane, isooctane and hexane), cosurfactant dosage (1.5-
cr
4.5 g) and reagents concentration (1.0-0.25 mol/L) by characterization with Scanning Electron
us
Microscopy. Previous results showed that size can be significantly increased by increasing reaction time and cosurfactant dosage. Nature of solvent also significantly affected particles size, whilst, as
an
above mentioned, the concentration of reagents had opposite effects depending on the system nature.
M
In this work we continued investigating the synthesis of MnCO3 nanoparticles by addressing different aspects such as the influence of post thermal treatment, water-surfactant molar ratio and reagents concentration in other different conditions, by using X-Ray, SEM and Transmission
Material and methods
te
2.
d
Electron Microscopy for the characterization of obtained nanoparticles.
Ac ce p
All chemicals used for this work (MnCl2, (NH4)2CO3, cyclohexane, n-pentanol, ethanol) were analytical grade reagents purchased from Kanto Chemicals Co. CTAB (99+%) was purchased from ACROS ORGANICS. A quaternary water in oil microemulsion of cyclohexane (20 g) + npenthanol (1.5 g) - CTAB (2g) – Water (MnCl2 and (NH4)2CO3 solutions) was employed as reaction medium. Pentanol was added as co-surfactant in order to increase the micelle permeability [28]. As shown in figure 1, MnCl2 and (NH4)2CO3 solutions were added dropwise in their respective organic phases until they became optically transparent. Then, after the mixing of the two initial microemulsions, reaction (1) occurred in MnCl2 + (NH4)2CO3 → MnCO3 + 2NH4Cl
(1)
Reaction was carried out at 30 °C, under stirring, for 2 hours. As showed in table 1 water-surfactant molar ratio (w) was investigated in the range 5-7.5 by keeping constant the amount of surfactant (2 g) and by modifying the amounts of aqueous
Page 5 of 31
solutions in the microemulsions: 0.49, 0.62 and 0.74 g of aqueous solutions were added in their respective systems in order to reach respectively w=5, w=6.25 and w=7.5. Concentration of reactants (MnCl2 and (NH4)2CO3) was investigated in the range 0.25-1.0 M. After precipitation, the insoluble product MnCO3 (KsMnCO32.24 x 10-11) was collected by
ip t
centrifugation and it washed twice by ethanol and then twice by distilled water. When the thermal treatment was tested, it was carried out in a teflon-lined autoclave reactor at 120°C for 4 hours
cr
after mixing. Size and morphology of the as synthesized particles were estimated by Scanning
us
Electron Microscope (JEOL JSM-5510LV) and by Transmission Electron Microscope (JEOL JEM2100 in ISSP, Univ. Tokyo) operating at 200 kV. Image J software was used for image analysis in
an
order to determine the size distribution of produced particles. For each sample, particle size distribution was evaluated by analyzing four TEM images resulting in a number of analyzed
M
particles included between 550 and 1150. Samples for microscopic observations were prepared by adding acetone and by dispersing the particles in an ultrasonic bath. After ultrasonication, a drop of the suspension was put on a copper grid and dried at air before microscope observation. Samples
d
were also analyzed by MacScience X-ray diffractometer M03X (XRD), with Cu Kα radiation (λ =
Ac ce p
of products.
te
1.5406 Ǻ) operating at 40 kV and 20 mA emission condition in order to determine the composition
Analysis of Particle Size
Particles were manually measured one by one by using first a graphic software (Adobe Illustrator) to draw the particles contours and to remove the original image from the background and then by using a software to analyze the resulting black/white image (ImageJ) by measuring the size of each particles whose contours had been preliminary drawn. In order to make sure the image analysis was representative, the first step in the image analysis was the determination of the minimum number of nanoparticles that should be counted in order to have representative sample statistics (i.e. mean size and standard deviation). This was done empirically by estimating mean and standard deviation for samples of nanoparticle with increasing numerosity and determining the minimum number of nanoparticles for which no significant variation was observed for these statistics.
Page 6 of 31
3.
