Femtosecond laser filament-assisted AgI-type pyrotechnic nucleant-induced water condensation in cloud chamber

AgI-type pyrotechnics are widely used in the field of weather modification, as a kind of artificial ice nuclei. However, their precipitation yield remains an intensively studied area. In this paper, we present a study of AgI-type pyrotechnic nucleant-induced water condensation promoted by femtosecond laser filaments in a cloud chamber. It is found that when 50-ml sample was irradiated by the laser filaments, the particles condensed on the glass slide are more soluble and slightly larger (5–15 μm). The irradiation of the laser filament on the nucleant rarely induces the generation of particles of sizes larger than 1 μm; however, it increases the decay time of particles from 13 to 18 min by the creation of numerous small particles. The amount of snow on the cold bottom plate increases by 4.2–13.1% in 2 h, compared to that without the irradiation of the laser filament. These results are associated with the production of high-concentration HNO3 by the laser filament. The concentration of HNO3 in the melt water increases by more than ten times when the sample was irradiated by the laser filaments. © 2018 Optical Society of America under the terms of the OSA Open Access Publishing Agreement


Introduction
Artificial weather modification has always been an aspiration of people for the control of disastrous weather, such as hail, flood, heavy snowfall, or drought.AgI is an important manmade ice nucleus for the catalyzing of cold clouds in weather modification, at temperatures from −5 to −20 °C [1][2][3].Owing to its high melting point, as a kind of inorganic compound, it volatilizes only at its flame temperature.Therefore, in practical applications, AgI is typically prepared for pyrotechnics by mixing with flammable compounds.Further, in order to increase the nucleation effectiveness (the number of particles nucleated per gram of nucleant) of the AgI pyrotechnic in the cloud, hygroscopic salts are mixed as well.The nucleation effectiveness of the AgI pyrotechnic mixed with KCl or KI is very high, reaching 10 12 -10 13 /g [4][5][6][7].The associated mechanism starts with the Bergeron process [8] by the addition of artificial ice nuclei.The artificial ice nuclei can be rapidly activated if there is sufficient amount of supercooled water in the cold cloud.The activation time is approximately a few seconds to 20 min.
Femtosecond laser filament is a new technique developed in recent years to induce water condensation, which has been demonstrated both in a cloud chamber and in the atmosphere [9][10][11][12][13].During the propagation of an intense laser beam in air and dispersive media, a long filament with a constant diameter of ~100 μm is generated when a dynamic balance is achieved between self-focusing, induced by the Kerr lens and the defocusing effect induced by the high-density plasma [14][15][16][17].The intensity inside the filament is high and clamped at ~5 × 10 13 W/cm 2 in air [18,19].Such a high intensity can ionize air molecules and generate a large number of molecules and ion fragments.The plasma density is in the range of ∼10 14 -10 16 cm −3 , depending on the focusing condition [20][21][22].Under atmospheric conditions, due to the existence of water and other compounds, highly hygroscopic acids, such as HNO 3 or salt in the form of NH 3 NO 3 are finally created following the complex oxidation reactions [12,23,24].The resulting concentration of HNO 3 is in the ppm range [12,23].The formation of HNO 3 has been identified as one of the major pathways to increase the total aerosol mass, for atmospheric relative humidity (RH) values higher than 70% [10,25].The aerosol particle diameters generated by laser filaments are ranging from a few nanometers to ~400 nm [26].The laser filament also generates a strong shock wave, which shatters or vaporizes the large particles around the filament into small ones.These small particles promote the secondary ice multiplication, with a supersaturation relative to ice, in the neighboring region created by shock wave vaporization [11,27].Water condensation by illumination with ultraviolet light has also been reported [28][29][30].However, atmosphere has strong absorbance of ultraviolet light, which limits its remote application.Powerful femtosecond laser pulses can propagate over long distances of several hundred meters [31][32][33] even to several kilometers [34] in air by filamentation.Femtosecond laser filament has also been demonstrated to be able to be transmitted through clouds with a high optical thickness of 3.2 (5% transmission), strong extended turbulence, or dense fogs [35][36][37].Therefore, it can be considered as a novel driving source for water condensation in the atmosphere.
The efficiency of water condensation has always been in the focus of attention in weather modification.When the AgI-type pyrotechnic is combined with the irradiation of a femtosecond laser filament, the water condensation yield is expected to increase, considering that laser filaments can provide both the highly hygroscopic acid and the strong airflow for the growth of AgI condensation nuclei.Therefore, in this work, the femtosecond laser filament-assisted AgI-type pyrotechnic nucleant-induced water condensation was investigated in a cloud chamber, where the kind of chosen AgI pyrotechnic is commonly used in the field of artificial weather modification.The side Mie scattering of particles, water condensation, laser-induced side spectra, and particle number and size were presented for the cases with and without the assistance of the laser filament.The results indicate that the femtosecond laser filament has a positive effect on hygroscopicity, and of the number density and size of pyrotechnically generated aerosol particles.The

