Dr Reidopened the discussion of Professor Shiratani’s paper: You stated that there is a temperature gradient of 25K cm−1. How do you measure this or is it based on simulation?Professor Shirataniresponded: The electrode temperature was measured with a thermocouple. The gas temperature between the electrodes was deduced using tiny thermo-labels hanging between the electrodes.Professor Goreesaid: You mentioned that particles rise towards your upper horizontal electrode due to a thermophoresis force. This force arises from a temperature gradient in the gas. Did you purposefully apply a temperature difference to the two electrodes to produce this temperature gradient, or does it arise for another reason?Professor Shiratanireplied: We did not give intentionally produce the temperature gradient. The powered electrode was heated by bombarding ions flowing out of the plasma, whereas the temperature of the grounded electrodes was kept at 358 K. The temperature rise of the powered electrode led to the temperature gradient between the powered electrode and the grounded electrodes.Professor Rühlasked: How did you determine the particle size and the width of the size distribution? How are the charges distributed in the particles? Are there multiply charged particles?Professor Shirataniresponded: The average size of particles was measured by anin-situlaser light scattering method. The particle size distribution was obtained byex-situTEM observation of particles collected on TEM mesh.We have not measured the charge distribution of nanoparticles. However, we can deduce the average charge on a nanoparticle from quasi-neutral conditions of plasmas. The average charge is in a range of −0.01 e to −0.15 e as shown in Fig. 5 of our paper, and therefore most nanoparticles are neutral and the rest have a charge of e−. We expect few multiply charged particles in our case.Professor Goreeremarked: How did you determine the size of particles in your time-series graphs of particle diameter? Did you do this by collecting them after turning the plasma off and sizing them using electron microscopy?Professor Shiratanianswered: We determined the size of particles using anin-situlaser light scattering method. The experimental setup and the method are described in ref. 13 of our paper.11 S. Nunomura, M. Kita, K. Koga, M. Shiratani and Y. Watanabe,J. Appl. Phys. 2006,99, 083302.Dr Reidasked: How do you determine the size of the nanoparticles? Can you give us more detail?Professor Shiratanianswered: The size and density of the nanoparticles are determined from their thermal coagulation that takes place after turning off the discharge. Thermal coagulation usually takes place after turning off the discharge; because during the discharge, particles are charged negatively and their coagulation is strongly suppressed by their mutual Coulomb repulsive force. The details were described in ref. 13 of our paper.11 S. Nunomura M. Kita, K. Koga, M. Shiratani and Y. Watanabe,J. Appl. Phys. 2006,99, 083302.Professor Kerstencommented: Do you have any information on the time evolution of the electron temperature during particle growth? (see Fig. 5 of your paper).Professor Shiratanianswered: At the time of the discussion, we had not carried out optical emission spectroscopic measurements which give information on the time evolution of the electron temperature. After the Faraday Discussion, we carried out such measurements. The results show that intensity of Ar emissions from the plasma tend to increase with time. Moreover, a ratio of two Ar emission lines varies with time, which indicates the electron temperature tends to increase during particle growth.Professor Stoffelsstated: A remark on the question by Professor Kersten: Considering the electron density decreases, it is expected that the electron temperature increases. In Ar/SiH4mixtures this is observed as a significant increase of emission of Ar lines. Determination of the excitation temperature, based on the Ar emission lines also shows an electron temperature increase.Professor Shiratanianswered: We have carried out some optical spectroscopic measurements and based on the results the time evolution of the excitation temperature during particle growth in our study is similar to that in Ar/SiH4mixture discharges.Professor Stoffelscommented: Is the average charge enough and/or fluctuating fast enough that collective effects are important? Does a cloud of 1–2 nm particles behave as a Coulomb liquid?Professor Shiratanianswered: Fast fluctuation of charge on particles is important to obtain collective effects of particles. For instance, the suppression of thermal coagulation occurs even when a fraction of particles are charged. It is known that the charge on particles always fluctuates due to the random attachment of electrons and ions, therefore, particles undergo Coulomb repulsion in a certain period of time. The suppression conditions for the thermal coagulation are considered to be that the charging frequencyfchof particles is higher than the collision frequencyfp–pamong them.Professor Goreeopened the discussion of Dr Pavlů’s paper: You reported that your spherical gold particles become elongated as they are reduced in size by sputtering when exposed to a directional ion beam. I would have expected that rotation of particles suspended in your Paul trap would average out the directional effects of the beam. Can you speculate why this occurs?Dr Pavlůreplied: The question is very complex starting with the uncertain shape of the grain—the only parameter we observe is the temporal change of the grain capacitance with respect to the sphere of the same mass (Fig. 1), thus the ellipsoidal shape is a speculation so far.1We can consider two possibilities:(i) A very simple intuitive mechanical approach is that the irregular (for any reason) grain is turning towards the ion beam with the smallest area (like a weather-vane in a wind).