Results and Discussion
3.1
Morphology and particle size characterization by image analysis Sample produced in condition 0 of table 1 (w=20 and C=0.5 M) denoted the formation of cubic
ip t
nanoparticles with mean size of 300 nm (Figure 2a). After thermal treatment (at 120°C for 4 h) shape remained unaltered (Figure 2b), while a clear increase of average particle size occurred
cr
(average size of 500 nm). The understanding of the mechanism which determined the increase of
us
particle size goes beyond the purpose of this paper but it might be due to coalescence and/or Ostwald Ripening since the final particle size was about twice than without thermal treatment
an
(suggesting two different phases might have come together) and because it was clearly temperature dependent, as showed by the pioneering works done by Lifshitz and Slyozov [29].
M
Aiming to produce nanoparticles smaller than 100 nm, further tests were performed without any thermal treatments as done elsewhere [27]. In addition, reduced w values were chosen due to the known effect of this parameter on particlesaverage size.
d
Experimental conditions for further tests were arranged according to a factorial design with two
te
factors (water-surfactant molar ratio and concentration of reagents) as listed in Table 1.
Ac ce p
As shown by TEM images and particle size distribution elaborated from them (figures 3-8), water-surfactant molar ratio (w) exhibited a relevant influence on particle size and polydispersity (evaluated as the standard deviation around the mean size) and also on particles shape. Regarding the histograms in figure 3-8, it should be noted that whereas particles were grouped as >x nm, it was because they were too few to be included to their real size group and only by grouping them together the histogram bar could be observed. Of course there are larger particles but in spite of being rather visible just because larger, when counted among the other particles they were just few (way less than 10) out 500-1150 counted particles. That is why just an observation without analysis can lead to mistakes. For w = 5 (CMn=0.25 M) particles had mostly a globular shape (fig. 3) with mean size 6.5 nm and with narrow dispersion (standard deviation 1.4 nm). By increasing water/surfactant molar ratio up to 6.25 (CMn=0.25 M)(fig. 4), particles become mainly rhombohedral (and also cubic) and their
Page 7 of 31
size and dispersion increased (mean size 71 nm and standard deviation 24 nm). When working at w = 7.5 (CMn=0.25 M) cubic nanoparticles were obtained (fig. 5) with increased particles size (mean size 89 nm) and dispersion (37 nm) with respect to lower levels of w. TEM images clearly showed the isotropic growth of particles according to spherical and cubic
ip t
shapes, even though literature works also reported anisotropic growth in presence of cationic surfactants, due to assembling phenomena on negatively charged surface of forming particles [18,
cr
30, 31].
us
As for the concentration of reagents, for the different levels of w, very similar results were obtained (Figures 6-8) in terms of shape, mean size and dispersion. Significance of observed
an
differences among mean size and dispersion for some conditions has been statistically evaluated in the following analysis (paragraph 3.3). X-ray characterization
M
3.2
Characterization by X-Ray diffraction was reported in Figures 9 for different w levels at CMn=1M. All reflection peaks can be indexed as a rhomb-centered hexagonal (rch) phase (space
d
group R3c(167)) of MnCO3 with lattice constants α=4.790 Ǻ and c=15.69 Ǻ (JCPDS card No 44-
te
1472). However, the XRD pattern of the sample obtained at w=5, the lowest water-surfactant molar
Ac ce p
ratio (Figure 9c), it also showed (peaks marked with *) the presence of a cubic structure of Mn2O3 (JPCDS No. 01-071-0636). Since all reactions were carried out by working under nitrogen atmosphere, it can be supposed that the reduced size of nanoparticles obtained in such condition and thus its high surface area, favored the oxidation of surface portions. In addition it should be mentioned that, for the lowest surfactant molar ratio, a modest amount of product was obtained giving a low signal to noise ratio of the XRD spectra (figure 9c). Average crystallite size (τ) was determined from full width at half maxima (FWHM) of peak 113 and by using the Scherer equation [32]: (2) where k is a dimensionless shape factor (0.9), λ is the X-ray wavelength (0.154 nm), β is the line broadening at half the maximum intensity, and θ is the Bragg angle. Samples obtained for w= 5, w=6.25 and w=7.5 (CMn=1 M) showed an average crystallite size respectively of 20, 35 and 45 nm. Page 8 of 31
These values are in agreement with mean size values obtained by TEM images that for the same conditions are 10, 66 and 89 nm, respectively. In particular, for the sample obtained at w=5, by considering the standard deviation due to sample polydispersity (2.6 nm), a 99% confidence interval of 10±7.8 is obtained, meaning that we have the 99% of probability that the true value of
ip t
the mean size of the population is included in this range. This consideration puts in evidence that similar predictions were obtained from image analysis and XRD, but also show the importance of
cr
evaluating sample polydispersity for nanoparticles population. Image analysis by TEM can be used
us
for such aim conversely to XRD spectra which, according to the elaboration proposed here, can only estimate the crystallite size. Resuming, the comparison between mean size by TEM and
an
crystallite size by XRD denoted that the first method give similar (w=5) or larger (w=6.25 and 7.5) values than the second. Slight underestimates for w=6.25 and w=7.5 could be due to
M
polycrystallinity of nanoparticles in which different crystal cells aggregate giving larger visible
3.3 Statistical analysis
d
particles in TEM images.