NO
− concentration in the melt water increases by more than ten times when 25-or 50-ml sample was irradiated by the laser filaments.

Experimental setup
A schematic of the experimental setup is shown in Fig. 1.Regenerative 800-nm Ti:sapphire laser pulses of 29 fs/6.7 mJ/1 kHz were used in our experiments.The laser pulses were focused by an f = 400 mm lens, and then propagated collinearly with the probe laser into the cloud chamber.The probe laser was a continuous wave (CW) 532-nm laser beam with an output power of 2.5 W, focused by an f = 700 mm cylindrical lens, after its diameter was expanded.The side Mie scattering pattern of the airflow was recorded using a digital camera (Nikon D7000, Nikon Corporation, Japan).The dimensions of the cloud chamber were 0.5 m (length) × 0.5 m (width) × 0.2 m (height).A vertical temperature gradient was maintained in the chamber by using a refrigerating machine to cool the bottom base plate to a set temperature of −50 °C, while the top of the cloud chamber was cooled by a layer of dry ice particles with a thickness of 40 mm.The dry ice was placed into a glass tray with the same size of that of the top plate of the cloud chamber.An air gap of 10 mm was maintained between the glass tray and the top plate of the chamber in order to avoid very low temperatures in the cloud chamber.During the experiments, the walls of the cloud chamber and the top of the glass tray were covered by 2-cm-thick insulation materials.A reservoir filled with distilled water was mounted at a height of 17 cm above the cold bottom plate inside the chamber.The quantity of water vapor inside the chamber was controlled by adjusting the electric current on a heating wire submerged in the water.The height of the laser axis relative to the cold bottom plate of the cloud chamber was set to be 30 mm, where the RH reached a maximum value of ~72% at approximately −14.1 °C, based on the measurements by the temperature and humidity sensors.The experiments with different cases were carried out after the cloud chamber was cooled for 1 h.The fluorescence signals of filament were detected through a side window (25 × 25 mm, 1.2-mm-thick, fused silica) on the cloud chamber, which was imaged by two lenses into the slit of a grating spectrometer (Shamrock 303i, Andor).A laser particle counter [six channels (0.3, 0.5, 0.7, 1.0, 2.0, and 5.0 μm), BCJ-1D, Suzhou Huada Instrument Factory] was used to measure the number density and size of the particles around the laser filament by sampling a certain amount of airflow, using a tube inserted from a hole in the center on the top plate of the cloud chamber.The pumping tube was located at a height of 1 cm above the filament center.The record time of per sample was 30-40min, and sampling air was at a rate of 2.83 L/min.
The AgI-type pyrotechnic was provided by Chinese Academy of Meteorological Science.Each pyrotechnic was pressed into a small cylinder with a mass of ~1 g.It mainly consisted of NH 4 ClO 4 , KCl, AgI (3% by mass of the mixture), phenolic resin, and CuI 2 .We can only provide the mass percent of AgI because the patent protection of the manufacturer.An abundance of aerosol particles were produced when AgI-type pyrotechnic was burned.The formed AgI nucleants are composed of KCl covered with AgI and few CuI 2 particles.The diameter of the larger particle was in the range of 100-600 nm, and for the smaller ones, typically in the range of tens of nanometers based on the transmission electron microscope (TEM) measurements from Chinese Academy of Meteorological Science.The mean cube root diameter of the aerosol particles is 0.2472 μm.Electron microscopy observations indicated that the aerosol particles had various shapes, such as spherical, hexagonal, and irregular.Occasionally, one or more small particles adhere to a larger one, which is a typical way of merging.The order of magnitude of the nucleating effectiveness is 10 15 /g at −15 °C and 10 14 /g at −7 °C.In the experiments, firstly a part of the pyrotechnic, weighed 0.3675 g, was burned in a closed cubic chamber of 1 m 3 to dilute the concentration, then 25-ml or 50-ml sample was injected into the experimental cloud chamber using an injection syringe.
We also used a microscope (XSP-11C, Shanghai Optical Instrument Factory No.1) to observe the condensate particles on the glass slide around the filament.Initially, the glass slide was cooled in a foam box with some dry ice particles for 1 min, then immediately inserted into the cloud chamber at a height of 2 cm above the cold plate for 5 min for the condensation of particles.Finally it was removed from the cloud chamber and immediately placed on the cooled stage of the microscope for observation.The magnification of the objective lens was 10 × and its numerical aperture (NA) was 0.25.The nucleant produced by the burning of AgI-type pyrotechnic was slowly injected into the cloud chamber from a small hole in the center of the top plate of the cloud chamber, after the cloud chamber was cooled for 1 h.The side Mie scattering images of the background following the 1-h cooling and the injection of 50-ml sample are shown in Figs.2(a) and 2(b), respectively.It can be seen, that due to the lower temperature, several background particles exist in the cloud chamber.The size of the larger particles was about 200 μm, and a downdraft was also observed due to the much lower temperature on the top of the cloud chamber (about −25.6 °C).Once the 50-ml sample was injected, as shown in Fig. 2(b), a part of them immediately condensed into very large particles with a size of a few mm.They fell onto the cold bottom plate with a velocity of ~24 cm/s.The very large particles became invisible after 3 min from the injection of 50-ml sample owing to precipitation [Fig.2(c)], which were replaced by more particles of slightly larger sizes than that of the background particles.When the laser filament was introduced into the cloud chamber, a vortex motion was also induced around it [Fig.2(c)].However, the strength of vortex is lower than that of background downdraft; therefore, it only affects the airflow around the filament.Comparing Fig. 2(c) to Fig. 2(d), it is seen that the particle concentration is lower with the laser filament only than that with both nucleant and the irradiation of the laser filament although the pattern of airflow is similar.It is also concluded that the vortices in Fig. 2(c) was induced by the laser filament.The size of the particles is much smaller around the filament [Figs.2(c) and 2(d)], which is due to the laser-filament induced heating, vaporization, and shattering by the shock wave, as well as the diffusion by airflow motion.Therefore, a weak scattering is visible in this case.