(ii) The second approach is taking into account the interaction of the space charge of the ion beam with the quadrupole term of the electric field of the grain. However, there is a key question of the conductivity of the grain—the described idea is valid for metal grains but for dielectrics it would probably lead to an increased rotation of the grain because of the dipole moment. This approach is described in our paper, however, without any quantitative measurements in our case. The question which of these alternatives dominates remains open so far. Nevertheless, our preliminary calculations show that the second interaction (ii) could be reasonably strong to stop grain rotation.Note here that, no matter whether (i) or (ii) is right, a similar effect would be observed wherever the grain is exposed to the directional stream of ions/atoms,e.g., the dust grains in the solar wind.1 J. Pavlů, I. Richterova, Z. Nemecek, J. Safrankova, and J. Wild,IEEE Trans. Plasma Sci., 2007,35(2), 297–302.Temporal evolution of the grain relative capacitance–the indication of shape change.1Modified with permission © 2007 IEEE.Professor Signorellasked: Would a comparison with sputtering of deposited particles help to elucidate the origin of the shape change?Dr Pavlůreplied: The origin of the grain shape change is known—the sputtering process is more efficient for grazing impacts. The question is how can the grain be stopped from rotating. Though we suppose we can carry out such an experiment, we don't think it will bring any new light to the changing shape feature we've observed.The major role of the grain–beam interaction would be overprinted by a rigid connection of the grain to the substrate.Professor Davidovitsasked: What is the astrophysical evidence for the sputtering results you observed?Dr Pavlůanswered: There exists some indirect observation of sputtering. Or, perhaps it is better to say, there exist phenomena that are likely to be explained by the sputtering of dust grains. One of these is the presence of so called pick-up ions. They are ions of heavy species which do not fit to the solar wind neither by their abundances nor by their energies. One of the possible sources of these ions could be sputtering of dust followed by ionization.1Mukai and Schwehm2calculated the critical distance from the Sun where sputtering of grains dominates their sublimation. They have shown that life-times of interplanetary dust grains are controlled by sputtering by solar wind.Another indirect evidence comes from modeling of Saturn's E-ring where sputtering is supposed to play a significant role. Our calculations, based on sputtering of 1 μm ice grain by O+ions (number density 5 cm−3, mean energy 100 eV), predict sputtering to half mass in the range of tens of years, sputtering to half diameter in less then hundred years, and total grain erosion in about one or two hundred years. The model of Juracet al.3clearly shows that the speed of grain erosion due to sputtering would be similar—1 μm grain in about 50 years.1 I. Mann, H. Kimura, D. A. Biesecker, B. T. Tsurutani, E. Grün, R. B. McKibben, J.-C. Liou, R. M. MacQueen, T. Mukai, M. Guhathakurta and P. Lamy,Space Sci. Rev., 2004,110, 269–305.2 T. Mukai and G. Schwehm,Astron. Astrophys., 1981,95, 373–382.3 S. Jurac, R. E. Johnson and J. D. Richardson,Icarus, 2001,149(2), 384–396.Professor Rühlasked: Is the emission of secondary electrons the only and the most important charging mechanism? Would you expect to obtain comparable results from a photon impact experiment?Dr Pavlůreplied: Definitely not, however, in several regions of space or in particular laboratory experiments hot electrons are present and secondary electron emission becomes important.Recent experiments by Grimmet al.1have shown that secondary emission plays an indispensable role when dust grains are charged by UV or soft X-ray irradiation because most of the electrons leaving the irradiated surface are secondary electrons excited inside the target by primary photoelectrons or Auger electrons.The photoemission and secondary emission become similar if the grain size is comparable with the penetration depth of the beam electrons into the grain. In such a case, a great majority of beam electrons deposits a part of their energy inside the grain and leaves it without changing its charge (like the photons do). However, there is an important difference in maximum potential of the grain that can be achieved by charging with a photon or electron