te
From the preliminary analysis of the method, it emerged that representative values of both
Ac ce p
mean size and standard deviation required at least the counting of 500 nanoparticles. Then counting more than this number ensures the choice of figures does not affect the final estimates. After this preliminary empirical estimate of minimum number of nanoparticles in the counting, the error variability of data (SSE) was supposed to be due to both the execution of image analysis (SSIMAG) and random experimental errors (SSPE). SSE=SSIMAG+SSPE
Different replicates (three) of the same experiments were then performed (to estimate variability due to random error) and the same images were analyzed independently by the same operator (to estimate variability due to image analysis execution). Statistical analysis of variance denoted that variability due to execution of image analysis is not statistically significant with respect to variability due to random errors.
Page 9 of 31
Experimental tests were performed according to a full factorial design with two factors, A=w and B=CMn (the molar concentration of the manganese source), the first at three levels (5, 6.25 and 7.5) and the second at two levels (0.25 and 1.0 M). Different statistics were evaluated for particles populations obtained for each treatment (i.e. different operating conditions used in factorial design).
ip t
In particular mean (μ) and standard deviation (S) of particles size (xi) were evaluated.
These statistics were reported for each treatment in Figure 10, from where it can be noted that
cr
w is the most influencing factor for all statistics, meaning that, even for a small increase of w in the
us
investigated range, an increase of both mean (Fig. 10a) and standard deviation (Fig. 10b) was observed. Thus, by increasing water-surfactant molar ratio, both size and heterogeneity of particle
an
population tend to increase. On the other hand, by increasing the concentration of reagents, a less relevant effect on particle mean size and standard deviation was observed.
M
These qualitative observations about mean values were confirmed by ANOVA [33]. Estimate of variance component due to error (MSE) was obtained by independent replicated tests and it takes into account the variability associated to random experimental errors, the
d
variability due to image choice for statistic determination, and the variability due to image analysis
te
uncertainties. In this context eight TEM images from two replicates (four images for each replicate)
Ac ce p
of the same treatments were used to estimate variability due to random errors. ANOVA put in evidence that only A factor (w ratio) has a significant effect on mean size, whilst B factor (concentration of reagents) and AB interaction are not significant (99% confidence level). Estimates of the effects were evaluated considering the higher and lower levels for each factor (w = 5 and 7.5 and CMn= 0.25 and 1 M) and all effects were reported in figure 11 along with the confidence bands ±tα/2,n-1√(MSE/n), where α=0.01 and n=8 (dashed lines). Estimates of the effects which are larger than the confidence bands are those statistically significant (99% confidence level). Standard deviations in the different operating conditions were compared by considering the corresponding variances (the square of standard deviations) and performing statistical tests. In particular, F tests were carried out by comparing variances at fixed values of reagent concentration
Page 10 of 31
for various w and at fixed w for the two levels of reagent concentration. These statistical tests (99% confidence level) denoted that: -
For both concentration levels, dispersion of particle size for w=5 is significantly lower than at w=6.25 and w=7.5, while variances are the same for these high levels of w. Then, an
ip t
increase of w from 5 to 6.25 determined an increase of polydispersity which remained unaltered passing from w=6.25 to w=7.5.