Experimental results
Figures 3(a)-3(d) indicate microscope images of the particles condensed on the glass slide for different cases.It was found that the condensate particles appeared like small ice particles with sizes in the range of 4-6 μm when the condensation was not promoted by the laser filament and pyrotechnically generated nucleant, as shown in Fig. 3(a).In addition, the condensate particles were less dense.When the 50-ml sample was added [Fig.3(b)], more particles with sizes in the range of 4-11 μm condensed on the glass slide.Because small-sized particles dominate on Fig. 3(b), a small number of large-sized particles (at the bottom of the image) are not easy to distinguish from the connected particles.There were several black dots inside the particles (probably aerosol particles); however, we could not recognize them clearly at the applied resolution.For the case when 50-ml sample was injected firstly and then the laser filament was shot [Fig.3(c)], the condensate particles became more noticeable and appeared to be similar to slightly frozen liquid droplets.The sizes of particles were in the range of 5-15 μm.Compared to Fig. 3(b), the number of large-sized particles increased evidently, and the interval of particles also became large.Moreover, the particles became liquid droplets linking together faster than those in the other two cases after the glass slide was moved to the cooled stage of the microscope from the cloud chamber.This is probably due to the generation of highly hygroscopic HNO 3 by the femtosecond laser filament.The particles condensed on the glass slide by using only the laser filament [Fig.3(d)] exhibited similar characteristics to Fig. 3(c) except that the number of large-sized particles increased when both AgI nucleant and the laser filament were used.In Figs.4(a)-4(d), the shape and the mass of snow on the cold bottom plate are compared for the cases when no laser filament and pyrotechnically generated nucleant were used (background), when only 50-ml sample was used, when the laser filament interacted with 50ml sample, and when the laser filament irradiated only, respectively.For the case without the laser filament and the injection of 50-ml sample [Fig.4(a)], the snow on the cold bottom plate was ramiform.The size of the main trunk of these structures was ~1.6 mm.When the 50-ml sample was added, as shown in Fig. 4(b), the condensation became more efficient and the stacked particles were visible on the cold plate, consequently the ramiform structure of particles could not be recognized easily.In this case the size of snow particles was ~2.1 mm.The ramiform structure of the condensate snow became noticeable again when 50-ml sample was injected followed by the shot of the laser filament [Fig.4(c)].The length of the main trunk of the structures was also slightly longer, approximately 2.3 mm.This is due to the production of more small sized-particles [26], as well as the turbulence disturbance by the laser filament [12,38,39].For the case with the laser filament only, the shape of snow was similar to that when the laser filament irradiated AgI nucleant, but the length of main trunk of snow structures were slightly shorter.The large-sized snow particles became visible when AgI nucleant was added [Fig.4(c)].The snow on the entire bottom plate of the cloud chamber for 2 h was collected and weighed.It was found that the weight of snow increased by 4.2-13.1% when the laser filament interacted with 50-ml sample, compared to that in the case when only 50-ml sample was added.The same experiments were also performed for the injection of 25-ml sample for comparison, and similar results were obtained.The concentration of several ions in the melt water was measured by an ion chromatograph (Dionex ICS-5000 + ), and summarized in Table 1.It can be seen that the 3 NO − concentration increases by more than ten times for the condensation by the laser filament irradiating on 25ml or 50-ml sample than that when the same amount of nucleant were added only.Compared to this, the irradiation of the laser filament has a negligible effect on the concentration of other ions, including Cl − , K + , and I − .This indicates that the production of hygroscopic HNO 3 plays an important role in the laser assisted AgI-type pyrotechnically generated nucleant-induced water condensation.a For I − , < 0.3 indicates that the concentration has not been detected because the lower limit of the measurement was 0.3 mg/L for the ion chromatograph.