beam of the same energy. Whereas an upper limit of the grain potential charged by photons is slightly less than the photon energy,1it is only one half of the energy of the electron beam. The upper limit of the grain potential is given by the highest energy of the electrons leaving the grain. This energy is roughly equal to the photon energy in case of photoemission. On the other hand, the primary electron from the beam brings its charge into the grain and, in an ideal case, it can spend half of its energy for excitation of the secondary electrons and both of them can then escape from the grain charged below a half of the beam energy.Our estimation of the equilibrium potential of ≈1 μm glass grain as a function of the energy of primary particles (electrons or photons) is sketched inFig. 2. The thin lines show theoretical limits of the grain potential for photons (dashed) and electrons (solid). The solid part of the thick gray curve shows the charging by photons estimated from Grimmet al.1The dotted continuation of this line is our estimation of further increase of the grain potential taking account of potential limitations by the ion field emission.2The solid part of the black curve is consistent with the measurement on 1.2 μm SiO2spheres shown in our paper and the dotted part is our estimation of a further potential evolution. We expect the same limit of the surface potential due to the ion field emission as for photon charging. From the sketch it follows that the above described “half-energy limit” cannot be reached for grains in the micrometer range of size.1 M. Grimm, B. Langer, S. Schlemmer, T. Lischke, U. Becker, W. Widdra, D. Gerlich, R. Flesch and E. Rühl,Phys. Rev. Lett., 2006,96(6), 066801.2 M. Jerab, I. Richterova, J. Pavlů, J. Safrankova and Z. Nemecek,IEEE Trans. Plasma Sci., 2007,35(2), 292–296.A sketch of possible equilibrium surface potential of about 1 μm glass dust grain as a function of the primary electron (black line)/photon (gray line) energy.Professor Rudichasked: How would surface corrugation affect the process and conclusion with respect to the real space-borne grains?Dr Pavlůanswered: All of the aforementioned processes have their size dependences. Space-born grains are often highly irregular exhibiting a lot of tips and other features.Most of these can be described by a characteristic dimension (curvature of the tip, thickness of flakes,etc.) which determines the charging process.Experiments with the spherical grains of similar diameter to this characteristic dimension could shed light on these processes. Sharp edges will play a major role when dealing with emissions due to an intense electric field.Also the secondary electron emission from thinner parts of dust grains can be significantly enhanced. Although we tend to use spherical grains, some of the results can be easily extended to irregular grains. We've reported several features of grain clusters that can simulate space-born dust.Professor Masonsaid: Role of secondary electrons in Interstellar dust chemistry; production of secondary electrons fromicecovered grains which subsequently leads to complex chemistry in the ice—such that UV + ion irradiation produce similar products.Dr Pavlůreplied: We fully agree with Professor Mason. The surfaces of grains serve as small chemical laboratories in space.Professor Kerstenasked: Do you observe the effect of ion drag (particle movement) if you switch on the ion beam? Do you have to tune the particle trap in order to stabilize the particles?Dr Pavlůreplied: The trap is tuned so that the secular frequency of the grain is well within the stability region.1,2Moreover, the grain motion is stabilized by a damping feedback loop which uses the position signal to reduce the amplitude of the grain oscillation in particular directions using the position signal.3When the ion beam is switched on we often observe grain “shaking” or being shifted a little bit outside the trap center. But it doesn't get far away from the center for a long time being firmly trapped in the potential well and actively stabilized thus the drag would be compensated by both of these.1 W. Paul and H. Steinwedel,Apparatus for separating charged particles of different specific charges,Ger. Pat. 944 900, Jun. 28, 1956.US Pat. 2 939 952, Jun. 7, 1960.2 R. F. Wuerker, H. Shelton and R. V. Langmuir,J. Appl. Phys., 1959,30(3), 342–349.3 I. Cermak, J. Pavlů, P. Zilavy, Z. Nemecek, J. Safrankova and I. Richterova, inWDS'04 Proceedings of Contributed Papers: Part II–Physics of Plasmas and Ionized Media, ed. J. Safrankova, Matfyzpress, Prague, 2004, pp. 279–286.Professor Raycommented: How do you keep the particle stable? By adjusting frequency or AC voltage or both?Dr Pavlůanswered: In principle, both of these parameters can (and sometimes must) be changed.However, the frequency is usually adjusted automatically in order to keep an optimum ratio of the quadrupole AC frequency and secular frequency of the grain motion that follows from the mathematical solution of the trap. In the present setup, we can tune the trap frequency over about two orders of magnitude and the voltage about one order of magnitude. Consequently, the range ofQ/mwe can trap