As for the effect of the concentration of reagents, different results were obtained depending
cr
-
us
on w level. In particular, for w=5 and w=6.25 a significant increase of variance was observed by increasing the concentration of reagents, whilst for w=7.5 by increasing the
an
concentration a statistically significant decrease of variance was observed. As for the effect of water/surfactant molar ratio on nanoparticle mean size, it can be explained
M
by considering that this factor determines an increase of micelle size, an then an increase of nanoparticle size growing inside them due to the templating properties and physical constrains offered by micelles during nanoparticle growth.
d
As for the effect of w on polydispersity (variances of particle size), it was not specifically
te
addressed in previous literature: it could be argued that as micelles are larger they can contain a
Ac ce p
wide range of nanoparticles below their size being a physical constrain for nanoparticles during their growth.
As for reagent concentration, the experimental results reported in this work comprehended a test condition already tested (w=6.25) [25] plus other two levels (w=5 and w=7.25). The intermediate level already studied presented the same behavior found in the past: the increase of reagent concentration in the range 0.25-1M did not alter the average size of particles but a significant increase of variance was observed. The reproducibility of this result obtained in different experimental campaigns with different image characterization (TEM instead of SEM) gave strength to such finding. For the other two levels of w, the effect of reagent concentration on particle size remained the same, i.e. by increasing the concentration in the investigated range no significant effect on mean size was observed. Nevertheless different effects were found for variances: positive effect for w=5 and negative effect for w=7.25.
Page 11 of 31
As for the lack of the effect of reagent concentration on particle mean size, by considering the non-monotonous trend predicted by simulations [21], it could be reasoned that, for this system, the investigated concentrations are near the critical concentration (maximum) of the curve “nanoparticle size” versus “reagent concentration” being the concentration of reagents scarcely
ip t
influencing nanoparticle mean size.
As for the opposite effect of reagent concentration on nanoparticle polydispersity, it is in
cr
agreement with previously reported experimental results showing both the increase and the
us
decrease of polydispersity by increasing the concentration of reagents [14, 22, 23, 24]. As in the case of mean size, it is evident the concentration of reagents can play a different role in
an
nanoparticle growth in microemulsions, thus requiring further experimentation in order to complete the mapping of experimental behavior for different operating conditions.
M
It should be noted that average size of nanoparticles ranged from few nanometers up to 100 nm. In particular, for w=7.5, only an average of 2-3% of particle population were larger than 150 nm, whilst for w=5 and w=6.25, even more than respectively 99% and about 80% of particles were
d
smaller than 100 nm. Thus, since the size of individual microemulsion droplets for CTAB systems
te
can range from 10 to 200 nm depending on system specific composition and operating conditions
Ac ce p
[34], only a very small percentage of particle population could be considered too large to be contained in the micelles. Therefore, it is likely that coalescence of primary nanoparticles occurred mostly inside the micelles, whilst a further stacking in rhombohedral structures might occurred in a second phase out of the micelle cores, resulting into the larger particles.
4
Conclusions
Nanoparticles of rhomb-centered hexagonal phase of manganese carbonate were obtained by a microemulsion-mediated route at room temperature without any post-thermal treatment. Operating conditions allowing to obtain the smaller particles with globular shape were w = 5 and CMn = 0.25 M. Experimental data and statistical analysis revealed that, in the investigated range, water-surfactant molar ratio had a significant positive effects on both particle size and particle dispersion, whilst the
Page 12 of 31
concentration of reagents did not show any significant effect on average size and opposite effects on variances (depending on w value), thus requiring a further experimentation in order to complete the mapping of experimental behavior for different compositions of the system.
ip t
Acknowledgement
This study was supported in part by KAKENHI 22110503 for Scientific-Research on Innovative
us
for the Promotion of Science and Technology (MEXT, Japan).
cr
Areas (Frontier Science of interactions between plasmas and nano-interfaces) and Strategic Funds
The use of the facilities of the Materials Design and Characterization Laboratory at the Institute for
an
Solid State Physics, University of Tokyo, is gratefully acknowledged.