Discussion
In order to clarify the condensation mechanism of the irradiation of laser filament on pyrotechnically generated nucleant, the side fluorescence spectrum was investigated, as shown in Figs.5(a)-5(c).The identification of spectral lines is obtained from Ref [40][41][42].As can be seen in Fig. 5 (a), the shape of the spectrum in the range of 250-550 nm and 400-700 nm begins to change when the dose of sample increases to 50 ml and 25 ml, respectively.An intense plasma continuum appears at this point.The shape of the spectrum is nearly the same for different doses except for the enhanced strength once the critical dose is surpassed.Several characteristic peaks of N 2 and 2 N + superposed on the wide strong plasma continuum in the range of 250-550 nm, indicating the strong ionization of N 2 by the laser filament.In addition, the spectral line of O 3 315.6 nm was also included.In Fig. 5(b), it is seen that the peak at 414.2 nm of N 2 disappeared with increasing doses.This is owing to that the spectra at 750-850-nm band (the laser wavelength is 800 nm) reached a saturated intensity at the integral time of 10 ms when the injected dose was larger than 25 ml, which just located at the right edge of sampled band and so smoothed the peak at 414.2 nm.It is seen that in Fig. 5(a), the peak at 414.2 nm of N 2 was always visible. Figure 5(d) indicated the dependence of spectral intensity of 391.4 nm for 2 N + on the injected doses of nucleant.It is seen that the spectral intensity of 391.4 nm for 2 N + rose rapidly when the dose of nucleant was larger than 25 ml.Other spectral intensity of 315.6 nm for O 3 also became much stronger as the increase of the injected doses of nucleant, as shown in Fig. 5(a).This indicated that more charged ions and O 3 were produced due to the addition of nucleant.Several atomic spectral lines of Ag, I, and K cannot be identified from the wide plasma peeks (I-V) as shown in Fig. 5(b).Peeks I and III are at ~531 nm, which covers several atomic spectral lines of Ag, I, and K or molecular spectral lines of CuI.The position of peek II is at ~517 nm, where the atomic spectral lines of I II 516.1 nm, 517.6 nm, and Ag II 519.8 nm are involved.Peeks IV and V are positioned at ~527 nm, where the atomic spectral lines of I II 526.9 nm and K XII 527.7 nm are included [40,41].This is induced by the intense avalanche ionization between the interaction of the laser filament with the nucleant.The active silver atoms, ions and nanoparticles acting as efficient catalysts [43][44][45], probably enhance the selective oxidation reaction of active nitrogen/nitrous oxides toward the creation of N 2 O 5 .Therefore, HNO 3 with much higher concentration was produced as shown in Table 1.+ on the injected doses of nucleant.For comparison, the data of burning AgI pyrotechnic directly has also been added (the black arrow).The identification of spectral lines is obtained from Ref [40][41][42].The spectra were collected by a grating spectrometer (Shamrock 303i, Andor) with a 300 grooves/mm grating, where the width of the slit was fixed at 100 µm.In pane (a), the integral time was 500 ms for 0-, 25-, and 50-ml sample, 10 ms for the cases of 100-and 200-ml sample, and when AgI pyrotechnic was burned by the laser filament directly.In pane (b), the integral time was 500 ms for 0-and 25-ml sample, 10 ms for the cases of 50-, 100-, and 200-ml sample, and when AgI pyrotechnic was burned by the laser filament directly.
The approximate number density and size of the particles were also measured using a laser particle counter, as shown in Fig. 6(a)-6(c).During the process of measurement, the aerosols were pumped continuously into the instrument, without any additional air and nucleant.It is noted that the number density of particles decayed to that of the initial background slowly over time.However, there is a significant difference in the decay time for the cases with and without the shooting of laser filament [Figs.6(b) and 6(c)].Here, for a given particle size, the decay time of particles is defined as the time interval from the addition of pyrotechnically generated nucleant/laser shooting to the time when the number density of particles decays to that of the initial background, that is the noise level.The particle decay depends on particle deposition, evaporation, coalescence, and etc. Laser filament-induced heating and shock wave lead to particle evaporation [27].The intense airflow motion induced by laser filament accelerates particle coalescence to grow into large-sized particles until sedimentation.Additionally, the adhesion of particles to the walls of chamber also contributes to particle decay.The decay time of particles was ~16 min when only the laser filament irradiation was used, where the highest particle number was ~3.2 × 10 6 /m 3 at the size of 0.3 μm, and the number density decreased with the increase of the particle diameter.The injection of nucleant only increased the particle number by one order of magnitude to ~8.5 × 10 7 /m 3 at the size of 0.3 μm, however, the decay time reduced to 13 min, as shown in Fig. 6(b).The decrease of decay time is associated the quick sedimentation of AgI pyrotechnic nucleant.Most of AgI pyrotechnic nucleant would grow into large-sized particles rapidly once it was injected into the cold cloud chamber, as shown in Fig. 2(b).When laser filament irradiated the AgI pyrotechnic nucleant, owing to the creation of lots of new