spans over five orders of magnitude (approximately from 4 × 10−3to 4 × 101C kg−1). For details on the trap principle and setup see Wuerkeret al.1and Cermaket al.21 R. F. Wuerker, H. Shelton and R. V. Langmuir,J. Appl. Phys., 1959,30(3), 342–349.2 I. Cermak, J. Pavlů, P. Zilavy, Z. Nemecek, J. Safrankova and I. Richterova, inWDS'04 Proceedings of Contributed Papers: Part II –Physics of Plasmas and Ionized Media, ed. J. Safrankova, Matfyzpress, Prague, 2004, pp. 279–286.Professor Goreeopened the discussion of Professor Kersten’s paper: When you turn off the plasma and the particles then fall down, but more slowly than they would in free fall: how long does it take for the plasma to disappear in this afterglow, do you think that the particles retain a residual charge, and are they exposed to an electric field?Professor Kerstenreplied: I am sure that the particles retain a relatively large residual charge. We can levitate the particles after the plasma is switched-off in front of our adaptive electrode due to a bias voltage at the pixels of this electrode. Hence, the still charged particles are trapped in the electric field of the biased surface even without plasma.Professor Stoffelsasked: How does the charge on a particle vary as it moves and is it a valid approximation to keep the charge constant in the derivation of eqn (10) and (11) of your paper?Professor Kerstenresponded: This is a legitimate question. The charge might vary with the height above the electrode,e.g. with the particle position in the sheath. The assumption of a constant charge is only very rough. Therefore, the measurements of the falling particles differs remarkably from the calculation, see Fig. 10 of our paper. Only at the beginning—where the change in the field is rather small—the assumption of a constant particle charge is valid. In principle, for each positionzin the sheath, a new charge has to be calculated. We hope that we can do this by further evaluation of the fall experiments in the near future. On the other hand, for the falling experiments, where the plasma is switched-off, the particles retain almost their original charge.Dr Reidasked: In this paper and those preceding we have seen that spatial variations in parameters such as temperature, electron density,etc. can be pronounced. There are also different timescales associated with different physical and chemical processes. Can you describe the degree of spatial and temporal resolution required to probe the range of processes important in plasma chemistry?Professor Kerstenanswered: There are different spatial and time scales in dusty plasmas. The rf-cycles of the plasma are in the order of nanoseconds and the particle charging is in the order of μs. The motion of the charged particles to follow changes in the field structure of our adaptive electrode is in the order of a few tens of Hz, which we can observe. The chemical processes resulting in etching or deposition of thin films at the particle surface is even longer. The spatial resolution of the particle position is about 10 to 100 μm compared to typical plasma structures such as the sheath in front of the electrodes or substrates which is commonly in the order of mm to cm.Professor Goreecommented: Is the slowest time scale of all the inertial time scale for particle motion?Professor Kerstenreplied: The lowest time scale is given by the response of the charged test particles to changes in the electric field which the particles can follow,e.g. in the order of some Hz.Dr Reidsaid: Temperature measurements were made in Professor Shiratani’s paper1within the plasma with a thermocouple. You propose to use microparticles which will give a higher degree of spatial resolution in the sheath. How do these two length scales for measurements compare?1 Masaharu Shiratani, Kazunori Koga, Shinya Iwashita and Syota Nunomura,Faraday Discuss., 2008,137, DOI: 10.1039/b704910bProfessor Kerstenreplied: A thermocouple in the plasma acts like a macroscopic probe which is shielded by a sheath. By this method, we would certainly disturb the plasma. For some applications this is not a problem, since plasmas in contact with solids (e.g. substrate processing) always interact with the sheath.However, if we are interested in the energy fluxes in the plasma or in the sheath, respectively, we want to know the temperature field at the probe position without any disturbance. Therefore, tiny test particles are more appropriate for this purpose.Professor Davidovitscommented: Although the particle is levitated it is still in (microscopic) motion about a mean position. Does this give us kinetic (temperature) information about the particle and plasma?Professor Kerstenresponded: The kinetic temperature of a plasma crystal might be in the order of tens of eV and can be determined by the assessment of the microscopic particle motion about its equilibrium position. This motion can be excited by field oscillations, by laser radiation, gas flows,etc.On the other hand, the real particle temperature (surface temperature) differs remarkably from the kinetic temperature and is much lower (e.g. ambient temperature to about 300 °C). We try to measure this temperature by fluorescence.The