[1]
M
References
M. Baldi, E. Finocchio, F. Milella, G. Busca, Evaluation of the mechanism of the oxy-
dehydrogenation of propane over manganese oxide, Appl. Catal. B 16 (1998) 43-51. E.R. Stobbe,
B.A. De Boer,
J.W. Geus, The reduction and oxidation behaviour of
d
[2]
J.K. Chang, W.C. Tsai, Material characterization and electrochemical performance of
Ac ce p
[3]
te
manganese oxides. Catal. Today 47 (1999) 161-167.
hydrous manganese oxide electrodes for use in electrochemical pseudocapacitors, J. Electrochem. Soc. 150, (2003) 1333–1338. [4]
S. Lei, K. Tang, Z. Fang, Q. Liu, Q. Zheng, Peparation of α-Mn2O3 and MnO from thermal
decomposition of MnCO3 and control of morphology, Mater. Lett. 60 (2006) 53-56. [5]
X. Shen, Z. Ji, H. Miao, J. Yang, K. Chen, Hydrothermal synthesis of MnCO3 nanorods and
their thermal transformation into Mn2O3 and Mn3O4nanorods with single crystalline structure, J. Alloy Compd. 509 (2011) 5672-5676. [6] L.X. Yang, Y.J. Zhu, H. Tong, W.W. Wang, Submicrocubes and highly oriented assemblies of MnCO3 synthesized by ultrasound agitation method and their thermal transformation to nanoporous Mn2O3, Ultrason. Sonochem. 14 (2007) 259–265.
Page 13 of 31
[7] H. Hu, J.Y. Xu, H. Yang, J. Liang, S. Yang, H. Wu, Morphology-controlled hydrothermal synthesis of MnCO3 hierarchical superstructures with Schiff base as stabilizer, Mater. Res. Bull. 46 (2011) 1908–1915. [8]
S. Lei, Z. Liang, L. Zhou, K. Tang, Synthesis and morphological control of MnCO3 and
ip t
Mn(OH)2 by a complex homogeneous precipitation method, Mater. Chem. Phys. 113 (2009) 445– 450.
M. Sanchez-Dominguez, K. Pemartin, M. Boutonnet, Preparation of inorganic nanoparticles
cr
[9]
us
in oil-in-water microemulsions: A soft and versatile approach, Curr. Opin. Colloid. In. 17 (2012) 297–305.
an
[10] X. Wang, Y. Li, Hydrothermal reduction route to Mn(OH)2 and MnCO3 nanocrystals, Mater. Chem. Phys. 82 (2003) 419–422.
M
[11] I. Danielsson, B. Lindman, The definition of Microemulsion, Colloids and Surface 3 (1981) 391-392.
[12] A. Vàzquez-Olmos, R. Redòn, A.L. Fernàndez-Osorio, J.M. Saniger, Room-temperature
d
synthesis of Mn3O4 nanorods, Appl. Phys. A 81 (2005) 1131–1134.
te
[13] M. Ethayaraja, K. Dutta, D. Muthukumaran, R. Bandyopadhyaya, Nanoparticle Formation in
Ac ce p
Water-in-Oil Microemulsions: Experiments, Mechanism and Monte Carlo Simulation, Langmuir 23 (2007) 3418-3423.
[14] C. Tojo, M.C. Blanco, F. Rivadulla, M.A. Lopez-Quintela, Kinetics of the Formation of Particles in Microemulsions, Langmuir 13 (1997) 1970-1977. [15] J. Eastoe, M.J. Hollamby, L. Hudson, Recent advances in nanoparticle synthesis withreversed micelles, Adv. Colloid Interfac. 128-130 (2006) 5-15. [16] A.K. Ganguli, A. Ganguly, S. Vaidya, Microemulsion-based synthesis of nanocrystalline materials, Chem. Soc. Rev. 39 (2010) 474-485. [17] M.A. Michaels, S. Sherwood, M. Kidwell, M.J. Allsbrook, S.A. Morrison, S.C. Rutan, E.E. Capenter, Quantitative model for prediction of hydrodynamic size of nonionic reverse micelles, J. Colloid Interf. Sci. 311 (2007) 70-76.