particles, as well as laser shock wave shattered the large particles to small ones, the decay time of particles increased from ~13 min to ~18 min by the comparison of Figs.6(b) with 6(c).It is seen that the irradiation of the laser filament hardly induced the generation of particles of sizes larger than 1 μm [Fig.6(a)].This is in agreement with the reported results, indicating that the sizes of particle created by the laser filament ranging mainly from a few to hundreds of nanometers [26].Laser filaments induce the generation of ionization, hygroscopic species, local heating, and shock wave.These processes play roles in different phases of water condensation.Laser ionization generates hygroscopic HNO 3 , charged ions and electrons into cloud, a fraction of which could create new cloud condensation nuclei (CCN) by themselves or mixing with /attaching to the existing aerosol particles.Due to the addition of these substances, the supersaturation degree required for nucleation is lowered greatly [25,46], so that the nucleating of particles becomes much easier.It is found that the required supersaturation degree for hygroscopic HNO 3 is much lower than that for charged ions, and for positive ions is higher than that for negative ions [25,46].Local heating induced by laser filaments also generates intense airflow [13], which promotes the growth of small-sized particles into largesized ones [12].The mixing of cold and warm air with a large temperature gradient creates a supersaturated condition inside the vortices, where the particles can continuously grow by both the condensation and the coagulation of particles with different sizes [47,48].Laser filament-induced shock wave could break down or vaporize a fraction of large particles around the filament, which produces lots of small-sized particles [11,27].These small-sized particles would trigger the secondary particle multiplication and increase the total amount of water condensation through the subsequent growth.
In our present experiments, when the laser filament irradiated the AgI nucleant, intense ionization would occur.Lots of electrons, charged ions (such as 2 N + , Ag + , I -, K + and Cl -), molecules and cluster fragments were generated through the intense impact ionization.Most free electrons would recombine finally with ions, but a small fraction of electrons and ions remained.Electrons tend to attach to O 2 molecules and form 2 O − [49].These charged ions are favourable to the nucleating of aerosols, and so new CCN are created.Due to the addition of AgI nucleant, the concentration of charged ions induced by the laser filament also increased strongly, which is indicated by Figs.5(a), 5(b) and 5(d).The laser filament also induced the ionization of air molecules in the cloud chamber, and generated extremely high concentration of O 3 [Fig.5(a)] and nitric oxide (NO, NO 2 and etc.), leading to the formation of HNO 3 in the ppm-range via a series of complex photo-chemical reactions [23].The active Ag, Ag + and nanoparticles generated by the laser filament irradiating AgI nucleant also enhanced the selective oxidation reaction of active nitrogen/nitrous oxides toward the creation of N 2 O 5 , finally HNO 3 by acting as efficient catalysts [43][44][45], and so the concentration of HNO 3 increased by more than ten times than that with the laser filament only.In addition, highly hygroscopic KCl was also produced by the irradiation of the laser filament on the nucleant.A certain amount of these hygroscopic salts, including HNO 3 and KCl, attaching to the surface of the existing aerosols or themselves would become another fraction of the initial CCN.As said above, the hygroscopic molecules are more effective than charged ions for particle nucleating, which are also known to contribute to the laser-induced condensation process by stabilizing and sustaining the growth of particles [25].In the experiments, the concentration of HNO 3 in the melt water was much higher than that of K and Cl ions when AgI nucleant was irradiated by the laser filament, as shown in Table 1, therefore, it is reasonable to believe that most of the "newly created" particles in our experiments were from highly hygroscopic HNO 3 .The hygroscopic CCN were activated and grew up subsequently in the environment of abundant water vapor or droplets.As a consequence, the condensate particles had larger sizes and were highly soluble, as shown in Fig. 3(c).For the roles of airflow and shock wave, it is seen that the laser-induced vortices [Figs.2(c) and 2(d)] were not so regular as those in our previous experiments [39], moreover, the numerous small-sized particles shattered or vaporized by shock wave were not observed evidently around the laser filament.Therefore, we believe that hygroscopic HNO 3 , instead of airflow and shock wave induced by laser filament played a dominant role in the increase of snow amount on the bottom plate for the case when the laser filament irradiated on the nucleant.In addition, the HNO 3 production with a high concentration by the interaction of the laser filament with the nucleant also indicate that the laser filament can be presented as a novel method of adding hygroscopic compounds to AgI-type pyrotechnics directly.As N 2 is abundant in the atmosphere, the amount of HNO 3 can be controlled flexibly by the laser irradiation time.Thus, the amount of water condensation and precipitation can be controlled accordingly.