energy of the plasma (e.g. mainly of the electrons), which is typically in the order of some eV, has to be determined by other methods like Langmuir-probe measurements.Professor Goreesaid: As a comment, I'll mention that the experimenter can track the motion of particles and compute their velocities. This allows computing their mean-square velocity, which is interpreted as a kinetic temperature for the particles. This kinetic temperature can be as high as tens of thousands of Kelvin, even when the surface of the particles remain at room temperature. The particle's surface temperature can be vastly different from its kinetic temperature.Professor Kerstenreplied: This is completely right.Professor Shiratanicommented: When we want to measure local electric field at a certain position, we must put a microparticle in the positionQnot only in thez-direction, but also they-direction. How accurate do you determine the position in 3D space?Professor Kerstenreplied: If we assume a homogeneous field and potential distribution along the electrode, in the first approximation we have only to take into consideration thez-direction. In the 2Dx–y-plane the change above the electrode is small compared to thez-direction. That means we can use a 1D-approximation. However, at the edge of the electrode or if we have a distinguished sheath structure of our adaptive electrode, the other directions also have to be considered. We can determine the particle position with an accuracy of about 10 to 100 μm. The pixels of the adaptive electrode have a diameter of 7 mm. Mostly, we apply a symmetric pattern in order to get a symmetric sheath structure for particle confinement and handling.Professor Masonasked: A key question for chemistry of microparticles is at which size the particle may exhibit bulk like properties—do your experiments give any clue to this? 100 nm? 100 μm?Professor Kerstenanswered: This really is always a key question. The transition between the different stages are not sharp. The synthesis of particles in a plasma is by radicals/moleculesviaclusters and nanoparticles to microparticles. In the stage of cluster/nanoparticle the charge can vary. They might be positively charged by secondary electron emission, negatively charged by electron attachment, or even neutral. These fluctuations are important for the coagulation of clusters/nanoparticles to larger ones. In this period, the particles do not exhibit real bulk properties. If they are grown to microparticles they are permanently negatively charged and now they are more likely to exhibit bulk properties.Professor Changasked: Are you aware of any work in the area of droplet spectroscopy that results in the changing of colors as a function of temperature? I remember there was some literature about an inorganic liquid that changes color as a function of temperature and this could be used as a temperature sensor in the droplet form.Professor Kerstenanswered: Thank you for this interesting advice. We are looking for particles which exhibit a temperature-dependent fluorescence in order to use them as tiny temperature sensors in the plasma or plasma sheath, respectively.Ms Kahanopened the discussion of Dr King’s paper: How does the hygroscopicity of the unoxidized aqueous solutions compare to that of pure water?Dr Kingreplied: The best way to answer this question is to draw the Köhler curves for a pure water droplet and a 10 µm diameter droplet of an aqueous solution of sodium benzoate (0.086 mol dm−3). As can be seen fromFig. 3the Köhler curve for the sodium benzoate solution is much more hygroscopic than the pure water droplet and will form a stable micron sized droplet at a high RH, whereas the pure water droplet will evaporate without a supersaturation of water vapour. Please note the form of the Köhler equation1is very simple and any effect for surface tension, density, solution non-ideality, or insoluble mass at small diameters has not been considered.1 John H. Seinfeld and Spyros N. Pandis, Atmospheric Chemistry and Physics: From Air Pollution to Climate Change, Wiley Interscience, New York, 1st edn, 1986.Köhler curves for a pure water droplet and a 10 µm diameter droplet of an aqueous solution of sodium benzoate (0.086 mol dm−3).Ms Kahansaid: With respect to the lack of droplet growth observed during ozonation of α-pinene, is it possible that volatilization of oxidation products might be occurring?Dr Kinganswered: Yes, this is indeed possible and we at present do not investigate gas-phase products. As an example we monitored the size and Raman spectra of a mixed 8 μm particle of nonanoic acid and aqueous sodium chloride trapped in laser tweezers. The Raman spectra clearly demonstrated that within a few minutes the nonanoic acid had evaporated from the surface of the aqueous droplet; the particle size did not change after the nonanoic acid had evaporated.Dr Reidasked: Knowledge of the gas phase composition is crucial in interpreting changes in involatile solute composition within the trapped droplet. Changes in relative humidity, can, for example, lead to significant changes in wet particle size and solute concentration. To what extent have you included in your analysis any potential