Page 14 of 31
[18] M.P. Pileni, Control of the size and shape of inorganic nanocrystals at various scales from nano to macrodomains, J. Phys. Chem. 111 (2007) 9019-9038. [19] M. Fernandez-Garcıa, X. Wang, C. Belver, A. Iglesias-Juez, J.C. Hanson, J.A. Rodriguez, Ca doping of nanosize Ce-Zr and Ce-Tb solid solutions: Structural and electronic effects, Chem. Mater.
ip t
17, 4181-4193.
[20] Z. Ye, M. Tan, G. Wang, J. Yuan, Preparation, characterization and application of fluorescent
cr
terbium complex-doped zirconia nanoparticles, J. Fluoresc. 15 (2005) 499-505
us
[21] R.A. Kumar, G. Hota, A. Mehra, K.C. Khilar, Modeling of Nanoparticles Formation by Mixing of Two Reactive Microemulsions, AIChE J. 50 (2004) 1556-1567.
micelles, Colloids Surf. A 19 (1996) 123-131.
an
[22] J. Eastoe, S. Stebbing, J. Dalton, R.K. Heenan, Preparation of colloidal cobalt using reversed
M
[23] I. Lisiecki, M.P. Pileni, Synthesis of well-defined and low size distribution cobalt nanocrystals: the limited influence of reverse micelles, Langmuir 19 (2003) 9486-9489. [24] M.P. Pileni, The role of soft colloidal templates in controlling the size and shape of inorganic
d
nanocrystals, Nat. Mater. 2 (2003) 145-150.
te
[25] F. Pagnanelli, G. Granata, E. Moscardini, L. Toro, Synthesis of MnCO3 nanoparticles by
Ac ce p
microemulsions: quantitative evaluation of the effect of operating conditions on particle size distribution, J. Nanopart. Res. 15 (2013) 1887-1898. [26] M.J. Aragón, C. Pérez-Vicente, J.L. Tirado, Submicronic particles of manganese carbonate prepared in reverse micelles: A new electrode material for lithium-ion batteries, Electrochem. Commun. 9 (2007) 1744-1748.
[27] X. Wu, M. Cao, H. Lu, X. He, C. Hu, Microemulsion-mediated solvothermal synthesis and morphological evolution of MnCO3 nanocrystals, J. Nanosci. Nanotechno. 6 (2006) 2123-2128. [28] G. Palazzo, F. Lopez, M. Giustini, G. Colafemmina, A. Ceglie, Role of the cosurfactant in the CTAB/water/n-pentanol/n-hexane
water-in-oil
microemulsion.
1.
Pentanol
effect
on
the
microstructure, J. Phys. Chem. B 107 (2003) 1924-1931. [29] I.M. Lifshitz, V.V. Slyozov, The kinetics of precipitation from supersaturated solid, J. Phys. Chem. Solids 19 (1961), 19-35.
Page 15 of 31
[30] Y. Liu, Z. Liu, G. Wang, Preparation of Mn3O4 nanowires by calcining the precursor powders synthesized in a novel inverse microemulsion, Appl. Phys. A-Mater. 76 (2003) 1117–1120. [31] S. Vaidya, P. Rastogi, S. Agarwal, S.K. Gupta, T. Ahmad, A.M. Antonelli, K.V. Ramanujachary, S.E. Lofland, A.K. Ganguli, Nanospheres, nanocubes, and nanorods of nickel
ip t
oxalate: Control of shape and size by surfactant and solvent, J. Phys. Chem. C 112 (2008)12610– 12615.
cr
[32] B.D. Cullity, Elements of X-ray diffraction. Addison-Wesley, Reading, 1978.