Conclusion
We investigated the AgI-type pyrotechnic nucleant-induced water condensation promoted by a femtosecond laser filament in a cloud chamber.It is found that when the nucleant were irradiated by laser filaments, the condensate particles on the glass slide become highly soluble, and their sizes become slightly larger, in the range of 5-15 μm.The snow particles condensed on the cold bottom plate were ramiform, but their microstructure can be identified more clearly compared to the case without the irradiation of laser filaments.The length of the main trunk of the ramiform structures also becomes longer, approximately 2.3 mm.Measurements using a laser particle counter indicate that particles of sizes larger than 1 μm are rarely produced by the irradiation of laser filaments on the nucleant; however, due to the generation of a large amount of small sized particles, the laser filaments increase the decay time of particles from ~13 min to ~18 min.The amount of snow on the cold bottom plate increases by 4.2-13.1% compared to that without the irradiation of laser filaments.These results are associated with the production of high-concentration HNO 3 .The concentration of

Fig. 1 .
Fig. 1.Schematic of the experimental setup.The femtosecond laser beam (red line) was focused by an f = 400 mm lens, then it passed through an entrance window (25 mm × 25 mm, 1.2-mm-thick, fused silica), and exited at the opposite side through an exit window (25 mm × 25 mm, 1.2-mm-thick, fused silica) onto a beam block.The probe laser beam (green line) was used for the in situ light scattering measurements after its diameter was expanded using f = 300 mm and f = 30 mm lenses, and then focused by an f = 700 mm cylindrical lens.