errors in reactant concentration resulting from uncertainty in relative humidity?Dr Kingresponded: Kinetics of the reaction could in effect be carried out in two regimes of Relative Humidity (RH). In the first regime the generation of aerosol using an ultrasonic nebuliser wets all surfaces inside the cell. The local RH inside the cell is larger than the RH of the gas stream entering the cell. A trapped particle quickly equilibrates with the local RH in the cell (1–2 min). The particle size is then stable in the cell for approximately an hour as the local RH in the cell is maintained by evaporation from the wet cell walls. However, once the water on the cell walls is exhausted the particle size is controlled by the RH of the incoming gas-stream. Thus, as the cell dries out the particle often shrinks over a period of 5 min and then equilibrates with the new RH. In this second RH regime the particle is again stable. Kinetic runs are often committed in one of these regimes of RH, and any kinetic run that has a changing RH (observed visually by changes in the background on the microscope camera) is discarded. The concentration of reactant is “measured” by using the peak area of the Raman signals for the reactant and knowledge of the initial reactant concentration from the nebulising solution and particle size, both of which are monitored during the initial trapping process. Raman spectra are collected over a maximum of 10 s.Dr Reidcommented: As the cell dries out, particularly during the reaction, can this lead to errors in your assessment of changing reactant and product concentrations?Dr Kingreplied: Kinetics of the reaction could in effect be carried out in two regimes of Relative Humidity (RH). In first regime the generation of aerosol using an ultrasonic nebuliser wets all surfaces inside the cell. The local RH inside the cell is larger than the RH of the gas stream entering the cell. A trapped particle quickly equilibrates with the local RH in the cell (1–2 min). The particle size is then stable in the cell for approximately an hour as the local RH in the cell is maintained by evaporation from the wet cell walls. However, once the water on the cell walls is exhausted the particle size is controlled by the RH of the incoming gas-stream. Thus, as the cell dries out the particle often shrinks over a period of 5 min and then equilibrates with the new RH. In this second RH regime the particle is again stable. Kinetic runs are often committed in one of these regimes of RH, and any kinetic run that has a changing RH (observed visually by change in background on the microscope camera) is discarded. The concentration of reactant is “measured” by using the peak area of the Raman signals for the reactant and knowledge of the initial reactant concentration from the nebulising solution and particle size, both of which are monitored during the initial trapping process. Raman spectra are collected over a maximum of 10 s.Professor Rühlasked: There are probably other oxidizing agents which might react with the organic compounds. This includes OH. What is the mixing ratio of OH radicals near the particles? Are the rate constants for the reactions with OH radicals known, and can these processes be of importance?Dr Kingreplied: In the experiments described in the paper there are no gas-phase hydroxyl radicals present. Hydroxyl radical production in the experiment was avoided by passing dry oxygen through the mercury discharge lamp. In the paper we raise the possibility that aqueous-phase hydroxyl radicals may be produced by secondary chemistry to explain the lack of observed products, but we also demonstrate in the paper that any hydroxyl radicals that may be formed in solution will not compete with the ozone reaction under study.Dr Kriegercommented: Suggestion: To use a test particle as a sensor for relative humidity and then repeat the measurements under identical conditions to ensure the relative humidity at the particle position.Dr Kingresponded: To use a reference particle trapped next to the reactant particle as a measurement and a probe of fluctuations in relative humidity is a good idea (e.g. Mitchemet al.1). However, this can limit the chemical systems that can be studied as it is extremely difficult to obtain two chemically different particles in two different traps without contaminating either particle during the nebulisation process. Such contamination may change the measurement of the uptake coefficient or Köhler behavior of the particle.1 Laura Mitchem, Rebecca J. Hopkins, Jariya Buajarern, Andrew D. Ward and Jonathan P. Reid,Chem. Phys. Lett., 2006,432(1–3), 362–366.Ms Kahanasked: We have observed that aromatic hydrocarbons in aqueous solutions and in organic films undergo ozonation at the surface rather than in the bulk. Could a surface reaction help explain the fast reaction observed for benzoate in water?Dr Kingreplied: We find no evidence for a surface reaction in our studies, but further experiments could be undertaken, as highlighted by Dr Ammann (see below), for further confirmation. A difference between your experiments1and ours is the solubility of the aromatic species studied. Benzoate is very soluble, especially relative to PAHs. Benzoate also has negligible surface activity, and diffuses easily through water.1 T. F. Kahan, N.