[34] G. Garg, P.A. Hassan, S.K. Kulshreshtha,
us
[33] D.C. Montgomery, Design and analysis of experiments, McGraw Hill, Milano, 2005. Dynamic light scattering studies of rod-like
Ac ce p
te
d
M
an
micelles in dilute and semidilute regime, Colloid Surf A 275 (2006)161–167
Page 16 of 31
CMn [M]
w
0
0.50
20.00
1
0.25
5.00
2
0.25
6.25
3
0.25
7.50
4
1.00
5.00
5
1.00
6.25
6
1.00
7.50
Ac ce p
te
d
M
an
us
cr
test
ip t
Table 1: investigated factors of full factorial experimentation
Page 17 of 31
Flowchart for the synthesis of nano-crystalline manganese carbonate
Figure 2
SEM images of MnCO3 particles obtained for w=20, 0.5 M at 30°C (a) and for w=20, 0.5 M at 90 °C for 4 hours (b)
Figure 3
Bright field TEM image of MnCO3 particles obtained for w=5, 0.25 M at 30 °C (a) and their size distribution (b).
Figure 4
Bright field TEM image of MnCO3 particles obtained for w=6.25, 0.25 M at 30 °C (a) and their size distribution (b).
Figure 5
Bright field TEM image of MnCO3 particles obtained for w=7. 5, 0.25 M at 30 °C (a) and their size distribution (b).
Figure 6
Bright field TEM image of MnCO3 particles obtained for w=5, 1 M at 30 °C (a) and their size distribution (b).
Figure 7
Bright field TEM image of MnCO3 particles obtained for w=6.25, 1 M at 30 °C (a) and their size distribution (b).
Figure 8
Bright field TEM image of MnCO3 particles obtained for w=7. 5, 1 M at 30 °C (a) and their size distribution (b).
Figure 9
X-ray diffraction patterns of nanocrystalline MnCO3 obtained for w=7.5 (a), w= 6.25 (b) and w=5 (c)
Figure 10
Mean size (a) and standard deviation (b) in the different treatments
Figure 11
Estimates of effect with confidence bands on mean size.
Ac ce p
te
d
M
an
us
cr
ip t
Figure 1
Page 18 of 31
Highlights MnCO3 nanoparticles were produced by inverse microemulsion at room temperature. Water-surfactant molar ratio and reactant concentrations were investigated.
Water-surfactant molar ratio has positive effect on both size and polydispersity.
Ac ce p
te
d
M
an
us
cr
Reactant concentration influences only polydispersity.
ip t
Effects on mean size and polydispersity were assessed by statistical analysis.
Page 19 of 31
Graphical Abstract (for review)
Ac
ce pt
ed
M
an
us
cr
ip t
Graphical abstract
Page 20 of 31
Figure
Ac
ce pt
ed
M
an
us
cr
ip t
Fig. 1
Page 21 of 31
Ac
ce pt
ed
M
an
us
cr
ip t
Fig. 2
Page 22 of 31
Ac
ce pt
ed
M
an
us
cr
ip t
Fig. 3
Page 23 of 31
90
d>130 nm
120
110
us
an
M
16
100
12
80
ed
14
70
60
6
50
8
40
10
30
d<30 nm
frequency (%)
ce pt
Ac
cr
ip t
Fig. 4
4
2
0
Page 24 of 31
Ac
ce pt
ed
M
an
us
cr
ip t
Fig. 5
Page 25 of 31
Ac
ce pt
ed
M
an
us
cr
ip t
Fig. 6
Page 26 of 31
Ac
ce pt
ed
M
an
us
cr
ip t
Fig. 7
Page 27 of 31
Ac
ce pt
ed
M
an
us
cr
ip t
Fig. 8
Page 28 of 31
Ac
ce pt
ed
M
an
us
cr
ip t
Fig. 9
Page 29 of 31
Fig 10
120
ip t
80 60
cr
CMn=0.25 M 40
CMn=1 M
us
mean size (nm)
100
20
4.5
5
5.5
6
6.5
7
7.5
40
ed
30
15 10 5
ce pt
25
Ac
standard deviation (nm)
35
20
8
M
w
an
0
CMn=0.25 M CMn=1 M
0
4.5
5
5.5
6
6.5
7
7.5
8
w
Page 30 of 31
Ac
ce pt
ed
M
an
us
cr
ip t
Fig. 11
Page 31 of 31