Fig. 2 .
Fig. 2. Side Mie scattering images of the airflow and particles: (a) background after cooling for 1 h, (b) injection of 50-ml sample from the hole on the top plate of the cloud chamber following the 1-h cooling, (c) laser filament irradiating after the injection of 50-ml sample, (d) irradiating by the laser filament only after the cloud chamber was cooled for 1 h.The scale is the same in the four images.The arrow in (c) indicates propagation direction of the laser.The laser propagation direction is the same for (c) and (d).Pane (a) was captured by a Nikon D7000 camera with a shutter speed S = 1/50 s, an f number F = 3.2, and a light sensitivity ISO = 1600.Panes (b) -(d) were captured by video frames (40 laser shots, F = 3.2, ISO = 1600).The two light green regions labelled "downdraft" in (a) are the blurry images of two edges of the opposite window.The white horizontal line at the bottom of each image in (a)-(d) is the precipitated snow and condensate frost at the cold bottom plate of the cloud chamber.

Fig. 3 .
Fig. 3. Microscope images of particles condensed on the glass slide: (a) without laser filament and pyrotechnically generated nucleant (background), (b) with the addition of 50-ml sample, (c) when 50-ml sample was injected at first and the laser filament was shot subsequently, and (d) with the irradiation of the laser filament only.The scale is the same in all four images.

Fig. 4 .
Fig. 4. Snow on the cold bottom plate: (a) background without pyrotechnically generated nucleant and laser filament, (b) addition of 50-ml sample only, (c) laser filament irradiation after the addition of 50-ml sample, and (d) irradiation by the laser filament only.The scale is the same in all four images.The exposure conditions of the camera were S = 1/50 s, F = 5, and ISO = 1600 for (a), (b) and (d), S = 1/50 s, F = 5, ISO = 1000 for (c).

Fig. 5 . 2 N+
Fig. 5. Side fluorescence spectra when the laser filament irradiated on pyrotechnically generated nucleant with different doses, including 0-(red lines), 25-(green lines), 50-(blue lines), 100-(black lines), 200-ml (cyan lines) sample, and burning 0.3675 g AgI pyrotechnic by the laser filament directly [pink lines].The plot legends are same in (a) and (b).(c) The enlarged spectra of (a) at 391.4 nm for 2 N + (marked with the black arrows).(d) The dependence of spectral intensity of 391.4 nm for 2 N

Table 1 . Concentration (mg/L) of several ions in the melt water
3 N O −