Density and mass of particles

Effective density of individual particles

A wide range of nanoparticles with different sizes and densities were produced in this thesis and an overview of them can be seen in Figure 4.4. The most diverse selection of generation methods was used inPaper I, where DOS particles were produced with an evaporation-condensation generator, NaCl by atomizing and then drying aqueous salt solution, titania by thermally decomposing bubbled TTIP vapor and silver with a typical tubular furnace setup. The generator used in the nanoparticle production inPaper II, Paper IIIand Paper IVwas the LFS, producing alumina (Al2O3), titania and silver particles, respectively.

The structure of the generated solid particles can be altered from the state that they come out in straight from the generator. Here, the morphology was controlled by using either a residence time chamber or a sintering furnace to change the effective density through agglomeration and sintering, respectively. InPaper I, the large agglomerates

4.2. Density and mass of particles 27

Figure 4.4: Produced particle morphologies. The relevant paper is listed in the legend. Lines connecting different points in the figure show the range of particles produced in any given instance, for example the DOS particles were generated for the calibration of DENSMO with the whole particle diameter range.

and the large spherical particles were produced by first letting the aerosol agglomerate, growing the diameter of the particles. The large agglomerates could then be sintered to more spherical morphology. InPaper IV, the LFS generated nanoparticles are guided through a residence tube, in which the gas temperature stays higher longer, sintering the silver particles spherical. This does not happen without the residence tube, as is reported in several previous studies (Paper II,Aromaa et al. (2012); Rostedt et al. (2009)).

DENSMO was used in Paper IandPaper IVin conjunction with the combined use of SMPS and ELPI. A comparison of the effective density measurement results given by these methods are shown in Figure 4.5. The measured effective densities fall mostly inside the estimated uncertainty of±20%, but a few measurements with greater deviation show the effect of particle bounce, which is caused by problematic particle morphologies and, possibly, the impactor loading due to high concentration.

Figure 4.5: Effective density measurements made with DENSMO. The results are fromPaper IandPaper IV. The SMPS-ELPI method was used as the reference. Dashed lines denote±20%

uncertainty. (Adapted fromPaper I)

Figure 4.6: Real-time measurements from transient processes: (a) the sintering of silver nanoparticles and (b) generating silver nanoparticles with LFS. The sintering took place during a continuous heating from 20^◦C to 300^◦C illustrated with the dark background, which in the LFS generation shows the start and end of the synthesis process. (Adapted fromPaper I)

All in all, DENSMO is a capable measurement instrument when the effective density is the relevant value. As a monitor, the effective density measured by DENSMO can be used alone to see important stages of the synthesis process. Figure 4.6 shows the results of (a) the sintering of silver agglomerates inPaper Iand (b) one synthesis cycle of silver nanoparticles inPaper IV. In the case of LFS synthesis, a clear transient period can be seen after the ignition of the LFS, showing as a steady increase in the effective density in Figure 4.6 (b), which is caused by stabilizing temperature.

Mass size distribution of nanoparticles

The reduction of residual particles is achieved in Paper II by the addition of ethyl-hexanoic acid (EHA), which aids the evaporation of the precursor liquid. TEM images of the produced powder are shown in Figure 4.7. The effect the addition of EHA has on the distribution of mass to the residual mode and on the nanoparticle mode is clearly visible. Without the addition of EHA, several micron-wide residual particles can be seen to contain most of the alumina that was collected. Previously, the residual removal has been achieved by adding excessive amount of EHA to the precursor solution (e.g. Mädler et al., 2006; Rosebrock et al., 2013), however as little as 5% was shown to be enough. As the EHA is added the coverage of the nanoparticles can be seen to increase dramatically on the TEM images.

The transfer of mass from the residual mode to the nanoparticle mode increases the number concentration of the nanoparticles only to a certain point, until the coagulation of the nanoparticles starts to increase the particle diameter with the cost of number concentration. This can be seen in the aerosol measurement results in Figure 4.8, where number size distributions from SMPS and ELPI are shown.

4.2. Density and mass of particles 29

Figure 4.7: Powder optimization by addition of EHA to remove residual particles from the produced alumina nanoparticle powder. The absence of the residual particles is evident after the addition of 5% of EHA, and higher amounts do not change the nanoparticle size any further.

The difference in the amount of nanoparticles can be explained by the image position on the grid and the collection time of the sample. (Paper II)

EHA0%

dN/dlogda,b(109 cm−3 ) ^{ELPI+ (da)}_{SMPS (d}

b)

Figure 4.8: The change in the number-size distribution of alumina particles as a function of added EHA. The ELPI+ and SMPS distributions overlapping here indicate an effective density of close to one. Additionally, the number concentration and the mean particle diameter do not change noticeably after the addition of 5% of EHA. (Paper II)