-O. A. Kwamena, D. J. Donaldson,Atmos. Environ., 2006,40(19), 3448–3459.Dr Ammannsaid:(1) With regard to the experiments with sodium benzoate, it is assumed that the particles are liquid at the relative humidities of the experiment. How was this assured? Have the hygroscopic properties of sodium benzoate been explored? What is the deliquescence humidity?(2) Attempts are presented to model ozone uptake to the particles based on different assumptions to allow conclusions about the limiting process. Especially to differentiate bulkvs. surface kinetics, measuring uptake kinetics as a function of ozone partial pressure would allow a better resolution of this issue than the approach as presented.Dr Kingresponded:(1) Previous experiments where we purposely dehydrated concentrated aqueous salt solutions held in an optical trap clearly show phase transitions. Using the microscope we observe the previous spherical liquid particles develop angular features and can image trapped particles which are a crystalline solid covered unevenly with some liquid. The resulting solid particles often leave the trap quickly and result in a crystalline solid on the cell window. The experimental Köhler curve for sodium benzoate was not measured by this method as it was not necessary and is time-consuming (∼24 h). A calculated curve was used to predict concentrations of the nebulising solution that would result in stable particles of the right size to trap in the laser tweezers.(2) The authors agree that this is one strategy to elucidate a surfaceversusbulk mechanism. Another strategy would be to use the suggestion of Smithet al1and to measure the uptake coefficient as a function of the particle size and thus particle area. In the experiments we undertook we did not have enough experimental runs at different particle sizes or ozone partial pressures to fully elucidate the mechanism. It should be noted that the kinetic decay of the concentration of the organic reactants studied all followed the mathematical formulation for a bulk reaction. As unpublished data for other compounds reacting with ozone show kinetic decays of reactant that clearly indicate surface only reactions, it is possible to distinguish between these two mechanisms with this type of experiment.1 Geoffrey D. Smith, Ephraim Woods, III, Cindy L. DeForest, Tomas Baer and Roger E. Miller,J. Phys. Chem. A, 2002,106, 8085–8095.Dr McGloinremarked: Is fluorescence from the glass in the sample cell a problem when collecting Raman spectra?Dr Kinganswered: We have had no problems with fluorescence from the glass windows on the cell. These are thin microscope glass cover slips from VWR International. They are pre-treated in conc. nitric acid and glued to the aluminium cell. During experimental runs the windows are regularly (every 1–2 h) cleaned in Decon-90 solutions and rinsed with high purity solvents: water, acetone and methanol. The cell and windows are regularly conditioned with high levels of ozone gas (>1 ppm) in oxygen.Dr Reidasked: With the extension of optical trapping to droplets containing aromatic chromophores, can you estimate the degree of heating of your droplets, or do you see the results of heating?Dr Kinganswered: Heating of the samples is an important aspect to consider when selecting chemical systems to study. Any compound (or reaction product) which absorbs at the laser wavelength is unsuitable for study. Experiments with compounds that do absorb light at the laser wavelength (substituted phenols in basic solutions can be problematic) are not stably trapped and will evaporate (often quickly) under conditions that extended Köhler theory would suggest are stable. Non-aqueous pure compounds trapped in an optical trap which absorb at the laser wavelength rapidly and obviously degrade. In another comment I suggested that the trapping of leuco dyes (which exhibit thermochromism) may demonstrate laser heating as a problem or not. Detailed laser heating calculations have not been attempted as the absorption cross-section at the laser wavelength would need to be known for the compounds we have studied. These absorption cross-sections are lower that any practical means we have of measuring them.Professor Jaenickecommented: To what extent would your results apply to the atmosphere? You have claimed influence on cloud properties!Dr Kinganswered: The systems studied were meant to be proxies for ozone reactions with HULIS (benzoate in water) and biogenic aerosol (α-pinene in alkane). To the authors’ knowledge this is the first paper demonstrating the use of Laser Raman tweezers to measure kinetics and uptake coefficients and the systems are kept reasonably simple so they may be understood. Future studies will include more atmospherically complex systems. It should be noted that in a previous study we demonstrated that the oxidation of a droplet of oleic acid and sea water by ozone lead to a dramatic increase in hygroscopicity and thus a change in cloud microphysical properties.11 M. D. King, K. C. Thompson and A. D. M. Ward,J. Am. Chem. Soc., 2004,126(51), 16710–16711.